@comment{THIS FILE IS GENERATED AUTOMATICALLY - DO NOT EDIT}

@misc{AbadietalTENSOR_FLOW-15,
        title = {{TensorFlow}: {L}arge-Scale Machine Learning on Heterogeneous Distributed Systems},
       author = {Martin Abadi and Ashish Agarwal and Paul Barham and Eugene Brevdo and Zhifeng Chen and Craig Citro and Greg Corrado and Andy Davis and Jeffrey Dean and Matthieu Devin and Sanjay Ghemawat and Ian Goodfellow and Andrew Harp and Geoffrey Irving and Michael Isard and Yangqing Jia and Rafal Jozefowicz and Lukasz Kaiser and Manjunath Kudlur and Josh Levenberg and Dan Man\'{e} and Rajat Monga and Sherry Moore and Derek Murray and Chris Olah and Mike Schuster and Jonathon Shlens and Benoit Steiner and Ilya Sutskever and Kunal Talwar and Paul Tucker and Vincent Vanhoucke and Vijay Vasudevan and Fernanda Vi\'{e}gas and Oriol Vinyals and Pete Warden and Martin Wattenberg and Martin Wicke and Yuan Yu and Xiaoqiang Zheng},
         year = 2015,
 howpublished = {Tensor Flow Technical Report},
          url = {https:tensorflow.org},
     abstract = {TensorFlow is an interface for expressing machine learning algorithms, and an implementation for executing such algorithms. A computation expressed using TensorFlow can be executed with little or no change on a wide variety of heterogeneous systems, ranging from mobile devices such as phones and tablets up to large-scale distributed systems of hundreds of machines and thousands of computational devices such as GPU cards. The system is flexible and can be used to express a wide variety of algorithms, including training and inference algorithms for deep neural network models, and it has been used for conducting research and for deploying machine learning systems into production across more than a dozen areas of computer science and other fields, including speech recognition, computer vision, robotics, information retrieval, natural language processing, geographic information extraction, and computational drug discovery. This paper describes the TensorFlow interface and an implementation of that interface that we have built at Google. The TensorFlow API and a reference implementation were released as an open-source package under the Apache 2.0 license in November, 2015 and are available at www.tensorflow.org.}
}
@article{AbbottandNelsonNATURE-00,
       author = {Abbott, L. F. and Nelson, S. B.},
        title = {Synaptic plasticity: taming the beast},
      journal = {Nature Neuroscience},
         year = {2000},
       volume = {3},
        pages = {1178-1183},
}
@article{AbbottandRegehrNATURE-04,
       author = {Abbott, L. F. and Regehr, W. G.},
        title = {Synaptic computation},
      journal = {Nature},
         year = {2004},
       volume = {431},
       number = {7010},
        pages = {796-803},
     abstract = {Neurons are often considered to be the computational engines of the brain, with synapses acting solely as conveyers of information. But the diverse types of synaptic plasticity and the range of timescales over which they operate suggest that synapses have a more active role in information processing. Long-term changes in the transmission properties of synapses provide a physiological substrate for learning and memory, whereas short-term changes support a variety of computations. By expressing several forms of synaptic plasticity, a single neuron can convey an array of different signals to the neural circuit in which it operates.}
}
@article{AbbottetalNATURE-NEUROSCIENCE-16,
       author = {Abbott, Larry F. and DePasquale, Brian and Memmesheimer, Raoul-Martin},
        title = {Building functional networks of spiking model neurons},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {19},
        issue = {3},
         year = {2016},
        pages = {350-355},
     abstract = {Most of the networks used by computer scientists and many of those studied by modelers in neuroscience represent unit activities as continuous variables. Neurons, however, communicate primarily through discontinuous spiking. We review methods for transferring our ability to construct interesting networks that perform relevant tasks from the artificial continuous domain to more realistic spiking network models. These methods raise a number of issues that warrant further theoretical and experimental study.},
}
@inproceedings{AbdallaandPickettICAS-11,
       author = {Abdalla, H. and Pickett, M.D.},
    booktitle = {IEEE International Symposium on Circuits and Systems},
        title = {{SPICE} modeling of memristors},
         year = 2011,
        pages = {1832-1835},
     abstract = {In this paper we present a {SPICE} model for the titanium dioxide memristor device from its modeling equations as described in and compare the {SPICE} simulations to the experimental data.},
}
@book{Abeles1991,
        title = {Corticonics: Neural Circuits of the Cerebral Cortex},
       author = {Moshe Abeles},
         year = 1991,
    publisher = {Cambridge University Press},
}
@article{AbelesSCIENCE-04,
       author = {Abeles, Moshe},
        title = {Neuroscience. {T}ime is precious},
      journal = {Science},
         year = 2004,
       volume = 304,
        issue = 5670,
        pages = {523-524},
}
@book{AbelsonandSussmanSICP-96,
       author = {Harold Abelson and Gerald Jay Sussman},
        title = {Structure and Interpretation of Computer Programs},
    publisher = {Second edition, MIT Press},
      address = {Cambridge, MA},
         year = 1996,
}
@techreport{AbiteboulelalTR-87,
       author = {Abiteboul, Serge and Paris Kanellakis and Gosta Grahne},
        title = {On the Representation and Querying of Sets of Possible Worlds},
        issue = {CS-87-03},
  institution = {Brown University Department of Computer Science},
         year = 1987,
}
@misc{AbolafiaAIT-21,
        title = {Primer to Algorithmic Information Theory},
       author = {Daniel Abolafia},
 howpublished = {Circulating Draft},
         year = {2021},
     abstract = {This writeup captures what I believe are the core insights I gained from my studies in algorithmic information theory (AIT). From what I can tell, this presentation is idiosyncratic and possibly heterodox. I leave out plenty of material that would be considered core in a textbook, and I focus on conceptual payloads that I do not see mentioned. The purpose of this document is primarly a reference for my future self, should I ever need a refresher, and secondarily as an artifact signifying my understanding of this topic. However, it may be helpful to others a primer to AIT, though I did not optimize the writing for comprehension by a general audience and this writeup is fairly technical some places. Being comfortable with calculus (infinite sums and limits), probability theory (probability measures and measure spaces), and theory of computation (Turing machines, computability, countable sets), will help a lot. Familiarity with Shannon information theory will also help.}
}
@article{AbolafiaetalCoRR-18,
       author = {Daniel A. Abolafia and Mohammad Norouzi and Quoc V. Le},
        title = {Neural Program Synthesis with Priority Queue Training},
      journal = {CoRR},
       volume = {arXiv:1801.03526},
         year = {2018},
     abstract = {We consider the task of program synthesis in the presence of a reward function over the output of programs, where the goal is to find programs with maximal rewards. We employ an iterative optimization scheme, where we train an RNN on a dataset of K best programs from a priority queue of the generated programs so far. Then, we synthesize new programs and add them to the priority queue by sampling from the RNN. We benchmark our algorithm, called priority queue training (or PQT), against genetic algorithm and reinforcement learning baselines on a simple but expressive Turing complete programming language called BF. Our experimental results show that our simple PQT algorithm significantly outperforms the baselines. By adding a program length penalty to the reward function, we are able to synthesize short, human readable programs.}
}
@article{AbolafiaetalCoRR-20,
        title = {Towards Modular Algorithm Induction},
       author = {Daniel A. Abolafia and Rishabh Singh and Manzil Zaheer and Charles Sutton},
      journal = {CoRR},
       volume = {arXiv:2003.04227},
         year = {2020},
     abstract = {We present a modular neural network architecture Main that learns algorithms given a set of input-output examples. Main consists of a neural controller that interacts with a variable-length input tape and learns to compose modules together with their corresponding argument choices. Unlike previous approaches, Main uses a general domain-agnostic mechanism for selection of modules and their arguments. It uses a general input tape layout together with a parallel history tape to indicate most recently used locations. Finally, it uses a memoryless controller with a length-invariant self-attention based input tape encoding to allow for random access to tape locations. The Main architecture is trained end-to-end using reinforcement learning from a set of input-output examples. We evaluate Main on five algorithmic tasks and show that it can learn policies that generalizes perfectly to inputs of much longer lengths than the ones used for training.}
}
@article{AbramsetalEJN-13,
       author = {Abrams, Daniel A and Ryali, Srikanth and Chen, Tianwen and Chordia, Parag and Khouzam, Amirah and Levitin, Daniel J and Menon, Vinod},
        title = {Inter-subject synchronization of brain responses during natural music listening},
      journal = {European Journal of Neuroscience},
       volume = 37,
        issue = 9,
         year = 2013,
        pages = {1458-1469},
     abstract = {Music is a cultural universal and a rich part of the human experience. However, little is known about common brain systems that support the processing and integration of extended, naturalistic 'real-world' music stimuli. We examined this question by presenting extended excerpts of symphonic music, and two pseudomusical stimuli in which the temporal and spectral structure of the Natural Music condition were disrupted, to non-musician participants undergoing functional brain imaging and analysing synchronized spatiotemporal activity patterns between listeners. We found that music synchronizes brain responses across listeners in bilateral auditory midbrain and thalamus, primary auditory and auditory association cortex, right-lateralized structures in frontal and parietal cortex, and motor planning regions of the brain. These effects were greater for natural music compared to the pseudo-musical control conditions. Remarkably, inter-subject synchronization in the inferior colliculus and medial geniculate nucleus was also greater for the natural music condition, indicating that synchronization at these early stages of auditory processing is not simply driven by spectro-temporal features of the stimulus. Increased synchronization during music listening was also evident in a right-hemisphere fronto-parietal attention network and bilateral cortical regions involved in motor planning. While these brain structures have previously been implicated in various aspects of musical processing, our results are the first to show that these regions track structural elements of a musical stimulus over extended time periods lasting minutes. Our results show that a hierarchical distributed network is synchronized between individuals during the processing of extended musical sequences, and provide new insight into the temporal integration of complex and biologically salient auditory sequences.},
}
@article{AbramsonetalCoRR-20,
        title = {Imitating Interactive Intelligence},
       author = {Josh Abramson and Arun Ahuja and Arthur Brussee and Federico Carnevale and Mary Cassin and Stephen Clark and Andrew Dudzik and Petko Georgiev and Aurelia Guy and Tim Harley and Felix Hill and Alden Hung and Zachary Kenton and Jessica Landon and Timothy Lillicrap and Kory Mathewson and Alistair Muldal and Adam Santoro and Nikolay Savinov and Vikrant Varma and Greg Wayne and Nathaniel Wong and Chen Yan and Rui Zhu},
      journal = {CoRR},
       volume = {arXiv:2012.05672},
         year = {2020},
     abstract = {A common vision from science fiction is that robots will one day inhabit our physical spaces, sense the world as we do, assist our physical labours, and communicate with us through natural language. Here we study how to design artificial agents that can interact naturally with humans using the simplification of a virtual environment. This setting nevertheless integrates a number of the central challenges of artificial intelligence (AI) research: complex visual perception and goal-directed physical control, grounded language comprehension and production, and multi-agent social interaction. To build agents that can robustly interact with humans, we would ideally train them while they interact with humans. However, this is presently impractical. Therefore, we approximate the role of the human with another learned agent, and use ideas from inverse reinforcement learning to reduce the disparities between human-human and agent-agent interactive behaviour. Rigorously evaluating our agents poses a great challenge, so we develop a variety of behavioural tests, including evaluation by humans who watch videos of agents or interact directly with them. These evaluations convincingly demonstrate that interactive training and auxiliary losses improve agent behaviour beyond what is achieved by supervised learning of actions alone. Further, we demonstrate that agent capabilities generalise beyond literal experiences in the dataset. Finally, we train evaluation models whose ratings of agents agree well with human judgement, thus permitting the evaluation of new agent models without additional effort. Taken together, our results in this virtual environment provide evidence that large-scale human behavioural imitation is a promising tool to create intelligent, interactive agents, and the challenge of reliably evaluating such agents is possible to surmount.}
}
@article{AbreuetalCoRR-19,
       author = {Jader Abreu and Luis Fred and David Mac{\^{e}}do and Cleber Zanchettin},
        title = {Hierarchical Attentional Hybrid Neural Networks for Document Classification},
      journal = {CoRR},
       volume = {arXiv:1901.06610},
         year = {2019},
     abstract = {Document classification is a challenging task with important applications. The deep learning approaches to the problem have gained much attention recently. Despite the progress, the proposed models do not incorporate the knowledge of the document structure in the architecture efficiently and not take into account the contexting importance of words and sentences. In this paper, we propose a new approach based on a combination of convolutional neural networks, gated recurrent units, and attention mechanisms for document classification tasks. The main contribution of this work is the use of convolution layers to extract more meaningful, generalizable and abstract features by the hierarchical representation. The proposed method in this paper improves the results of the current attention-based approaches for document classification.}
}
@article{AcharyaandShuklaJNSBM-12,
       author = {Acharya, Sourya and Shukla, Samarth},
        title = {Mirror neurons: Enigma of the metaphysical modular brain},
      journal = {Journal of Natural Science, Biology and Medicine},
         year = {2012},
    publisher = {Medknow Publications \& Media Pvt Ltd},
       volume = {3},
        issue = {2},
        pages = {118-124},
     abstract = {Mirror neurons are one of the most important discoveries in the last decade of neuroscience. These are a variety of visuospatial neurons which indicate fundamentally about human social interaction. Essentially, mirror neurons respond to actions that we observe in others. The interesting part is that mirror neurons fire in the same way when we actually recreate that action ourselves. Apart from imitation, they are responsible for myriad of other sophisticated human behavior and thought processes. Defects in the mirror neuron system are being linked to disorders like autism. This review is a brief introduction to the neurons that shaped our civilization.}
}
@article{AdamaszeketalCEREBELLUM-17,
        title = {Consensus paper: cerebellum and emotion},
       author = {Adamaszek, Michael and D'Agata, Federico and Ferrucci, Roberta and Habas, Christophe and Keulen, S and Kirkby, KC and Leggio, Maria and Mari{\"e}n, Peter and Molinari, Marco and Moulton, E and others},
      journal = {The Cerebellum},
       volume = {16},
       number = {2},
        pages = {552-576},
         year = {2017},
    publisher = {Springer}
}
@inproceedings{Adamic1999,
       author = {Lada A. Adamic},
        title = {{The Small World Web}},
    booktitle = {Research and Advanced Technology for Digital Libraries},
    publisher = {Springer-Verlag},
       editor = {S. Abiteboul and A.-M. Vercoustre},
        pages = {443-452},
         year = 1999,
}
@book{Adams1995,
      author = {Douglas Adams},
       title = {Hitchhiker's Guide to the Galaxy},
   publisher = {Ballantine Books},
     address = {New York, NY},
        year = 1995,
}
@article{AdamsandGaythercoleIJLCD-00,
       author = {Anne-Marie Adams, Susan. E. Gathercole},
        title = {Limitations in working memory: implications for language development},
      journal = {International Journal of Language \& Communication Disorders},
       volume = {35},
       number = {1},
        pages = {95-116},
         year = {2000},
    publisher = {Taylor \& Francis},
     abstract = {In this study, the proposal that individual differences in spoken language acquisition may be due to limitations in short-term memory abilities was investigated within a working memory framework. The relationship between speech production skills and working memory abilities was examined in two groups of 4-yearold children, matched for non-verbal ability but who had either relatively good or poor non-word repetition skills. Children with better non-word repetition skills produced speech that comprised a wider repertoire of words, on average longer utterances and a greater range of syntactic constructions than did children with relatively poor non-word repetition skills. The significant association found between these indices of language development and verbal short-term memory span assessed with non-spoken recall, suggested that this relationship was not merely due to the common output requirements of the language and memory tasks. Inconsistent associations between language performance and two tasks of visuo-spatial short-term memory precluded firm conclusions being drawn regarding the specificity of the relationship to the phonological domain. Cognitive mechanisms that may underlie the association between spoken language development and working memory skills are discussed.}
}
@article{AdamsandHortonNEUROSCIENTIST-09,
       author = {Adams, D. L. and Horton, J. C.},
        title = {Ocular dominance columns: enigmas and challenges},
      journal = {Neuroscientist},
         year = 2009,
       volume = 15,
        issue = 1,
        pages = {62-77},
     abstract = {In some mammalian species, geniculocortical afferents serving each eye are segregated in layer 4C of striate cortex into stripes called ocular dominance columns. Having described the complete pattern of ocular dominance columns in the human brain, the authors enumerate here the principal enigmas that confront future investigators. Probably the overarching challenge is to explain the function, if any, of ocular dominance columns and why they are present in some species and not others. A satisfactory solution must account for the enormous natural variation, even within the same species, among individuals in column expression, pattern, periodicity, and alignment with other components of the functional architecture. Another major priority is to explain the development of ocular dominance columns. It has been established clearly that they form without visual experience, but the innate signals that guide their segregation and maturation are unknown. Experiments addressing the role of spontaneous retinal activity have yielded contradictory data. These studies must be reconciled, to pave the way for new insights into how columnar structure is generated in the cerebral cortex.}
}
@article{AdamsandTsienARP-93,
       author = {Adams, S R and Tsien, R Y},
        title = {Controlling Cell Chemistry with Caged Compounds},
      journal = {Annual Review of Physiology},
       volume = 55,
        issue = 1,
         year = 1993,
        pages = {755-784},
}
@article{AdamsetalMS-88,
       author = {Adams, J. and Balas, E. and Zawack, D.},
        title = {The Shifting Bottleneck Procedure for Job Shop Scheduling},
      journal = {Management Science},
       volume = 34,
        issue = 3,
         year = 1988,
        pages = {391-401},
}
@inproceedings{AdeletalICML-18,
        title = {Discovering Interpretable Representations for Both Deep Generative and Discriminative Models},
       author = {Adel, Tameem and Ghahramani, Zoubin and Weller, Adrian},
    booktitle = {Proceedings of the 35th International Conference on Machine Learning},
       editor = {Dy, Jennifer and Krause, Andreas},
    publisher = {PMLR},
         year = {2018},
        pages = {50-59},
     abstract = {Interpretability of representations in both deep generative and discriminative models is highly desirable. Current methods jointly optimize an objective combining accuracy and interpretability. However, this may reduce accuracy, and is not applicable to already trained models. We propose two interpretability frameworks. First, we provide an interpretable lens for an existing model. We use a generative model which takes as input the representation in an existing (generative or discriminative) model, weakly supervised by limited side information. Applying a flexible and invertible transformation to the input leads to an interpretable representation with no loss in accuracy. We extend the approach using an active learning strategy to choose the most useful side information to obtain, allowing a human to guide what "interpretable" means. Our second framework relies on joint optimization for a representation which is both maximally informative about the side information and maximally compressive about the non-interpretable data factors. This leads to a novel perspective on the relationship between compression and regularization. We also propose a new interpretability evaluation metric based on our framework. Empirically, we achieve state-of-the-art results on three datasets using the two proposed algorithms.}
}
@incollection{AdelsonandBergenCMVP-91,
        title = {The Plenoptic Function and the Elements of Early Vision},
       author = {Edward H. Adelson and James R. Bergen},
    booktitle = {Computational Models of Visual Processing},
       editor = {M. Landy and J. A. Movshon},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1991,
        pages = {3-20},
     abstract = {What are the elements of early vision? This question might be taken to mean, "What are the fundamental atoms of vision?" and might be variously answered in terms of such candidate structures as edges, peaks, corners, and so on. In this chapter we adopt a rather different point of view and ask the question, What are the fundamental substances of vision? This distinction is important because we wish to focus on the first steps in extraction of visual information. At this level it is premature to talk about discrete objects, even such simple ones as edges and corners.}
}
@article{AdelsonandBergenJOSA-85,
       author = {Edward H. Adelson and James R. Bergen},
        title = {Spatiotemporal energy models for the perception of motion},
      journal = {Journal of the Optical Society of America A},
        issue = 2,
    publisher = {Optical Society of America},
       volume = 2,
        pages = {284-299},
         year = 1985,
     abstract = {A motion sequence may be represented as a single pattern in x-y-t space; a velocity of motion corresponds to a three-dimensional orientation in this space. Motion information can be extracted by a system that responds to the oriented spatiotemporal energy. We discuss a class of models for human motion mechanisms in which the first stage consists of linear filters that are oriented in space-time and tuned in spatial frequency. The outputs of quadrature pairs of such filters are squared and summed to give a measure of motion energy. These responses are then fed into an opponent stage. Energy models can  be built from elements that are consistent with known physiology  and psychophysics, and they permit a qualitative understanding of a variety of motion phenomena.},
}
@article{AdelsonandMovshonNATURE-82,
       author = {Adelson, E. and Movshon, J.},
        title = {Phenomenal coherence of moving visual patterns},
      journal = {Nature},
       volume = 300,
        pages = {523-525},
         year = 1982,
}
@article{AdlemanSCIENCE-94,
        title = {Molecular Computation of Solutions to Combinatorial Problems},
       author = {Leonard M. Adleman},
      journal = {Science},
       volume = 266,
        issue = 5187,
         year = 1994,
        pages = {1021-1024},
}
@inproceedings{AdlerandMitzenmacher,
       author = {M. Adler and M. Mitzenmacher},
        title = {Towards compressing {W}eb graphs},
    booktitle = {Proceedings of the 2001 Data Compression Conference},
        pages = {203-212},
         year = 2001,
          url = {ftp://ftp.cs.umass.edu/pub/techrept/techreport/2000},
}
@article{AfeketalSCIENCE-11,
       author = {Afek, Yehuda and Alon, Noga and Barad, Omer and Hornstein, Eran and Barkai, Naama and Bar-Joseph, Ziv},
        title = {A Biological Solution to a Fundamental Distributed Computing Problem},
      journal = {Science},
       volume = 331,
        issue = 6014,
        pages = {183-185},
         year = 2011,
     abstract = {Computational and biological systems are often distributed so that processors (cells) jointly solve a task, without any of them receiving all inputs or observing all outputs. Maximal independent set (MIS) selection is a fundamental distributed computing procedure that seeks to elect a set of local leaders in a network. A variant of this problem is solved during the development of the fly's nervous system, when sensory organ precursor (SOP) cells are chosen   By studying SOP selection, we derived a fast algorithm for MIS  selection that combines two attractive features. First, processors do  not need to know their degree; second, it has an optimal message  complexity while only using one-bit messages. Our findings suggest  that simple and efficient algorithms can be developed on the basis of  biologically derived insights.},
}
@article {AflaloetalJoN-06,
       author = {Aflalo, Tyson N. and Graziano, Michael S. A.},
        title = {Possible Origins of the Complex Topographic Organization of Motor Cortex: Reduction of a Multidimensional Space onto a Two-Dimensional Array},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {26},
       number = {23},
         year = {2006},
        pages = {6288--6297},
     abstract = {We propose that some of the features of the topographic organization in motor cortex emerge from a competition among several conflicting mapping requisites. These competing requisites include a somatotopic map of the body, a map of hand location in space, and a partitioning of cortex into regions that emphasize different complex, ethologically relevant movements. No one type of map fully explains the topography; instead, all three influences (and perhaps others untested here) interact to form the topography. A standard algorithm (Kohonen network) was used to generate an artificial motor cortex array that optimized local continuity for these conflicting mapping requisites. The resultant hybrid map contained many features seen in actual motor cortex, including the following: a rough, overlapping somatotopy; a posterior strip in which simpler movements were represented and more somatotopic segregation was observed, and an anterior strip in which more complex, multisegmental movements were represented and the somatotopy was less segregated; a clustering of different complex, multisegmental movements into specific subregions of cortex that resembled the arrangement of subregions found in the monkey; three hand representations arranged on the cortex in a manner similar to the primary motor, dorsal premotor, and ventral premotor hand areas in the monkey; and maps of hand location that approximately matched the maps observed in the monkey.},
}
@inproceedings{AgarwalandTriggsECCV-06,
       author = {Ankur Agarwal and Bill Triggs},
        title = {Hyperfeatures - multilevel local coding for visual recognition},
    booktitle = {Proceedings of the 9th European Conference on Computer Vision},
    publisher = {Springer},
        pages = {30-43},
         year = 2006,
}
@article{AgarwaletalSCIENCE-14,
       author = {Agarwal, Gautam and Stevenson, Ian H. and Ber\`{e}nyi, Antal and Mizuseki, Kenji and Buzs\`{a}ki, Gy\"{o}rgy and Sommer, Friedrich T.},
        title = {Spatially Distributed Local Fields in the Hippocampus Encode Rat Position},
      journal = {Science},
       volume = 344,
        issue = 6184,
         year = 2014,
        pages = {626-630},
     abstract = {Although neuronal spikes can be readily detected from extracellular recordings, synaptic and subthreshold activity remains undifferentiated within the local field potential (LFP). In the hippocampus, neurons discharge selectively when the rat is at certain locations, while LFPs at single anatomical sites exhibit no such place-tuning. Nonetheless, because the representation of position is sparse and distributed, we hypothesized that spatial information can be recovered from multiple-site LFP recordings. Using high-density sampling of LFP and computational methods, we show that the spatiotemporal structure of the theta rhythm can encode position as robustly as neuronal spiking populations. Because our approach exploits the rhythmicity and sparse structure of neural activity, features found in many brain regions, it is useful as a general tool for discovering distributed LFP codes.},
}
@inproceedings{AggarwalandYuPODS-98,
       author = {Aggarwal, Charu C. and Yu, Philip S.},
        title = {A new framework for itemset generation},
    booktitle = {Proceedings of the seventeenth ACM SIGACT-SIGMOD-SIGART symposium on Principles of database systems},
     location = {Seattle, Washington, United States},
    publisher = {ACM},
         year = 1998,
        pages = {18-24},
}
@article{AgietalNEUROGENETICS-14,
       author = {Agi, Egemen and Langen, Marion and Altschuler, Steven J. and Wu, Lani F. and Zimmermann, Timo and Hiesinger, Peter Robin},
        title = {The Evolution and Development of Neural Superposition},
      journal = {Journal of Neurogenetics},
       volume = 28,
        issue = 0,
         year = 2014,
        pages = {216-232},
     abstract = {Visual systems have a rich history as model systems for the discovery and understanding of basic principles underlying neuronal connectivity. The compound eyes of insects consist of up to thousands of small unit eyes that are connected by photoreceptor axons to set up a visual map in the brain. The photoreceptor axon terminals thereby represent neighboring points seen in the environment in neighboring synaptic units in the brain. Neural superposition is a special case of such a wiring principle, where photoreceptors from different unit eyes that receive the same input converge upon the same synaptic units in the brain. This wiring principle is remarkable, because each photoreceptor in a single unit eye receives different input and each individual axon, among thousands others in the brain, must be sorted together with those few axons that have the same input. Key aspects of neural superposition have been described as early as 1907. Since then neuroscientists, evolutionary and developmental biologists have been fascinated by how such a complicated wiring principle could evolve, how it is genetically encoded, and how it is developmentally realized. In this review article, we will discuss current ideas about the evolutionary origin and developmental program of neural superposition. Our goal is to identify in what way the special case of neural superposition can help us answer more general questions about the evolution and development of genetically hard-wired synaptic connectivity in the brain.},
}
@techreport{AgoginoandRamamurthiTR-88,
       author = {Agogino, A. M. and Ramamurthi, K.},
        title = {Real-Time Influence Diagrams for Monitoring and Controlling Mechanical Systems},
  institution = {Department of Mechanical Engineering, University of California, Berkeley},
         year = 1988,
}
@incollection{AgoginoandRege87,
       author = {Alice M. Agogino and Ashutosh Rege},
        title = {IDES: Influence Diagram Based Expert System},
    booktitle = {Mathematical Modelling},
    publisher = {Pergamon Press},
         year = 1987,
        pages = {227-233},
}
@article{AgoginoetalMSSP-88,
       author = {Agogino, Alice M. and Srinivas, Sampath and Schneider, Kenneth M.},
        title = {Multiple Sensor Expert System for Diagnostic Reasoning, Monitoring and Control of Mechanical Systems},
      journal = {Mechanical Systems and Signal Processing},
       volume = 2,
        pages = {165-185},
         year = 1988,
}
@inproceedings{AgostaSPIE-91,
       author = {J. M. Agosta},
        title = {An example of a {B}ayes network of relations among visual features},
    booktitle = {Proceedings of the {SPIE} Conference on Advances in Intelligent Robotic Systems},
 organization = {SPIE},
         year = 1991,
}
@inproceedings{AgostaUAI-91,
       author = {J. M. Agosta},
        title = {Conditional Inter-Causally Independent Node Distributions, a Property of Noisy-{OR} Models},
    booktitle = {Seventh Conference on Uncertainty in Artificial Intelligence},
        pages = {9-16},
         year = 1991,
}
@techreport{AgrePhD,
       author = {Agre, Philip E.},
        title = {The Dynamics of Everyday Life},
        issue = {AI-TR 1085},
  institution = {{MIT} {AI} Laboratory},
         year = 1988,
}
@inproceedings{AgreandChapmanAAAI-87,
       author = {Agre, Philip E. and Chapman, David},
        title = {Pengi: An Implementation of a Theory of Activity},
    booktitle = {Proceedings AAAI-87},
     location = {Seattle, Washington},
 organization = {AAAI},
         year = 1987,
        pages = {268-272},
}
@article{AgueraetalNC-03,
        title = {Computation in a single neuron: Hodgkin and Huxley revisited},
       author = {Ag\"{u}era y Arcas, B. and Fairhall, A.L. and Bialek, W.},
      journal = {Neural Computation},
       volume = 15,
        issue = 8,
        pages = {1715-49},
         year = 2003,
     abstract = {A spiking neuron "computes" by transforming a complex dynamical input into a train of action potentials, or spikes. The computation performed by the neuron can be formulated as dimensional reduction, or feature detection, followed by a nonlinear decision function over the low-dimensional space. Generalizations of the reverse correlation technique with white noise input provide a numerical strategy for extracting the relevant low-dimensional features from experimental data, and information theory can be used to evaluate the quality of the low-dimensional approximation. We apply these methods to analyze the simplest biophysically realistic model neuron, the Hodgkin-Huxley (HH) model, using this system to illustrate the general methodological issues. We focus on the features in the stimulus that trigger a spike, explicitly eliminating the effects of interactions between spikes. One can approximate this triggering "feature space" as a two-dimensional linear subspace in the high-dimensional space of input histories, capturing in this way a substantial fraction of the mutual information between inputs and spike time. We find that an even better approximation, however, is to describe the relevant subspace as two dimensional but curved; in this way, we can capture 90\% of the mutual information even at high time resolution. Our analysis provides a new understanding of the computational properties of the HH model. While it is common to approximate neural behavior as "integrate and fire," the HH model is not an integrator nor is it well described by a single threshold.}
}
@inproceedings{AguireetalIWANN-03,
       author = {Carlos Aguirre and Doris Campos and Pedro Pascual and Eduardo Serrano},
        title = {Effects of Different Connectivity Patterns in a Model of Cortical Circuits},
    booktitle = {Proceedings of the IWANN},
     crossref = {conf/iwann/2003-1},
         year = 2003,
        pages = {78-85},
          url = {AguireetalIWANN-03.pdf},
}
@article{AhmedandMandicPRE-11,
        title = {Multivariate multiscale entropy: A tool for complexity analysis of multichannel data},
       author = {Ahmed, Mosabber Uddin and Mandic, Danilo P},
      journal = {Physical Review E},
    publisher = {APS},
       volume = {84},
       number = {6},
        pages = {061918},
         year = {2011},
     abstract = {This work generalizes the recently introduced univariate multiscale entropy (MSE) analysis to the multivariate case. This is achieved by introducing multivariate sample entropy (MSampEn) in a rigorous way, in order to account for both within- and cross-channel dependencies in multiple data channels, and by evaluating it over multiple temporal scales. The multivariate MSE (MMSE) method is shown to provide an assessment of the underlying dynamical richness of multichannel observations, and more degrees of freedom in the analysis than standard MSE. The benefits of the proposed approach are illustrated by simulations on complexity analysis of multivariate stochastic processes and on real-world multichannel physiological and environmental data.}
}
@book{AhoHopcroftandUllman1974,
       author = {Aho, Alfred V. and Hopcroft, John E. and Ullman, Jeffrey D.},
        title = {The Design and Analysis of Computer Algorithms},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1974,
}
@book{AhoHopcroftandUllman1983,
       author = {Aho, Alfred V. and Hopcroft, John E. and Ullman, Jeffrey D.},
        title = {Data Structures and Algorithms},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1983,
}
@book{Ahoetal1987awk,
        title = {The {AWK} programming language},
       author = {Aho, Alfred V and Kernighan, Brian W and Weinberger, Peter J},
         year = {1987},
    publisher = {Addison-Wesley Longman Publishing Co., Inc.}
}
@article{AhrensetalNATURE-12,
       author = {Ahrens, Misha B. and Li, Jennifer M. and Orger, Michael B. and Robson, Drew N. and Schier, Alexander F. and Engert, Florian and Portugues, Ruben},
        title = {Brain-wide neuronal dynamics during motor adaptation in zebrafish},
      journal = {Nature},
         year = {2012},
    publisher = {Nature Publishing Group},
       volume = {485},
        issue = {7399},
        pages = {471-477},
     abstract = {A fundamental question in neuroscience is how entire neural circuits generate behaviour and adapt it to changes in sensory feedback. Here we use two-photon calcium imaging to record the activity of large populations of neurons at the cellular level, throughout the brain of larval zebrafish expressing a genetically encoded calcium sensor, while the paralysed animals interact fictively with a virtual environment and rapidly adapt their motor output to changes in visual feedback. We decompose the network dynamics involved in adaptive locomotion into four types of neuronal response properties, and provide anatomical maps of the corresponding sites. A subset of these signals occurred during behavioural adjustments and are candidates for the functional elements that drive motor learning. Lesions to the inferior olive indicate a specific functional role for olivocerebellar circuitry in adaptive locomotion. This study enables the analysis of brain-wide dynamics at single-cell resolution during behaviour.}
}
@article{AhrensetalNATURE-13,
        title = {Whole-brain functional imaging at cellular resolution using light-sheet microscopy},
       author = {Ahrens, Misha B and Orger, Michael B and Robson, Drew N and Li, Jennifer M and Keller, Philipp J},
      journal = {Nature Methods},
       volume = 10,
        issue = 5,
        pages = {413-420},
         year = 2013,
    publisher = {Nature Publishing Group}
}
@inproceedings{AielloetalSTOC-00,
       author = {W. Aiello and F. Chung and L. Lu},
        title = {A random graph model for massive graphs},
    booktitle = {Proceedings of the 32nd Annual ACM Symposium on Theory of Computing},
        pages = {171-180},
         year = 2000,
}
@article{AimoneCACM-19,
       author = {James B. Aimone},
        title = {Neural Algorithms and Computing Beyond Moore's Law},
      journal = {Communications of the {ACM}},
       volume = {62},
       number = {4},
         year = 2019,
        pages = {832-843},
          url = {https://cacm.acm.org/magazines/2019/4/235577-neural-algorithms-and-computing-beyond-moores-law/fulltext},
}
@article{AimoneetalNEURON-09,
       author = {Aimone, J. B. and Wiles, J. and Gage, F. H.},
        title = {Computational influence of adult neurogenesis on memory encoding},
      journal = {Neuron},
       volume = {61},
       number = {2},
         year = {2009},
        pages = {187-202},
     abstract = {Adult neurogenesis in the hippocampus leads to the incorporation of thousands of new granule cells into the dentate gyrus every month, but its function remains unclear. Here, we present computational evidence that indicates that adult neurogenesis may make three separate but related contributions to memory formation. First, immature neurons introduce a degree of similarity to memories learned at the same time, a process we refer to as pattern integration. Second, the extended maturation and change in excitability of these neurons make this added similarity a time-dependent effect, supporting the possibility that temporal information is included in new hippocampal memories. Finally, our model suggests that the experience-dependent addition of neurons results in a dentate gyrus network well suited for encoding new memories in familiar contexts while treating novel contexts differently. Taken together, these results indicate that new granule cells may affect hippocampal function in several unique and previously unpredicted ways.}
}
@article{AimoneetalNEURON-11,
       author = {James B. Aimone and Wei Deng and Fred H. Gage},
        title = {Resolving New Memories: A Critical Look at the Dentate Gyrus, Adult Neurogenesis, and Pattern Separation},
      journal = {Neuron},
       volume = {70},
       number = {4},
         year = {2011},
        pages = {589-596},
     abstract = {Recently, investigation of new neurons in memory formation has focused on a specific function-pattern separation. However, it has been difficult to reconcile the form of separation tested in behavioral tasks with how it is conceptualized according to computational and electrophysiology perspectives. Here, we propose a memory resolution hypothesis that considers the unique information contributions of broadly tuned young neurons and highly specific mature neurons and describe how the fidelity of memories can relate to spatial and contextual discrimination.}
}
@article {AimonetalBIORXIV-15,
       author = {Aimon, Sophie and Katsuki, Takeo and Grosenick, Logan and Broxton, Michael and Deisseroth, Karl and Greenspan, Ralph J.},
        title = {Activity sources from fast large-scale brain recordings in adult Drosophila},
      journal = {bioRxiv},
          url = {http://biorxiv.org/content/2016/09/12/033803},
         year = {2015},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {We provide a method to record near-whole brain activity in behaving adult flies and extract signals from specific anatomical structures. We image pan-neuronal calcium or voltage sensors fluorescence at high speed with light field microscopy. We then apply computational methods to extract functional maps, and find that their characteristic shape match the anatomy of sub-neuropile regions and sometimes small population of neurons. Associated time series are also consistent with the literature.},
}
@article{AjiandMcElieceTIT-00,
       author = {Srinivas M. Aji and Robert J. McEliece},
        title = {The generalized distributive law},
      journal = {IEEE Transactions on Information Theory},
       volume = 46,
        issue = 2,
         year = 2000,
        pages = {325-343},
}
@article{AkalaletalLM-06,
       author = {Akalal, David-Benjamin G. and Wilson, Curtis F. and Zong, Lin and Tanaka, Nobuaki K. and Ito, Kei and Davis, Ronald L.},
        title = {Roles for Drosophila mushroom body neurons in olfactory learning and memory},
      journal = {Learning and Memory},
         year = {2006},
    publisher = {Cold Spring Harbor Laboratory Press},
       volume = {13},
        issue = {5},
        pages = {659-668},
     abstract = {Olfactory learning assays in Drosophila have revealed that distinct brain structures known as mushroom bodies (MBs) are critical for the associative learning and memory of olfactory stimuli. However, the precise roles of the different neurons comprising the MBs are still under debate. The confusion surrounding the roles of the different neurons may be due, in part, to the use of different odors as conditioned stimuli in previous studies. We investigated the requirements for the different MB neurons, specifically the {\^I}{\textpm}/{\^I}{\texttwosuperior} versus the {\^I}{\textthreesuperior} neurons, and whether olfactory learning is supported by different subsets of MB neurons irrespective of the odors used as conditioned stimuli. We expressed the rutabaga (rut)-encoded adenylyl cyclase in either the {\^I}{\textthreesuperior} or {\^I}{\textpm}/{\^I}{\texttwosuperior} neurons and examined the effects on restoring olfactory associative learning and memory of rut mutant flies. We also expressed a temperature-sensitive shibire (shi) transgene in these neuron sets and examined the effects of disrupting synaptic vesicle recycling on Drosophila olfactory learning. Our results indicate that although we did not detect odor-pair-specific learning using GAL4 drivers that primarily express in {\^I}{\textthreesuperior} neurons, expression of the transgenes in a subset of {\^I}{\textpm}/{\^I}{\texttwosuperior} neurons resulted in both odor-pair-specific rescue of the rut defect as well as odor-pair-specific disruption of learning using shi(ts1).},
}
@article{AkemannetalNATURE-METHODS-10,
       author = {Akemann, Walther and Mutoh, Hiroki and Perron, Amelie and Rossier, Jean and Knopfel, Thomas},
        title = {Imaging brain electric signals with genetically targeted voltage-sensitive fluorescent proteins},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = {7},
        issue = {8},
         year = {2010},
        pages = {643-649},
}
@article{AkerboometalNEUROSCIENCE-12,
       author = {Akerboom J and Chen TW and Wardill TJ and Tian L and Marvin JS and Mutlu S and Calder\'{o}n NC and Esposti F and Borghuis BG and Sun XR and Gordus A and Orger MB and Portugues R and Engert F and Macklin JJ and Filosa A and Aggarwal A and Kerr RA and Takagi R and Kracun S and Shigetomi E and Khakh BS and Baier H and Lagnado L and Wang SS and Bargmann CI and Kimmel BE and Jayaraman V and Svoboda K and Kim DS and Schreiter ER and Looger LL},
        title = {Optimization of a GCaMP calcium indicator for neural activity imaging},
      journal = {The Journal of Neuroscience},
       volume = 32,
        issue = 40,
         year = 2012,
        pages = {13819-13840},
     abstract = {Genetically encoded calcium indicators (GECIs) are powerful tools for systems neuroscience. Recent efforts in protein engineering have significantly increased the performance of GECIs. The state-of-the art single-wavelength GECI, GCaMP3, has been deployed in a number of model organisms and can reliably detect three or more action potentials in short bursts in several systems in vivo. Through protein structure determination, targeted mutagenesis, high-throughput screening, and a battery of in vitro assays, we have increased the dynamic range of GCaMP3 by severalfold, creating a family of "GCaMP5" sensors. We tested GCaMP5s in several systems: cultured neurons and astrocytes, mouse retina, and in vivo in Caenorhabditis chemosensory neurons, Drosophila larval neuromuscular junction and adult antennal lobe, zebrafish retina and tectum, and mouse visual cortex. Signal-to-noise ratio was improved by at least 2- to 3-fold. In the visual cortex, two GCaMP5 variants detected twice as many visual stimulus-responsive cells as GCaMP3. By combining in vivo imaging with electrophysiology we show that GCaMP5 fluorescence provides a more reliable measure of neuronal activity than its predecessor GCaMP3. GCaMP5 allows more sensitive detection of neural activity in vivo and may find widespread applications for cellular imaging in general},
}
@article{AlbertinetalNATURE-15,
       author = {Albertin, Caroline B. and Simakov, Oleg and Mitros, Therese and Wang, Z. Yan and Pungor, Judit R. and Edsinger-Gonzales, Eric and Brenner, Sydney and Ragsdale, Clifton W. and Rokhsar, Daniel S.},
        title = {The octopus genome and the evolution of cephalopod neural and morphological novelties},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = {524},
        issue = {7564},
         year = {2015},
        pages = {220-224},
     abstract = {Coleoid cephalopods (octopus, squid and cuttlefish) are active, resourceful predators with a rich behavioural repertoire. They have the largest nervous systems among the invertebrates and present other striking morphological innovations including camera-like eyes, prehensile arms, a highly derived early embryogenesis and a remarkably sophisticated adaptive colouration system. To investigate the molecular bases of cephalopod brain and body innovations, we sequenced the genome and multiple transcriptomes of the California two-spot octopus, Octopus bimaculoides. We found no evidence for hypothesized whole-genome duplications in the octopus lineage. The core developmental and neuronal gene repertoire of the octopus is broadly similar to that found across invertebrate bilaterians, except for massive expansions in two gene families previously thought to be uniquely enlarged in vertebrates: the protocadherins, which regulate neuronal develop- ment, and the C2H2 superfamily of zinc-finger transcription factors. Extensive messenger RNA editing generates transcript and protein diversity in genes involved in neural excitability, as previously described, as well as in genes participating in a broad range of other cellular functions. We identified hundreds of cephalopod-specific genes, many of which showed elevated expression levels in such specialized structures as the skin, the suckers and the nervous system. Finally, we found evidence for large-scale genomic rearrangements that are closely associated with transposable element expansions. Our analysis suggests that substantial expansion of a handful of gene families, along with extensive remodelling of genome linkage and repetitive content, played a critical role in the evolution of cephalopod morphological innovations, including their large and complex nervous systems.}
}
@book{Albertsetal2002,
        title = {Molecular Biology of the Cell. 4th edition},
       author = {Bruce Alberts and Alexander Johnson and Julian Lewis and Martin Raff and Keith Roberts and Peter Walter},
    publisher = {Garland Science},
      address = {New York},
         year = {2002},
          url = {http://www.ncbi.nlm.nih.gov/books/NBK26910/},
}

@article{AlbinetalNEUROSCIENCE-89,
       author = {Albin, R. L. and Young, A. B. and Penney, J. B.},
        title = {The functional anatomy of basal ganglia disorders},
      journal = {Trends in Neurosciences},
       volume = 12,
         year = 1989,
        pages = {366-375},
     abstract = {Basal ganglia disorders are a heterogeneous group of clinical syndromes with a common anatomic locus within the basal ganglia. To account for the variety of clinical manifestations associated with insults to various parts of the basal ganglia we propose a model in which specific types of basal ganglia disorders are associated with changes in the function of subpopulations of striatal projection neurons. This model is based on a synthesis of experimental animal and post-mortem human anatomic and neuro chemical data. Hyperkinetic disorders, which are characterized by an excess of abnormal movements, are postulated to result from the selective impairment of striatal neurons projecting to the lateral globus pallidus. Hypokinetic phenomenology seen in a wide variety of basal disorders, such as Parkinson's disease, are hypothesized to result from acomplex series of changes in the activity of striatal projection neuron subpopulations resulting in an increase in basal ganglia output. This model suggests that the activity of subpopulations of striatal projection neurons is differentially regulated by striatal afferents and that different striatal projection neuron subpopulations may mediate different aspects of motor control.},
}
@article{Albus75,
       author = {Albus, James S.},
        title = {A New Approach to Manipulator Control: The Cerebellar Model Articulation Controller (CMAC)},
      journal = {Journal of Dynamic Systems, Measurement, and Control},
       volume = 97,
         year = 1975,
        pages = {270-277},
}
@book{AlbusBYTE-81,
       author = {Albus, James S.},
        title = {Brains, Behavior, and Robotics},
    publisher = {BYTE Books, Subsidiary of McGraw Hill},
      address = {Petersborough, New Hampshire},
         year = 1981,
}
@article{AlbusMB-71,
       author = {Albus, James S.},
        title = {A Theory of Cerebellar Functions},
      journal = {Mathematical Biology},
       volume = 10,
         year = 1971,
        pages = {25-61},
}
@inproceedings{AlbusetalSME-84,
       author = {Albus, James S. and Barbera, Anthony J. and Fitzgerald, M. L.},
        title = {A Control System for an Automated Manufacturing Research Facility},
    booktitle = {Proceedings of Robots 8},
 organization = {Robotics International of SME},
     location = {Detroit, Michigan},
         year = 1984,
        pages = {13-28-13-34},
}
@article{AlcamiandHadyCoRR-19,
       author = {Pepe Alcami and Ahmed El Hady},
        title = {Axonal Computations},
      journal = {CoRR},
       volume = {arXiv:1905.02007},
         year = {2019},
     abstract = {Axons functionally link the somato-dendritic compartment to synaptic terminals. Structurally and functionally diverse, they accomplish a central role in determining the delays and reliability with which neuronal ensembles communicate. By combining their active and passive biophysical properties, they ensure a plethora of physiological computations. In this review, we revisit the biophysics of generation and propagation of electrical signals in the axon, their complex interplay, and their rich dynamics. We further place the computational abilities of axons in the context of intracellular and intercellular coupling. We discuss how, by means of sophisticated biophysical mechanisms, axons expand the repertoire of axonal computation, and thereby, of neural computation.},
}
@article{Alderson-DayandFernyhoughAPA-15,
       author = {Alderson-Day, Ben and Fernyhough, Charles},
        title = {Inner Speech: Development, Cognitive Functions, Phenomenology, and Neurobiology},
      journal = {Psychological bulletin},
    publisher = {American Psychological Association},
       volume = {141},
        issue = {5},
         year = {2015},
        pages = {931-965},
     abstract = {Inner speech-also known as covert speech or verbal thinking-has been implicated in theories of cognitive development, speech monitoring, executive function, and psychopathology. Despite a growing body of knowledge on its phenomenology, development, and function, approaches to the scientific study of inner speech have remained diffuse and largely unintegrated. This review examines prominent theoretical approaches to inner speech and methodological challenges in its study, before reviewing current evidence on inner speech in children and adults from both typical and atypical populations. We conclude by considering prospects for an integrated cognitive science of inner speech, and present a multicomponent model of the phenomenon informed by developmental, cognitive, and psycholinguistic considerations. Despite its variability among individuals and across the life span, inner speech appears to perform significant functions in human cognition, which in some cases reflect its developmental origins and its sharing of resources with other cognitive processes.},
}
@article{Alderson-DayetalNBR-15,
        title = {Hearing voices in the resting brain: A review of intrinsic functional connectivity research on auditory verbal hallucinations},
       author = {Ben Alderson-Day and Simon McCarthy-Jones and Charles Fernyhough},
      journal = {Neuroscience and Biobehavioral Reviews},
       volume = {55},
         year = {2015},
        pages = {78-87},
     abstract = {Resting state networks (RSNs) are thought to reflect the intrinsic functional connectivity of brain regions. Alterations to RSNs have been proposed to underpin various kinds of psychopathology, including the occurrence of auditory verbal hallucinations (AVH). This review outlines the main hypotheses linking AVH and the resting state, and assesses the evidence for alterations to intrinsic connectivity provided by studies of resting fMRI in AVH. The influence of hallucinations during data acquisition, medication confounds, and movement are also considered. Despite a large variety of analytic methods and designs being deployed, it is possible to conclude that resting connectivity in the left temporal lobe in general and left superior temporal gyrus in particular are disrupted in AVH. There is also preliminary evidence of atypical connectivity in the default mode network and its interaction with other RSNs. Recommendations for future research include the adoption of a common analysis protocol to allow for more overlapping datasets and replication of intrinsic functional connectivity alterations.}
}
@article{Alderson-DayetalSCAN-16,
       author = {Alderson-Day, B. and Weis, S. and McCarthy-Jones, S. and Moseley, P. and Smailes, D. and Fernyhough, C. },
        title = {{{T}he brain's conversation with itself: neural substrates of dialogic inner speech}},
      journal = {Social Cognitive Affective Neuroscience},
         year = {2016},
       volume = {11},
       number = {1},
        pages = {110-120},
     abstract = {Inner speech has been implicated in important aspects of normal and atypical cognition, including the development of auditory hallucinations. Studies to date have focused on covert speech elicited by simple word or sentence repetition, while ignoring richer and arguably more psychologically significant varieties of inner speech. This study compared neural activation for inner speech involving conversations ('dialogic inner speech') with single-speaker scenarios ('monologic inner speech'). Inner speech-related activation differences were then compared with activations relating to Theory-of-Mind (ToM) reasoning and visual perspective-taking in a conjunction design. Generation of dialogic (compared with monologic) scenarios was associated with a widespread bilateral network including left and right superior temporal gyri, precuneus, posterior cingulate and left inferior and medial frontal gyri. Activation associated with dialogic scenarios and ToM reasoning overlapped in areas of right posterior temporal cortex previously linked to mental state representation. Implications for understanding verbal cognition in typical and atypical populations are discussed.}
}
@inproceedings{AleliunasetalFOCS-78,
       author = {Aleliunas, Romas and Karp, Richard M. and Lipton, Richard J. and Lovasz, Laszlo and Rackoff, Charles},
        title = {Random Walks, Universal Traversal Sequences, and the Complexity of Maze Problems},
    booktitle = {Proceedings of the 20th Symposium on the Foundations of Computer Science},
         year = 1979,
        pages = {218-223},
}
@article{AlemietalCoRR-16,
       author = {Alexander A. Alemi and Fran{\c{c}}ois Chollet and Geoffrey Irving and Christian Szegedy and Josef Urban},
        title = {DeepMath - Deep Sequence Models for Premise Selection},
      journal = {CoRR},
       volume = {arXiv:1606.04442},
         year = {2016},
     abstract = {We study the effectiveness of neural sequence models for premise selection in automated theorem proving, one of the main bottlenecks in the formalization of mathematics. We propose a two stage approach for this task that yields good results for the premise selection task on the Mizar corpus while avoiding the handengineered features of existing state-of-the-art models. To our knowledge, this is the first time deep learning has been applied to theorem proving on a large scale.},
}
@article{AlemietalCoRR-16b,
       author = {Alexander A. Alemi and Ian Fischer and Joshua V. Dillon and Kevin Murphy},
        title = {Deep Variational Information Bottleneck},
      journal = {CoRR},
       volume = {arXiv:1612.00410},
         year = {2016},
     abstract = {We present a variational approximation to the information bottleneck of Tishby et al. (1999). This variational approach allows us to parameterize the information bottleneck model using a neural network and leverage the reparameterization trick for efficient training. We call this method "Deep Variational Information Bottleneck", or Deep VIB. We show that models trained with the VIB objective outperform those that are trained with other forms of regularization, in terms of generalization performance and robustness to adversarial attack.}
}
@misc{Alessio_SetzuPhD-07,
        title = {A Framework for Financial Markets Modeling and Simulation},
       author = {Alessio Setzu},
 howpublished = {\url{http://www.diee.unica.it/DRIEI/tesi/19_setzu.pdf}},
         year = {2007},
}
@article{AlexDISCOVER-18,
        title = {How Did Human Language Evolve? Scientists Still Don't Know},
       author = {Bridget Alex},
      journal = {Discover},
         year = {2018},
      comment = {https://www.discovermagazine.com/planet-earth/how-did-human-language-evolve-scientists-still-dont-know},
}
@article{AlexanderandBrownNATURE-18,
       author = {Alexander, William H.and Brown, Joshua W.},
        title = {Frontal cortex function as derived from hierarchical predictive coding},
      journal = {Scientific Reports},
         year = {2018},
       volume = {8},
        issue = {1},
        pages = {3843},
     abstract = {The frontal lobes are essential for human volition and goal-directed behavior, yet their function remains unclear. While various models have highlighted working memory, reinforcement learning, and cognitive control as key functions, a single framework for interpreting the range of effects observed in prefrontal cortex has yet to emerge. Here we show that a simple computational motif based on predictive coding can be stacked hierarchically to learn and perform arbitrarily complex goal-directed behavior. The resulting Hierarchical Error Representation (HER) model simulates a wide array of findings from fMRI, ERP, single-units, and neuropsychological studies of both lateral and medial prefrontal cortex. By reconceptualizing lateral prefrontal activity as anticipating prediction errors, the HER model provides a novel unifying account of prefrontal cortex function with broad implications for understanding the frontal cortex across multiple levels of description, from the level of single neurons to behavior.},
}
@article{AlexanderandBrownSR-18,
       author = {Alexander, William H. and Brown, Joshua W.},
        title = {Frontal cortex function as derived from hierarchical predictive coding},
      journal = {Scientific Reports},
    publisher = {Nature Publishing Group},
         year = {2018},
       volume = {8},
        issue = {1},
        pages = {3843},
     abstract = {The frontal lobes are essential for human volition and goal-directed behavior, yet their function remains unclear. While various models have highlighted working memory, reinforcement learning, and cognitive control as key functions, a single framework for interpreting the range of effects observed in prefrontal cortex has yet to emerge. Here we show that a simple computational motif based on predictive coding can be stacked hierarchically to learn and perform arbitrarily complex goal-directed behavior. The resulting Hierarchical Error Representation (HER) model simulates a wide array of findings from fMRI, ERP, single-units, and neuropsychological studies of both lateral and medial prefrontal cortex. By reconceptualizing lateral prefrontal activity as anticipating prediction errors, the HER model provides a novel unifying account of prefrontal cortex function with broad implications for understanding the frontal cortex across multiple levels of description, from the level of single neurons to behavior.},
}
@article{AlexanderetalARN-86,
       author = {Alexander, G. E. and DeLong, M. R. and Strick, P. L.},
        title = {Parallel organization of functionally segregated circuits linking basal ganglia and cortex},
      journal = {Annual Review Neuroscience},
       volume = {9},
         year = {1986},
        pages = {357-381}
}
@inproceedings{AlexeetalCVPR-10,
        title = {What is an object?},
       author = {Bogdan Alexe and Thomas Deselaers and Vittorio Ferrari},
    booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE},
         year = 2010,
        pages = {73-80},
}
@misc{Algorithms_in_Nature,
       author = {Ziv Bar-Joseph [http://www.cs.cmu.edu/~zivbj/] and Matt Ruffalo},
        title = {CMU 02317: Algorithms in Nature},
         year = {2015},
 howpublished = {http://www.algorithmsinnature.org/},
     abstract = {Computer science and biology have shared a long history together. For many years, computer scientists have designed algorithms to process and analyze biological data (e.g. microarrays), and likewise, biologists have discovered several operating principles that have inspired new optimization methods (e.g. neural networks). Recently, these two directions have been converging based on the view that biological processes are inherently algorithms that nature has designed to solve computational problems. \\ This website documents new studies that have taken a joint computational­biological approach to study the algorithmic properties of biological processes across all levels of life (molecular, cellular, and organism).},
}
@article{AlivisatosetalNANO-13,
        title = {Nanotools for Neuroscience and Brain Activity Mapping},
       author = {Alivisatos, A. Paul and Andrews, Anne M. and Boyden, Edward S. and Chun, Miyoung and Church, George M. and Deisseroth, Karl and Donoghue, John P. and Fraser, Scott E. and Lippincott-Schwartz, Jennifer and Looger, Loren L. and Masmanidis, Sotiris and McEuen, Paul L. and Nurmikko, Arto V. and Park, Hongkun and Peterka, Darcy S. and Reid, Clay and Roukes, Michael L. and Scherer, Axel and Schnitzer, Mark and Sejnowski, Terrence J. and Shepard, Kenneth L. and Tsao, Doris and Turrigiano, Gina and Weiss, Paul S. and Xu, Chris and Yuste, Rafael and Zhuang, Xiaowei},
      journal = {ACS Nano},
       volume = 7,
        issue = 3,
         year = 2013,
        pages = {1850-1866},
}
@article{AlivisatosetalNEURON-12,
        title = {The Brain Activity Map Project and the Challenge of Functional Connectomics},
       author = {A. Paul Alivisatos and Miyoung Chun and George M. Church and  Ralph J. Greenspan and Michael L. Roukes and and Rafael Yuste},
      journal = {Neuron},
       volume = {74},
        issue = {6},
         year = 2012,
        pages = {970-974},
     abstract = {The function of neural circuits is an emergent property that arises from the coordinated activity of large numbers of neurons. To capture this, we propose launching a large-scale, international public effort, the Brain Activity Map Project, aimed at reconstructing the full record of neural activity across complete neural circuits. This technological challenge could prove to be an invaluable step toward understanding fundamental and pathological brain processes.},
}
@article{AlivisatosetalSCIENCE-13,
       author = {Alivisatos, A. Paul and Chun, Miyoung and Church, George M. and Deisseroth, Karl and Donoghue, John P. and Greenspan, Ralph J. and McEuen, Paul L. and Roukes, Michael L. and Sejnowski, Terrence J. and Weiss, Paul S. and Yuste, Rafael},
        title = {The Brain Activity Map},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {339},
       number = {6125},
        pages = {1284-1285},
         year = {2013},
     abstract = {Neuroscientists have made impressive advances in understanding the microscale function of single neurons and the macroscale activity of the human brain. One can probe molecular and biophysical aspects of individual neurons and also view the human brain in action with magnetic resonance imaging (MRI) or magnetoencephalography (MEG). However, the mechanisms of perception, cognition, and action remain mysterious because they emerge from the real-time interactions of large sets of neurons in densely interconnected, widespread neural circuits.}
}
@article{AljadeffetalNEURON-16,
       author = {Aljadeff, J. and Lansdell, B. J. and Fairhall, A. L. and Kleinfeld, D. },
        title = {{{A}nalysis of {N}euronal {S}pike {T}rains, {D}econstructed}},
      journal = {Neuron},
         year = {2016},
       volume = {91},
       number = {2},
        pages = {221-259},
     abstract = {As information flows through the brain, neuronal firing progresses from encoding the world as sensed by the animal to driving the motor output of subsequent behavior. One of the more tractable goals of quantitative neuroscience is to develop predictive models that relate the sensory or motor streams with neuronal firing. Here we review and contrast analytical tools used to accomplish this task. We focus on classes of models in which the external variable is compared with one or more feature vectors to extract a low-dimensional representation, the history of spiking and other variables are potentially incorporated, and these factors are nonlinearly transformed to predict the occurrences of spikes. We illustrate these techniques in application to datasets of different degrees of complexity. In particular, we address the fitting of models in the presence of strong correlations in the external variable, as occurs in natural sensory stimuli and in movement. Spectral correlation between predicted and measured spike trains is introduced to contrast the relative success of different methods.}
}
@article{AllamanisandBrockschmidtCoRR-17,
       author = {Miltiadis Allamanis and Marc Brockschmidt},
        title = {SmartPaste: Learning to Adapt Source Code},
      journal = {CoRR},
       volume = {arXiv:1705.07867},
         year = {2017},
     abstract = {Deep Neural Networks have been shown to succeed at a range of natural language tasks such as machine translation and text summarization. While tasks on source code (ie, formal languages) have been considered recently, most work in this area does not attempt to capitalize on the unique opportunities offered by its known syntax and structure. In this work, we introduce SmartPaste, a first task that requires to use such information. The task is a variant of the program repair problem that requires to adapt a given (pasted) snippet of code to surrounding, existing source code. As first solutions, we design a set of deep neural models that learn to represent the context of each variable location and variable usage in a data flow-sensitive way. Our evaluation suggests that our models can learn to solve the SmartPaste task in many cases, achieving 58.6\% accuracy, while learning meaningful representation of variable usages.}
}
@article{AllamanisetalCoRR-17,
       author = {Miltiadis Allamanis and Marc Brockschmidt and Mahmoud Khademi},
        title = {Learning to Represent Programs with Graphs},
      journal = {CoRR},
       volume = {arXiv:1711.00740},
         year = {2017},
     abstract = {Learning tasks on source code (i.e., formal languages) have been considered recently, but most work has tried to transfer natural language methods and does not capitalize on the unique opportunities offered by code's known syntax. For example, long-range dependencies induced by using the same variable or function in distant locations are often not considered. We propose to use graphs to represent both the syntactic and semantic structure of code and use graph-based deep learning methods to learn to reason over program structures.\\
                 In this work, we present how to construct graphs from source code and how to scale Gated Graph Neural Networks training to such large graphs. We evaluate our method on two tasks: VarNaming, in which a network attempts to predict the name of a variable given its usage, and VarMisuse, in which the network learns to reason about selecting the correct variable that should be used at a given program location. Our comparison to methods that use less structured program representations shows the advantages of modeling known structure, and suggests that our models learn to infer meaningful names and to solve the VarMisuse task in many cases. Additionally, our testing showed that VarMisuse identifies a number of bugs in mature open-source projects.}
}
@article{AllamanisetalCoRR-17b,
       author = {Miltiadis Allamanis and Earl T. Barr and Premkumar T. Devanbu and Charles A. Sutton},
        title = {A Survey of Machine Learning for Big Code and Naturalness},
      journal = {CoRR},
       volume = {arXiv:1709.06182},
         year = {2017},
     abstract = {Research at the intersection of machine learning, programming languages, and software engineering has recently taken important steps in proposing learnable probabilistic models of source code that exploit code's abundance of patterns. In this article, we survey this work. We contrast programming languages against natural languages and discuss how these similarities and differences drive the design of probabilistic models. We present a taxonomy based on the underlying design principles of each model and use it to navigate the literature. Then, we review how researchers have adapted these models to application areas and discuss cross-cutting and application-specific challenges and opportunities.},
}
@inproceedings{AllamanisetalCoRR-17c,
       author = {Allamanis, Miltiadis and Chanthirasegaran, Pankajan and Kohli, Pushmeet and Sutton, Charles},
    booktitle = {Proceedings Interational Conference on Machine Learning (ICML)},
      journal = {CoRR},
        title = {Learning Continuous Semantic Representations of Symbolic Expressions},
         year = {2017},
     abstract = {Combining abstract, symbolic reasoning with continuous neural reasoning is a grand challenge of representation learning. As a step in this direction, we propose a new architecture, called neural equivalence networks, for the problem of learning continuous semantic representations of algebraic and logical expressions. These networks are trained to represent semantic equivalence, even of expressions that are syntactically very different. The challenge is that semantic representations must be computed in a syntax-directed manner, because semantics is compositional, but at the same time, small changes in syntax can lead to very large changes in semantics, which can be difficult for continuous neural architectures. We perform an exhaustive evaluation on the task of checking equivalence on a highly diverse class of symbolic algebraic and boolean expression types, showing that our model significantly outperforms existing architectures.}
}
@inproceedings{AllamanisetalICLR-18,
        title = {Learning to Represent Programs with Graphs},
       author = {Miltiadis Allamanis and Marc Brockschmidt and Mahmoud Khademi},
    booktitle = {International Conference on Learning Representations},
         year = {2018},
     abstract = {Learning tasks on source code (i.e., formal languages) have been considered recently, but most work has tried to transfer natural language methods and does not capitalize on the unique opportunities offered by code's known semantics. For example, long-range dependencies induced by using the same variable or function in distant locations are often not considered. We propose to use graphs to represent both the syntactic and semantic structure of code and use graph-based deep learning methods to learn to reason over program structures. In this work, we present how to construct graphs from source code and how to scale Gated Graph Neural Networks training to such large graphs. We evaluate our method on two tasks: VARNAMING, in which a network attempts to predict the name of a variable given its usage, and VARMISUSE, in which the network learns to reason about selecting the correct variable that should be used at a given program location. Our comparison to methods that use less structured program representations shows the advantages of modeling known structure, and suggests that our models learn to infer meaningful names and to solve the VARMISUSE task in many cases. Additionally, our testing showed that VARMISUSE identifies a number of bugs in mature open-source projects.},
}
@book{Allen1992,
       author = {Allen, James F.},
        title = {Natural Language Understanding, Second Edition},
    publisher = {Benjamin-Cummings},
      address = {Redwood City, California},
         year = 1994,
}
@article{AllenAIJ-84,
       author = {Allen, James F.},
        title = {Towards a General Theory of Action and Time},
      journal = AIJ,
       volume = 23,
         year = 1984,
        pages = {123-154},
}
@article{AllenCACM-83,
       author = {Allen, James F.},
        title = {Maintaining knowledge about temporal intervals},
      journal = {Communications of the {ACM}},
       volume = 26,
         year = 1983,
        pages = {832-843},
}
@inproceedings{AllenandHayesIJCAI-85,
       author = {Allen, James F. and Hayes, Patrick J.},
        title = {A Common-Sense Theory of Time},
    booktitle = {Proceedings {IJCAI} 9},
     location = {Los Angeles, California},
  institution = {IJCAII},
         year = 1985,
}
@inproceedings{AllenandKoomenIJCAI-83,
       author = {Allen, James F. and Koomen, J. A.},
        title = {Planning Using a Temporal World Model},
    booktitle = {Proceedings {IJCAI} 8, Karlsruhe, West Germany},
 organization = {IJCAII},
         year = 1983,
        pages = {741-747},
}
@book{Allenetal1990,
       editor = {Allen, James F. and Hendler, James and Tate, Austin},
        title = {Readings in Planning},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1990,
}
@book{Allenetal1991,
       author = {J. F. Allen and H. A. Kautz and R. N. Pelavin and J. D. Tenenberg},
        title = {Reasoning about Plans},
    publisher = {Morgan-Kaufmann},
      address = {San Francisco, California},
         year = 1991,
}
@article{AllenetalVR-93,
       author = {Allen, D. and Banks, M.S. and Norcia, A.M.},
        title = {Does chromatic sensitivity develop more slowly than luminance sensitivity?},
      journal = {Vision Research},
       volume = 33,
        pages = {2553-2562},
         year = 1993,
      comment = {Allen, D. and Banks, M.S. and Norcia, A.M. (1993). Does chromatic sensitivity develop more slowly than luminance sensitivity? Vision Research, 33, 2553-2562.}
}
@article{AllmanetalARP-14,
       author = {Allman, M. J. and Teki, S. and Griffiths, T. D. and Meck, W. H.},
        title = {Properties of the internal clock: first- and second-order principles of subjective time},
      journal = {Annual Review of Psychology},
         year = 2014,
       volume = 65,
        pages = {743-771},
     abstract = {Humans share with other animals an ability to measure the passage of physical time and subjectively experience a sense of time passing. Subjective time has hallmark qualities, akin to other senses, which can be accounted for by formal, psychological, and neurobiological models of the internal clock. These include first-order principles, such as changes in clock speed and how temporal memories are stored, and second-order principles, including timescale invariance, multisensory integration, rhythmical structure, and attentional time-sharing. Within these principles there are both typical individual differences--influences of emotionality, thought speed, and psychoactive drugs--and atypical differences in individuals affected with certain clinical disorders (e.g., autism, Parkinson's disease, and schizophrenia). This review summarizes recent behavioral and neurobiological findings and provides a theoretical framework for considering how changes in the properties of the internal clock impact time perception and other psychological domains.}
}
@article{AllmanetalBSF-10,
        title = {The von {E}conomo neurons in fronto-insular and anterior cingulate cortex in great apes and humans},
       author = {Allman, J.M. and Tetreault, N.A. and Hakeem, A.Y. and Manaye, K.F. and Semendeferi, K. and Erwin, J.M. and Park, S. and Goubert, V. and Hof, P.R.},
      journal = {Brain Structure and Function},
       volume = 214,
        issue = {5-6},
         year = 2010,
        pages = {495-517},
     abstract = {The von Economo neurons (VENs) are large bipolar neurons located in fronto-insular (FI) and anterior cingulate cortex in great apes and humans, but not other primates. We performed stereological counts of the VENs in FI and LA (limbic anterior, a component of anterior cingulate cortex) in great apes and in humans. The VENs are more numerous in humans than in apes, although one gorilla approached the lower end of the human range. We also examined the ontological development of the VENs in FI and LA in humans. The VENs first appear in small numbers in the 36th week post-conception, are rare at birth, and increase in number during the first 8 months after birth. There are significantly more VENs in the right hemisphere than in the left in FI and LA in postnatal brains of apes and humans. This asymmetry in VEN numbers may be related to asymmetries in the autonomic nervous system. The activity of the inferior anterior insula, which contains FI, is related to physiological changes in the body, decision-making, error recognition, and awareness. The VENs appear to be projection neurons, although their targets are unknown. We made a preliminary study of the connections of FI cortex based on diffusion tensor imaging in the brain of a gorilla. The VEN-containing regions connect to the frontal pole as well as to other parts of frontal and insular cortex, the septum, and the amygdala. It is likely that the VENs in FI are projecting to some or all of these structures and relaying information related to autonomic control, decision-making, or awareness. The VENs selectively express the bombesin peptides neuromedin B (NMB) and gastrin releasing peptide (GRP) which are also expressed in another population of closely related neurons, the fork cells. NMB and GRP signal satiety. The genes for NMB and GRP are expressed selectively in small populations of neurons in the insular cortex in mice. These populations may be related to the VEN and fork cells and may be involved in the regulation of appetite. The loss of these cells may be related to the loss of satiety signaling in patients with fronto-temporal dementia who have damage to FI. The VENs and fork cells may be morphological specializations of an ancient population of neurons involved in the control of appetite present in the insular cortex in all mammals. We found that the protein encoded by the gene DISC1 (disrupted in schizophrenia) is preferentially expressed by the VENs. DISC1 has undergone rapid evolutionary change in the line leading to humans, and since it suppresses dendritic branching it may be involved in the distinctive VEN morphology.},
}
@article{AllmanetalNEUROSCIENTIST-02,
        title = {Two phylogenetic specializations in the human brain},
       author = {John Allman and Atiya Hakeem and Karli Watson},
      journal = {Neuroscientist},
       volume = 8,
        issue = 4,
         year = 2002,
        pages = {335-346},
     abstract = {In this study, two anatomical specializations of the brain in apes and humans are considered. One of these is a whole cortical area located in the frontal polar cortex (Brodmann's area 10), and the other is a morphologically distinctive cell type, the spindle neuron of the anterior cingulate cortex. The authors suggest that the spindle cells may relay to other parts of the brain-especially to area 10, the outcome of processing within the anterior cingulate cortex. This relay conveys the motivation to act. It particularly concerns the recognition of having committed an error that leads to the initiation of adaptive responses to these adverse events so as to reduce error commission. This capacity is related to the development of self-control as an individual matures and gains social insight. Although the anterior cingulate deals with the individual's immediate response to changing conditions, area 10 is involved in the retrieval of memories from the individual's past experience and the capacity to plan adaptive responses. The authors suggest that these neurobehavioral specializations are crucial aspects of intelligence as defined as the capacity to make adaptive responses to changing conditions. The authors further hypothesize that these specializations facilitated the evolution of the unique capacity for the intergenerational transfer of the food and information characteristic of human extended families.}
}
@article{AllowayCEREBRAL-CORTEX-08,
       author = {Alloway, Kevin D.},
        title = {{Information Processing Streams in Rodent Barrel Cortex: The Differential Functions of Barrel and Septal Circuits}},
      journal = {Cerebral Cortex},
       volume = {18},
       number = {5},
         year = {2007},
        pages = {979-989},
     abstract = {Rodent somatosensory cortex contains an isomorphic map of the mystacial whiskers in which each whisker is represented by neuronal populations, or barrels, that are separated from each other by intervening septa. Separate afferent pathways convey somatosensory information to the barrels and septa that represent the input stages for 2 partially segregated circuits that extend throughout the other layers of barrel cortex. Whereas the barrel-related circuits process spatiotemporal information generated by whisker contact with external objects, the septa-related circuits encode the frequency and other kinetic features of active whisker movements. The projection patterns from barrel cortex indicate that information processed by the septa-related circuits is used both separately and in combination with information from the barrel-related circuits to mediate specific functions. According to this theory, outputs from the septal processing stream modulate the brain regions that regulate whisking behavior, whereas both processing streams cooperate with each other to identify external stimuli encountered by passive or active whisker movements. This theoretical view prompts several testable hypotheses about the coordination of neuronal activity during whisking behavior. Foremost among these, motor brain regions that control whisker movements are more strongly coordinated with the septa-related circuits than with the barrel-related circuits.},
}
@inproceedings{AlmetalNLP-05,
       author = {Alm, Cecilia Ovesdotter and Roth, Dan and Sproat, Richard},
        title = {Emotions from text: machine learning for text-based emotion prediction},
    booktitle = {Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing},
         year = 2005,
        pages = {579-586},
    publisher = {Association for Computational Linguistics},
     abstract = {In addition to information, text contains attitudinal, and more specifically, emotional content. This paper explores the text-based emotion prediction problem empirically, using supervised machine learning with the SNoW learning architecture. The goal is to classify the emotional affinity of sentences in the narrative domain of children's fairy tales, for subsequent usage in appropriate expressive rendering of text-to-speech synthesis. Initial experiments on a preliminary data set of 22 fairy tales show encouraging results over a naïve baseline and BOW approach for classification of emotional versus non-emotional contents, with some dependency on parameter tuning. We also discuss results for a tripartite model which covers emotional valence, as well as feature set alternations. In addition, we present plans for a more cognitively sound sequential model, taking into consideration a larger set of basic emotions.}
}
@book{Aloimonos1993,
       editor = {Yiannis Aloimonos},
        title = {Active Perception},
         year = 1993,
    publisher = {Lawrence Erlbaum Associates, Inc.},
      address = {Mahwah, NJ, USA},
}
@inproceedings{AloimonosetalICCV-87,
       author = {Aloimonos, J. and Bandyopadhyay, A. and Weiss, I.},
        title = {Active Vision},
    booktitle = {Proceedings of the First International Conference on Computer Vision},
         year = 1987,
        pages = {35-55},
}
@article{AlonetalCoRR-18,
       author = {Uri Alon and Meital Zilberstein and Omer Levy and Eran Yahav},
        title = {code2vec: Learning Distributed Representations of Code},
      journal = {CoRR},
       volume = {arXiv:1803.09473},
         year = {2018},
     abstract = {We present a neural model for representing snippets of code as continuous distributed vectors. The main idea is to represent code as a collection of paths in its abstract syntax tree, and aggregate these paths, in a smart and scalable way, into a single fixed-length code vector, which can be used to predict semantic properties of the snippet. We demonstrate the effectiveness of our approach by using it to predict a method's name from the vector representation of its body. We evaluate our approach by training a model on a dataset of 14M methods. We show that code vectors trained on this dataset can predict method names from files that were completely unobserved during training. Furthermore, we show that our model learns useful method name vectors that capture semantic similarities, combinations, and analogies. Comparing previous techniques over the same data set, our approach obtains a relative improvement of over 75\%, being the first to successfully predict method names based on a large, cross-project, corpus.}
}
@article{AlonetalCoRR-18b,
       author = {Uri Alon and Omer Levy and Eran Yahav},
        title = {code2seq: Generating Sequences from Structured Representations of Code},
      journal = {CoRR},
       volume = {arXiv:1808.01400},
         year = {2018},
     abstract = {The ability to generate natural language sequences from source code snippets can be used for code summarization, documentation, and retrieval. Sequence-to-sequence (seq2seq) models, adopted from neural machine translation (NMT), have achieved state-of-the-art performance on these tasks by treating source code as a sequence of tokens. We present CODE2SEQ: an alternative approach that leverages the syntactic structure of programming languages to better encode source code. Our model represents a code snippet as the set of paths in its abstract syntax tree (AST) and uses attention to select the relevant paths during decoding, much like contemporary NMT models. We demonstrate the effectiveness of our approach for two tasks, two programming languages, and four datasets of up to 16M examples. Our model significantly outperforms previous models that were specifically designed for programming languages, as well as general state-of-the-art NMT models.},
}
@inproceedings{AlonetalPOPL-19,
       author = {Alon, Uri and Zilberstein, Meital and Levy, Omer and Yahav, Eran},
        title = {Code2vec: Learning Distributed Representations of Code},
    publisher = {Association for Computing Machinery},
    booktitle = {Proceedings of the ACM on Programming Languages},
      address = {New York, NY, USA},
       volume = {3},
         year = {2019},
     abstract = {We present a neural model for representing snippets of code as continuous distributed vectors (code embeddings). The main idea is to represent a code snippet as a single fixed-length code vector, which can be used to predict semantic properties of the snippet. To this end, code is first decomposed to a collection of paths in its abstract syntax tree. Then, the network learns the atomic representation of each path while simultaneously learning how to aggregate a set of them.  We demonstrate the effectiveness of our approach by using it to predict a method's name from the vector representation of its body. We evaluate our approach by training a model on a dataset of 12M methods. We show that code vectors trained on this dataset can predict method names from files that were unobserved during training. Furthermore, we show that our model learns useful method name vectors that capture semantic similarities, combinations, and analogies.  A comparison of our approach to previous techniques over the same dataset shows an improvement of more than 75\%, making it the first to successfully predict method names based on a large, cross-project corpus. Our trained model, visualizations and vector similarities are available as an interactive online demo at http://code2vec.org. The code, data and trained models are available at https://github.com/tech-srl/code2vec.},
}
@inproceedings{AltermanAAAI-86,
       author = {Alterman, Richard},
        title = {An Adaptive Planner},
    booktitle = {Proceedings AAAI-86},
     location = {Philadelphia, Pennsylvania},
 organization = {AAAI},
         year = 1986,
        pages = {65-69},
}
@misc{AltosaarVAE-REVIEW-17,
        title = {Understanding Variational Autoencoders (VAEs) from two perspectives: deep learning and graphical models},
       author = {Jaan Altoaar},
 howpublished = {\url{https://jaan.io/what-is-variational-autoencoder-vae-tutorial/}},
         year = {2017},
}
@inproceedings{AltunetalICML-03,
       author = {Altun, Y. and Tsochantaridis, I. and Hofmann, T.},
        title = {Hidden {M}arkov Support Vector Machines},
    booktitle = {International Conference on Machine Learning (ICML)},
         year = 2003,
}
@inproceedings{AluretalSTOC-95,
       author = {Alur, Rajeev and Courcoubetis, Costas and Yannakakis, Mihalis},
        title = {Distinguishing Tests for Nondeterministic and Probabilistic Machines},
    booktitle = {Proceedings of the 27th Annual {ACM} Symposium on Theory of Computing},
     location = {Las Vegas, Nevada},
         year = 1995,
}
@article{Alvarez-MelisandJaakkolaCoRR-17,
       author = {David Alvarez{-}Melis and Tommi S. Jaakkola},
        title = {A causal framework for explaining the predictions of black-box sequence-to-sequence models},
      journal = {CoRR},
       volume = {arXiv:1707.01943},
         year = {2017},
     abstract = {We interpret the predictions of any black-box structured input-structured output model around a specific input-output pair. Our method returns an "explanation" consisting of groups of input-output tokens that are causally related. These dependencies are inferred by querying the black-box model with perturbed inputs, generating a graph over tokens from the responses, and solving a partitioning problem to select the most relevant components. We focus the general approach on sequence-to-sequence problems, adopting a variational autoencoder to yield meaningful input perturbations. We test our method across several NLP sequence generation tasks.}
}
@inproceedings{AlvarezetalICAART-19,
        title = {Contextual Action with Multiple Policies Inverse Reinforcement Learning for Behavior Simulation},
       author = {Nahum Alvarez and Itsuki Noda},
    booktitle = {11th International Conference on Agents and Artificial Intelligence},
         year = {2019},
     abstract = {Machine learning is a discipline with many simulator-driven applications oriented to learn behavior. However, behavior simulation it comes with a number of associated difficulties, like the lack of a clear reward function, actions that depend of the state of the actor and the alternation of different policies. We present a method for behavior learning called Contextual Action Multiple Policy Inverse Reinforcement Learning (CAMP-IRL) that tackles those factors. Our method allows to extract multiple reward functions and generates different behavior profiles from them. We applied our method to a large scale crowd simulator using intelligent agents to imitate pedestrian behavior, making the virtual pedestrians able to switch between behaviors depending of the goal they have and navigating efficiently across unknown environments.}
}
@article{AmalricandDehaenePNAS-16,
       author = {Amalric, Marie and Dehaene, Stanislas},
        title = {Origins of the brain networks for advanced mathematics in expert mathematicians},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {113},
       number = {18},
         year = {2016},
        pages = {4909-4917},
     abstract = {Our work addresses the long-standing issue of the relationship between mathematics and language. By scanning professional mathematicians, we show that high-level mathematical reasoning rests on a set of brain areas that do not overlap with the classical left-hemisphere regions involved in language processing or verbal semantics. Instead, all domains of mathematics we tested (algebra, analysis, geometry, and topology) recruit a bilateral network, of prefrontal, parietal, and inferior temporal regions, which is also activated when mathematicians or nonmathematicians recognize and manipulate numbers mentally. Our results suggest that high-level mathematical thinking makes minimal use of language areas and instead recruits circuits initially involved in space and number. This result may explain why knowledge of number and space, during early childhood, predicts mathematical achievement.The origins of human abilities for mathematics are debated: Some theories suggest that they are founded upon evolutionarily ancient brain circuits for number and space and others that they are grounded in language competence. To evaluate what brain systems underlie higher mathematics, we scanned professional mathematicians and mathematically naive subjects of equal academic standing as they evaluated the truth of advanced mathematical and nonmathematical statements. In professional mathematicians only, mathematical statements, whether in algebra, analysis, topology or geometry, activated a reproducible set of bilateral frontal, Intraparietal, and ventrolateral temporal regions. Crucially, these activations spared areas related to language and to general-knowledge semantics. Rather, mathematical judgments were related to an amplification of brain activity at sites that are activated by numbers and formulas in nonmathematicians, with a corresponding reduction in nearby face responses. The evidence suggests that high-level mathematical expertise and basic number sense share common roots in a nonlinguistic brain circuit.},
}
@article{AmaraletalPBR-17,
       author = {Amaral, D. G. and Scharfman, H.E. and Lavenex, P.},
        title = {The dentate gyrus: fundamental neuroanatomical organization (dentate gyrus for dummies)},
      journal = {Progress in Brain Research},
         year = {2007},
       volume = {163},
        pages = {3-22},
     abstract = {The dentate gyrus is a simple cortical region that is an integral portion of the larger functional brain system called the hippocampal formation. In this review, the fundamental neuroanatomical organization of the dentate gyrus is described, including principal cell types and their connectivity, and a summary of the major extrinsic inputs of the dentate gyrus is provided. Together, this information provides essential information that can serve as an introduction to the dentate gyrus--a "dentate gyrus for dummies."}
}
@article{AmaraletalPBR-90,
       author = {Amaral, {David G} and Norio Ishizuka and Brenda Claiborne},
        title = {Chapter Neurons, numbers and the hippocampal network},
    publisher = {Elsevier},
       volume = {83},
         year = {1990},
        pages = {1-11},
      journal = {Progress in Brain Research},
     abstract = {Anatomists involved with studies of the hippocampal formation are being prodded by computational modelers and physiologists who demand detailed and quantitative information concerning hippocampal neurons and circuits. The beautiful camera lucida drawings of old, and the elegant descriptions of dendritic form that accompanied them are giving way to computer-reconstructed and three-dimensionally analyzed cells with rigorous determination of dendritic lengths and volumes, branching pattern and spine distribution. We will review certain quantitative aspects of hippocampal organization in the rat based on a survey of available literature and on our own intracellular labeling studies of granule cells of the dentate gyrus and pyramidal cells of the hippocampus. Some of the potential implications of these data for hippocampal information processing will be discussed.},
}
@article{AmbroseetalNEURON-16,
       author = {Ambrose, R. E. and Pfeiffer, B. E. and Foster, D. J.},
        title = {{{R}everse {R}eplay of {H}ippocampal {P}lace {C}ells {I}s {U}niquely {M}odulated by {C}hanging {R}eward}},
      journal = {Neuron},
       volume = {91},
       number = {5},
         year = {2016},
        pages = {1124-1136},
     abstract = {Hippocampal replays are episodes of sequential place cell activity during sharp-wave ripple oscillations (SWRs). Conflicting hypotheses implicate awake replay in learning from reward and in memory retrieval for decision making. Further, awake replays can be forward, in the same order as experienced, or reverse, in the opposite order. However, while the presence or absence of reward has been reported to modulate SWR rate, the effect of reward changes on replay, and on replay direction in particular, has not been examined. Here we report divergence in the response of forward and reverse replays to changing reward. While both classes of replays were observed at reward locations, only reverse replays increased their rate at increased reward or decreased their rate at decreased reward, while forward replays were unchanged. These data demonstrate a unique relationship between reverse replay and reward processing and point to a functional distinction between different directions of replay.}
}
@article{AmitetalPAMI-04,
     author = {Y. Amit and D. Geman and X. Fan},
      title = {A coarse-to-fine strategy for multi-class shape detection},
    journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence}, 
     volume = 26,
      issue = 12,
      pages = {1606-1621},
       year = 2004,
}
@article{AmosandSchlieperAPC-05,
        title = {Microtubules and maps},
       author = {Amos, Linda A. and Schlieper, Daniel},
      journal = {Advances in protein chemistry},
       volume = {71},
         year = {2005},
        pages = {257-298},
     abstract = {Microtubules are very dynamic polymers whose assembly and disassembly is determined by whether their heterodimeric tubulin subunits are in a straight or curved conformation. Curvature is introduced by bending at the interfaces between monomers. Assembly and disassembly are primarily controlled by the hydrolysis of guanosine triphosphate (GTP) in a site that is completed by the association of two heterodimers. However, a multitude of associated proteins are able to fine-tune these dynamics so that microtubules are assembled and disassembled where and when they are required by the cell. We review the recent progress that has been made in obtaining a glimpse of the structural interactions involved.},
}
@article{AmosetalCoRR-18,
       author = {Brandon Amos and Laurent Dinh and Serkan Cabi and Thomas Roth{\"{o}}rl and Sergio Gomez Colmenarejo and Alistair Muldal and Tom Erez and Yuval Tassa and Nando de Freitas and Misha Denil},
        title = {Learning Awareness Models},
      journal = {CoRR},
       volume = {arXiv:1804.06318},
         year = {2018},
     abstract = {We consider the setting of an agent with a fixed body interacting with an unknown and uncertain external world. We show that models trained to predict proprioceptive information about the agent's body come to represent objects in the external world. In spite of being trained with only internally available signals, these dynamic body models come to represent external objects through the necessity of predicting their effects on the agent's own body. That is, the model learns holistic persistent representations of objects in the world, even though the only training signals are body signals. Our dynamics model is able to successfully predict distributions over 132 sensor readings over 100 steps into the future and we demonstrate that even when the body is no longer in contact with an object, the latent variables of the dynamics model continue to represent its shape. We show that active data collection by maximizing the entropy of predictions about the body---touch sensors, proprioception and vestibular information---leads to learning of dynamic models that show superior performance when used for control. We also collect data from a real robotic hand and show that the same models can be used to answer questions about properties of objects in the real world. Videos with qualitative results of our models are available at this {\urlh{https://goo.gl/mZuqAV}{URL}}.}
}
@article{AmuntsetalNI-07,
        title = {Cytoarchitecture of the cerebral cortex --- more than localization},
       author = {Amunts, K. and Schleicher, A. and Zilles, K.},
       volume = 37,
        issue = 4,
      journal = {NeuroImage},
         year = 2007,
        pages = {1061-1065},
}
@article{AmuntsetalSCIENCE-13,
       author = {Amunts, K. and Lepage, C. and Borgeat, L. and Mohlberg, H. and Dickscheid, T. and Rousseau, M.-E. and Bludau, S. and Bazin, P.-L. and Lewis, L. B. and Oros-Peusquens, A.-M. and Shah, N. J. and Lippert, T. and Zilles, K. and Evans, A. C.},
        title = {BigBrain: An Ultrahigh-Resolution {3D} Human Brain Model},
      journal = {Science},
       volume = 340,
        issue = 6139,
      address = {Washington, DC},
    publisher = {American Association for the Advancement of Science},
        pages = {1472-1475},
         year = 2013,
}
@techreport{AnandanLetovskyMjolsnessTR-89,
        title = {Connectionist Variable Binding by Optimization},
       author = {Anandan, P. and Letovsky, Stanley and Mjolsness, Eric},
         year = 1989,
  institution = {Yale Technical Report DCS-RR-711}
}
@inproceedings{AnanthanarayananetalSC-09,
       author = {Ananthanarayanan, Rajagopal and Esser, Steven K. and Simon, Horst D. and Modha, Dharmendra S.},
        title = {The cat is out of the bag: cortical simulations with 10$^9$ neurons, 10$^{13}$ synapses},
    booktitle = {Proceedings of the Conference on High Performance Computing Networking, Storage and Analysis},
         year = 2009,
        pages = {63:1-63:12},
     abstract = {In the quest for cognitive computing, we have built a mas- sively parallel cortical simulator, C2, that incorporates a number of innovations in computation, memory, and communication. Using C2 on LLNL's Dawn Blue Gene/P supercomputer with 147, 456 CPUs and 144 TB of main memory, we report two cortical simulations---at unprecedented scale---that effectively saturate the entire memory capacity and refresh it at least every simulated second. The first simulation consists of 1.6 billion neurons and 8.87 trillion synapses with experimentally-measured gray matter thalamocortical connectivity. The second simulation has 900 million neurons and 9 trillion synapses with probabilistic connectivity. We demonstrate nearly perfect weak scaling and attractive strong scaling. The simulations, which incorporate phenomenological spiking neurons, individual learning synapses, axonal delays, and dynamic synaptic channels, exceed the scale of the cat cortex, marking the dawn of a new era in the scale of cortical simulations.}
}
@article{AnanthaswamyNEW-SCIENTIST-09,
        title = {Language may be key to theory of mind},
       author = {Anil Ananthaswamy},
      journal = {New Scientist},
         year = 2009,
     abstract = {How blind and deaf people approach a cognitive test regarded as a milestone in human development has provided clues to how we deduce what others are thinking. Understanding another person's perspective, and realising that it can differ from our own, is known as theory of mind.},
}
@incollection{AnastassiouandShai3MDB-16,
       author = {Costas A. Anastassiou and Adam S. Shai},
        title = {Psyche, Signals and Systems},
    booktitle = {Micro-, Meso- and Macro-Dynamics of the Brain},
       editor = {Buzs\`{a}ki, Gy\"{o}rgy and Christen, Yves},
    publisher = {Springer New York},
      address = {New York, NY},
         year = {2016},
        pages = {107-156},
}
@article{AnastassiouetalNATURE-11,
        title = {Ephaptic coupling of cortical neurons},
       author = {Anastassiou, Costas A. and Perin, Rodrigo and Markram, Henry and Koch, Christof},
      journal = {Nature Neuroscience},
       volume = 14,
        issue = 2,
         year = 2011,
        pages = {217-223},
     abstract = {The electrochemical processes that underlie neural function manifest themselves in ceaseless spatiotemporal field fluctuations. However, extracellular fields feed back onto the electric potential across the neuronal membrane via ephaptic coupling, independent of synapses. The extent to which such ephaptic coupling alters the functioning of neurons under physiological conditions remains unclear. To address this question, we stimulated and recorded from rat cortical pyramidal neurons in slices with a 12-electrode setup. We found that extracellular fields induced ephaptically mediated changes in the somatic membrane potential that were less than 0.5 mV under subthreshold conditions. Despite their small size, these fields could strongly entrain action potentials, particularly for slow (<8 Hz) fluctuations of the extracellular field. Finally, we simultaneously measured from up to four patched neurons located proximally to each other. Our findings indicate that endogenous brain activity can causally affect neural function through field effects under physiological conditions.},
}
@inproceedings{AndHof04,
       author = {S. Andrews  and T. Hofmann},
        title = {Multiple Instance Learning via Disjunctive Programming Boosting},
    booktitle = {Advances in Neural Information Processing (NIPS 16)},
         year = 2004,
} 
@article{Anderson03,
       author = {Anderson, J.A.},
        title = {Arithmetic on a Parallel Computer: Perception Versus Logic},
      journal = {Brain and Mind},
    publisher = {Kluwer Academic Publishers},
       volume = 4,
        pages = {169-188},
         year = 2003,
}
@incollection{Anderson93,
       author = {Anderson, J.A.},
        title = {The {BSB} Model: {A} simple nonlinear autoassociative neural network},
    booktitle = {Associative Neural Memories},
       editor = {M. Hassoun},
    publisher = {Oxford University Press},
      address = {New York, NY},
         year = 1993,
}
@incollection{AndersonandMartinABA-16,
        title = {{Chapter 6 - Interareal Connections of the Macaque Cortex: How Neocortex Talks to Itself}},
       author = {J.C. Anderson and K.A.C. Martin},
    booktitle = {Axons and Brain Architecture},
       editor = {Kathleen S. Rockland},
    publisher = {Academic Press},
      address = {San Diego},
         year = {2016},
        pages = {117-134},
     abstract = {The neocortex forms a single sheet that is subdivided into 100 or more different functional areas in primates. One unifying concept is that the local circuits in the neocortex share the same basic organization and that the functional differences between cortical areas arise from differences in their afferent input-both from subcortical regions like the thalamus and other regions of neocortex. While investigations in the macaque monkey have provided us with the most complete and quantitative map of the interareal connections for any species, our knowledge of the organization of synaptic connections of these interareal pathways is still quite incomplete. We have investigated the macaque visual cortex in search of the design principles of the synaptic relationships formed by long-distance interareal projections. The pyramidal cells, which form the interareal projections, contribute only a tiny fraction of all the extrinsic excitatory synapses in a given target area. One lamina is the principal target, but frequently the axons form synapses in a secondary lamina. Most interareal synapses are formed with the spines of pyramidal cells. The GABAergic smooth cell dendrites constitute 0-20\% of the target population. To explain why the tiny numbers of excitatory synapses that connect two areas are so effective, we suggest that the local recurrent circuit amplifies spatio-temporal patterns of excitation generated by long-distance connections-both thalamic and interareal. The specific computations within an area are a consequence of specific patterns of excitatory input from the long-distance projections that are shaped by excitatory and inhibitory activity evoked in the local circuit.}
}
@article{AndersonandSchoolerPSYCHOLOGY-SCIENCE-91,
       author = {John R. Anderson and Lael J. Schooler},
        title = {Reflections of the Environment in Memory},
      journal = {Psychology Science},
       volume = {2},
        issue = {6},
         year = {1991},
     abstract = {Availability of human memories for specific items shows reliable relationships to frequency, recency, and pattern of prior exposures to the item. These relationships have defied a systematic theoretical treatment. Empirical data about a number of environmental sources (New York Times, parental speech, electronic mail) are examined to show that the probability that a memory will be needed also shows reliable relationships to frequency, recency, and pattern of prior exposures. Moreover, the environmental relationships are the same as the memory relationships. It is argued that human memory has the form it does because it is adapted to these environmental relationships. Models for both the environment and human memory are described.}
}
@article{AndersonandSutton97,
       author = {Anderson, J.A. and Sutton, J.P.},
        title = {If we compute faster, do we understand better?},
      journal = {Behavior Ressearch Methods, Instruments and Computers},
    publisher = {Psychonomic Society},
       volume = 29,
        pages = {67-77},
         year = 1997,
}
@article{AndersonetalAPP-14,
       author = {Anderson, G. M. and Foulsham, T. and Nasiopoulos, E. and Chapman, C. S. and Kingstone, A.},
        title = {{{H}ide and seek: the theory of mind of visual concealment and search}},
      journal = {Attention and Perception Psychophysics},
       volume = {76},
       number = {4},
         year = {2014},
        pages = {907-913},
     abstract = {Researchers have investigated visual search behavior for almost a century. During that time, few studies have examined the cognitive processes involved in hiding items rather than finding them. To investigate this, we developed a paradigm that allowed participants to indicate where they would hide (or find) an item that was to be found (or hidden) by a friend or a foe. We found that (i) for friends more than foes, participants selected the pop-out item in the display, and (ii) when the display was homogeneous, they selected nearby and corner items. These behaviors held for both hiding and finding, although hide and find behaviors were not identical. For pop-out displays, decision times were unusually long when hiding an item from a foe. These data converge on the conclusion that the principles of search and concealment are similar, but not the same. They also suggest that this paradigm will provide researchers a powerful method for investigating theory of mind in adults.}
}
@article{AndersonetalNEUROSCIENCE-11,
       author = {Emily B. Anderson and Jude F. Mitchell and John H. Reynolds},
        title = {Attentional Modulation of Firing Rate Varies with Burstiness across Putative Pyramidal Neurons in Macaque Visual Area V4},
      journal = {The Journal of Neuroscience},
       volume = 31,
        issue = 30,
         year = 2011,
        pages = {10983-10992},
     abstract = {One of the most well established forms of attentional modulation is an increase in firing rate when attention is directed into the receptive field of a neuron. The degree of rate modulation, however, can vary considerably across individual neurons, especially among broad spiking neurons (putative pyramids). We asked whether this heterogeneity might be correlated with a neuronal response property that is used in intracellular recording studies to distinguish among distinct neuronal classes: the burstiness of the neuronal spike train. We first characterized the burst spiking behavior of visual area V4 neurons and found that this varies considerably across the population, but we did not find evidence for distinct classes of burst behavior. Burstiness did, however, vary more widely across the class of neurons that shows the greatest heterogeneity in attentional modulation, and within that class, burstiness helped account for differences in attentional modulation. Among these broad spiking neurons, rate modulation was primarily restricted to bursty neurons, which as a group showed a highly significant increase in firing rate with attention. Furthermore, every bursty broad spiking neuron whose firing rate was significantly modulated by attention exhibited an increase in firing rate. In contrast, non-bursty broad spiking neurons exhibited no net attentional modulation, and, although some individual neurons did show significant rate modulation, these were divided among neurons showing increases and decreases. These findings show that macaque area V4 shows a range of bursting behavior and that the heterogeneity of attentional modulation can be explained, in part, by variation in burstiness.}
}
@article{AndersonetalNLM-16,
       author = {Anderson, Michael C. and Bunce, Jamie G. and Barbas, Helen},
        title = {Prefrontal-hippocampal pathways underlying inhibitory control over memory},
      journal = {Neurobiology of learning and memory},
         year = {2016},
       volume = {134},
        issue = {Pt A},
        pages = {145-161},
     abstract = {A key function of the prefrontal cortex is to support inhibitory control over behavior. It is widely believed that this function extends to stopping cognitive processes as well. Consistent with this, mounting evidence establishes the role of the right lateral prefrontal cortex in a clear case of cognitive control: retrieval suppression. Retrieval suppression refers to the ability to intentionally stop the retrieval process that arises when a reminder to a memory appears. Functional imaging data indicate that retrieval suppression involves top-down modulation of hippocampal activity by the dorsolateral prefrontal cortex, but the anatomical pathways supporting this inhibitory modulation remain unclear. Here we bridge this gap by integrating key findings about retrieval suppression observed through functional imaging with a detailed consideration of relevant anatomical pathways observed in non-human primates. Focusing selectively on the potential role of the anterior cingulate cortex, we develop two hypotheses about the pathways mediating interactions between lateral prefrontal cortex and the medial temporal lobes during suppression, and their cellular targets: the entorhinal gating hypothesis, and thalamo-hippocampal modulation via the nucleus reuniens. We hypothesize that whereas entorhinal gating is well situated to stop retrieval proactively, thalamo-hippocampal modulation may interrupt an ongoing act of retrieval reactively. Isolating the pathways that underlie retrieval suppression holds the potential to advance our understanding of a range of psychiatric disorders characterized by persistent intrusive thoughts. More broadly, an anatomical account of retrieval suppression would provide a key model system for understanding inhibitory control over cognition.},
}
@book{Andersson88,
       author = {Andersson, Russell L.},
        title = {A Robot Ping-Pong Player: Experiment in Real-Time Intelligent Control},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1988,
}
@inproceedings{AndradeetalNIPS-16,
       author = {Daniel Andrade and Bing Bai and Ramkumar Rajendran and Yotaro Watanabe},
        title = {Analogy-Based Reasoning with Memory Networks for Future Prediction},
    booktitle = {Proceedings of the Workshop on Cognitive Computation: Integrating neural and symbolic approaches 2016 co-located with the 30th Annual Conference on Neural Information Processing Systems},
         year = {2016},
     abstract = {Making predictions about what might happen in the future is important for reacting adequately in many situations. For example, observing that "Man kidnaps girl" may have the consequence that "Man kills girl". While this is part of common sense reasoning for humans, it is not obvious how machines can learn and generalize over such knowledge automatically. The order of event's textual occurrence in documents offers a clue to acquire such knowledge automatically. Here, we explore another clue, namely, logical and temporal relations of verbs from lexical resources. We argue that it is possible to generalize to unseen events, by using the entailment relation between two events expressed as (subject, verb, object) triples. We formulate our hypotheses of analogy-based reasoning for future prediction, and propose a memory network that incorporates our hypotheses. Our evaluation for predicting the next future event shows that the proposed model can be competitive to (deep) neural networks and rankSVM, while giving interpretable answers.}
}
@article{AndreasetalCoRR-15,
       author = {Jacob Andreas and Marcus Rohrbach and Trevor Darrell and Dan Klein},
        title = {Deep Compositional Question Answering with Neural Module Networks},
      journal = {CoRR},
       volume = {arXiv:1511.02799},
         year = {2015},
     abstract = {Visual question answering is fundamentally compositional in nature---a question like "where is the dog?" shares substructure with questions like "what color is the dog?" and "where is the cat?" This paper seeks to simultaneously exploit the representational capacity of deep networks and the compositional linguistic structure of questions. We describe a procedure for constructing and learning *neural module networks*, which compose collections of jointly-trained neural "modules" into deep networks for question answering. Our approach decomposes questions into their linguistic substructures, and uses these structures to dynamically instantiate modular networks (with reusable components for recognizing dogs, classifying colors, etc.). The resulting compound networks are jointly trained. We evaluate our approach on two challenging datasets for visual question answering, achieving state-of-the-art results on both the VQA natural image dataset and a new dataset of complex questions about abstract shapes.}
}
@article{AndreasetalICML-17,
       author = {Jacob Andreas and Dan Klein and Sergey Levine},
        title = {Modular Multitask Reinforcement Learning with Policy Sketches},
      journal = {CoRR},
       volume = {arXiv:1611.01796},
         year = {2016},
     abstract = {We describe a framework for multitask deep reinforcement learning guided by policy sketches. Sketches annotate tasks with sequences of named subtasks, providing information about high-level structural relationships among tasks but not how to implement them---specifically not providing the detailed guidance used by much previous work on learning policy abstractions for RL (e.g. intermediate rewards, subtask completion signals, or intrinsic motivations). To learn from sketches, we present a model that associates every subtask with a modular subpolicy, and jointly maximizes reward over full task-specific policies by tying parameters across shared subpolicies. Optimization is accomplished via a decoupled actor--critic training objective that facilitates learning common behaviors from multiple dissimilar reward functions. We evaluate the effectiveness of our approach in three environments featuring both discrete and continuous control, and with sparse rewards that can be obtained only after completing a number of high-level subgoals. Experiments show that using our approach to learn policies guided by sketches gives better performance than existing techniques for learning task-specific or shared policies, while naturally inducing a library of interpretable primitive behaviors that can be recombined to rapidly adapt to new tasks.}
}
@article{Andreassenetal-94,
       author = {S. Andreassen and J. Benn and R. Hovorka and K.G. Olesen and E.R. Carson},
        title = {A probabilistic approach to glucose prediction and insulin dose adjustment: description of metabolic model and pilot evaluation study},
      journal = {Computer Methods and Programs in Biomedicine},
        pages = {153-165},
         year = 1994,
}
@inproceedings{AndreassenetalAAAI-87,
       author = {Andreassen, S. and Woldbye, M. and Falck, B. and Andersen, S.},
        title = {MUNIN: A Causal Probabilistic Network for Interpretion of Electromygraphic Findings},
    booktitle = {Proceedings AAAI-87},
     location = {Seattle, Washington},
 organization = {AAAI},
        pages = {121-124},
         year = 1987,
}
@article{AndresetalMIA-12,
       author = {Bj{\"o}rn Andres and Ullrich K{\"o}the and Thorben Kr{\"o}ger and Moritz Helmstaedter and Kevin L. Briggman and Winfried Denk and Fred A. Hamprecht},
        title = {{3D} segmentation of {SBFSEM} images of neuropil by a graphical model over supervoxel boundaries},
      journal = {Medical Image Analysis},
       volume = 16,
        issue = 4,
         year = 2012,
        pages = {796-805},
}
@inproceedings{AndrewsetalNIPS-03,
       author = {Stuart Andrews and Ioannis Tsochantaridis and Thomas Hofmann},
        title = {Support vector machines for multiple-instance learning},
    booktitle = {Advances in Neural Information Processing Systems 15},
    publisher = {MIT Press},
        pages = {561-568},
         year = 2003,
}
@article{AndrieuetalML-03,
       author = {Christophe Andrieu C. and Nando de Freitas Nando and Arnaud Doucet and Michael I. Jordan},
        title = {An Introduction to {MCMC} for Machine Learning},
      journal = {Machine Learning},
       volume = 50,
         year = 2003,
        pages = {5-43},
     abstract = {This purpose of this introductory paper is threefold. First, it introduces the Monte Carlo method with emphasis on probabilistic machine learning. Second, it reviews the main building blocks of modern Markov chain Monte Carlo simulation, thereby providing and introduction to the remaining papers of this special issue. Lastly, it discusses new interesting research horizons.},
}
@article{AndrychowiczetalCoRR-17,
       author = {Marcin Andrychowicz and Filip Wolski and Alex Ray and Jonas Schneider and Rachel Fong and Peter Welinder and Bob McGrew and Josh Tobin and Pieter Abbeel and Wojciech Zaremba},
        title = {Hindsight Experience Replay},
      journal = {CoRR},
       volume = {arXiv:1707.01495},
         year = {2017},
     abstract = {Dealing with sparse rewards is one of the biggest challenges in Reinforcement Learning (RL). We present a novel technique called Hindsight Experience Replay which allows sample-efficient learning from rewards which are sparse and binary and therefore avoid the need for complicated reward engineering. It can be combined with an arbitrary off-policy RL algorithm and may be seen as a form of implicit curriculum. We demonstrate our approach on the task of manipulating objects with a robotic arm. In particular, we run experiments on three different tasks: pushing, sliding, and pick-and-place, in each case using only binary rewards indicating whether or not the task is completed. Our ablation studies show that Hindsight Experience Replay is a crucial ingredient which makes training possible in these challenging environments. We show that our policies trained on a physics simulation can be deployed on a physical robot and successfully complete the task.}
}
@article{AngeliandManningEMNLP-14,
        title = {NaturalLI: Natural logic inference for common sense reasoning},
       author = {Angeli, Gabor and Manning, Christopher D.},
      journal = {Proceedings of the Conference on Empirical Methods in Natural Language Processing},
         year = 2014,
     abstract = {Common-sense reasoning is important for AI applications, both in NLP and many vision and robotics tasks. We propose NaturalLI: a Natural Logic inference system for inferring common sense facts - for instance, that cats have tails or tomatoes are round - from a very large database of known facts. In addition to being able to provide strictly valid derivations, the system is also able to produce derivations which are only likely valid, accompanied by an associated confidence. We both show that our system is able to capture strict Natural Logic inferences on the FraCaS test suite, and demonstrate its ability to predict common sense facts with 49\% recall and 91\% precision.}
}
@unpublished{AngeliandManningNATLOG-14,
         note = {Tutorial at Conference on Empirical Methods in Natural Language Processing},
       author = {Angeli, Gabor and Manning, Christopher D.},
        title = {Natural logic inference for common sense reasoning},
          url = {http://emnlp2014.org/material/presentation-EMNLP2014059.pdf},
         year = 2014,
        month = {September},
}
@article{AngelucciandBressloff06,
       author = {J. S. Lund and A. Angelucci and P. C. Bressloff},
        title = {The contribution of feedforward, lateral and feedback connections to the classical receptive field center and extra-classical receptive field surround of primate {V1} neurons},
      journal = {Prog. Brain Ressearch},
       volume = 154,
        pages = {93-121},
         year = 2006,
}
@article{AngluinIC-78,
       author = {Angluin, Dana},
        title = {On the Complexity of Minimum Inference of Regular Sets},
      journal = {Information and Control},
       volume = 39,
        pages = {337-350},
         year = 1978,
}
@article{AngluinIC-87,
       author = {Angluin, Dana},
        title = {Learning Regular Sets from Queries and Counterexamples},
      journal = {Information and Computation},
       volume = 75,
         year = 1987,
        pages = {87-106},
     abstract = {The problem of identifying an unknown regular set from examples of its members and nonmembers is addressed. It is assumed that the regular set is presented by a minimally adequate Teacher, which can answer membership queries about the set and can also test a conjecture and indicate whether it is equal to the unknown set and provide a counterexample if not. (A counterexample is a string in the symmetric difference of the correct set and the conjectured set.) A learning algorithm L* is described that correctly learns any regular set from any minimally adequate Teacher in time polynomial in the number of states of the minimum dfa for the set and the maximum length of any counterexample provided by the Teacher. It is shown that in a stochastic setting the ability of the Teacher to test conjectures may be replaced by a random sampling oracle, EX( ). A polynomial-time learning algorithm is shown for a particular problem of context-free language identification.}
}
@article{AnonymousICLR-18a,
        title = {Neural Program Search: Solving Programming Tasks from Description and Examples},
       author = {Anonymous},
      journal = {International Conference on Learning Representations (Accepted)},
         year = {2018},
          url = {https://openreview.net/forum?id=B1KJJf-R-},
     abstract = {We present a Neural Program Search, an algorithm to generate programs from natural language description and a small number of input / output examples. The algorithm combines methods from Deep Learning and Program Synthesis fields by designing rich domain-specific language (DSL) and defining efficient search algorithm guided by a Seq2Tree model on it. To evaluate the quality of the approach we also present a semi-synthetic dataset of descriptions with test examples and corresponding programs. We show that our algorithm significantly outperforms sequence-to-sequence model with attention baseline.}
}
@article{AnonymousICLR-18c,
        title = {Integrating Episodic Memory into a Reinforcement Learning Agent Using Reservoir Sampling},
       author = {Anonymous},
      journal = {Submitted to International Conference on Learning Representations},
         year = {2018},
          url = {https://openreview.net/forum?id=ByJDAIe0b},
     abstract = {Episodic memory is a psychology term which refers to the ability to recall specific events from the past. We suggest one advantage of this particular type of memory is the ability to easily assign credit to a specific state when remembered information is found to be useful. Inspired by this idea, and the increasing popularity of external memory mechanisms to handle long-term dependencies in deep learning systems, we propose a novel algorithm which uses a reservoir sampling procedure to maintain an external memory consisting of a fixed number of past states. The algorithm allows a deep reinforcement learning agent to learn online to preferentially remember those states which are found to be useful to recall later on. Critically this method allows for efficient online computation of gradient estimates with respect to the write process of the external memory. Thus unlike most prior mechanisms for external memory it is feasible to use in an online reinforcement learning setting.}
}
@article{AnonymousICLR-18d,
        title = {Memory Architectures in Recurrent Neural Network Language Models},
       author = {Anonymous},
      journal = {International Conference on Learning Representations},
         year = {2018},
          url = {https://openreview.net/forum?id=SkFqf0lAZ},
     abstract = {We compare and analyze sequential, random access, and stack memory architectures for recurrent neural network language models. Our experiments on the Penn Treebank and Wikitext-2 datasets show that stack-based memory architectures consistently achieve the best performance in terms of held out perplexity. We also propose a generalization to existing continuous stack models (Joulin \& Mikolov,2015; Grefenstette et al., 2015)  to allow a variable number of pop operations more naturally that further improves performance. We further evaluate these language models in terms of their ability to capture non-local syntactic dependencies on a subject-verb agreement dataset  (Linzen et al., 2016) and establish new state of the art results using memory augmented language models. Our results demonstrate the value of stack-structured memory for explaining the distribution of words in natural language, in line with linguistic theories claiming a context-free backbone for natural language.}
}    
@article{AnonymousICLR-18e,
        title = {Memory-based Parameter Adaptation},
       author = {Anonymous},
      journal = {International Conference on Learning Representations (Accepted)},
         year = {2018},
          url = {https://openreview.net/forum?id=rkfOvGbCW},
     abstract = {Deep neural networks have excelled on a wide range of problems, from vision to language and game playing. Neural networks very gradually incorporate information into weights as they process data, requiring very low learning rates. If thetraining distribution shifts, the network is slow to adapt, and when it does adapt, ittypically performs badly on the training distribution before the shift. Our method, Memory-based Parameter Adaptation, stores examples in memory and then usesa context-based lookup to directly modify the weights of a neural network. Much higher learning rates can be used for this local adaptation, reneging the need formany iterations over similar data before good predictions can be made. As our method is memory-based, it alleviates several shortcomings of neural networks, such as catastrophic forgetting, fast, stable acquisition of new knowledge, learning with an imbalanced class labels, and fast learning during evaluation. We demonstrate this on a range of supervised tasks: large-scale image classification andlanguage modelling.}
}
@article{Ansaklisetal89,
       author = {Ansaklis, P. J. and Passino, K. M. and Wang, S. J.},
        title = {Towards Intelligent Control Systems: Architectures and Fundamental Issues},
      journal = {Journal of Intelligent and Robotic Systems},
       volume = 1,
         year = 1989,
        pages = {315-342},
}
@article{AnsetalCSS-02,
       author = {Ans, Bernard and Rousset, St\`{e}phane and French, Robert and Musca, Serban},
        title = {Preventing Catastrophic Interference in Multiple-Sequence Learning Using Coupled Reverberating Elman Networks},
         year = {2002},
        pages = {},
      journal = {Proceedings of the 24th Annual Meeting of the Cognitive Science Society},
     abstract = {Everyone agrees that real cognition requires much more than static pattern recognition. In particular, it requires the ability to learn sequences of patterns (or actions) But learning sequences really means being able to learn multiple sequences, one after the other, without the most recently learned ones erasing the previously learned ones. But if catastrophic interference is a problem for the sequential learning of individual patterns, the problem is amplified many times over when multiple sequences of patterns have to be learned consecutively, because each new sequence consists of many linked patterns. In this paper we will present a connectionist architecture that would seem to solve the problem of multiple sequence learning using pseudopatterns.},
}
@article{AokietalNEURON-13,
       author = {Tazu Aoki and Masae Kinoshita and Ryo Aoki and Masakazu Agetsuma and Hidenori Aizawa and Masako Yamazaki and Mikako Takahoko and Ryunosuke Amo and Akiko Arata and Shin-ichi Higashijima and Takashi Tsuboi and Hitoshi Okamoto},
        title = {Imaging of Neural Ensemble for the Retrieval of a Learned Behavioral Program},
      journal = {Neuron},
       volume = {78},
       number = {5},
         year = {2013},
        pages = {881-894},
     abstract = {The encoding of long-term associative memories for learned behaviors is a fundamental brain function. Yet, how behavior is stably consolidated and retrieved in the vertebrate cortex is poorly understood. We trained zebrafish in aversive reinforcement learning and measured calcium signals across their entire brain during retrieval of the learned response. A discrete area of dorsal telencephalon that was inactive immediately after training became active the next day. Analysis of the identified area indicated that it was specific and essential for long-term memory retrieval and contained electrophysiological responses entrained to the learning stimulus. When the behavioral rule changed, a rapid spatial shift in the functional map across the telencephalon was observed. These results demonstrate that the retrieval of long-term memories for learned behaviors can be studied at the whole-brain scale in behaving zebrafish in vivo. Moreover, the findings indicate that consolidated memory traces can be rapidly modified during reinforcement learning.}
}
@article{ApostolicoandBejeranoCB-00,
       author = {Alberto Apostolico and Gill Bejerano},
        title = {Optimal amnesic probabilistic automata, or, how to learn and classify proteins in linear time and space},
      journal = {Computational Biology},
       volume = 7,
        issue = {3-4},
        pages = {381-393},
         year = 2000,
}
@article{AppelandHaken76,
       author = {Appel, K. I. and Haken, W.},
        title = {Every Planar Map is Four Colorable},
      journal = {Bulletin of the American Mathematical Society},
       volume = 82,
        pages = {711-712},
         year = 1976,
}
@inproceedings{AppuswamyetalCIDR-19,
       author = {Raja Appuswamy and Kevin Le Brigand and Pascal Barbry and Marc Antonini and Olivier Madderson and Paul Freemont and James McDonald and Thomas Heinis},
        title = {OligoArchive: Using {DNA} in the {DBMS} storage hierarchy},
    booktitle = {Proceedings of the 8th Conference on Innovative Data Systems Research (CIDR)},
     location = {Asilomar, CA},
         year = {2019},
     abstract = {The demand for data-driven decision making coupled with need to retain data to meet regulatory compliance requirements has resulted in a rapid increase in the amount of archival data stored by enterprises. As data generation rate far outpaces the rate of improvement in storage density of media like HDD and tape, researchers have started investigating new architectures and media types that can store such "cold", infrequently accessed data at very low cost. Synthetic DNA is one such storage media that has received some attention recently due to its high density and durability. In this paper, we investigate the problem of integrating DNA in the database storage hierarchy. More specifically, we ask the following two questions: (i) how can database knowledge help optimize DNA encoding and decoding? and (ii) how can biochemical mechanisms used for DNA manipulation be used to perform in-vitro, near-data SQL query processing? In answering these questions, we present OligoArchive, an architecture for using DNA-based storage system as the archival tier of a relational database. We demonstrate that OligoArchive can be realized in practice by building archiving and recovery tools (pg_oligo_dump and pg_oligo_restore) for PostgreSQL that perform schema-aware encoding and decoding of relational data on DNA, and using these tools to archive a 12KB TPC-H database to DNA, perform in-vitro computation, and restore it back again.}
}
@article{AptJACM-82,
       author = {Apt, Krzysztof and {van Emden}, M. H.},
        title = {Contributions to the Theory of Logic Programming},
      journal = {Journal of the {ACM}},
       volume = 29,
         year = 1982,
        pages = {841-862},
}
@article{ApthorpeetalCoRR-16,
        title = {Automatic Neuron Detection in Calcium Imaging Data Using Convolutional Networks},
       author = {Noah J. Apthorpe and Alexander J. Riordan and Rob E. Aguilar and Jan Homann and Yi Gu and David W. Tank and H. Sebastian Seung},
      journal = {CoRR},
       volume = {arXiv:1606.07372},
         year = {2016},
     abstract = {Calcium imaging is an important technique for monitoring the activity of thousands of neurons simultaneously. As calcium imaging datasets grow in size, automated detection of individual neurons is becoming important. Here we apply a supervised learning approach to this problem and show that convolutional networks can achieve near-human accuracy and superhuman speed. Accuracy is superior to the popular PCA/ICA method based on precision and recall relative to ground truth annotation by a human expert. These results suggest that convolutional networks are an efficient and flexible tool for the analysis of large-scale calcium imaging data.}
}
@inproceedings{AradhyeetalICDM-09,
       author = {Hrishikesh Aradhye and George Toderici and Jay Yagnik},
        title = {{Video2Text}: Learning to Annotate Video Content},
    booktitle = {{ICDM} Workshops 2009 {IEEE} International Conference on Data Mining},
         year = 2009,
        pages = {144-151},
}
@article{ArdeschetalPNAS-19,
       author = {Ardesch, Dirk Jan and Scholtens, Lianne H. and Li, Longchuan and Preuss, Todd M. and Rilling, James K. and van den Heuvel, Martijn P.},
        title = {Evolutionary expansion of connectivity between multimodal association areas in the human brain compared with chimpanzees},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {116},
       number = {14},
         year = {2019},
        pages = {7101-7106},
     abstract = {Comparative connectomics provides a powerful framework for studying cross-species differences in brain network architecture, offering important insights into the origin of human brain function. The present study highlights key differences between the human and chimpanzee connectome that have arisen since the divergence from our last common ancestor. Comparative analysis suggests an evolutionary shift in the human connectome toward investment of neural resources in global integration of multimodal information and enhanced functional specialization, potentially supporting the enhancement of complex cognitive function during human evolution. Identification of human connectome adaptations has broad implications for our fundamental understanding of human brain function and may contribute to our knowledge of human-specific mental disorders that involve macroscale changes to the brain's wiring architecture. The development of complex cognitive functions during human evolution coincides with pronounced encephalization and expansion of white matter, the brain's infrastructure for region-to-region communication. We investigated adaptations of the human macroscale brain network by comparing human brain wiring with that of the chimpanzee, one of our closest living primate relatives. White matter connectivity networks were reconstructed using diffusion-weighted MRI in humans (n = 57) and chimpanzees (n = 20) and then analyzed using network neuroscience tools. We demonstrate higher network centrality of connections linking multimodal association areas in humans compared with chimpanzees, together with a more pronounced modular topology of the human connectome. Furthermore, connections observed in humans but not in chimpanzees particularly link multimodal areas of the temporal, lateral parietal, and inferior frontal cortices, including tracts important for language processing. Network analysis demonstrates a particularly high contribution of these connections to global network integration in the human brain. Taken together, our comparative connectome findings suggest an evolutionary shift in the human brain toward investment of neural resources in multimodal connectivity facilitating neural integration, combined with an increase in language-related connectivity supporting functional specialization.},
}
@book{Arenkiel2014neural,
        title = {Neural Tracing Methods: Tracing Neurons and Their Connections},
       author = {Arenkiel, B.R.},
         year = {2014},
    publisher = {Springer New York}
}
@incollection{Arentoftetletal87,
       author = {Aarup, M. and Arentoft, M. M. and Parrod, Y. and Stader, J. and Stokes, I. and Vadon, H.},
        title = {{OPTIMUM-AIV}: A Knowledge Based Planning and Scheduling System for Spacecraft {AIV}},
       editor = {Fox, M. and Zweben, M.},
    booktitle = {Knowledge Based Scheduling},
    publisher = {Morgan Kaufmann},
         year = 1987,
}
@article {ArielandRyanPHYSIOLOGY-12,
       author = {Ariel, Pablo and Ryan, Timothy A.},
        title = {New Insights Into Molecular Players Involved in Neurotransmitter Release},
      journal = {Physiology},
    publisher = {American Physiological Society},
       volume = {27},
       number = {1},
         year = {2012},
        pages = {15-24},
     abstract = {The strength of a synapse can profoundly influence network function. How this strength is set at the molecular level is a key question in neuroscience. Here, we review a simple model of neurotransmission that serves as a convenient framework to discuss recent studies on RIM and synaptotagmin.},
}     
@inproceedings{ArisoyetalNAACL-HLT-12,
     title = {Deep Neural Network Language Models},
    author = {Arisoy, Ebru and Sainath, Tara N. and Kingsbury, Brian and Ramabhadran, Bhuvana},
 booktitle = {Proceedings of the {NAACL-HLT} 2012 Workshop: Will We Ever Really Replace the {N}-gram Model? On the Future of Language Modeling for {HLT}},
 publisher = {Association for Computational Linguistics},
      year = 2012,
     pages = {20-28},
}
@inproceedings{ArkinIEEE-87,
       author = {Arkin, Ronald R.},
        title = {Motor Schema Based Navigation for a Mobile Robot: An Approach to Programming by Behavior},
    booktitle = {Proceedings of the {IEEE} International Conference on Robotics and Automation},
         year = 1987,
        pages = {264-271},
}
@article{ArmananzasandAscoliTiN-15,
       author = {Rub\'{e}n Arma\~{n}anzas and Giorgio A. Ascoli},
        title = {Towards the automatic classification of neurons},
      journal = {Trends in Neurosciences},
       volume = {38},
       number = {5},
        pages = {307-318},
         year = {2015},
     abstract = {The classification of neurons into types has been much debated since the inception of modern neuroscience. Recent experimental advances are accelerating the pace of data collection. The resulting growth of information about morphological, physiological, and molecular properties encourages efforts to automate neuronal classification by powerful machine learning techniques. We review state-of-the-art analysis approaches and the availability of suitable data and resources, highlighting prominent challenges and opportunities. The effective solution of the neuronal classification problem will require continuous development of computational methods, high-throughput data production, and systematic metadata organization to enable cross-laboratory integration.}
}
@book{ArnoldandGosling98,
       author = {Ken Arnold and James Gosling},
        title = {The Java Programming Language},
    publisher = {Second edition, Addison-Wesley},
      address = {Reading, MA},
         year = 1998,
}
@article{AronetalJOURNAL-NEUROSCIENCE-16,
       author = {Aron, Adam R. and Herz, Damian M. and Brown, Peter and Forstmann, Birte U. and Zaghloul, Kareem},
        title = {Frontosubthalamic Circuits for Control of Action and Cognition},
      journal = {Journal of Neuroscience},
       volume = {36},
       number = {45},
         year = {2016},
        pages = {11489-11495},
    publisher = {Society for Neuroscience},
     abstract = {The subthalamic nucleus (STN) of the basal ganglia appears to have a potent role in action and cognition. Anatomical and imaging studies show that different frontal cortical areas directly project to the STN via so-called hyperdirect pathways. This review reports some of the latest findings about such circuits, including simultaneous recordings from cortex and the STN in humans, single-unit recordings in humans, high-resolution fMRI, and neurocomputational modeling. We argue that a major function of the STN is to broadly pause behavior and cognition when stop signals, conflict signals, or surprise signals occur, and that the fronto-STN circuits for doing this, at least for stopping and conflict, are dissociable anatomically and in terms of their spectral reactivity. We also highlight recent evidence for synchronization of oscillations between prefrontal cortex and the STN, which may provide a preferential "window in time" for single neuron communication via long-range connections.},
}
@article{AronovetalNATURE-17,
       author = {Aronov, Dmitriy and Nevers, Rhino and Tank, David W.},
        title = {Mapping of a non-spatial dimension by the hippocampal-entorhinal circuit},
      journal = {Nature},
       volume = {543},
        issue = {7647},
         year = {2017},
        pages = {719-722},
     abstract = {During spatial navigation, neural activity in the hippocampus and the medial entorhinal cortex (MEC) is correlated to navigational variables such as location, head direction, speed, and proximity to boundaries. These activity patterns are thought to provide a map-like representation of physical space. However, the hippocampal-entorhinal circuit is involved not only in spatial navigation, but also in a variety of memory-guided behaviours. The relationship between this general function and the specialized spatial activity patterns is unclear. A conceptual framework reconciling these views is that spatial representation is just one example of a more general mechanism for encoding continuous, task-relevant variables. Here we tested this idea by recording from hippocampal and entorhinal neurons during a task that required rats to use a joystick to manipulate sound along a continuous frequency axis. We found neural representation of the entire behavioural task, including activity that formed discrete firing fields at particular sound frequencies. Neurons involved in this representation overlapped with the known spatial cell types in the circuit, such as place cells and grid cells. These results suggest that common circuit mechanisms in the hippocampal-entorhinal system are used to represent diverse behavioural tasks, possibly supporting cognitive processes beyond spatial navigation.},
}
@article{AroraetalCoRR-15,
        title = {Random Walks on Context Spaces: Towards an Explanation of the Mysteries of Semantic Word Embeddings},
       author = {Sanjeev Arora and Yuanzhi Li and Yingyu Liang and Tengyu Ma and Andrej Risteski},
      journal = {CoRR},
       volume = {arXiv:1502.03520},
         year = 2015,
     abstract = {The papers of Mikolov et al. 2013 as well as subsequent works have led to dramatic progress in solving word analogy tasks using semantic word embeddings. This leverages linear structure that is often found in the word embeddings, which is surprising since the training method is usually nonlinear. There were attempts --notably by Levy and Goldberg and Pennington et al.-- to explain how this linear structure arises. The current paper points out the gaps in these explanations and provides a more complete explanation using a loglinear generative model for the corpus that directly models the latent semantic structure in words. The novel methodological twist is that instead of trying to fit the best model parameters to the data, a rigorous mathematical analysis is performed using the model priors to arrive at a simple closed form expression that approximately relates co-occurrence statistics and word embeddings. This expression closely corresponds to --and a bit simpler than-- the existing training methods, and leads to good solutions to analogy tasks. Empirical support is provided also for the validity of the modeling assumptions. This methodology of letting some mathematical analysis substitute for some of the computational difficulty may be useful in other settings with generative models.},
}
@inproceedings{AroraetalPMLR-15,
        title = {Simple, Efficient, and Neural Algorithms for Sparse Coding},
       author = {Sanjeev Arora and Rong Ge and Tengyu Ma and Ankur Moitra},
    booktitle = {Proceedings of The 28th Conference on Learning Theory},
       editor = {Peter Gr\"{u}nwald and Elad Hazan and Satyen Kale},
    publisher = {PMLR},
       volume = {40},
         year = {2015},
        pages = {113-149},
     abstract = {Sparse coding is a basic task in many fields including signal processing, neuroscience and machine learning where the goal is to learn a basis that enables a sparse representation of a given set of data, if one exists. Its standard formulation is as a non-convex optimization problem which is solved in practice by heuristics based on alternating minimization.  Recent work has resulted in several algorithms for sparse coding with provable guarantees, but somewhat surprisingly these are outperformed by the simple alternating minimization heuristics. Here we give a general framework for understanding alternating minimization which we leverage to analyze existing heuristics and to design new ones also with provable guarantees. Some of these algorithms seem implementable on simple neural architectures, which was the original motivation of Olshausen and Field in introducing sparse coding. We also give the first efficient algorithm for sparse coding that works almost up to the information theoretic limit for sparse recovery on incoherent dictionaries. All previous algorithms that approached or surpassed this limit run in time exponential in some natural parameter. Finally, our algorithms improve upon the sample complexity of existing approaches. We believe that our analysis framework will have applications in other settings where simple iterative algorithms are used}
}
@book{Artificial_Intelligence_Index_Annual_Report_2017,
        title = {Artificial Intelligence Index Annual Report},
       author = {Yoav Shoham and Raymond Perrault and Erik Brynjolfsson and Jack Clark},
      comment = {Includes statistics on the number of attendees at IJCAI conferences},
         year = {2017},
    publisher = {Stanford University},
}
@article{ArumugametalCoRR-17,
       author = {Dilip Arumugam and Siddharth Karamcheti and Nakul Gopalan and Lawson L. S. Wong and Stefanie Tellex},
        title = {Accurately and Efficiently Interpreting Human-Robot Instructions of Varying Granularities},
      journal = {CoRR},
       volume = {arXiv:1704.06616},
         year = {2017},
     abstract = {Humans can ground natural language commands to tasks at both abstract and fine-grained levels of specificity. For instance, a human forklift operator can be instructed to perform a high-level action, like "grab a pallet" or a low-level action like "tilt back a little bit." While robots are also capable of grounding language commands to tasks, previous methods implicitly assume that all commands and tasks reside at a single, fixed level of abstraction. Additionally, methods that do not use multiple levels of abstraction encounter inefficient planning and execution times as they solve tasks at a single level of abstraction with large, intractable state-action spaces closely resembling real world complexity. In this work, by grounding commands to all the tasks or subtasks available in a hierarchical planning framework, we arrive at a model capable of interpreting language at multiple levels of specificity ranging from coarse to more granular. We show that the accuracy of the grounding procedure is improved when simultaneously inferring the degree of abstraction in language used to communicate the task. Leveraging hierarchy also improves efficiency: our proposed approach enables a robot to respond to a command within one second on 90\% of our tasks, while baselines take over twenty seconds on half the tasks. Finally, we demonstrate that a real, physical robot can ground commands at multiple levels of abstraction allowing it to efficiently plan different subtasks within the same planning hierarchy.}
}

@inproceedings{AryaetalSODA-94,
    author = {S. Arya and D. Mount and N. Netanyahu and R. Silverman and A. Wu},
     title = {An optimal algorithm for approximate nearest neighbor searching},
 booktitle = {Proceedings of the 5th ACM-SIAM Symposium on Discrete Algorithms},
     pages = {573-582},
      year = 1994,
},
@article{AsanoetalSCIENCE-15,
       author = {Asano, Shoh and Fukuda, Yoshiyuki and Beck, Florian and Aufderheide, Antje and F\"{o}rster, Friedrich and Danev, Radostin and Baumeister, Wolfgang},
        title = {A molecular census of {26S} proteasomes in intact neurons},
      journal = {Science},
       volume = 347,
        issue = 6220,
        pages = {439-442},
         year = 2015,
     abstract = {The 26S proteasome is a key player in eukaryotic protein quality control and in the regulation of numerous cellular processes. Here, we describe quantitative in situ structural studies of this highly dynamic molecular machine in intact hippocampal neurons. We used electron cryotomography with the Volta phase plate, which allowed high fidelity and nanometer precision localization of 26S proteasomes. We undertook a molecular census of single- and double-capped proteasomes and assessed the conformational states of individual complexes. Under the conditions of the experiment--that is, in the absence of proteotoxic stress--only 20\% of the 26S proteasomes were engaged in substrate processing. The remainder was in the substrate-accepting ground state. These findings suggest that in the absence of stress, the capacity of the proteasome system is not fully used.},
}
@book{Ashby1957cybernetics,
        title = {An introduction to cybernetics},
       author = {Ashby, W. Ross},
       series = {University paperbacks},
         year = {1964},
    publisher = {Methuen}
}
@incollection{AshbyandValentinCATEGORIES-17,
        title = {Multiple Systems of Perceptual Category Learning: Theory and Cognitive Tests},
       author = {F. Gregory Ashby and Vivian V. Valentin},
    booktitle = {Handbook of Categorization in Cognitive Science (Second Edition)},
       editor = {Cohen, Henri and Lefebvre, Claire},
    publisher = {Elsevier},
      address = {San Diego},
         year = {2017},
        pages = {157-188},
     abstract = {The first neurobiologically detailed theory of multiple systems in category learning, called COVIS, was originally conceived in 1998. COVIS, which is now well established, postulates two systems that compete throughout learning--a frontal-based declarative system that uses logical reasoning and depends on working memory and executive attention, and a basal ganglia-mediated system that uses procedural-learning. The procedural system can learn a wide variety of category structures, but it learns in a slow incremental fashion and is highly dependent on reliable and immediate feedback. In contrast, the declarative rule-based (RB) system can learn a fairly small set of category structures quickly--specifically, those structures that can be learned via a logical reasoning process. These two systems learn simultaneously, but as long as \{RB\} strategies lead to successful performance, the declarative system inhibits the procedural system. This theory is described in detail and a variety of cognitive behavioral and cognitive neuroscience experiments are reviewed that test some parameter-free a priori predictions made by COVIS.}
}
@book{Ashworth82,
       author = {Ashworth, M. J.},
        title = {Feedback Design of Systems with Significant Uncertainty},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1982,
}
@inproceedings{AslamCOLT-90,
       author = {Aslam, Javed A. and Rivest, Ronald L.},
        title = {Inferring Graphs from Walks},
    booktitle = {Proceedings of the Third Annual {ACM} Conference on Computational
                Learning Theory},
         year = 1990,
}
@article{AsoetallELIFE-14,
       author = {Aso, Y. and Hattori, D. and Yu, Y. and Johnston, R. M. and Iyer, N. A. and Ngo, T. T. and Dionne, H. and Abbott, L. F. and Axel, R. and Tanimoto, H.  nd Rubin, G. M.},
        title = {The neuronal architecture of the mushroom body provides a logic for associative learning},
      journal = {Elife},
       volume = {3},
         year = {2014},
        pages = {e04577},
     abstract = {We identified the neurons comprising the Drosophila mushroom body (MB), an associative center in invertebrate brains, and provide a comprehensive map describing their potential connections. Each of the 21 MB output neuron (MBON) types elaborates segregated dendritic arbors along the parallel axons of approximately 2000 Kenyon cells, forming 15 compartments that collectively tile the MB lobes. MBON axons project to five discrete neuropils outside of the MB and three MBON types form a feedforward network in the lobes. Each of the 20 dopaminergic neuron (DAN) types projects axons to one, or at most two, of the MBON compartments. Convergence of DAN axons on compartmentalized Kenyon cell-MBON synapses creates a highly ordered unit that can support learning to impose valence on sensory representations. The elucidation of the complement of neurons of the MB provides a comprehensive anatomical substrate from which one can infer a functional logic of associative olfactory learning and memory.}
}
@article{AsoulinGLOSSA-16,
       author = {Eran Asoulin},
        title = {Language as an Instrument of Thought},
      journal = {Glossa: {A} Journal of General Linguistics},
       number = {1},
         year = {2016},
       volume = {1},
        pages = {1-23},
}
@article{AsplundetalPC-10,
       author = {Asplund, Maria and Nyberg, Tobias and Ingan\"{a}s, Olle},
        title = {Electroactive polymers for neural interfaces},
      journal = {Polymer Chemistry},
       volume = 1,
        issue = 9,
         year = 2010,
        pages = {1374-1391},
    publisher = {The Royal Society of Chemistry},
     abstract = {Development of electroactive conjugated polymers, for the purpose of recording and eliciting signals in the neural systems in humans, can be used to fashion the interfaces between the two signalling systems of electronics and neural systems. The design of desirable chemical, mechanical and electrical properties in the electroactive polymer electrodes, and the means of integration of these into biological systems, are here reviewed.}
}
@article{AsricanetalFiNC-13,
       author = {Asrican, Brent and Augustine, George J. and Berglund, Ken and Chen, Susu and Chow, Nick and Deisseroth, Karl and Feng, Guoping and Gloss, Bernd and Hira, Riichiro and Hoffmann, Carolin and Kasai, Haruo and Katarya, Malvika and Kim, Jinsook and Kudolo, John and Lee, Li Ming and Lo, Shun Qiang and Mancuso, James and Matsuzaki, Masanori and Nakajima, Ryuichi and Qiu, Li and Tan, Gregory and Tang, Yanxia and Ting, Jonathan T. and Tsuda, Sachiko and Wen, Lei and Zhang, Xuying and Zhao, Shengli},
        title = {Next-generation transgenic mice for optogenetic analysis of neural circuits},
      journal = {Frontiers in Neural Circuits},
    publisher = {Frontiers Media S.A.},
       volume = {7},
         year = {2013},
        pages = {160-160},
     abstract = {Here we characterize several new lines of transgenic mice useful for optogenetic analysis of brain circuit function. These mice express optogenetic probes, such as enhanced halorhodopsin or several different versions of channelrhodopsins, behind various neuron-specific promoters. These mice permit photoinhibition or photostimulation both in vitro and in vivo. Our results also reveal the important influence of fluorescent tags on optogenetic probe expression and function in transgenic mice.}
}
@book{Astrom89,
       author = {{\AA}str\"{o}m, Karl J. and Wittenmark, Bj\"{o}rn},
        title = {Adaptive Control},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1989,
}
@article{AtanasijevicetalPNAS-06,
        title = {Calcium-sensitive {MRI} contrast agents based on superparamagnetic iron oxide nanoparticles and calmodulin},
       author = {Atanasijevic, T. and Shusteff, M. and Fam, P. and Jasanoff, A.},
      journal = {Proceedings of the National Academy of Science},
       volume = 103,
        issue = 40,
         year = 2006,
        pages = {14707-14712},
     abstract = {We describe a family of calcium indicators for magnetic resonance imaging (MRI), formed by combining a powerful iron oxide nanoparticle-based contrast mechanism with the versatile calcium-sensing protein calmodulin and its targets. Calcium-dependent protein-protein interactions drive particle clustering and produce up to 5-fold changes in T2 relaxivity, an indication of the sensors' potency. A variant based on conjugates of wild-type calmodulin and the peptide M13 reports concentration changes near 1 microM Ca(2+), suitable for detection of elevated intracellular calcium levels. The midpoint and cooperativity of the response can be tuned by mutating the protein domains that actuate the sensor. Robust MRI signal changes are achieved even at nanomolar particle concentrations (&lt;1 microM in calmodulin) that are unlikely to buffer calcium levels. When combined with technologies for cellular delivery of nanoparticulate agents, these sensors and their derivatives may be useful for functional molecular imaging of biological signaling networks in live, opaque specimens.}
}
@article{AtanceandONeillLEARNING-AND-EMOTION-05,
       author = {Atance, Cristina and O'Neill, Daniela},
        title = {The emergence of episodic future thinking in humans},
      journal = {Learning and Motivation},
       volume = {36},
         year = {2005},
        pages = {126-144},
     abstract = {In this paper, we discuss the construct of episodic future thinking. We have previously defined episodic future thinking as the ability to project oneself into the future to pre-experience an event (Atance \& O'Neill, 2001). We distinguish this type of thinking about the future from that which is largely based on a script of how an event routinely unfolds (e.g., a restaurant or birthday party script). This distinction is related to the episodic/semantic distinction that has been applied to memory (Tulving, 1972). We discuss tasks, both verbal and nonverbal, that we have developed for young children, and that we believe assess episodic future thinking. Based on our findings from these tasks, we conclude that episodic future thinking emerges between 3 and 4 years of age. Throughout the paper, we attempt to specify the nature of the projection associated with episodic future thinking by elaborating upon how children's behavior in the tasks we discuss, as well as their behaviors in other contexts, are a reflection of this projection.}
}
@article{AtencioandSchreinerPLoS-10,
       author = {Atencio, C. A. and Schreiner, C. E.},
        title = {Columnar connectivity and laminar processing in cat primary auditory cortex},
      journal = {{PLoS} ONE},
         year = 2010,
       volume = 5,
        issue = 3,
        pages = {e9521},
     abstract = {Background: Radial intra- and interlaminar connections form a basic microcircuit in primary auditory cortex (AI) that extracts acoustic information and distributes it to cortical and subcortical networks. Though the structure of this microcircuit is known, we do not know how the functional connectivity between layers relates to laminar processing. Methodology \& Findings: We studied the relationships between functional connectivity and receptive field properties in this columnar microcircuit by simultaneously recording from single neurons in cat AI in response to broadband dynamic moving ripple stimuli. We used spectrotemporal receptive fields (STRFs) to estimate the relationship between receptive field parameters and the functional connectivity between pairs of neurons. Interlaminar connectivity obtained through cross-covariance analysis reflected a consistent pattern of information flow from thalamic input layers to cortical output layers. Connection strength and STRF similarity were greatest for intralaminar neuron pairs and in supragranular layers and weaker for interlaminar projections. Interlaminar connection strength co-varied with several STRF parameters: feature selectivity, phase locking to the stimulus envelope, best temporal modulation frequency, and best spectral modulation frequency. Connectivity properties and receptive field relationships differed for vertical and horizontal connections. Conclusions/Significance: Thus, the mode of local processing in supragranular layers differs from that in infragranular layers. Therefore, specific connectivity patterns in the auditory cortex shape the flow of information and constrain how spectrotemporal processing transformations progress in the canonical columnar auditory microcircuit.},
}
@article{AthalyeetalCOiN-20,
       author = {Vivek R Athalye and Jose M Carmena and Rui M Costa},
        title = {Neural reinforcement: re-entering and refining neural dynamics leading to desirable outcomes},
      journal = {Current Opinion in Neurobiology},
       volume = {60},
         year = {2020},
        pages = {145-154},
     abstract = {How do organisms learn to do again, on-demand, a behavior that led to a desirable outcome? Dopamine-dependent cortico-striatal plasticity provides a framework for learning behavior's value, but it is less clear how it enables the brain to re-enter desired behaviors and refine them over time. Reinforcing behavior is achieved by re-entering and refining the neural patterns that produce it. We review studies using brain-machine interfaces which reveal that reinforcing cortical population activity requires cortico-basal ganglia circuits. Then, we propose a formal framework for how reinforcement in cortico-basal ganglia circuits acts on the neural dynamics of cortical populations. We propose two parallel mechanisms: i) fast reinforcement which selects the inputs that permit the re-entrance of the particular cortical population dynamics which naturally produced the desired behavior, and ii) slower reinforcement which leads to refinement of cortical population dynamics and more reliable production of neural trajectories driving skillful behavior on-demand.}
}
@book{Athans66,
       author = {Athans, Michael and Falb, Peter L.},
        title = {Optimal Control: An Introduction to the Theory and Its Applications},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1966,
}
@article{AtiyaandParlosNN-00,
       author = {Amir F. Atiya and Alexander G. Parlos},
        title = {New Results on Recurrent Network Training: Unifying the Algorithms and Accelerating Convergence},
      journal = {{IEEE Transactions on Neural Networks}},
       volume = 11,
        issue = 3,
         year = 2000,
        pages = {697-709}
}
@incollection{AtkinsonHUMAN-MEMORY-68,
        title = {Human Memory: {A} Proposed System and its Control Processes},
       author = {R.C. Atkinson and R.M. Shiffrin},
       editor = {Kenneth W. Spence and Janet Taylor Spence},
    booktitle = {Psychology of Learning and Motivation},
    publisher = {Academic Press},
       volume = {2},
        pages = {89-195},
         year = {1968},
     abstract = {This chapter presents a general theoretical framework of human memory and describes the results of a number of experiments designed to test specific models that can be derived from the overall theory. This general theoretical framework categorizes the memory system along two major dimensions. The first categorization distinguishes permanent, structural features of the system from control processes that can be readily modified or reprogrammed at the will of the subject. The second categorization divides memory into three structural components: the sensory register, the short-term store, and the long-term store. Incoming sensory information first enters the sensory register, where it resides for a very brief period of time, then decays and is lost. The short-term store is the subject's working memory; it receives selected inputs from the sensory register and also from long-term store. The chapter also discusses the control processes associated with the sensory register. The term control process refers to those processes that are not permanent features of memory, but are instead transient phenomena under the control of the subject; their appearance depends on several factors such as instructional set, the experimental task, and the past history of the subject.}
}
@article{AtkinsonetalCoRR-18,
       author = {Craig Atkinson and Brendan McCane and Lech Szymanski and Anthony Robins},
        title = {Pseudo-Rehearsal: Achieving Deep Reinforcement Learning without Catastrophic Forgetting},
      journal = {CoRR},
       volume = {arXiv:1812.02464},
         year = {2018},
     abstract = {Neural networks can achieve extraordinary results on a wide variety of tasks. However, when they attempt to sequentially learn a number of tasks, they tend to learn the new task while destructively forgetting previous tasks. One solution to this problem is pseudo-rehearsal, which involves learning the new task while rehearsing generated items representative of previous tasks. Our model combines pseudo-rehearsal with a deep generative model and a dual memory system, resulting in a method that does not demand additional storage requirements as the number of tasks increase. Our model iteratively learns three Atari 2600 games while retaining above human level performance on all three games and performing as well as a set of networks individually trained on the tasks. This result is achieved without revisiting or storing raw data from past tasks. Furthermore, previous state-of-the-art solutions demonstrate substantial forgetting compared to our model on these complex deep reinforcement learning tasks.}
}
@article{AttiasNC-99,
       author = {Hagai Attias},
        title = {Independent Factor Analysis},
      journal = {Neural Computation},
       volume = 11,
        issue = 4,
        pages = {803-851},
         year = 1999,
}
@article{AttwelletalNEURON-02,
       author = {Phillip J.E. Attwell and Samuel F. Cooke and Christopher H. Yeo},
        title = {Cerebellar Function in Consolidation of a Motor Memory},
      journal = {Neuron},
       volume = {34},
       number = {6},
         year = {2002},
        pages = {1011-1020},
     abstract = {Several forms of motor learning, including classical conditioning of the eyeblink and nictitating membrane response (NMR), are dependent upon the cerebellum, but it is not known how motor memories are stored within the cerebellar circuitry. Localized infusions of the GABAA agonist muscimol were used to target putative consolidation processes by producing reversible inactivations after NMR conditioning sessions. Posttraining inactivations of eyeblink control regions in cerebellar cortical lobule HVI completely prevented conditioning from developing over four sessions. In contrast, similar inactivations of eyeblink control regions in the cerebellar nuclei allowed conditioning to develop normally. These findings provide evidence that there are critical posttraining memory consolidation processes for eyeblink conditioning mediated by the cerebellar cortex.}
}	 
@article{AtwoodandCooperIN-96,
       author = {Atwood, H.L. and Cooper, R.L.},
        title = {Synaptic diversity and differentiation: Crustacean neuromuscular junctions},
      journal = {Invertebrate Neuroscience},
    publisher = {Springer-Verlag},
       volume = {1},
       number = {4},
         year = {1996},
        pages = {291-307},
     abstract = {Crustacean motor neurons exhibit a wide range of synaptic responses. Tonically active neurons generally produce small excitatory postsynaptic potentials (EPSPs) at low impulse frequencies, and are able to release much more transmitter as the impulse frequency increases. Phasic neurons typically generate large EPSPs in their target cells, but have less capability for frequency facilitation, and undergo synaptic depression during maintained activity. These differences depend in part upon the neuron's ongoing levels of activity; phasic neurons acquire physiological and morphological features of tonic neurons when their activity level is altered. Molecules responsible for adaptation to activity can be sought in single identified phasic neurons with current techniques. The fact that both phasic and tonic neurons innervate the same target muscle fibers is evidence for presynaptic determination of synaptic properties, but there is also evidence for postsynaptic determination of specific properties of different endings of a single neuron. The occurrence of high- and low-output endings of the same tonic motor neurons on different muscle fibers suggests a target-specific influence on synaptic properties. Structural variation of synapses on individual terminal varicosities leads to the hypothesis that individual synapses have different probabilities for release of transmitter. We hypothesize that structurally complex synapses have a higher probability for release than the less complex synapses. This provides an explanation for the larger quantal contents of 'high-output' terminals (where the proportion of complex synapses is higher), and also a mechanism for progressive recruitment of synapses during frequency facilitation.}
}
@article{AtwoodandTowsleyNIPS-16,
       author = {James Atwood and Don Towsley},
        title = {Diffusion-Convolutional Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1511.02136},
         year = {2015},
     abstract = {We present diffusion-convolutional neural networks (DCNNs), a new model for graph-structured data. Through the introduction of a diffusion-convolution operation, we show how diffusion-based representations can be learned from graph-structured data and used as an effective basis for node classification. DCNNs have several attractive qualities, including a latent representation for graphical data that is invariant under isomorphism, as well as polynomial-time prediction and learning that can be represented as tensor operations and efficiently implemented on the GPU. Through several experiments with real structured datasets, we demonstrate that DCNNs are able to outperform probabilistic relational models and kernel-on-graph methods at relational node classification tasks.}
}
@incollection{Avedyan1995cmac,
       author = {Aved'yan, Eduard},
       editor = {Mason, J. and Parks, P. C.},
        title = {The Cerebellar Model Articulation Controller ({CMAC})},
    booktitle = {Learning Systems},
         year = {1995},
    publisher = {Springer London},
      address = {London},
        pages = {110-119},
     abstract = {In 1975 Albus [1], [2] presented a new approach to the solution of the robot control problem by means of a controller he called a CMAC (Cerebellar Model Articulation Controller). An application of this approach may be found, for example, in [3]. Albus' approach was effectively implemented by Ers{\"u} and Millitzer in 1984 (see, for example [4], [5]), and they named their system AMS (Associate Memory System). AMS was implemented in the LERNAS adaptive system for control of dynamic objects [4].},
}
@article{AverbecketalNATURE-06,
       author = {Averbeck, B. B. and Latham, P. E. and Pouget, A.},
        title = {Neural correlations, population coding and computation},
      journal = {Nature Reviews Neuroscience},
       volume = {7},
       number = {5},
         year = {2006},
        pages = {358-366},
     abstract = {How the brain encodes information in population activity, and how it combines and manipulates that activity as it carries out computations, are questions that lie at the heart of systems neuroscience. During the past decade, with the advent of multi-electrode recording and improved theoretical models, these questions have begun to yield answers. However, a complete understanding of neuronal variability, and, in particular, how it affects population codes, is missing. This is because variability in the brain is typically correlated, and although the exact effects of these correlations are not known, it is known that they can be large. Here, we review studies that address the interaction between neuronal noise and population codes, and discuss their implications for population coding in general.}
}
@book{Axelrod84,
       author = {Robert Axelrod},
        title = {The Evolution of Cooperation},
    publisher = {Basic Books},
      address = {New York, NY},
         year = 1984,
}
@article{AycinenaSTANFORD-19,
        title = {Hierarchical Relational Reinforcement Learning},
       author = {Aycinena, Meg},
         year = {2019},
     abstract = {Experiments with relational reinforcement learning, a combination of reinforcement learning and inductive logic programming, are presented. This technique offers greater expressive power than that offered by traditional reinforcement learning. We use it to find general solutions in the blocks world domain. We discuss some difficulties associated with relational reinforcement learning, specifically its decreased effectiveness when presented with more complex problems. Finally, we explore ways in which hierarchical methods and learning subroutines may be able to overcome some of these weaknesses.}
}
@article{AzaretalICML-12,
        title = {On the Sample Complexity of Reinforcement Learning with a Generative Model},
       author = {Mohammad Gheshlaghi Azar and Remi Munos and Bert Kappen},
      comment = {Proceedings of the 29th International Conference on Machine Learning2012)},
      journal = {CoRR},
       volume = {arXiv:1206.6461},
         year = {2012},
     abstract = {We consider the problem of learning the optimal action-value function in the discounted-reward Markov decision processes (MDPs). We prove a new PAC bound on the sample-complexity of model-based value iteration algorithm in the presence of the generative model, which indicates that for an MDP with N state-action pairs and the discount factor \gamma\in[0,1) only O(N\log(N/\delta)/((1-\gamma)^3\epsilon^2)) samples are required to find an \epsilon-optimal estimation of the action-value function with the probability 1-\delta. We also prove a matching lower bound of \Theta (N\log(N/\delta)/((1-\gamma)^3\epsilon^2)) on the sample complexity of estimating the optimal action-value function by every RL algorithm. To the best of our knowledge, this is the first matching result on the sample complexity of estimating the optimal (action-) value function in which the upper bound matches the lower bound of RL in terms of N, \epsilon, \delta and 1/(1-\gamma). Also, both our lower bound and our upper bound significantly improve on the state-of-the-art in terms of 1/(1-\gamma).}
}
@article{AzevedoetalJCN-09,
       author = {Azevedo, Frederico A.C. and Carvalho, Ludmila R.B. and Grinberg, Lea T. and Farfel, Jos\'{e} Marcelo and Ferretti, Renata E.L. and Leite, Renata E.P. and Filho, Wilson Jacob and Lent, Roberto and Herculano-Houzel, Suzana},
        title = {Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain},
      journal = {The Journal of Comparative Neurology},
       volume = {513},
       number = {5},
         year = {2009},
        pages = {532-541},
}
@article{Azuma67,
       author = {K. Azuma},
        title = {Weighted sums of certain dependent random variables},
      journal = {Tohoku Mathematical Journal},
       volume = 19,
        pages = {357-367},
         year = 1967,
}
@unpublished{BGL,
       author = {Jeremy Siek and Lie-Quan Lee and Andrew Lumsdaine},
        title = {{BOOST} {G}raph {L}ibrary},
          url = {http://www.boost.org/},
         year = 2003,
}
@unpublished{BOW,
       author = {Andrew Kachites McCallum},
        title = {Bow: A toolkit for statistical language modeling, text retrieval, classification and clustering},
          url = {http://www.cs.cmu.edu/~mccallum/bow},
         year = 2002,
}
@article{BaacetalSCIENCE_REPORTS-12,
        title = {Carbon-Nanotube Optoacoustic Lens for Focused Ultrasound Generation and High-Precision Targeted Therapy},
       author = {Baac, Hyoung Won and Ok, Jong G. and Maxwell, Adam and Lee, Kyu-Tae and Chen, Yu-Chih and Hart, A. John and Xu, Zhen and Yoon, Euisik and Guo, L. Jay},
      journal = {Scientific Reports},
       volume = 2,
        issue = 989,
        pages = {1-8},
         year = 2012,
}
@incollection{BaandCaruanaNIPS-14,
        title = {Do Deep Nets Really Need to be Deep?},
       author = {Ba, Jimmy and Caruana, Rich},
    booktitle = {Advances in Neural Information Processing Systems 27},
       editor = {Z. Ghahramani and M. Welling and C. Cortes and N. D. Lawrence and K. Q. Weinberger},
        pages = {2654-2662},
         year = {2014},
    publisher = {Curran Associates, Inc.},
}
@book{Baars1988,
        title = {A cognitive theory of consciousness},
       author = {Baars, B. J.},
         year = {1988},
    publisher = {Cambridge University Press},
      address = {New York, NY},
}
@article{BaarsetalFiP-13,
       author = {Baars, Bernard and Franklin, Stan and Ramsy, Thomas},
        title = {Global Workspace Dynamics: Cortical "Binding and Propagation" Enables Conscious Contents},
      journal = {Frontiers in Psychology},
       volume = {4},
        pages = {200},
         year = {2013},
     abstract = {A global workspace (GW) is a functional hub of binding and propagation in a population of loosely coupled signaling elements. In computational applications, GW architectures recruit many distributed, specialized agents to cooperate in resolving focal ambiguities. In the brain, conscious experiences may reflect a GW function. For animals, the natural world is full of unpredictable dangers and opportunities, suggesting a general adaptive pressure for brains to resolve focal ambiguities quickly and accurately. GW theory aims to understand the differences between conscious and unconscious brain events. In humans and related species the cortico-thalamic (C-T) core is believed to underlie conscious aspects of perception, thinking, learning, feelings of knowing (FOK), felt emotions, visual imagery, working memory, and executive control. Alternative theoretical perspectives are also discussed. The C-T core has many anatomical hubs, but conscious percepts are unitary and internally consistent at any given moment. Over time, conscious contents constitute a very large, open set. {\it{This suggests that a brain-based GW capacity cannot be localized in a single anatomical hub. Rather, it should be sought in a functional hub - a dynamic capacity for binding and propagation of neural signals over multiple task-related networks, a kind of neuronal cloud computing.}} In this view, conscious contents can arise in any region of the C-T core when multiple input streams settle on a winner-take-all equilibrium. {\it{The resulting conscious gestalt may ignite an any-to-many broadcast, lasting ~100-200ms, and trigger widespread adaptation in previously established networks.}} To account for the great range of conscious contents over time, the theory suggests an open repertoire of binding coalitions that can broadcast via theta/gamma or alpha/gamma phase coupling, like radio channels competing for a narrow frequency band. {\it{Conscious moments are thought to hold only 1-4 unrelated items; this small focal capacity may be the biological price to pay for global access.}} Visuotopic maps in cortex specialize in features like color, retinal size, motion, object identity, and egocentric/allocentric framing, so that a binding coalition for the sight of a rolling billiard ball in nearby space may resonate among activity maps of LGN, V1-V4, MT, IT, as well as the dorsal stream. {\it{Spatiotopic activity maps can bind into coherent gestalts using adaptive resonance (reentry).}} Single neurons can join a dominant coalition by phase tuning to regional oscillations in the 4-12Hz range. Sensory percepts may bind and broadcast from posterior cortex, while non-sensory FOKs may involve prefrontal and frontotemporal areas. The anatomy and physiology of the hippocampal complex suggest a GW architecture as well. In the intact brain the hippocampal complex may support conscious event organization as well as episodic memory storage.}
}
@book{Bacchus90,
       author = {Bacchus, Fahiem},
        title = {Representing and Reasoning with Probabilistic Knowledge: A Logical Approach to Probabilities},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1990,
}
@incollection{BacchusetalKR-89,
       author = {Bacchus, Fahiem and Tenenberg, Josh and Koomen, Johannes A.},
        title = {A Non-Reified Temporal Logic},
     crossref = {KR-89},
        pages = {2-10},
}
@article{BachandJordanJLMR-03,
       author = {Francis Bach and Michael Jordan},
        title = {Learning spectral clustering, with application to speech separation},
      journal = {Journal of Machine Learning Research},
       volume = 7,
        pages = {1963-2001},
         year = 2006,
}
@incollection{BachandJordanNIPS-03,
       author = {Francis Bach and Michael Jordan},
        title = {Learning spectral clustering},
    booktitle = {Advances in Neural Information Processing Systems},
       editor = {Sebastian Thrun and Lawrence Saul},
         year = 2003,
}
@techreport{BachrachPhD,
       author = {Bachrach, Jonathan R.},
        title = {Connectionist Modeling and Control of Finite State Environments},
        issue = {92-6},
  institution = {University of Massachusetts at Amherst Department of Computer and Information Science},
         year = 1992,
}
@misc{BacigalupiWINDUP-STORIES-09,
        title = {Stories From the World of The Windup Girl},
       author = {Paolo Bacigalupi},
         year = {2009},
 howpublished = {Originally published in The Magazine of Fantasy \& Science Fiction}
}
@inproceedings{BackstromandKleinIJCAI-91,
       author = {B\"{a}ckstr\"{o}m, C. and Klein, I.},
        title = {Parallel Non-Binary Planning in Polynomial Time},
    booktitle = {Proceedings {IJCAI} 12},
     location = {Sydney, Australia},
 organization = {IJCAII},
         year = 1991,
        pages = {268-273},
}
@article{BadcocketalCABN-19,
       author = {Badcock, P. B.  and Friston, K. J.  and Ramstead, M. J. D.  and Ploeger, A.  and Hohwy, J.},
        title = {{{T}he hierarchically mechanistic mind: an evolutionary systems theory of the human brain, cognition, and behavior}},
      journal = {Cognitive Affect Behavior Neuroscience},
         year = {2019},
       volume = {19},
       number = {6},
        pages = {1319-1351},
      abstract = {The purpose of this review was to integrate leading paradigms in psychology and neuroscience with a theory of the embodied, situated human brain, called the Hierarchically Mechanistic Mind (HMM). The HMM describes the brain as a complex adaptive system that functions to minimize the entropy of our sensory and physical states via action-perception cycles generated by hierarchical neural dynamics. First, we review the extant literature on the hierarchical structure of the brain. Next, we derive the HMM from a broader evolutionary systems theory that explains neural structure and function in terms of dynamic interactions across four nested levels of biological causation (i.e., adaptation, phylogeny, ontogeny, and mechanism). We then describe how the HMM aligns with a global brain theory in neuroscience called the free-energy principle, leveraging this theory to mathematically formulate neural dynamics across hierarchical spatiotemporal scales. We conclude by exploring the implications of the HMM for psychological inquiry.}
}
@article{BaddeleyARP-12,
       author = {Baddeley, Alan},
        title = {Working Memory: Theories, Models, and Controversies},
      journal = {Annual Review of Psychology},
       volume = {63},
       number = {1},
        pages = {1-29},
         year = {2012},
     abstract = {I present an account of the origins and development of the multicomponent approach to working memory, making a distinction between the overall theoretical framework, which has remained relatively stable, and the attempts to build more specific models within this framework. I follow this with a brief discussion of alternative models and their relationship to the framework. I conclude with speculations on further developments and a comment on the value of attempting to apply models and theories beyond the laboratory studies on which they are typically based.}
}
@article{BaddeleyNETWORK-96,
      title = {Searching for filters with `interesting' output distributions: an uninteresting direction to explore?}, 
     author = {Roland Baddeley}, 
    journal = {Network: Computation in Neural Systems}, 
     volume = 7, 
      issue = 1, 
       year = 1996, 
      pages = {409-421}, 
}
@article{BaddeleyQJoEP-86,
       author = {Alan Baddeley},
        title = {Modularity, Mass-Action and Memory},
      journal = {The Quarterly Journal of Experimental Psychology Section A},
       volume = {38},
       number = {4},
         year = {1986},
        pages = {527-533},
     abstract = {Joaqu\'{i}n Fuster wrote in Profrontal Cortex that "Baddeley defined working memory as a temporary memory store for information needed to perform a task or to solve a problem in the short term. [...] In my view, working memory is indeed a form of attention: sustained attention focused on an executive cognit for the processing of prospective action".},
}
@incollection{BaddeleyWMA-93,
       author = {Alan Baddeley},
        title = {Working Memory or Working Attention?},
    booktitle = {Attention: Selection, Awareness, and Control},
    publisher = {Oxford University Press},
       editor = {A. D. Baddeley and Lawrence Weiskrantz},
         year = {1993},
}
@incollection{BaddeleyandHitchRALM-74,
        title = {Working memory},
       author = {Alan Baddeley and Hitch, {Graham James}},
    booktitle = {Recent Advances in Learning and Motivation},
       editor = {G.A. Bower},
    publisher = {Academic Press},
       volume = {8},
         year = {1974},
        pages = {47-90},
}
@article{BaddeleyetalPR-98,
       author = {Baddeley, A. and Gathercole, S. and Papagno, C.},
        title = {The phonological loop as a language learning device},
      journal = {Psychological Review},
         year = {1998},
       volume = {105},
       number = {1},
        pages = {158-173},
     abstract = {A relatively simple model of the phonological loop (A. D. Baddeley, 1986), a component of working memory, has proved capable of accommodating a great deal of experimental evidence from normal adult participants, children, and neuropsychological patients. Until recently, however, the role of this subsystem in everyday cognitive activities was unclear. In this article the authors review studies of word learning by normal adults and children, neuropsychological patients, and special developmental populations, which provide evidence that the phonological loop plays a crucial role in learning the novel phonological forms of new words. The authors propose that the primary purpose for which the phonological loop evolved is to store unfamiliar sound patterns while more permanent memory records are being constructed. Its use in retaining sequences of familiar words is, it is argued, secondary.}
}
@article{BadreTiCS-08,
        title = {Cognitive control, hierarchy, and the rostro-caudal organization of the frontal lobes},
       author = {David Badre},
      journal = {Trends in Cognitive Sciences},
       volume = {12},
       number = {5},
        pages = {193-200},
         year = {2008},
     abstract = {Cognitive control supports flexible behavior by selecting actions that are consistent with our goals and appropriate for our environment. The prefrontal cortex (PFC) has an established role in cognitive control, and research on the functional organization of PFC promises to contribute to our understanding of the architecture of control. A recently popular hypothesis is that the rostro-caudal axis of PFC supports a control hierarchy whereby posterior-to-anterior PFC mediates progressively abstract, higher-order control. This review discusses evidence for a rostro-caudal gradient of function in PFC and the theories proposed to account for these results, including domain generality in working memory, relational complexity, the temporal organization of behavior and abstract representational hierarchy. Distinctions among these frameworks are considered as a basis for future research.}
}
@article{BadreandDEspositoJCN-07,
       author = {Badre, D. and D'Esposito, M.},
        title = {Functional magnetic resonance imaging evidence for a hierarchical organization of the prefrontal cortex},
      journal = {Journal Cognitive Neuroscience},
         year = {2007},
       volume = {19},
       number = {12},
        pages = {2082-2099},
     abstract = {The prefrontal cortex (PFC) is central to flexible and organized action. Recent theoretical and empirical results suggest that the rostro-caudal axis of the frontal lobes may reflect a hierarchical organization of control. Here, we test whether the rostro-caudal axis of the PFC is organized hierarchically, based on the level of abstraction at which multiple representations compete to guide selection of action. Four functional magnetic resonance imaging (fMRI) experiments parametrically manipulated the set of task-relevant (a) responses, (b) features, (c) dimensions, and (d) overlapping cue-to-dimension mappings. A systematic posterior to anterior gradient was evident within the PFC depending on the manipulated level of representation. Furthermore, across four fMRI experiments, activation in PFC subregions was consistent with the sub- and superordinate relationships that define an abstract representational hierarchy. In addition to providing further support for a representational hierarchy account of the rostro-caudal gradient in the PFC, these data provide important empirical constraints on current theorizing about control hierarchies and the PFC.}
}
@article{BadreandFrankCEREBRAL-CORTEX-12,
       author = {Badre, D. and Frank, M. J.},
        title = {{M}echanisms of hierarchical reinforcement learning in cortico-striatal circuits {II}: evidence from f{MRI}},
      journal = {Cerebral Cortex},
         year = {2012},
       volume = {22},
       number = {3},
        pages = {527-536},
     abstract = {The frontal lobes may be organized hierarchically such that more rostral frontal regions modulate cognitive control operations in caudal regions. In our companion paper (Frank MJ, Badre D. 2011. Mechanisms of hierarchical reinforcement learning in corticostriatal circuits I: computational analysis. 22:509-526), we provide novel neural circuit and algorithmic models of hierarchical cognitive control in cortico-striatal circuits. Here, we test key model predictions using functional magnetic resonance imaging (fMRI). Our neural circuit model proposes that contextual representations in rostral frontal cortex influence the striatal gating of contextual representations in caudal frontal cortex. Reinforcement learning operates at each level, such that the system adaptively learns to gate higher order contextual information into rostral regions. Our algorithmic Bayesian "mixture of experts" model captures the key computations of this neural model and provides trial-by-trial estimates of the learner's latent hypothesis states. In the present paper, we used these quantitative estimates to reanalyze fMRI data from a hierarchical reinforcement learning task reported in Badre D, Kayser AS, D'Esposito M. 2010. Frontal cortex and the discovery of abstract action rules. Neuron. 66:315--326. Results validate key predictions of the models and provide evidence for an individual cortico-striatal circuit for reinforcement learning of hierarchical structure at a specific level of policy abstraction. These findings are initially consistent with the proposal that hierarchical control in frontal cortex may emerge from interactions among nested cortico-striatal circuits at different levels of abstraction.}
}
@article{BadreandNeeTiCS-18,
       author = {David Badre and Derek Evan Nee},
        title = {Frontal Cortex and the Hierarchical Control of Behavior},
      journal = {Trends in Cognitive Sciences},
       volume = {22},
       number = {2},
         year = {2018},
        pages = {170-188},
     abstract = {The frontal lobes are important for cognitive control, yet their functional organization remains controversial. An influential class of theory proposes that the frontal lobes are organized along their rostrocaudal axis to support hierarchical cognitive control. Here, we take an updated look at the literature on hierarchical control, with particular focus on the functional organization of lateral frontal cortex. Our review of the evidence supports neither a unitary model of lateral frontal function nor a unidimensional abstraction gradient. Rather, separate frontal networks interact via local and global hierarchical structure to support diverse task demands.}
}
@article{BadreandWagnerNEURON-04,
       author = {Badre, D. and Wagner, A. D.},
        title = {Selection, integration, and conflict monitoring; assessing the nature and generality of prefrontal cognitive control mechanisms},
      journal = {Neuron},
       volume = {41},
       number = {3},
         year = {2004},
        pages = {473-487},
     abstract = {Prefrontal cortex (PFC) supports flexible behavior by mediating cognitive control, though the elemental forms of control supported by PFC remain a central debate. Dorsolateral PFC (DLPFC) is thought to guide response selection under conditions of response conflict or, alternatively, may refresh recently active representations within working memory. Lateral frontopolar cortex (FPC) may also adjudicate response conflict, though others propose that FPC supports higher order control processes such as subgoaling and integration. Anterior cingulate cortex (ACC) is hypothesized to upregulate response selection by detecting response conflict; it remains unclear whether ACC functions generalize beyond monitoring response conflict. The present fMRI experiment directly tested these competing theories regarding the functional roles of DLPFC, FPC, and ACC. Results reveal dissociable control processes in PFC, with mid-DLPFC selectively mediating resolution of response conflict and FPC further mediating subgoaling/integration. ACC demonstrated a broad sensitivity to control demands, suggesting a generalized role in modulating cognitive control.},
}
@article{BaetalCoRR-16,
       author = {Jimmy Ba and Geoffrey Hinton and Volodymyr Mnih and Joel Z. Leibo and Catalin Ionescu},
        title = {Using Fast Weights to Attend to the Recent Past},
      journal = {CoRR},
       volume = {arXiv:1610.06258},
         year = {2016},
     abstract = {Until recently, research on artificial neural networks was largely restricted to systems with only two types of variable: Neural activities that represent the current or recent input and weights that learn to capture regularities among inputs, outputs and payoffs. There is no good reason for this restriction. Synapses have dynamics at many different time-scales and this suggests that artificial neural networks might benefit from variables that change slower than activities but much faster than the standard weights. These "fast weights" can be used to store temporary memories of the recent past and they provide a neurally plausible way of implementing the type of attention to the past that has recently proved very helpful in sequence-to-sequence models. By using fast weights we can avoid the need to store copies of neural activity patterns.}
}
@article{BaetalCoRR-16b,
       author = {Jimmy Lei Ba and Jamie Ryan Kiros and Geoffrey E. Hinton},
        title = {Layer Normalization},
      journal = {CoRR},
       volume = {arXiv:1607.06450},
         year = {2016},
     abstract = {Training state-of-the-art, deep neural networks is computationally expensive. One way to reduce the training time is to normalize the activities of the neurons. A recently introduced technique called batch normalization uses the distribution of the summed input to a neuron over a mini-batch of training cases to compute a mean and variance which are then used to normalize the summed input to that neuron on each training case. This significantly reduces the training time in feed-forward neural networks. However, the effect of batch normalization is dependent on the mini-batch size and it is not obvious how to apply it to recurrent neural networks. In this paper, we transpose batch normalization into layer normalization by computing the mean and variance used for normalization from all of the summed inputs to the neurons in a layer on a single training case. Like batch normalization, we also give each neuron its own adaptive bias and gain which are applied after the normalization but before the non-linearity. Unlike batch normalization, layer normalization performs exactly the same computation at training and test times. It is also straightforward to apply to recurrent neural networks by computing the normalization statistics separately at each time step. Layer normalization is very effective at stabilizing the hidden state dynamics in recurrent networks. Empirically, we show that layer normalization can substantially reduce the training time compared with previously published techniques.}
}
@inproceedings{BaetalICLR-15,
       author = {Jimmy Lei Ba and Volodymyr Mnih and Koray Kavukcuoglu},
        title = {Multiple Object Recognition with Visual Attention},
    booktitle = {Submitted to International Conference on Learning Representations},
         year = 2015,
       volume = {arXiv:1412.7755},
     abstract = {We present an attention-based model for recognizing multiple objects in images. The proposed model is a deep recurrent neural network trained with reinforcement learning to attend to the most relevant regions of the input image. We show that the model learns to both localize and recognize multiple objects despite being given only class labels during training. We evaluate the model on the challenging task of transcribing house number sequences from Google Street View images and show that it is both more accurate than the state-of-the-art convolutional networks and uses fewer parameters and less computation.}
}
@article{BahdanauetalCoRR-14,
        title = {Neural machine translation by jointly learning to align and translate},
       author = {Dzmitry Bahdanau and Kyunghyun Cho and Yoshua Bengio},
      journal = {CoRR},
       volume = {arXiv:1409.0473},
         year = 2014,
          url = {http://arxiv.org/abs/1409.0473},
     abstract = {Neural machine translation is a recently proposed approach to machine translation. Unlike the traditional statistical machine translation, the neural machine translation aims at building a single neural network that can be jointly tuned to maximize the translation performance. The models proposed recently for neural machine translation often belong to a family of encoder-decoders and encodes a source sentence into a fixed-length vector from which a decoder generates a translation. In this paper, we conjecture that the use of a fixed-length vector is a bottleneck in improving the performance of this basic encoder-decoder architecture, and propose to extend this by allowing a model to automatically (soft-)search for parts of a source sentence that are relevant to predicting a target word, without having to form these parts as a hard segment explicitly. With this new approach, we achieve a translation performance comparable to the existing state-of-the-art phrase-based system on the task of English-to-French translation. Furthermore, qualitative analysis reveals that the (soft-)alignments found by the model agree well with our intuition.},
}
@article{BahdanauetalCoRR-15,
        title = {Neural machine translation by jointly learning to align and translate},
       author = {Dzmitry Bahdanau and KyungHyun Cho and Yoshua Bengio},
      journal = {CoRR},
       volume = {arXiv:1409.0473},
         year = 2015,
     abstract = {Neural machine translation is a recently proposed approach to machine translation. Unlike the traditional statistical machine translation, the neural machine translation aims at building a single neural network that can be jointly tuned to maximize the translation performance. The models proposed recently for neural machine translation often belong to a family of encoder-decoders and encode a source sentence into a fixed-length vector from which a decoder generates a translation. In this paper, we conjecture that the use of a fixed-length vector is a bottleneck in improving the performance of this basic encoder-decoder architecture, and propose to extend this by allowing a model to automatically (soft-)search for parts of a source sentence that are relevant to predicting a target word, without having to form these parts as a hard segment explicitly. With this new approach, we achieve a translation performance comparable to the existing state-of-the-art phrase-based system on the task of English-to-French translation. Furthermore, qualitative analysis reveals that the (soft-)alignments found by the model agree well with our intuition.}
}
@article{BahletalNEURON-15,
       author = {Bahl, A. and Serbe, E. and Meier, M. and Ammer, G. and Borst, A.},
        title = {{{N}eural {M}echanisms for {D}rosophila {C}ontrast {V}ision}},
      journal = {Neuron},
       volume = {88},
       number = {6},
         year = {2015},
        pages = {1240-1252},
     abstract = {Spatial contrast, the difference in adjacent luminance values, provides information about objects, textures, and motion and supports diverse visual behaviors. Contrast computation is therefore an essential element of visual processing. The underlying mechanisms, however, are poorly understood. In human psychophysics, contrast illusions are means to explore such computations, but humans offer limited experimental access. Via behavioral experiments in Drosophila, we find that flies are also susceptible to contrast illusions. Using genetic silencing techniques, electrophysiology, and modeling, we systematically dissect the mechanisms and neuronal correlates underlying the behavior. Our results indicate that spatial contrast computation involves lateral inhibition within the same pathway that computes motion of luminance increments (ON pathway). Yet motion-blind flies, in which we silenced downstream motion-sensitive neurons needed for optomotor behavior, have fully intact contrast responses. In conclusion, spatial contrast and motion cues are first computed by overlapping neuronal circuits which subsequently feed into parallel visual processing streams.}
}
@article{BaietalCoRR-18,
       author = {Shaojie Bai J. and Zico Kolter and Vladlen Koltun},
        title = {An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling},
      journal = {CoRR},
       volume = {arXiv:1803.01271},
         year = {2018},
     abstract = {For most deep learning practitioners, sequence modeling is synonymous with recurrent networks. Yet recent results indicate that convolutional architectures can outperform recurrent networks on tasks such as audio synthesis and machine translation. Given a new sequence modeling task or dataset, which architecture should one use? We conduct a systematic evaluation of generic convolutional and recurrent architectures for sequence modeling. The models are evaluated across a broad range of standard tasks that are commonly used to benchmark recurrent networks. Our results indicate that a simple convolutional architecture outperforms canonical recurrent networks such as LSTMs across a diverse range of tasks and datasets, while demonstrating longer effective memory. We conclude that the common association between sequence modeling and recurrent networks should be reconsidered, and convolutional networks should be regarded as a natural starting point for sequence modeling tasks.},
}
@article{BaietalCoRR-20,
       author = {Bai, Shaojie and Koltun, Vladlen and Kolter, J. Zico},
        title = {Multiscale Deep Equilibrium Models},
      journal = {CoRR},
       volume = {arxiv:2006.08656},
          url = {https://arxiv.org/pdf/2006.08656.pdf},
         year = {2020},
     abstract = {We propose a new class of implicit networks, the multiscale deep equilibrium model (MDEQ), suited to large-scale and highly hierarchical pattern recognition domains. An MDEQ directly solves for and backpropagates through the equilibrium points of multiple feature resolutions simultaneously, using implicit differentiation to avoid storing intermediate states (and thus requiring only O(1) memory consumption). These simultaneously-learned multi-resolution features allow us to train a single model on a diverse set of tasks and loss functions, such as using a single MDEQ to perform both image classification and semantic segmentation. We illustrate the effectiveness of this approach on two large-scale vision tasks: ImageNet classification and semantic segmentation on high-resolution images from the Cityscapes dataset. In both settings, MDEQs are able to match or exceed the performance of recent competitive computer vision models: the first time such performance and scale have been achieved by an implicit deep learning approach. The code and pre-trained models are at this https URL.}
}
@incollection{BaietalNIPS-19,
        title = {Deep Equilibrium Models},
       author = {Bai, Shaojie and Kolter, J. Zico and Koltun, Vladlen},
    booktitle = {Advances in Neural Information Processing Systems 32},
       editor = {H. Wallach and H. Larochelle and A. Beygelzimer and F. d{\'{e}}-Buc and E. Fox and R. Garnett},
    publisher = {Curran Associates, Inc.},
        pages = {690-701},
          url = {http://papers.neurips.cc/paper/8358-deep-equilibrium-models.pdf},
         year = {2019},
     abstract = {We present a new approach to modeling sequential data: the deep equilibrium model (DEQ). Motivated by an observation that the hidden layers of many existing deep sequence models converge towards some fixed point, we propose the DEQ approach that directly finds these equilibrium points via root-finding. Such a method is equivalent to running an infinite depth (weight-tied) feedforward network, but has the notable advantage that we can analytically backpropagate through the equilibrium point using implicit differentiation. Using this approach, training and prediction in these networks require only constant memory, regardless of the effective "depth" of the network. We demonstrate how DEQs can be applied to two state-of-the-art deep sequence models: self-attention transformers and trellis networks. On large-scale language modeling tasks, such as the WikiText-103 benchmark, we show that DEQs (1) often improve performance over these state-of-the-art models (for similar parameter counts); (2) have similar computational requirements to existing models; and (3) vastly reduce memory consumption (often the bottleneck for training large sequence models), demonstrating an up-to 88\% memory reduction in our experiments.},
}
@incollection{BaillargeonHCCD-02,
       author = {Baillargeon, Ren{\'{e}}e},
        title = {The Acquisition of Physical Knowledge in Infancy: A Summary in Eight Lessons},
    booktitle = {Blackwell Handbook of Childhood Cognitive Development, Edition: 1st},
       editor = {U. Goswami},
    publisher = {Blackwell},
         year = {2002},
        pages = {46-83},
}
@article{BaillargeonetalARP-16,
       author = {Renee Baillargeon and Rose M. Scott and Lin Bian},
        title = {Psychological Reasoning in Infancy},
      journal = {Annual Review of Psychology},
       volume = {67},
         year = {2016},
        pages = {159-86},
     abstract = {Adults routinely make sense of others' actions by inferring the mental states that underlie these actions. Over the past two decades, developmental researchers have made significant advances in understanding the origins of this ability in infancy. This evidence indicates that when infants observe an agent act in a simple scene, they infer the agent's mental states and then use these mental states, together with a principle of rationality (and its corollaries of efficiency and consistency), to predict and interpret the agent's subsequent actions and to guide their own actions toward the agent. In this review, we first describe the initial demonstrations of infants' sensitivity to the efficiency and consistency principles. We then examine how infants identify novel entities as agents. Next, we summarize what is known about infants' ability to reason about agents' motivational, epistemic, and counterfactual states. Finally, we consider alternative interpretations of these findings and discuss the current controversy about the relation between implicit and explicit psychological reasoning.}
}
@article{BailletNATURE-NEUROSCIENCE-17,
       author = {Baillet, Sylvain},
        title = {Magnetoencephalography for brain electrophysiology and imaging},
      journal = {Nature Neuroscience},
         year = {2017},
    publisher = {Nature Publishing Group},
       volume = {20},
        pages = {327-339},
     abstract = {We review the aspects that uniquely characterize magnetoencephalography (MEG) among the techniques available to explore and resolve brain function and dysfunction. While emphasizing its specific strengths in terms of millisecond source imaging, we also identify and discuss current practical challenges, in particular in signal extraction and interpretation. We also take issue with some perceived disadvantages of MEG, including the misconception that the technique is redundant with electroencephalography. Overall, MEG contributes uniquely to our deeper comprehension of both regional and large-scale brain dynamics: from the functions of neural oscillations and the nature of event-related brain activation, to the mechanisms of functional connectivity between regions and the emergence of modes of network communication in brain systems. We expect MEG to play an increasing and pivotal role in the elucidation of these grand mechanistic principles of cognitive, systems and clinical neuroscience.}
}
@inproceedings{BairdandMooreNIPS-99,
       author = {Baird, Leemon and Moore, Andrew},
        title = {Gradient Descent for General Reinforcement Learning},
    booktitle = {Advances in Neural Information Processing Systems 11},
    publisher = {{MIT} Press},
      address = {Cambridge, MA},
         year = 1999,
}
@article{BajcsyIEEE-88,
        title = {Active perception},
       author = {Bajcsy, Ruzena},
      journal = {Proceedings of the {IEEE}},
       volume = 76,
        issue = 8,
        pages = {966-972},
         year = 1988,
}
@inproceedings{BakerIJCAI-89,
       author = {Baker, Andrew B. and Ginsberg, Matthew L.},
        title = {Temporal Projection and Explanation},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
         year = 1989,
}
@article{BakerNATURE-11,
       author = {Baker, Monya},
        title = {Tissue models: A living system on a chip},
      journal = {Nature},
         year = {2011},
    publisher = {Nature Publishing Group},
       volume = {471},
        issue = {7340},
        pages = {661-665},
     abstract = {For more than a decade, researchers have been etching grooves into silicon and plastic wafers, filling the spaces with living cells, and hoping that the resulting devices will mimic biological systems such as the liver or gut. Scientists at the Wyss Institute for Biologically Inspired Engineering at Harvard University in Boston, Massachusetts, have created one of the most sophisticated devices so far: a lung on a chip that represents several types of tissue1. "We started with the simplest embodiment of human airway and capillary cells, and then introduced immune cells," says Donald Ingber, head of the institute. The chip holds a pair of microchannels separated by a flexible, porous 10-micrometre membrane. One channel contains air and a layer of epithelial cells such as those lining the tiniest air sacs in the lung; the other holds the type of cell that lines capillaries, along with flowing liquid to simulate blood. The set-up even models breathing: vacuum chambers attached to the channels simulate the mechanical forces that cells encounter as a person's chest expands and contracts.}
}
@incollection{BakerandTenenbaumPAIR-14,
       author = {Chris L. Baker and Joshua B. Tenenbaum},
        title = {Modeling Human Plan Recognition Using Bayesian Theory of Mind},
       editor = {Sukthankar, Gita and Geib, Christopher and Bui, Hung Hai and Pynadath, David and Goldman, Robert P.},
    booktitle = {Plan, Activity, and Intent Recognition: Theory and Practice},
         year = {2014},
    publisher = {Morgan Kaufmann Publishers Inc.},
      address = {San Francisco, CA, USA},
}
@article{BakeretalCOGNITION-09,
       author = {Baker, C. L. and Saxe, R. and Tenenbaum, J. B.},
        title = {Action understanding as inverse planning},
      journal = {Cognition},
       volume = {113},
       number = {3},
         year = {2009},
        pages = {329-349},
     abstract = {Humans are adept at inferring the mental states underlying other agents' actions, such as goals, beliefs, desires, emotions and other thoughts. We propose a computational framework based on Bayesian inverse planning for modeling human action understanding. The framework represents an intuitive theory of intentional agents' behavior based on the principle of rationality: the expectation that agents will plan approximately rationally to achieve their goals, given their beliefs about the world. The mental states that caused an agent's behavior are inferred by inverting this model of rational planning using Bayesian inference, integrating the likelihood of the observed actions with the prior over mental states. This approach formalizes in precise probabilistic terms the essence of previous qualitative approaches to action understanding based on an "intentional stance" [Dennett, D. C. (1987). The intentional stance. Cambridge, MA: MIT Press] or a "teleological stance" [Gergely, G., N\'{a}dasdy, Z., Csibra, G., \& Bir, S. (1995). Taking the intentional stance at 12 months of age. Cognition, 56, 165-193]. In three psychophysical experiments using animated stimuli of agents moving in simple mazes, we assess how well different inverse planning models based on different goal priors can predict human goal inferences. The results provide quantitative evidence for an approximately rational inference mechanism in human goal inference within our simplified stimulus paradigm, and for the flexible nature of goal representations that human observers can adopt. We discuss the implications of our experimental results for human action understanding in real-world contexts, and suggest how our framework might be extended to capture other kinds of mental state inferences, such as inferences about beliefs, or inferring whether an entity is an intentional agent.}
}
@inproceedings{BakkerandSchmidhuberIAS-04,
       author = {Bram Bakker and J\"{u}rgen Schmidhuber},
        title = {Hierarchical Reinforcement Learning Based on Subgoal Discovery and Subpolicy Specialization},
    booktitle = {Proceedings of the 8th Conference on Intelligent Autonomous Systems},
         year = {2004},
        pages = {438-445},
     abstract = {We introduce a new method for hierarchical reinforcement learning. High-level policies automatically discover subgoals; low-level policies learn to specialize on different subgoals. Subgoals are represented as desired abstract observations which cluster raw input data. High-level value functions cover the state space at a coarse level; low-level value functions cover only parts of the state space at a fine-grained level. Experiments show that this method outperforms several flat reinforcement learning methods in a deterministic task and in a stochastic task.}
}
@article{BakogluandMeindlVLSI-85,
       author = {Bakoglu, H.B. and Meindl, J.D.},
      journal = {Electron Devices, IEEE Transactions on},
        title = {Optimal interconnection circuits for {VLSI}},
         year = {1985},
       volume = {32},
       number = {5},
        pages = {903-909},
     abstract = {The propagation delay of interconnection lines is a major factor in determining the performance Of VLSI circuits because the RC time delay of these lines increases rapidly as chip size is increased and cross-sectional interconnection dimensions are reduced. In this paper, a model for interconnection time delay is developed that includes the effects of scaling transistor, interconnection, and chip dimensions. The delays of aluminum, WSi2, and polysilicon lines are compared, and propagation delays in future VLSI circuits are projected. Properly scaled multilevel conductors, repeaters, cascaded drivers, and cascaded driver/ repeater combinations are investigated as potential methods for reducing propagation delay. The model yields optimal cross-sectional interconnection dimensions and driver/repeater configurations that can lower propagation delays by more than an order of magnitude in MOSFET circuits.},
}
@inproceedings{BakshyWSDM-11,
       author = {Eytan Bakshy and Jake Hofman and Winter A. Mason and Duncan J. Watts},
        title = {Everyone's an Influencer: Quantifying Influence on Twitter},
    booktitle = {Proceedings of the 4th International Conference on Web Search and Data Mining},
     location = {Hong Kong},
         year = 2011,
}
@article {BaldassiPLoS-13,
       author = {Baldassi, Carlo and Alemi-Neissi, Alireza and Pagan, Marino and James J {DiCarlo} and Zecchina, Riccardo and Davide Zoccolan},
        title = {Shape similarity, better than semantic membership, accounts for the structure of visual object representations in a population of monkey inferotemporal neurons},
      journal = {{PLoS} Computational Biology},
       volume = 9,
         year = 2013,
        pages = {e1003167},
     abstract = {The anterior inferotemporal cortex (IT) is the highest stage along the hierarchy of visual areas that, in primates, processes visual objects. Although several lines of evidence suggest that IT primarily represents visual shape information, some recent studies have argued that neuronal ensembles in IT code the semantic membership of visual objects (i.e., represent conceptual classes such as animate and inanimate objects). In this study, we investigated to what extent semantic, rather than purely visual information, is represented in IT by performing a multivariate analysis of IT responses to a set of visual objects. By relying on a variety of machine-learning approaches (including a cutting-edge clustering algorithm that has been recently developed in the domain of statistical physics), we found that, in most instances, IT representation of visual objects is accounted for by their similarity at the level of shape or, more surprisingly, low-level visual properties. Only in a few cases we observed IT representations of semantic classes that were not explainable by the visual similarity of their members. Overall, these findings reassert the primary function of IT as a conveyor of explicit visual shape information, and reveal that low-level visual properties are represented in IT to a greater extent than previously appreciated. In addition, our work demonstrates how combining a variety of state-of-the-art multivariate approaches, and carefully estimating the contribution of shape similarity to the representation of object categories, can substantially advance our understanding of neuronal coding of visual objects in cortex.},
}
@article{BalduzzietalCoRR-14,
       author = {David Balduzzi and Hastagiri Vanchinathan and Joachim M. Buhmann},
        title = {Kickback cuts Backprop's red-tape: Biologically plausible credit assignment in neural networks},
      journal = {CoRR},
       volume = {arXiv:1411.6191},
         year = {2014},
     abstract = {Error backpropagation is an extremely effective algorithm for assigning credit in artificial neural networks. However, weight updates under Backprop depend on lengthy recursive computations and require separate output and error messages -- features not shared by biological neurons, that are perhaps unnecessary. In this paper, we revisit Backprop and the credit assignment problem. We first decompose Backprop into a collection of interacting learning algorithms; provide regret bounds on the performance of these sub-algorithms; and factorize Backprop's error signals. Using these results, we derive a new credit assignment algorithm for nonparametric regression, Kickback, that is significantly simpler than Backprop. Finally, we provide a sufficient condition for Kickback to follow error gradients, and show that Kickback matches Backprop's performance on real-world regression benchmarks.}
}
@book{Bale2014,
        title = {Beat the Heart Attack Gene: The Revolutionary Plan to Prevent Heart Disease, Stroke, and Diabetes},
       author = {Bale, B. and Doneen, A. and King, L. and Cool, L.C.},
         year = {2014},
    publisher = {John Wiley \& Sons Canada, Limited}
}
@article{BalintetalACTA-BIOMATERIALIA-14,
        title = {Conductive polymers: Towards a smart biomaterial for tissue engineering},
       author = {Richard Balint and Nigel J. Cassidy and Sarah H. Cartmell},
      journal = {Acta Biomaterialia},
       volume = 10,
        issue = 6,
         year = 2014,
        pages = {2341-2353},
     abstract = {Developing stimulus-responsive biomaterials with easy-to-tailor properties is a highly desired goal of the tissue engineering community. A novel type of electroactive biomaterial, the conductive polymer, promises to become one such material. Conductive polymers are already used in fuel cells, computer displays and microsurgical tools, and are now finding applications in the field of biomaterials. These versatile polymers can be synthesised alone, as hydrogels, combined into composites or electrospun into microfibres. They can be created to be biocompatible and biodegradable. Their physical properties can easily be optimized for a specific application through binding biologically important molecules into the polymer using one of the many available methods for their functionalization. Their conductive nature allows cells or tissue cultured upon them to be stimulated, the polymers' own physical properties to be influenced post-synthesis and the drugs bound in them released, through the application of an electrical signal. It is thus little wonder that these polymers are becoming very important materials for biosensors, neural implants, drug delivery devices and tissue engineering scaffolds. Focusing mainly on polypyrrole, polyaniline and poly(3,4-ethylenedioxythiophene), we review conductive polymers from the perspective of tissue engineering. The basic properties of conductive polymers, their chemical and electrochemical synthesis, the phenomena underlying their conductivity and the ways to tailor their properties (functionalization, composites, etc.) are discussed.}
}
@article{BallSIGSOFT-99,
       author = {Ball, Thoms},
        title = {The Concept of Dynamic Analysis},
      journal = {{SIGSOFT} Software Engineering Notes},
    publisher = {ACM},
      address = {New York, NY, USA},
       volume = {24},
       number = {6},
         year = {1999},
        pages = {216-234},
     abstract = {Dynamic analysis is the analysis of the properties of a running program. In this paper, we explore two new dynamic analyses based on program profiling: (a) Frequency Spectrum Analysis. We show how analyzing the frequencies of program entities in a single execution can help programmers to decompose a program, identify related computations, and find computations related to specific input and output characteristics of a program. (b) Coverage Concept Analysis. Concept analysis of test coverage data computes dynamic analogs to static control flow relationships such as domination, postdomination, and regions. Comparison of these dynamically computed relationships to their static counterparts can point to areas of code requiring more testing and can aid programmers in understanding how a program and its test sets relate to one another.}
}
@book{Ballard2015,
        title = {Brain Computation as Hierarchical Abstraction},
       author = {Ballard, Dana H.},
       series = {Computational Neuroscience Series},
         year = {2015},
    publisher = {MIT Press}
}
@book{Ballard97,
       author = {Ballard, Dana H.},
        title = {An Introduction to Natural Computation},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1997,
}
@inproceedings{BallardIJCAI-89,
       author = {Ballard, Dana H.},
        title = {Reference Frames for Animate Vision},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
         year = 1989,
        pages = {1635-1641},
}
@article{BallardandBrownCVGIP-92,
    author = {Dana H. Ballard and Christopher M. Brown},
     title = {Principles of animate vision},
   comment = {Computer Vision Graphics and Image Processing},
   journal = {CVGIP: Image Understanding},
    volume = 56,
     issue = 1,
     pages = {3-21},
      year = 1992,
}
@article{BallardetalNATURE-COMMUNICATIONS-19,
       author = {Ballard, Ian C. and Wagner, Anthony D. and McClure, Samuel M.},
        title = {Hippocampal pattern separation supports reinforcement learning},
      journal = {Nature Communications},
         year = {2019},
       volume = {10},
        issue = {1},
        pages = {1073},
     abstract = {Animals rely on learned associations to make decisions. Associations can be based on relationships between object features (e.g., the three leaflets of poison ivy leaves) and outcomes (e.g., rash). More often, outcomes are linked to multidimensional states (e.g., poison ivy is green in summer but red in spring). Feature-based reinforcement learning fails when the values of individual features depend on the other features present. One solution is to assign value to multi-featural conjunctive representations. Here, we test if the hippocampus forms separable conjunctive representations that enables the learning of response contingencies for stimuli of the form: AB+, B-, AC-, C+. Pattern analyses on functional MRI data show the hippocampus forms conjunctive representations that are dissociable from feature components and that these representations, along with those of cortex, influence striatal prediction errors. Our results establish a novel role for hippocampal pattern separation and conjunctive representation in reinforcement learning.},
}
@article{BallardetalPCCP-17,
       author = {Ballard, Andrew J. and Das, Ritankar and Martiniani, Stefano and Mehta, Dhagash and Sagun, Levent and Stevenson, Jacob D. and Wales, David J.},
        title = {Energy landscapes for machine learning},
      journal = {Physical Chemistry Chemical Physics},
         year = {2017},
       volume = {19},
        issue = {20},
        pages = {12585-12603},
    publisher = {The Royal Society of Chemistry},
     abstract = {Machine learning techniques are being increasingly used as flexible non-linear fitting and prediction tools in the physical sciences. Fitting functions that exhibit multiple solutions as local minima can be analysed in terms of the corresponding machine learning landscape. Methods to explore and visualise molecular potential energy landscapes can be applied to these machine learning landscapes to gain new insight into the solution space involved in training and the nature of the corresponding predictions. In particular, we can define quantities analogous to molecular structure, thermodynamics, and kinetics, and relate these emergent properties to the structure of the underlying landscape. This Perspective aims to describe these analogies with examples from recent applications, and suggest avenues for new interdisciplinary research.}
}
@article{BalletalFiN-11,
       author = {Ball, Gareth and Stokes, Paul R. and Rhodes, Rebecca A. and Bose, Subrata K. and Rezek, Iead and Wink, Alle-Meije and Lord, Louis-David and Mehta, Mitul A. and Grasby, Paul M. and Turkheimer, Federico E.},
        title = {Executive Functions and Prefrontal Cortex: A Matter of Persistence?},
      journal = {Frontiers in Systems Neuroscience},
         year = {2011},
    publisher = {Frontiers Research Foundation},
       volume = {5},
        pages = {3},
     abstract = {Executive function is thought to originates from the dynamics of frontal cortical networks. We examined the dynamic properties of the blood oxygen level dependent time-series measured with functional MRI (fMRI) within the prefrontal cortex (PFC) to test the hypothesis that temporally persistent neural activity underlies performance in three tasks of executive function. A numerical estimate of signal persistence, the Hurst exponent, postulated to represent the coherent firing of cortical networks, was determined and correlated with task performance. Increasing persistence in the lateral PFC was shown to correlate with improved performance during an n-back task. Conversely, we observed a correlation between persistence and increasing commission error --- indicating a failure to inhibit a prepotent response --- during a Go/No-Go task. We propose that persistence within the PFC reflects dynamic network formation and these findings underline the importance of frequency analysis of fMRI time-series in the study of executive functions.},
}
@article{BalogetalCoRR-16,
       author = {Matej Balog and Alexander L. Gaunt and Marc Brockschmidt and Sebastian Nowozin and Daniel Tarlow},
        title = {{DeepCoder}: Learning to Write Programs},
      journal = {CoRR},
       volume = {arXiv:1611.01989},
         year = {2016},
}
@article{BalogetalCoRR-20,
        title = {Neural Program Synthesis with a Differentiable Fixer},
       author = {Matej Balog and Rishabh Singh and Petros Maniatis and Charles Sutton},
         year = {2020},
      journal = {CoRR},
       volume = {arXiv:2006.10924},
     abstract = {We present a new program synthesis approach that combines an encoder-decoder based synthesis architecture with a differentiable program fixer. Our approach is inspired from the fact that human developers seldom get their program correct on the first attempt, and perform iterative testing-based program fixing to get to the desired program functionality. Similarly, our approach first learns a distribution over programs conditioned on an encoding of a set of input-output examples, and then iteratively performs fix operations using the differentiable fixer. The fixer takes as input the original examples and the current program's outputs on example inputs, and generates a new distribution over the programs with the goal of reducing the discrepancies between the current program outputs and the desired example outputs. We train our architecture end-to-end on the RobustFill domain, and show that the addition of the fixer module leads to a significant improvement on synthesis accuracy compared to using beam search},
}
@inproceedings{BalogetalICLR-17,
       author = {Matej Balog and Alexander L. Gaunt and Marc Brockschmidt and Sebastian Nowozin and Daniel Tarlow},
        title = {DeepCoder: Learning to Write Programs},
    booktitle = {International Conference on Learning Representations},
         year = 2017,
     abstract = {We develop a first line of attack for solving programming competition-style problems from input-output examples using deep learning. The approach is to train a neural network to predict properties of the program that generated the outputs from the inputs. We use the neural network's predictions to augment search techniques from the programming languages community, including enumerative search and an SMT-based solver. Empirically, we show that our approach leads to an order of magnitude speedup over the strong non-augmented baselines and a Recurrent Neural Network approach, and that we are able to solve problems of difficulty comparable to the simplest problems on programming competition websites.}
}
@article{BalstersetalNEUROIMAGE-09,
     title = {Evolution of the cerebellar cortex: {T}he selective expansion of prefrontal-projecting cerebellar lobules},
    author = {Balsters, J. H. and Cussans, E. and Diedrichsen, J. and Phillips, K. A. and Preuss, T. M. and Rilling, J. K. and Ramnani, N.},
   journal = {NeuroImage},
     issue = 3,
     pages = {2045-2052},
    volume = 49,
      year = 2010,
  abstract = {It has been suggested that interconnected brain areas evolve in tandem because evolutionary pressures act on complete functional systems rather than on individual brain areas. The cerebellar cortex has reciprocal connections with both the prefrontal cortex and motor cortex, forming independent loops with each. Specifically, in capuchin monkeys cerebellar cortical lobules Crus I and Crus II connect with prefrontal cortex  whereas the primary motor cortex connects with cerebellar lobules V, VI, VIIb, and VIIIa. Comparisons of extant primate species suggest that the prefrontal cortex has expanded more than cortical motor areas in human evolution. Given the enlargement of the prefrontal cortex relative to motor cortex in humans, our hypothesis would predict corresponding volumetric increases in the parts of the cerebellum connected to the prefrontal cortex, relative to cerebellar lobules connected to the motor cortex. We tested the hypothesis by comparing the volumes of cerebellar lobules in structural MRI scans in capuchins, chimpanzees and humans. The fractions of cerebellar volume occupied by Crus I and Crus II were significantly larger in humans compared to chimpanzees and capuchins. Our results therefore support the hypothesis that in the cortico-cerebellar system, functionally related structures evolve in concert with each other. The evolutionary expansion of these prefrontal-projecting cerebellar territories might contribute to the evolution of the higher cognitive functions of humans.},
}
@article{BaltzetalNSR-17,
       author = {Baltz, E. A. and Trask, E. and Binderbauer, M. and Dikovsky, M. and Gota, H. and Mendoza, R. and Platt, J. C. and Riley, P. F.},
        title = {Achievement of Sustained Net Plasma Heating in a Fusion Experiment with the Optometrist Algorithm},
      journal = {Scientific Reports},
    publisher = {Nature Publishing Group},
         year = {2017},
       volume = {7},
        issue = {1},
        pages = {6425},
     abstract = {Many fields of basic and applied science require efficiently exploring complex systems with high dimensionality. An example of such a challenge is optimising the performance of plasma fusion experiments. The highly-nonlinear and temporally-varying interaction between the plasma, its environment and external controls presents a considerable complexity in these experiments. A further difficulty arises from the fact that there is no single objective metric that fully captures both plasma quality and equipment constraints. To efficiently optimise the system, we develop the Optometrist Algorithm, a stochastic perturbation method combined with human choice. Analogous to getting an eyeglass prescription, the Optometrist Algorithm confronts a human operator with two alternative experimental settings and associated outcomes. A human operator then chooses which experiment produces subjectively better results. This innovative technique led to the discovery of an unexpected record confinement regime with positive net heating power in a field-reversed configuration plasma, characterised by a > 50\% reduction in the energy loss rate and concomitant increase in ion temperature and total plasma energy.},
}
@inproceedings{BalujaetalICASP-08,
     author = {Shumeet Baluja and Michele Covell and Sergey Ioffe},
      title = {Permutation grouping: intelligent Hash function design for audio {\&} image retrieval},
  booktitle = {{IEEE} International Conference on Acoustics, Speech, and Signal Processing},
       year = 2008,
      pages = {2137-2140},
}
@article{BangaloreetalASLP-08,
       author = {Bangalore, S. and Di Fabbrizio, G. and Stent, A.},
      journal = {IEEE Transactions on Audio, Speech and Language Processing},
        title = {Learning the Structure of Task-Driven Human Dialogs},
         year = 2008,
       volume = 16,
        issue = 7,
        pages = {1249-1259},
     abstract = {With the availability of large corpora of spoken dialog, it is now possible to use data-driven techniques to build and use models of task-oriented dialogs. In this paper, we use data-driven techniques to build task structures for individual dialogs, and use the dialog task structures for: dialog act classification, task/subtask classification, task/subtask prediction, and dialog act prediction. We evaluate our approach using a corpus of customer/agent dialogs from a catalog service domain. This paper demonstrates the feasibility of using corpora of human-human conversation to learn dialog models suitable for human-computer dialog applications.},
}
@article{BaninoetalCoRR-20,
        title = {MEMO: A Deep Network for Flexible Combination of Episodic Memories},
       author = {Andrea Banino and Adri{\`a} Puigdom{\`e}nech Badia and Raphael K{\"o}ster and Martin J. Chadwick and Vin{\'i}cius Flores Zambaldi and Demis Hassabis and Caswell Barry and Matthew M Botvinick and Dharshan Kumaran and Charles Blundell},
      journal = {CoRR},
         year = {2020},
       volume = {arXiv:2001.10913},
     abstract = {Recent research developing neural network architectures with external memory have often used the benchmark bAbI question and answering dataset which provides a challenging number of tasks requiring reasoning. Here we employed a classic associative inference task from the memory-based reasoning neuroscience literature in order to more carefully probe the reasoning capacity of existing memory-augmented architectures. This task is thought to capture the essence of reasoning -- the appreciation of distant relationships among elements distributed across multiple facts or memories. Surprisingly, we found that current architectures struggle to reason over long distance associations. Similar results were obtained on a more complex task involving finding the shortest path between nodes in a path. We therefore developed MEMO, an architecture endowed with the capacity to reason over longer distances. This was accomplished with the addition of two novel components. First, it introduces a separation between memories (facts) stored in external memory and the items that comprise these facts in external memory. Second, it makes use of an adaptive retrieval mechanism, allowing a variable number of "memory hops" before the answer is produced. MEMO is capable of solving our novel reasoning tasks, as well as match state of the art results in bAbI.}
}
@article{BaninoetalNATURE-18,
       author = {Banino, Andrea and Barry, Caswell and Uria, Benigno and Blundell, Charles and Lillicrap, Timothy and Mirowski, Piotr and Pritzel, Alexander and Chadwick, Martin J. and Degris, Thomas and Modayil, Joseph and Wayne, Greg and Soyer, Hubert and Viola, Fabio and Zhang, Brian and Goroshin, Ross and Rabinowitz, Neil and Pascanu, Razvan and Beattie, Charlie and Petersen, Stig and Sadik, Amir and Gaffney, Stephen and King, Helen and Kavukcuoglu, Koray and Hassabis, Demis and Hadsell, Raia and Kumaran, Dharshan},
        title = {Vector-based navigation using grid-like representations in artificial agents},
      journal = {Nature},
         year = {2018},
     abstract = {Deep neural networks have achieved impressive successes in fields ranging from object recognition to complex games such as Go. Navigation, however, remains a substantial challenge for artificial agents, with deep neural networks trained by reinforcement learning failing to rival the proficiency of mammalian spatial behaviour, which is underpinned by grid cells in the entorhinal cortex. Grid cells are thought to provide a multi-scale periodic representation that functions as a metric for coding space, and is critical for integrating self-motion (path integration), and planning direct trajectories to goals (vector-based navigation). Here we set out to leverage the computational functions of grid cells to develop a deep reinforcement learning agent with mammal-like navigational abilities. We first trained a recurrent network to perform path integration, leading to the emergence of representations resembling grid cells, as well as other entorhinal cell types. We then showed that this representation provided an effective basis for an agent to locate goals in challenging, unfamiliar, and changeable environments optimizing the primary objective of navigation through deep reinforcement learning. The performance of agents endowed with grid-like representations surpassed that of an expert human and comparison agents, with the metric quantities necessary for vector-based navigation derived from grid-like units within the network. Furthermore, grid-like representations enabled agents to conduct shortcut behaviours reminiscent of those performed by mammals. Our findings show that emergent grid-like representations furnish agents with a Euclidean spatial metric and associated vector operations, providing a foundation for proficient navigation. As such, our results support neuroscientific theories that see grid cells as critical for vector-based navigation, demonstrating that the latter can be combined with path-based strategies to support navigation in challenging environments.},
}
@article{BanksandBennettJOSA-88,
     author = {Banks, M.S. and Bennett, P.J.},
      title = {Optical and photoreceptor immaturities limit the spatial and chromatic vision of human neonates},
    journal = {Journal of the Optical Society of America A}, 
     volume = 5,
      pages = {2059-2079},
       year = 1988,
    comment = {Banks, M.S. and Bennett, P.J. (1988). Optical and photoreceptor immaturities limit the spatial and
               chromatic vision of human neonates. Journal of the Optical Society of America A, 5, 2059-2079.},
}
@inproceedings{BansaletalFOCS-06,
     author = {Bansal, Nikhil and Caprara, Alberto and Sviridenko, Maxim},
      title = {Improved approximation algorithms for multidimensional bin packing problems},
  booktitle = {Proceedings of the 47th Annual {IEEE} Symposium on Foundations of Computer Science},
  publisher = {IEEE Computer Society},
    address = {Washington, DC, USA},
       year = 2006,
      pages = {697-708},
}
@article{BansaletalNEUROPHYSIOLOGY-12,
    author = {Bansal, A.K. and Truccolo, W. and Vargas-Irwin, C.E. and Donoghue, J.P.},
     title = {Decoding {3D} reach and grasp from hybrid signals in motor and premotor cortices: spikes, multiunit activity, and local field potentials},
   journal = {Journal of Neurophysiology},
    volume = {107},
     issue = {5},
     pages = {1337-55},
      year = {2012},
  abstract = {Neural activity in motor cortex during reach and grasp movements shows modulations in a broad range of signals from single-neuron spiking activity (SA) to various frequency bands in broadband local field potentials (LFPs). In particular, spatiotemporal patterns in multiband LFPs are thought to reflect dendritic integration of local and interareal synaptic inputs, attentional and preparatory processes, and multiunit activity (MUA) related to movement representation in the local motor area. Nevertheless, the relationship between multiband LFPs and SA, and their relationship to movement parameters and their relative value as brain-computer interface (BCI) control signals, remain poorly understood. Also, although this broad range of signals may provide complementary information channels in primary (MI) and ventral premotor (PMv) areas, areal differences in information have not been systematically examined. Here, for the first time, the amount of information in SA and multiband LFPs was compared for MI and PMv by recording from dual 96-multielectrode arrays while monkeys made naturalistic reach and grasp actions. Information was assessed as decoding accuracy for {3D} arm end point and grip aperture kinematics based on SA or LFPs in MI and PMv, or combinations of signal types across areas. In contrast with previous studies with \hmlessthanorequal{}16 simultaneous electrodes, here ensembles of >16 units (on average) carried more information than multiband, multichannel LFPs. Furthermore, reach and grasp information added by various LFP frequency bands was not independent from that in SA ensembles but rather typically less than and primarily contained within the latter. Notably, MI and PMv did not show a particular bias toward reach or grasp for this task or for a broad range of signal types. For BCIs, our results indicate that neuronal ensemble spiking is the preferred signal for decoding, while LFPs and combined signals from PMv and MI can add robustness to BCI control.}
}
@article{BaoetalPAMI-16,
       author = {C. Bao and H. Ji and Y. Quan and Z. Shen},
      journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
        title = {Dictionary Learning for Sparse Coding: Algorithms and Convergence Analysis},
         year = {2016},
       volume = {38},
       number = {7},
        pages = {1356-1369},
     abstract = {In recent years, sparse coding has been widely used in many applications ranging from image processing to pattern recognition. Most existing sparse coding based applications require solving a class of challenging non-smooth and non-convex optimization problems. Despite the fact that many numerical methods have been developed for solving these problems, it remains an open problem to find a numerical method which is not only empirically fast, but also has mathematically guaranteed strong convergence. In this paper, we propose an alternating iteration scheme for solving such problems. A rigorous convergence analysis shows that the proposed method satisfies the global convergence property: the whole sequence of iterates is convergent and converges to a critical point. Besides the theoretical soundness, the practical benefit of the proposed method is validated in applications including image restoration and recognition. Experiments show that the proposed method achieves similar results with less computation when compared to widely used methods such as K-SVD.},
}
@book{Bar-Shalom,
       author = {Bar-Shalom, Yaakov and Fortmann, Thomas E.},
        title = {Tracking and Data Association},
    publisher = {Academic Press},
      address = {New York},
         year = 1988,
}
@article{Barabasi99,
       author = {Albert-Laszlo Barabasi and Reka Albert},
        title = {Emergence of Scaling in Random Networks},
      journal = {Science},
       volume = 286,
         year = 1999
}
@article{BarabasiandAlbert,
       author = {Albert-Laszlo Barabasi and Reka Albert},
        title = {Emergence of Scaling in Random Networks},
      journal = {Science},
       volume = {286},
         year = 1999,
	pages = {509-512},
}
@article{BaraketalPIN-13,
       author = {Omri Barak and David Sussillo and Ranulfo Romo and Misha Tsodyks and L.F. Abbott},
        title = {From fixed points to chaos: Three models of delayed discrimination},
      journal = {Progress in Neurobiology},
       volume = {103},
        pages = {214-222},
         year = {2013},
         note = {Conversion of Sensory Signals into Perceptions, Memories and Decisions},
      comment = {This manuscript argues persuasively that mixed selectivity, a signature of high-dimensional neural representations, is a fundamental component of the computational power of prefrontal cortex.},
     abstract = {Working memory is a crucial component of most cognitive tasks. Its neuronal mechanisms are still unclear despite intensive experimental and theoretical explorations. Most theoretical models of working memory assume both time-invariant neural representations and precise connectivity schemes based on the tuning properties of network neurons. A different, more recent class of models assumes randomly connected neurons that have no tuning to any particular task, and bases task performance purely on adjustment of network readout. Intermediate between these schemes are networks that start out random but are trained by a learning scheme. Experimental studies of a delayed vibrotactile discrimination task indicate that some of the neurons in prefrontal cortex are persistently tuned to the frequency of a remembered stimulus, but the majority exhibit more complex relationships to the stimulus that vary considerably across time. We compare three models, ranging from a highly organized line attractor model to a randomly connected network with chaotic activity, with data recorded during this task. The random network does a surprisingly good job of both performing the task and matching certain aspects of the data. The intermediate model, in which an initially random network is partially trained to perform the working memory task by tuning its recurrent and readout connections, provides a better description, although none of the models matches all features of the data. Our results suggest that prefrontal networks may begin in a random state relative to the task and initially rely on modified readout for task performance. With further training, however, more tuned neurons with less time-varying responses should emerge as the networks become more structured.}
}
@article{BaraniukandWakinFoCM-07,
       author = {Baraniuk, Richard G. and Wakin, Michael B.},
        title = {Random Projections of Smooth Manifolds},
      journal = {Foundations of Computational Mathematics},
    publisher = {Springer-Verlag New York},
       volume = 9,
        issue = 1,
         year = 2009,
        pages = {51-77}, 
     abstract = {We propose a new approach for nonadaptive dimensionality reduction of manifold-modeled data, demonstrating that a small number of random linear projections can preserve key information about a manifold-modeled signal. We center our analysis on the effect of a random linear projection operator, Phi, on a smooth well-conditioned K-dimensional submanifold. As our main theoretical contribution, we establish a sufficient number M of random projections to guarantee that, with high probability, all pairwise Euclidean and geodesic distances between points on manifold are well preserved under the mapping Phi. Our results bear strong resemblance to the emerging theory of Compressed Sensing (CS), in which sparse signals can be recovered from small numbers of random linear measurements. As in CS, the random measurements we propose can be used to recover the original data. Moreover, like the fundamental bound in CS, our requisite M is linear in the "information level" K and logarithmic in the ambient dimension N; we also identify a logarithmic dependence on the volume and conditioning of the manifold. In addition to recovering faithful approximations to manifold-modeled signals, however, the random projections we propose can also be used to discern key properties about the manifold. We discuss connections and contrasts with existing techniques in manifold learning, a setting where dimensionality reducing mappings are typically nonlinear and constructed adaptively from a set of sampled training data.},
}
@article{BarbasandGarcia-CabezasCOiN-16,
       author = {Barbas, H. and Garcia-Cabezas, M. A.},
        title = {How the prefrontal executive got its stripes},
      journal = {Current Opinion in Neurobiology},
       volume = {40},
         year = {2016},
        pages = {125-134},
     abstract = {Pathways from cortical and subcortical structures give the prefrontal cortex a panoramic view of the sensory environment and the internal milieu of motives and drives. The prefrontal cortex also receives privileged information from the output of the basal ganglia and cerebellum and innervates widely the inhibitory thalamic reticular nucleus that gates thalamo-cortical communication. Connections, in general, are strongly related to the systematic structural variation of the cortex that can be traced to development. Insights from development have profound implications for the special connections of the prefrontal cortex for executive control, learning and memory, and vulnerability in psychiatric and neurologic diseases.}
}
@inproceedings{BarfordandCrovella98,
       author = {P. Barford and M. Crovella},
        title = {Generating Representative {W}eb Workloads for Network and Server Performance Evaluation},
    booktitle = {Proceedings of ACM SIGMETRICS 1998},
         year = 1998,
}
@article{BargmannBIOESSAYS-13,
       author = {Bargmann, C.I.},
        title = {{B}eyond the connectome: how neuromodulators shape neural circuits},
      journal = {Bioessays},
         year = 2012,
       volume = 34,
        issue = 6,
        pages = {458-465},
}
@article{BargmannandMarderNATURE-13,
       author = {Bargmann, C.I. and Marder, E.},
        title = {{F}rom the connectome to brain function},
      journal = {Nature Methods},
         year = 2013,
       volume = 10,
        issue = 6,
        pages = {483-490},
}
@article{BarkerandOConnorCACM-98,
       author = {Barker, Virginia E. and O'Connor, Dennis E.},
        title = {Expert Systems for Configuration at Digital: XCON and Beyond},
      journal = {Communications of the {ACM}},
       volume = 32,
        issue = 3,
         year = 1989,
        pages = {298-319},
}
@incollection{Barlow61,
       author = {Barlow, Horace B.},
        title = {Possible principles underlying the transformations of sensory messages},
    booktitle = {Sensory Communication},
       editor = {Rosenblith, W. A.},
    publisher = {MIT Press},
      address = {Cambridge, MA},
        pages = {217-234},
         year = 1961
}
@article{BarlowNC-89,
       author = {Barlow, Horace B.},
        title = {Unsupervised Learning},
      journal = {Neural Computation},
       volume = 1,
         year = 1989,
        pages = {295-311},
     abstract = {What use can the brain make of the massive flow of sensory information that occurs without any associated rewards or punishments? This question is reviewed in the light of connectionist models of unsupervised learning and some older ideas, namely the cognitive maps and working models of Tolman and Craik, and the idea that redundancy is important for understanding perception (Attneave 1954), the physiology of sensory pathways (Barlow 1959), and pattern recognition (Watanabe 1960). It is argued that (1) The redundancy of sensory messages provides the knowledge incorporated in the maps or models. (2) Some of this knowledge can be obtained by observations of mean, variance, and covariance of sensory messages, and perhaps also by a method called "minimum entropy coding" (3) Such knowledge may be incorporated in a model of "what usually happens" with which incoming messages are automatically compared, enabling unexpected discrepancies to be immediately identified. (4) Knowledge of the sort incorporated into such a filter is a necessary prerequisite of ordinary learning, and a representation whose elements are independent makes it possible to form associations with logical functions of the elements, not just with the elements themselves.}
}
@article{BarnardetalJMLR-03,
    author = {Kobus Barnard and Pinar Duygulu and David A. Forsyth and Nando de Freitas and David M. Blei and Michael I. Jordan},
     title = {Matching Words and Pictures},
   journal = {Journal of Machine Learning Research},
   volume  = 3,
     pages = {1107-1135},
      year = 2003,
}
@book{Barnett82,
       author = {Barnett, V.},
        title = {Comparative Statistical Inference},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1982,
}
@article{BarnettetalPloS-12,
    author = {Barnett, L. and Platisa, J. and Popovic, M. and Pieribone, V.A. and Hughes, T.},
     title = {A fluorescent, optogenetic voltage sensor capable of resolving action potentials},
   journal = {{PLoS} ONE},
    volume = 7,
     issue = 9,
      year = 2012,
     pages = {e43454},
}
@book{Barrett2011,
     title = {Beyond the Brain: How Body and Environment Shape Animal and Human Minds},
    author = {Louise Barrett},
 publisher = {Princeton University Press},
      year = 2011,
}
@book{Barrett2017,
       author = {Barrett, L.F.},
        title = {How Emotions Are Made},
    publisher = {Pan Macmillan},
         year = {2017},
}
@article{BarrettSCAN-17a,
        title = {The theory of constructed emotion: an active inference account of interoception and categorization},
       author = {Barrett, Lisa Feldman},
      journal = {Social cognitive and affective neuroscience},
    publisher = {Oxford University Press},
       volume = {12},
       number = {1},
         year = {2017},
        pages = {1-23},
     abstract = {The science of emotion has been using folk psychology categories derived from philosophy to search for the brain basis of emotion. The last two decades of neuroscience research have brought us to the brink of a paradigm shift in understanding the workings of the brain, however, setting the stage to revolutionize our understanding of what emotions are and how they work. In this article, we begin with the structure and function of the brain, and from there deduce what the biological basis of emotions might be. The answer is a brain-based, computational account called the theory of constructed emotion.}
}
@article{BarrettSCAN-17b,
        title = {Functionalism cannot save the classical view of emotion},
       author = {Barrett, Lisa Feldman},
      journal = {Social cognitive and affective neuroscience},
    publisher = {Oxford University Press},
       volume = {12},
       number = {1},
         year = {2016},
        pages = {34-36},
        abstract = {Ralph's theoretical approach embodies the assumptions of the classical view, and are generally consistent with a cherished (Western) theory of human nature (Barrett, 2017). These assumptions allow the classical view to 'ring true' - to appear obvious and therefore pleasing - when compared to the seemingly complex and counter-intuitive constructionist approach. As an approach to understanding emotions, Ralph employs a version of functionalism. At the turn of the 20th century, functionalism was pressed into service to rescue the classical view of emotion in the face of mounting disconfirming evidence. And so it is today (for a list of disconfirming evidence, see Table 1 in Barrett, this issue). But functionalism brings with it a set of philosophical assumptions. When scientists like Ralph and I debate the nature of emotion, we are really deliberating over the validity of these assumptions. My goal here is to highlight them and offer my point of view in hopes of moving the dialogue forward. I discuss Ralph's classical view of emotion as an attempt to defend these assumptions while at the same time remaining responsive to the accumulating empirical evidence that calls them into doubt.},
}
@article{BarrettandWeldAIJ-93,
       author = {Barrett, A. and Weld, D.},
        title = {Partial Order Planning: Evaluating Possible Efficiency Gains},
      journal = AIJ,
       volume = 67,
         year = 1993,
        pages = {71-112},
}
@article{BarrettetalCoRR-15,
       author = {Daniel Paul Barrett and Scott Alan Bronikowski and Haonan Yu and Jeffrey Mark Siskind},
        title = {Robot Language Learning, Generation, and Comprehension},
      journal = {CoRR},
       volume = {arXiv:1508.06161},
         year = {2015},
     abstract = {We present a unified framework which supports grounding natural-language semantics in robotic driving. This framework supports acquisition (learning grounded meanings of nouns and prepositions from human annotation of robotic driving paths), generation (using such acquired meanings to generate sentential description of new robotic driving paths), and comprehension (using such acquired meanings to support automated driving to accomplish navigational goals specified in natural language). We evaluate the performance of these three tasks by having independent human judges rate the semantic fidelity of the sentences associated with paths, achieving overall average correctness of 94.6\% and overall average completeness of 85.6\%.}
}
@inproceedings{BarrettetalWSC-09,
       author = {Barrett, C.L. and Beckman, R.J. and Khan, M. and Kumar, V. and Marathe, M.V. and Stretz, P.E. and Dutta, T. and Lewis, B.},
        title = {Generation and Analysis of Large Synthetic Social Contact Networks},
    booktitle = {Proceedings of the 2009 Winter Simulation Conference},
     location = {Austin, Texas},
         year = {2009},
        pages = {1003-1014},
}
@article{BarrettoandSchnitzerCSHP-12,
        title = {In Vivo Optical Microendoscopy for Imaging Cells Lying Deep within Live Tissue},
       author = {Barretto, R.P.J. and Schnitzer, M.J.},
      journal = {Cold Spring Harbor Protocols},
       volume = 2012,
        issue = 10,
         year = 2012,
     abstract = {Although in vivo microscopy has been pivotal in enabling studies of neuronal structure and function in the intact mammalian brain, conventional intravital microscopy has generally been limited to superficial brain areas such as the olfactory bulb, the neocortex, or the cerebellar cortex. For imaging cells in deeper areas, this article discusses in vivo optical microendoscopy using gradient refractive index (GRIN) microlenses that can be inserted into tissue. Our general methodology is broadly applicable to many deep brain regions and areas of the body. Microendoscopes are available in a wide variety of optical designs, allowing imaging across a range of spatial scales and with spatial resolution that can now closely approach that offered by standard water-immersion microscope objectives. The incorporation of microendoscope probes into portable miniaturized microscopes allows imaging in freely behaving animals. When combined with the broad sets of available fluorescent markers, animal preparations, and genetically modified mice, microendoscopic methods enable sophisticated experimental designs for probing how cellular characteristics may underlie or reflect animal behavior and life experience, in healthy animals and animal models of disease.}
}
@article{BarrosoetaMICRO-03,
        title = {Web Search for a Planet: The Google Cluster Architecture},
       author = {Luiz Andr\'{e} Barroso and Jeffrey Dean and Urs H\"{o}lzle},
      journal = {{IEEE} Micro},
        pages = {22-28},
       volume = {23},
         year = {2003},
}
@article{BartoBradtkeandSinghAIJ-94,
       author = {Barto, Andrew G. and Bradtke, Steven J. and Singh, Satinder P.},
        title = {Learning to act using real-time dynamic programming},
      journal = AIJ,
       volume = {},
        pages = {},
         year = 1994,
}
@article{BartoIEEE-83,
       author = {Barto, Andrew G. and Sutton, Richard S. and Anderson, Charles W.},
        title = {Neuronlike Adaptive Elements That Can Solve Difficult Learning  Control Problems},
      journal = {{IEEE} Transactions on Systems, Man, and Cybernetics},
       volume = 13,
        issue = 5,
        pages = {835-846},
         year = 1983,
}
@incollection{BartoLSDM,
       author = {Barto, Andrew G. and Sutton, Richard S. and Watkins, Christopher J. C. H.},
        title = {Learning and Sequential Decision Making},
       editor = {Gabriel, Michael and Moore, John},
    booktitle = {Learning and Computational Neuroscience: Foundations of Adaptive Networks},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1990,
}
@article{BartoandMahadevan03,
       author = {A. Barto and S. Mahadevan},
        title = {Recent Advances in Hierarchical Reinforcement Learning},
      journal = {Discrete Event Dynamic Systems: Theory and Applications},
       volume = 13,
        pages = {41-77},
         year = 2003,
}
@article{BartoandSuttonBC-81,
       author = {Barto, Andrew G. and Sutton, Richard S.},
        title = {Landmark Learning: An Illustration of Associative Search},
      journal = {Biological Cybernetics},
       volume = 42,
         year = 1981,
        pages = {1-8},
}
@article{BartoletalELIFE-15,
       author = {Thomas M. Bartol and Cailey Bromer and Justin P. Kinney and Micheal A. Chirillo and Jennifer N. Bourne and Kristen M. Harris and Terrence J. Sejnowski},
        title = {Nanoconnectomic upper bound on the variability of synaptic plasticity},
      journal = {eLife},
    publisher = {eLife Sciences Publications Limited},
       volume = 10,
         year = 2015,
     abstract = {Information in a computer is quantified by the number of bits that can be stored and recovered. An important question about the brain is how much information can be stored at a synapse through synaptic plasticity, which depends on the history of probabilistic synaptic activity. The strong correlation between size and efficacy of a synapse allowed us to estimate the variability of synaptic plasticity. In an EM reconstruction of hippocampal neuropil we found single axons making two or more synaptic contacts onto the same dendrites, having shared histories of presynaptic and postsynaptic activity. The spine heads and neck diameters, but not neck lengths, of these pairs were nearly identical in size. We found that there is a minimum of 26 distinguishable synaptic strengths, corresponding to storing 4.7 bits of information at each synapse. Because of stochastic variability of synaptic activation the observed precision requires averaging activity over several minutes.}
}
@article{BartonandHarveyNATURE-00,
    author = {Barton, Robert A. and Harvey, Paul H.},
     title = {Mosaic evolution of brain structure in mammals},
   journal = {Nature},
    volume = 405,
      year = 2000,
     pages = {1055-1058},
}
@article{BartunovetalCoRR-18,
       author = {Sergey Bartunov and Adam Santoro and Blake A. Richards and Geoffrey E. Hinton and Timothy P. Lillicrap},
        title = {Assessing the Scalability of Biologically-Motivated Deep Learning Algorithms and Architectures},
      journal = {CoRR},
       volume = {arXiv:1807.04587},
         year = {2018},
     abstract = {The backpropagation of error algorithm (BP) is impossible to implement in a real brain. The recent success of deep networks in machine learning and AI, however, has inspired proposals for understanding how the brain might learn across multiple layers, and hence how it might approximate BP. As of yet, none of these proposals have been rigorously evaluated on tasks where BP-guided deep learning has proved critical, or in architectures more structured than simple fully-connected networks. Here we present results on scaling up biologically motivated models of deep learning on datasets which need deep networks with appropriate architectures to achieve good performance. We present results on the MNIST, CIFAR-10, and ImageNet datasets and explore variants of target-propagation (TP) and feedback alignment (FA) algorithms, and explore performance in both fully- and locally-connected architectures. We also introduce weight-transport-free variants of difference target propagation (DTP) modified to remove backpropagation from the penultimate layer. Many of these algorithms perform well for MNIST, but for CIFAR and ImageNet we find that TP and FA variants perform significantly worse than BP, especially for networks composed of locally connected units, opening questions about whether new architectures and algorithms are required to scale these approaches. Our results and implementation details help establish baselines for biologically motivated deep learning schemes going forward.}
}
@article{BasakNC-01,
       author = {Jayanta Basak},
        title = {Learning Hough transform: a neural network model},
      journal = {Neural Computing},
       volume = 13,
        issue = 3,
         year = 2001,
        pages = {651-76},
     abstract = {A single-layered Hough transform network is proposed that accepts image coordinates of each object pixel as input and produces a set of outputs that indicate the belongingness of the pixel to a particular structure (e.g., a straight line). The network is able to learn adaptively the parametric forms of the linear segments present in the image. It is designed for learning and identification not only of linear segments in two-dimensional images but also the planes and hyperplanes in the higher-dimensional spaces. It provides an efficient representation of visual information embedded in the connection weights. The network not only reduces the large space requirement, as in the case of classical Hough transform, but also represents the parameters with high precision.}
}
@misc{Basal_Gangli_Cerebellum_Interaction,
        title = {Cerebellum \& Basal Ganglia},
       author = {Tutis Vilis},
         year = {2017}
}
@article{BasariaetalNEJM-10,
     title = {Adverse Events Associated with Testosterone Administration},
    author = {Basaria, Shehzad and Coviello, Andrea D. and Travison, Thomas G. and Storer, Thomas W. and 
              Farwell, Wildon R. and Jette, Alan M. and Eder, Richard and Tennstedt, Sharon and Ulloor, 
              Jagadish and Zhang, Anqi and Choong, Karen and Lakshman, Kishore M. and Mazer, Norman A. 
              and Miciek, Renee and Krasnoff, Joanne and Elmi, Ayan and Knapp, Philip E. and Brooks, Brad 
              and Appleman, Erica and Aggarwal, Sheetal and Bhasin, Geeta and Hede-Brierley, Leif and 
              Bhatia, Ashmeet and Collins, Lauren and LeBrasseur, Nathan and Fiore, Louis D. and Bhasin, Shalender},
   journal = {New England Journal of Medicine},
    volume = 363,
     issue = 2,
     pages = {109-122},
      year = 2010,
       URL = {http://www.nejm.org/doi/full/10.1056/NEJMoa1000485},
  abstract = {The paper states that a total of 209 men (mean age, 74 years) were enrolled at the time the trial was terminated. Within this population there was a high prevalence of hypertension, diabetes, hyperlipidemia, and obesity. Patients were selected in part based on their having total serum testosterone level between 100 and 350 ng per deciliter. They were administered 100 mg of testosterone by transdermal gel applied once daily for six months with the dose adjusted by 15 mg up or down depending on whether the level was below 500 ng or above 1000 ng after an initial trial period. Among the participants. During the course of the study, the testosterone group had higher rates of cardiac, respiratory, and dermatologic events than did the placebo group. A total of 23 subjects in the testosterone group, as compared with 5 in the placebo group, had cardiovascular-related adverse events. The relative risk of a cardiovascular-related adverse event remained constant throughout the 6-month treatment period. As compared with the placebo group, the testosterone group had significantly greater improvements in leg-press and chest-press strength and in stair climbing while carrying a load. A greater proportion of men in the testosterone group than in the placebo group reported that they had received a diagnosis of hyperlipidemia or were taking a statin.},
}
@article{BashivanetalCoRR-19,
       author = {Pouya Bashivan and Martin Schrimpf and Robert Ajemian and Irina Rish and Matthew Riemer and Yuhai Tu},
        title = {Continual Learning with Self-Organizing Maps},
      journal = {CoRR},
       volume = {arXiv:1904.09330},
         year = {2019},
     abstract = {Despite remarkable successes achieved by modern neural networks in a wide range of applications, these networks perform best in domain-specific stationary environments where they are trained only once on large-scale controlled data repositories. When exposed to non-stationary learning environments, current neural networks tend to forget what they had previously learned, a phenomena known as catastrophic forgetting. Most previous approaches to this problem rely on memory replay buffers which store samples from previously learned tasks, and use them to regularize the learning on new ones. This approach suffers from the important disadvantage of not scaling well to real-life problems in which the memory requirements become enormous. We propose a memoryless method that combines standard supervised neural networks with self-organizing maps to solve the continual learning problem. The role of the self-organizing map is to adaptively cluster the inputs into appropriate task contexts - without explicit labels - and allocate network resources accordingly. Thus, it selectively routes the inputs in accord with previous experience, ensuring that past learning is maintained and does not interfere with current learning. Out method is intuitive, memoryless, and performs on par with current state-of-the-art approaches on standard benchmarks.}
}
@article {BashivanetalSCIENCE-19,
       author = {Bashivan, Pouya and Kar, Kohitij and DiCarlo, James J.},
        title = {Neural population control via deep image synthesis},
      journal = {Science},
       volume = {364},
       number = {6439},
         year = {2019},
    publisher = {American Association for the Advancement of Science},
     abstract = {To what extent are predictive deep learning models of neural responses useful for generating experimental hypotheses? Bashivan et al. took an artificial neural network built to model the behavior of the target visual system and used it to construct images predicted to either broadly activate large populations of neurons or selectively activate one population while keeping the others unchanged. They then analyzed the effectiveness of these images in producing the desired effects in the macaque visual cortex. The manipulations showed very strong effects and achieved considerable and highly selective influence over the neuronal populations. Using novel and non-naturalistic images, the neural network was shown to reproduce the overall behavior of the animals - neural responses. The pattern of light that strikes the eyes is processed and re-represented via patterns of neural activity in a 'deep' series of six interconnected cortical brain areas called the ventral visual stream. Visual neuroscience research has revealed that these patterns of neural activity underlie our ability to recognize objects and their relationships in the world. Recent advances have enabled neuroscientists to build ever more precise models of this complex visual processing. Currently, the best such models are particular deep artificial neural network (ANN) models in which each brain area has a corresponding model layer and each brain neuron has a corresponding model neuron. Such models are quite good at predicting the responses of brain neurons, but their contribution to an understanding of primate visual processing remains controversial.},
}
@article{BassoketalJEPLMC-12,
       author = {Bassok, Miriam and Holyoak, Keith and Dunbar, Kevin N.},
      journal = {Journal of Experimental Psychology. Learning, Memory \& Cognition},
       number = {2},
        pages = {261-263},
        title = {Introduction to the Special Section on the Neural Substrate of Analogical Reasoning and Metaphor Comprehension.},
       volume = {38},
         year = {2012},
     abstract = {The special section on the neural substrate of relational reasoning includes 4 articles that address the processes and brain regions involved in analogical reasoning (Green, Kraemer, Fugelsang, Gray, \& Dunbar, 2011; Maguire, McClelland, Donovan, Tiliman, \& Krawczyk, 2011) and in metaphor comprehension (Chettih, Durgin,\& Grodner, 2011; Prat, Mason,\& Just, 2011). We see this work as an example of how neuroscience approaches to cognition can lead to increased understanding of cognitive processes. In this brief introduction, we first situate the 4 articles in the context of prior cognitive neuroscience work on relational reasoning. We then highlight the main issues explored in these articles: different sources of complexity and difficulty in relational processing, potential differences between the roles of the 2 hemispheres, and the impact of individual differences in various cognitive abilities. The 4 articles illustrate a range of methodologies, including functional magnetic resonance},
}
@article{BastosandTaylorNATURE-COMMUNICATIONS-20,
       author = {Bastos, Amalia P. M. and Taylor, Alex H.},
        title = {Kea show three signatures of domain-general statistical inference},
      journal = {Nature Communications},
        volume = {11},
        issue = {1},
         year = {2020},
       pages = {828},
     abstract = "One key aspect of domain-general thought is the ability to integrate information across different cognitive domains. Here, we tested whether kea (Nestor notabilis) can use relative quantities when predicting sampling outcomes, and then integrate both physical information about the presence of a barrier, and social information about the biased sampling of an experimenter, into their predictions. Our results show that kea exhibit three signatures of statistical inference, and therefore can integrate knowledge across different cognitive domains to flexibly adjust their predictions of sampling events. This result provides evidence that true statistical inference is found outside of the great apes, and that aspects of domain-general thinking can convergently evolve in brains with a highly different structure from primates. This has important implications not only for our understanding of how intelligence evolves, but also for research focused on how to create artificial domain-general thought processes.",
}
@article{BasuetalJMLR-15,
       author = {Sumanta Basu and Ali Shojaie and George Michailidis},
        title = {Network Granger Causality with Inherent Grouping Structure},
      journal = {Journal of Machine Learning Research},
         year = {2015},
       volume = {16},
        pages = {417-453},
     abstract = {The problem of estimating high-dimensional network models arises naturally in the analysis of many biological and socio-economic systems. In this work, we aim to learn a network structure from temporal panel data, employing the framework of Granger causal models under the assumptions of sparsity of its edges and inherent grouping structure among its nodes. To that end, we introduce a group lasso regression regularization framework, and also examine a thresholded variant to address the issue of group misspecification. Further, the norm consistency and variable selection consistency of the estimates are established, the latter under the novel concept of direction consistency. The performance of the proposed methodology is assessed through an extensive set of simulation studies and comparisons with existing techniques. The study is illustrated on two motivating examples coming from functional genomics and financial econometrics.}
}
@techreport{BasyePhD,
       author = {Basye, Kenneth},
        title = {A Framework for Map Construction},
  institution = {Brown University Department of Computer Science},
         year = 1992,
      address = {Providence, RI},
}
@inproceedings{BasyeandDeanJPL-89,
       author = {Basye, Kenneth and Dean, Thomas},
        title = {Map Learning With Indistinguishable Locations},
    booktitle = {Proceedings of the {NASA} Conference on Telerobotics},
         year = 1989,
}
@inproceedings{BasyeandDeanUAI-89,
       author = {Basye, Kenneth and Dean, Thomas},
        title = {Map Learning with Indistinguishable Locations},
    booktitle = {Proceedings of the 5th Conference on Uncertainty in Artificial Intelligence},
 organization = {AUAI},
     location = {Windsor, Ontario},
         year = 1989,
        pages = {7-13},
}
@article{BasyeetalAIJ-95,
       author = {Basye, Kenneth and Dean, Thomas and Kaelbling, Leslie},
        title = {Learning Dynamics: {S}ystem Identification for Perceptually Challenged Agents},
      journal = AIJ,
       volume = 72,
        issue = {1-2},
        pages = {139-171},
         year = 1995,
}
@article{BasyeetalEXPERT-92,
       author = {Basye, Kenneth and Dean, Thomas and Kirman, Jak and Lejter, Moises},
        title = {A Decision-Theoretic Approach to Planning, Perception, and Control},
      journal = {{IEEE} Expert},
       volume = 7,
        issue = 4,
        pages = {58-65},
         year = 1992,
}
@inproceedings{BasyeetalIJCAI-89,
       author = {Basye, Kenneth and Dean, Thomas and Vitter, Jeffrey Scott},
        title = {Coping With Uncertainty in Map Learning},
    booktitle = {Proceedings {IJCAI}-89},
     location = {Detroit, Michigan},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
 organization = {IJCAI},
        pages = {663-668},
         year = 1989,
}
@article{BasyeetalML-97,
       author = {Basye, Kenneth and Dean, Thomas and Vitter, Jeffrey Scott},
        title = {Coping With Uncertainty in Map Learning},
      journal = {Machine Learning},
       volume = 29,
        issue = 1,
        pages = {65-88},
         year = 1997,
}
@book{Batchelor2017,
        title = {Secular Buddhism: Imagining the Dharma in an Uncertain World},
       author = {Batchelor, Stephen},
         year = {2017},
    publisher = {Yale University Press}
}
@article{BatesCACM-94,
    author = {Bates, Joseph},
     title = {The role of emotion in believable agents},
   journal = {Communications of the {ACM}},
    volume = 37,
     issue = 7,
      year = 1994,
     pages = {122-125},
 publisher = {ACM},
   address = {New York, NY},
}
@incollection{BatesRETHINKING-98,
        title = {Rethinking Innateness},
       author = {Elizabeth A. Bates},
    booktitle = {Rethinking Innateness: A Connectionist Perspective on Development},
       editor = {Elman, J.L. and Bates, E.A. and Johnson, M.H.},
       series = {A Bradford book},
         year = {1998},
    publisher = {A Bradford Book}
}
@book{Bateson2002mind,
        title = {Mind and Nature: A Necessary Unity},
       author = {Bateson, G.},
       series = {Advances in systems theory, complexity, and the human sciences},
         year = {2002},
    publisher = {Hampton Press}
}
@article{BattagliaetalCSH-04,
       author = {Battaglia, F. P. and Sutherland, G. R. and McNaughton, B. L.},
        title = {Hippocampal sharp wave bursts coincide with neocortical up-state transitions},
      journal = {Learning Memory},
       volume = {11},
       number = {6},
         year = {2004},
        pages = {697-704},
     abstract = {The sleeping neocortex shows nested oscillatory activity in different frequency ranges, characterized by fluctuations between "up-states" and "down-states." High-density neuronal ensemble recordings in rats now reveal the interaction between synchronized activity in the hippocampus and neocortex: Electroencephalographic sharp waves in the hippocampus were more probable during down-states than during up-states, and tended to coincide with transitions from down-states to up-states. The form of cortical activity fluctuations and their interactions with sharp waves depend on sleep depth: In deeper sleep stages, characterized by strong neocortical oscillation in the delta range or slower (approximately 0.8-4 Hz), sharp-wave-triggered peri-event time histograms (PETH) are consistent with a longer duration for down-states than for up-states. In lighter sleep, the sharp-wave-triggered PETH suggested longer up-states than down-states. These results highlight the interplay in the hippocampal/neocortical loop: Decreased neocortical input during down-states may be a factor in generation of sharp waves. In turn, sharp waves may facilitate down-to-up transitions. This interplay may reflect joint memory trace reactivation in the hippocampus and in the neocortex, possibly contributing to consolidation of long-term memory: Off-line reactivation of recent neural activity patterns in the hippocampus occurs during 50-100-msec electroencephalographic sharp waves, corresponding to pyramidal-cell population bursts. The neocortical up-states starting in correspondence with sharp waves may be influenced by the reactivated information carried by the hippocampal sharp wave.}
}
@article{BattagliaetalCoRR-16,
       author = {Peter W. Battaglia and Razvan Pascanu and Matthew Lai and Danilo Jimenez Rezende and Koray Kavukcuoglu},
        title = {Interaction Networks for Learning about Objects, Relations and Physics},
      journal = {CoRR},
       volume = {arXiv:1612.00222},
         year = {2016},
     abstract = {Reasoning about objects, relations, and physics is central to human intelligence, and a key goal of artificial intelligence. Here we introduce the interaction network, a model which can reason about how objects in complex systems interact, supporting dynamical predictions, as well as inferences about the abstract properties of the system. Our model takes graphs as input, performs object- and relation-centric reasoning in a way that is analogous to a simulation, and is implemented using deep neural networks. We evaluate its ability to reason about several challenging physical domains: n-body problems, rigid-body collision, and non-rigid dynamics. Our results show it can be trained to accurately simulate the physical trajectories of dozens of objects over thousands of time steps, estimate abstract quantities such as energy, and generalize automatically to systems with different numbers and configurations of objects and relations. Our interaction network implementation is the first general-purpose, learnable physics engine, and a powerful general framework for reasoning about object and relations in a wide variety of complex real-world domains.}
}
@article{BattagliaetalCoRR-18,
       author = {Peter W. Battaglia and Jessica B. Hamrick and Victor Bapst and Alvaro Sanchez-Gonzalez and Vinicius Zambaldi and Mateusz Malinowski and Andrea Tacchetti and David Raposo and Adam Santoro and Ryan Faulkner and Caglar Gulcehre and Francis Song and Andrew Ballard and Justin Gilmer and George Dahl and Ashish Vaswani and Kelsey Allen and Charles Nash and Victoria Langston and Chris Dyer and Nicolas Heess and Daan Wierstra and Pushmeet Kohli and Matt Botvinick and Oriol Vinyals and Yujia Li and Razvan Pascanu},
        title = {Relational inductive biases, deep learning, and graph networks},
      journal = {CoRR},
       volume = {arXiv:1806.01261},
         year = {2018},
     abstract = {Artificial intelligence (AI) has undergone a renaissance recently, making major progress in key domains such as vision, language, control, and decision-making. This has been due, in part, to cheap data and cheap compute resources, which have fit the natural strengths of deep learning. However, many defining characteristics of human intelligence, which developed under much different pressures, remain out of reach for current approaches. In particular, generalizing beyond one's experiences--a hallmark of human intelligence from infancy--remains a formidable challenge for modern AI. The following is part position paper, part review, and part unification. We argue that combinatorial generalization must be a top priority for AI to achieve human-like abilities, and that structured representations and computations are key to realizing this objective. Just as biology uses nature and nurture cooperatively, we reject the false choice between "hand-engineering" and "end-to-end" learning, and instead advocate for an approach which benefits from their complementary strengths. We explore how using relational inductive biases within deep learning architectures can facilitate learning about entities, relations, and rules for composing them. We present a new building block for the AI toolkit with a strong relational inductive bias--the graph network--which generalizes and extends various approaches for neural networks that operate on graphs, and provides a straightforward interface for manipulating structured knowledge and producing structured behaviors. We discuss how graph networks can support relational reasoning and combinatorial generalization, laying the foundation for more sophisticated, interpretable, and flexible patterns of reasoning. As a companion to this paper, we have released an open-source software library for building graph networks, with demonstrations of how to use them in practice.}
}
@inproceedings{BattagliaetalNIPS-16,
       author = {Battaglia, Peter and Pascanu, Razvan and Lai, Matthew and Rezende, Danilo Jimenez and Kavukcuoglu, Koray},
        title = {Interaction Networks for Learning About Objects, Relations and Physics},
    booktitle = {Proceedings of the 30th International Conference on Neural Information Processing Systems},
    publisher = {Curran Associates Inc.},
     location = {Barcelona, Spain},
         year = {2016},
        pages = {4509-4517},
     abstract = {Reasoning about objects, relations, and physics is central to human intelligence, and a key goal of artificial intelligence. Here we introduce the interaction network, a model which can reason about how objects in complex systems interact, supporting dynamical predictions, as well as inferences about the abstract properties of the system. Our model takes graphs as input, performs object- and relation-centric reasoning in a way that is analogous to a simulation, and is implemented using deep neural networks. We evaluate its ability to reason about several challenging physical domains: n-body problems, rigid-body collision, and non-rigid dynamics. Our results show it can be trained to accurately simulate the physical trajectories of dozens of objects over thousands of time steps, estimate abstract quantities such as energy, and generalize automatically to systems with different numbers and configurations of objects and relations. Our interaction network implementation is the first general-purpose, learnable physics engine, and a powerful general framework for reasoning about object and relations in a wide variety of complex real-world domains.},
}
@article{BattyetalICMLR-17,
       author = {Eleanor Batty and Josh Merel and Nora Brackbill and Alexander Heitman and Alexander Sher and Alan Litke and E.J. Chichilnisky and Liam Paninski},
        title = {Multilayer recurrent network models of primate retinal ganglion cell responses},
      journal = {Internation Conference on Machine Learning and Representation},
         year = {2017},
     abstract = {Developing accurate predictive models of sensory neurons is vital to understanding sensory processing and brain computations. The current standard approach to modeling neurons is to start with simple models and to incrementally add interpretable features. An alternative approach is to start with a more complex model that captures responses accurately, and then probe the fitted model structure to understand the neural computations. Here, we show that a multitask recurrent neural network (RNN) framework provides the flexibility necessary to model complex computations of neurons that cannot be captured by previous methods. Specifically, multilayer recurrent neural networks that share features across neurons outperform generalized linear models (GLMs) in predicting the spiking responses of parasol ganglion cells in the primate retina to natural images. The networks achieve good predictive performance given surprisingly small amounts of experimental training data. Additionally, we present a novel GLM-RNN hybrid model with separate spa- tial and temporal processing components which provides insights into the aspects of retinal processing better captured by the recurrent neural networks.},
}
@article{BauerDEVELOPMENTAL-REVIEW-15,
       author = {Bauer, Patricia J.},
        title = {Development of episodic and autobiographical memory: The importance of remembering forgetting},
      journal = {Developmental Review},
         year = {2015},
       volume = {38},
        pages = {146-166},
     abstract = {Some memories of the events of our lives have a long shelf-life-they remain accessible to recollection even after long delays. Yet many other of our experiences are forgotten, sometimes very soon after they take place. In spite of the prevalence of forgetting, theories of the development of episodic and autobiographical memory largely ignore it as a potential source of variance in explanation of age-related variability in long-term recall. They focus instead on what may be viewed as positive developmental changes, that is, changes that result in improvements in the quality of memory representations that are formed. The purpose of this review is to highlight the role of forgetting as an important variable in understanding the development of episodic and autobiographical memory. Forgetting processes are implicated as a source of variability in long-term recall due to the protracted course of development of the neural substrate responsible for transformation of fleeting experiences into memory traces that can be integrated into long-term stores and retrieved at later points in time. It is logical to assume that while the substrate is developing, neural processing is relatively inefficient and ineffective, resulting in loss of information from memory (i.e., forgetting). For this reason, focus on developmental increases in the quality of representations of past events and experiences will tell only a part of the story of how memory develops. A more complete account is afforded when we also consider changes in forgetting.},
}
@article{BaueretalCONNECTION-SCIENCE-15,
    author = {Johannes Bauer and Jorge D\'{a}vila-Chac\'{́o}n and Stefan Wermter},
     title = {Modeling development of natural multi-sensory integration using neural self-organisation and probabilistic population codes},
   journal = {Connection Science},
    volume = {27},
     issue = {4},
      year = {2015},
     pages = {358-376},
     abstract = {Humans and other animals have been shown to perform near-optimally in multi-sensory integration tasks. Probabilistic population codes (PPCs) have been proposed as a mechanism by which optimal integration can be accomplished. Previous approaches have focussed on how neural networks might produce PPCs from sensory input or perform calculations using them, like combining multiple PPCs. Less attention has been given to the question of how the necessary organisation of neurons can arise and how the required knowledge about the input statistics can be learned. In this paper, we propose a model of learning multi-sensory integration based on an unsupervised learning algorithm in which an artificial neural network learns the noise characteristics of each of its sources of input. Our algorithm borrows from the self-organising map the ability to learn latent-variable models of the input and extends it to learning to produce a PPC approximating a probability density function over the latent variable behind its (noisy) input. The neurons in our network are only required to perform simple calculations and we make few assumptions about input noise properties and tuning functions. We report on a neurorobotic experiment in which we apply our algorithm to multi-sensory integration in a humanoid robot to demonstrate its effectiveness and compare it to human multi-sensory integration on the behavioural level. We also show in simulations that our algorithm performs near-optimally under certain plausible conditions, and that it reproduces important aspects of natural multi-sensory integration on the neural level.}
}
@article{BaumeetalAMS-70,
       author = {Baum, L. C. and Petrie, T. and Soules, G and Weiss, N.},
        title = {A maximization technique occurring in the statistical analysis of probabilistic functions in {M}arkov chains},
      journal = {Annals of Mathematical Statistics},
       volume = 41,
        issue = 1,
        pages = {164-171},
         year = 1970,
}
@book{Baumetal00,
      author = {Baum, D. and Gasperi, M. and Hempel, R. and Villa, L.},
       title = {Extreme Mindstorms: An Advanced Guide to Lego Mindstorms},
   publisher = {Apress},
     address = {Berkeley, CA},
        year = 2000,
}
@article{BaumolTAER-86,
       author = {William J. Baumol},
        title = {Productivity Growth, Convergence, and Welfare: What the Long-Run Data Show},
      journal = {The American Economic Review},
       volume = {76},
        issue = {5},
         year = {1986},
        pages = {1072-1085},
}
@article{BaxterandBartlettJAIR-01,
       author = {J. Baxter and P. L. Bartlett},
        title = {Infinite-Horizon Gradient-Based Policy Search},
      journal = {Journal of Artificial Intelligence Research},
       volume = 15,
         year = 2001,
        pages = {319-350},
     abstract = {Gradient-based approaches to direct policy search in reinforcement learning have received much recent attention as a means to solve problems of partial observability and to avoid some of the problems associated with policy degradation in value-function methods. In this paper we introduce REINFORCE, a simulation-based algorithm for generating a biased estimate of the gradient of the average reward in Partially Observable Markov Decision Processes (POMDPs) controlled by parameterized stochastic policies. A similar algorithm was proposed by Kimura, Yamamura, and Kobayashi (1995). The algorithm's chief advantages are that it requires storage of only twice the number of policy parameters, uses one free parameter $\hmbeta{} \hmisin{} [0, 1]$ (which has a natural interpretation in terms of bias-variance trade-off), and requires no knowledge of the underlying state. We prove convergence of REINFORCE, and show how the correct choice of the parameter $\hmbeta{}$ is related to the mixing time of the controlled POMDP. We briefly describe extensions of REINFORCE to controlled Markov chains, continuous state, observation and control spaces, multiple-agents, higher-order derivatives, and a version for training stochastic policies with internal states. In a companion paper (Baxter, Bartlett, \& Weaver, 2001) we show how the gradient estimates generated by REINFORCE can be used in both a traditional stochastic gradient algorithm and a conjugate-gradient procedure to find local optima of the average reward.}
}
@article{BaxteretalCSHL-08,
       author = {Baxter, Mark G. and Browning, Philip GF and Mitchell, Anna S.},
        title = {Perseverative interference with object-in-place scene learning in rhesus monkeys with bilateral ablation of ventrolateral prefrontal cortex},
      journal = {Learning and Memory},
         year = {2008},
    publisher = {Cold Spring Harbor Laboratory Press},
       volume = {15},
        issue = {3},
        pages = {126-132},
     abstract = {Surgical disconnection of the frontal cortex and inferotemporal cortex severely impairs many aspects of visual learning and memory, including learning of new object-in-place scene memory problems, a monkey model of episodic memory. As part of a study of specialization within prefrontal cortex in visual learning and memory, we tested monkeys with bilateral ablations of ventrolateral prefrontal cortex in object-in-place scene learning. These monkeys were mildly impaired in scene learning relative to their own preoperative performance, similar in severity to that of monkeys with bilateral ablation of orbital prefrontal cortex. An analysis of response types showed that the monkeys with lesions were specifically impaired in responding to negative feedback during learning: The post-operative increase in errors was limited to trials in which the first response to each new problem, made on the basis of trial and error, was incorrect. This perseverative pattern of deficit was not observed in the same analysis of response types in monkeys with bilateral ablations of the orbital prefrontal cortex, who were equally impaired on trials with correct and incorrect first responses. This may represent a specific signature of ventrolateral prefrontal involvement in episodic learning and memory.},
}
@book{BazaraaJarvis90,
       author = {M.S. Bazaraa, J. J. Jarvis, H. D. Sherali},
        title = {Linear Programming and Network Flows},
    publisher = {John Wiley \& Sons},
      address = {New York},
         year = 1990,
}
@article{BealetalPLoS-11,
        title = {Automatic Compilation from High-Level Biologically-Oriented Programming Language to Genetic Regulatory Networks},
       author = {Beal, Jacob and Lu, Ting and Weiss, Ron},
      journal = {{PLoS} ONE},
    publisher = {Public Library of Science},
       volume = 6,
        issue = 8,
         year = 2011,
        pages = {e22490},
     abstract = {The field of synthetic biology promises to revolutionize our ability to engineer biological systems, providing important benefits for a variety of applications. Recent advances in DNA synthesis and automated DNA assembly technologies suggest that it is now possible to construct synthetic systems of significant complexity. However, while a variety of novel genetic devices and small engineered gene networks have been successfully demonstrated, the regulatory complexity of synthetic systems that have been reported recently has somewhat plateaued due to a variety of factors, including the complexity of biology itself and the lag in our ability to design and optimize sophisticated biological circuitry. [PAR] To address the gap between DNA synthesis and circuit design capabilities, we present a platform that enables synthetic biologists to express desired behavior using a convenient high-level biologically-oriented programming language, Proto. The high level specification is compiled, using a regulatory motif based mechanism, to a gene network, optimized, and then converted to a computational simulation for numerical verification. Through several example programs we illustrate the automated process of biological system design with our platform, and show that our compiler optimizations can yield significant reductions in the number of genes [PAR] Our platform provides a convenient and accessible tool for the automated design of sophisticated synthetic biological systems, bridging an important gap between DNA synthesis and circuit design capabilities. Our platform is user-friendly and features biologically relevant compiler optimizations, providing an important foundation for the development of sophisticated biological systems.},
}
@book{Bearetal2006,
    author = {Mark F. Bear and Barry Connors and Michael Paradiso},
     title = {Neuroscience: Exploring the Brain (Third Edition)},
 publisher = {Lippincott Williams \& Wilkins},
   address = {Baltimore, Maryland},
      year = 2006,
}
@book{Bearetal2007,
    author = {Mark F. Bear and Barry Connors and Michael Paradiso},
     title = {Neuroscience: Exploring the Brain (Third Edition)},
 publisher = {Lippincott Williams \& Wilkins},
   address = {Baltimore, Maryland},
      year = 2006,
}
@book{Bearetal2015,
    author = {Mark F. Bear and Barry Connors and Michael Paradiso},
     title = {Neuroscience: Exploring the Brain (Fifth Edition)},
 publisher = {Lippincott Williams \& Wilkins},
   address = {Baltimore, Maryland},
      year = 2015,
}
@article{BearetalCoRR-20,
        title = {Learning Physical Graph Representations from Visual Scenes},
       author = {Daniel M. Bear and Cairong Fan and Damian Mrowca and Yunzhu Li and Seth Alter and Aran Nayebi and Jeremy I Schwartz and Li Fei-Fei and Jiajun Wu and Joshua B. Tenenbaum and Daniel L. K. Yamins},
      journal = {CoRr},
         year = {2020},
       volume = {arXiv:2006.12373},
     abstract = {Convolutional Neural Networks (CNNs) have proved exceptional at learning representations for visual object categorization. However, CNNs do not explicitly encode objects, parts, and their physical properties, which has limited CNNs' success on tasks that require structured understanding of visual scenes. To overcome these limitations, we introduce the idea of Physical Scene Graphs (PSGs), which represent scenes as hierarchical graphs, with nodes in the hierarchy corresponding intuitively to object parts at different scales, and edges to physical connections between parts. Bound to each node is a vector of latent attributes that intuitively represent object properties such as surface shape and texture. We also describe PSGNet, a network architecture that learns to extract PSGs by reconstructing scenes through a PSG-structured bottleneck. PSGNet augments standard CNNs by including: recurrent feedback connections to combine low and high-level image information; graph pooling and vectorization operations that convert spatially-uniform feature maps into object-centric graph structures; and perceptual grouping principles to encourage the identification of meaningful scene elements. We show that PSGNet outperforms alternative self-supervised scene representation algorithms at scene segmentation tasks, especially on complex real-world images, and generalizes well to unseen object types and scene arrangements. PSGNet is also able learn from physical motion, enhancing scene estimates even for static images. We present a series of ablation studies illustrating the importance of each component of the PSGNet architecture, analyses showing that learned latent attributes capture intuitive scene properties, and illustrate the use of PSGs for compositional scene inference.}
}
@article{BeckerHIPPOCAMPUS-05,
       author = {Becker, S.},
        title = {A computational principle for hippocampal learning and neurogenesis},
      journal = {Hippocampus},
         year = {2005},
       volume = {15},
       number = {6},
        pages = {722-738},
     abstract = {In the three decades since Marr put forward his computational theory of hippocampal coding, many computational models have been built on the same key principles proposed by Marr: sparse representations, rapid Hebbian storage, associative recall and consolidation. Most of these models have focused on either the CA3 or CA1 fields, using "off-the-shelf" learning algorithms such as competitive learning or Hebbian pattern association. Here, we propose a novel coding principle that is common to all hippocampal regions, and from this one principal, we derive learning rules for each of the major pathways within the hippocampus. The learning rules turn out to have much in common with several models of CA3 and CA1 in the literature, and provide a unifying framework in which to view these models. Simulations of the complete circuit confirm that both recognition memory and recall are superior relative to a hippocampally lesioned model, consistent with human data. Further, we propose a functional role for neurogenesis in the dentate gyrus (DG), namely, to create distinct memory traces for highly similar items. Our simulation results support our prediction that memory capacity increases with the number of dentate granule cells, while neuronal turnover with a fixed dentate layer size improves recall, by minimizing interference between highly similar items.}
}
@article{BeckerandBergMD-01,
     title = {Neuroimaging in basal ganglia disorders: perspectives for transcranial ultrasound.},
    author = {Becker, G. and Berg, D.},
   journal = {Movement Disorders},
    volume = 16,
     issue = 1,
     pages = {23-32},
      year = 2001,
  abstract = {Transcranial sonography is a new diagnostic tool, allowing not only the evaluation of cerebral arteries but also the two-dimensional display of the brain parenchyma. In this review we will summarize basics of the application, the ultrasound anatomy of the brain and sonographic findings in some movement disorders. While in normal adults basal ganglia nuclei are hypoechogenic, they are hyperechogenic in certain basal ganglia disorders. In Parkinson's disease, for example, the substantia nigra can be depicted as a distinctly echogenic area. An elevated echogenicity of the lentiform nuclei was noticed in patients with primary adult-onset dystonia. In both disorders the altered echogenicity may arise from higher heavy metal tissue content (i.e. iron in Parkinson's disease and copper in primary dystonia). Our findings converge to the hypothesis that transcranial ultrasound sensitively detects pathological metal accumulation not identified by other neuroimaging techniques (CT and MRI) and therefore provides new insights in the diagnosis of basal ganglia disorders. The implications of these findings for the understanding of the pathogenesis and its usefulness for the early diagnosis of movement disorders are outlined.}
}
@article{BeckinghausenandSillitoeNEUROSCIENCE-LETTERS-19,
        title = {Insights into cerebellar development and connectivity},
       author = {Jaclyn Beckinghausen and Roy V. Sillitoe},
      journal = {Neuroscience Letters},
       volume = {688},
        pages = {2-13},
         year = {2019},
     abstract = {The cerebellum has a well-established role in controlling motor functions such coordination, balance, posture, and skilled learning. There is mounting evidence that it might also play a critical role in non-motor functions such as cognition and emotion. It is therefore not surprising that cerebellar defects are associated with a wide array of diseases including ataxia, dystonia, tremor, schizophrenia, dyslexia, and autism spectrum disorder. What is intriguing is that a seemingly uniform circuit that is often described as being "simple" should carry out all of these behaviors. Analyses of how cerebellar circuits develop have revealed that such descriptions massively underestimate the complexity of the cerebellum. The cerebellum is in fact highly patterned and organized around a series of parasagittal stripes and transverse zones. This topographic architecture partitions all cerebellar circuits into functional modules that are thought to enhance processing power during cerebellar dependent behaviors. What are arguably the most remarkable features of cerebellar topography are the developmental processes that produce them. This review is concerned with the genetic and cellular mechanisms that orchestrate cerebellar patterning. We place a major focus on how Purkinje cells control multiple aspects of cerebellar circuit assembly. Using this model, we discuss evidence for how "zebra-like" patterns in Purkinje cells sculpt the cerebellum, how specific genetic cues mediate the process, and how activity refines the patterns into an adult map that is capable of executing various functions. We also discuss how defective Purkinje cell patterning might impact the pathogenesis of neurological conditions.}
}
@article{BednyetalPNAS-09,
       author = {Bedny, Marina and Pascual-Leone, Alvaro and Saxe, Rebecca R.},
        title = {Growing up blind does not change the neural bases of Theory of Mind},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {106},
       number = {27},
        pages = {11312-11317},
         year = {2009},
     abstract = {Humans reason about the mental states of others; this capacity is called Theory of Mind (ToM). In typically developing adults, ToM is supported by a consistent group of brain regions: the bilateral temporoparietal junction (TPJ), medial prefrontal cortex (MPFC), precuneus (PC), and anterior temporal sulci (aSTS). How experience and intrinsic biological factors interact to produce this adult functional profile is not known. In the current study we investigate the role of visual experience in the development of the ToM network by studying congenitally blind adults. In experiment 1, participants listened to stories and answered true/false questions about them. The stories were either about mental or physical representations of reality (e.g., photographs). In experiment 2, participants listened to stories about people's beliefs based on seeing or hearing; people's bodily sensations (e.g., hunger); and control stories without people. Participants judged whether each story had positive or negative valance. We find that ToM brain regions of sighted and congenitally blind adults are similarly localized and functionally specific. In congenitally blind adults, reasoning about mental states leads to activity in bilateral TPJ, MPFC, PC, and aSTS. These brain regions responded more to passages about beliefs than passages about nonbelief representations or passages about bodily sensations. Reasoning about mental states that are based on seeing is furthermore similar in congenitally blind and sighted individuals. Despite their different developmental experience, congenitally blind adults have a typical ToM network. We conclude that the development of neural mechanisms for ToM depends on innate factors and on experiences represented at an abstract level, amodally.},
}
@article{BegleiterJAIR-04,
       author = {R. Begleiter and R. El-Yaniv and G. Yona},
        title = {On Prediction Using Variable Order {M}arkov Models},
      journal = {Journal of Artificial Intelligence Research},
         year = 2004,
       volume = 22,
        pages = {385-421},
}
@article{BehniaetalNATURE-14,
       author = {Behnia, Rudy and Clark, Damon A. and Carter, Adam G. and Clandinin, Thomas R. and Desplan, Claude},
        title = {Processing properties of {ON} and {OFF} pathways for Drosophila motion detection},
      journal = {Nature},
         year = {2014},
    publisher = {Nature Publishing Group},
}
@article{BehrendtFiP-13,
       author = {Behrendt, Ralf-Peter},
        title = {Conscious Experience and Episodic Memory: Hippocampus at the Crossroads},
      journal = {Frontiers in Psychology},
       volume = {4},
        pages = {304},
         year = {2013},
     abstract = {If an instance of conscious experience of the seemingly objective world around us could be regarded as a newly formed event memory, much as an instance of mental imagery has the content of a retrieved event memory, and if, therefore, the stream of conscious experience could be seen as evidence for ongoing formation of event memories that are linked into episodic memory sequences, then unitary conscious experience could be defined as a symbolic representation of the pattern of hippocampal neuronal firing that encodes an event memory -- a theoretical stance that may shed light into the mind-body and binding problems in consciousness research. Exceedingly detailed symbols that describe patterns of activity rapidly self-organizing, at each cycle of the θ rhythm, in the hippocampus are instances of unitary conscious experience that jointly constitute the stream of consciousness. Integrating object information (derived from the ventral visual stream and orbitofrontal cortex) with contextual emotional information (from the anterior insula) and spatial environmental information (from the dorsal visual stream), the hippocampus rapidly forms event codes that have the informational content of objects embedded in an emotional and spatiotemporally extending context. Event codes, formed in the CA3-dentate network for the purpose of their memorization, are not only contextualized but also allocentric representations, similarly to conscious experiences of events and objects situated in a seemingly objective and observer-independent framework of phenomenal space and time. Conscious perception is likely to be related to more fleeting and seemingly internal forms of conscious experience, such as autobiographical memory recall, mental imagery, including goal anticipation, and to other forms of externalized conscious experience, namely dreaming and hallucinations; and evidence pointing to an important contribution of the hippocampus to these conscious phenomena will be reviewed.}
}
@article{BejeranoetalBIOMETRICS-01,
       author = {Gill Bejerano and Yevgeny Seldin and Hanah Margalit and Naftali Tishby},
        title = {Markovian domain fingerprinting: statistical segmentation of protein sequences},
      journal = {Bioinformatics},
       volume = 17,
        issue = 10,
        pages = {927-934},
         year = 2001,
}
@article{BelatrecheetalAAA-10,
       author = {Ammar Belatreche and Liam Maguire and Martin McGinnity and Liam McDaid and Arfan Ghani},
        title = {Computing with Biologically Inspired Neural Oscillators: Application to Colour Image Segmentation},
      journal = {Advances in Artificial Intelligence},
         year = 2010,
        pages = 405073,
     abstract = {This paper investigates the computing capabilities and potential applications of neural oscillators, a biologically inspired neural model, to grey scale and colour image segmentation, an important task in image understanding and object recognition. A proposed neural system that exploits the synergy between neural oscillators and Kohonen self-organising maps (SOMs) is presented. It consists of a two-dimensional grid of neural oscillators which are locally connected through excitatory connections and globally connected to a common inhibitor. Each neuron is mapped to a pixel of the input image and existing objects, represented by homogenous areas, are temporally segmented through synchronisation of the activity of neural oscillators that are mapped to pixels of the same object. Self-organising maps form the basis of a colour reduction system whose output is fed to a 2D grid of neural oscillators for temporal correlation-based object segmentation. Both chromatic and local spatial features are used. The system is simulated in Matlab and its demonstration on real world colour images shows promising results and the emergence of a new bioinspired approach for colour image segmentation. The paper concludes with a discussion of the performance of the proposed system and its comparison with traditional image segmentation approaches.}
}
@article{BelgardetalNEURON-11,
     title = {A transcriptomic atlas of mouse neocortical layers}, 
    author = {Belgard, T. Grant and Marques, Ana C. and Oliver, Peter L. and Abaan, Hatice Ozel O. and Sirey, Tamara M. and Hoerder-Suabedissen, Anna and Garc\'{\i}a-Moreno, Fernando and Moln\'{a}r, Zolt\'{a}n and Margulies, Elliott H. and Ponting, Chris P.}, 
   journal = {Neuron}, 
    volume = 71, 
     issue = 4, 
 publisher = {Cell Press}, 
      year = 2011,
     pages = {605-616}, 
  abstract = {In the mammalian cortex, neurons and glia form a patterned structure across six layers whose complex cytoarchitectonic arrangement is likely to contribute to cognition. We sequenced transcriptomes from layers 1-6b of different areas (primary and secondary) of the adult (postnatal day 56) mouse somatosensory cortex to understand the transcriptional levels and functional repertoires of coding and noncoding loci for cells constituting these layers. A total of 5,835 protein-coding genes and 66 noncoding RNA loci are differentially expressed ("patterned") across the layers, on the basis of a machine-learning model (naive Bayes) approach. Layers 2-6b are each associated with specific functional and disease annotations that provide insights into their biological roles. This new resource (http://genserv.anat.ox.ac.uk/layers  greatly extends currently available resources, such as the Allen Mouse Brain Atlas and microarray data sets, by providing quantitative expression levels, by being genome-wide, by including novel loci, and by identifying candidate alternatively spliced
  transcripts that are differentially expressed across layers.},
}
@article{BelkinandNiyogiJCSS-08,
      title = {Towards a theoretical foundation for {L}aplacian-based manifold methods},
     author = {Mikhail Belkin and Partha Niyogi},
    journal = {Journal of Computer and Systems Science},
      issue = 8,
     volume = 74,
      pages = {1289-1308},
       year = 2008,
}
@techreport{Bell,
       author = {Bell, Colin E. and Tate, Austin},
        title = {Using Temporal Constraints to Restrict Search in a Planner},
        issue = {AIAI-TR-5},
  institution = {Artificial Intelligence Applications Institute, University of Edinburgh},
         year = 1985,
}
@book{BellandGemmell2009,
      title = {Total Recall: How the E-Memory Revolution Will Change Everything},
     author = {C. Gordon Bell, Jim Gemmell},
  publisher = {Dutton},
    address = {New York},
       year = 2009,
}
@article{BellandSejnowskiNC-95,
       author = {A.J. Bell and T.J. Sejnowski},
        title = {An information-maximization approach to blind separation and blind deconvolution},
      journal = {Neural Computation},
         year = 1995,
       volume = 7,
        pages = {1129-1159},
}
@book{Bellman57,
       author = {Bellman, Richard},
        title = {Dynamic Programming},
    publisher = {Princeton University Press},
         year = 1957,
}
@book{Bellman61,
       author = {Bellman, Richard},
        title = {Adaptive Control Processes},
    publisher = {Princeton University Press},
      address = {Princeton, New Jersey},
         year = 1961,
}
@book{Bellman62,
       author = {Bellman, Richard and Dreyfus, Stuart},
        title = {Applied Dynamic Programming},
    publisher = {Princeton University Press},
      address = {Princeton, New Jersey},
         year = 1962,
}
@article{BellmanCACM-61,
       author = {R. Bellman},
        title = {On the approximation of curves by line segments using dymanic programming},
      journal = {Communications of the {ACM}},
       volume = 4,
        issue = 6,
         year = 1961
}
@article{BelloetalCoRR-16,
       author = {Irwan Bello and Hieu Pham and Quoc V. Le and Mohammad Norouzi and Samy Bengio},
        title = {Neural Combinatorial Optimization with Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1611.09940},
         year = {2016},
     abstract = {This paper presents a framework to tackle combinatorial optimization problems using neural networks and reinforcement learning. We focus on the traveling salesman problem (TSP) and train a recurrent network that, given a set of city coordinates, predicts a distribution over different city permutations. Using negative tour length as the reward signal, we optimize the parameters of the recurrent network using a policy gradient method. We compare learning the network parameters on a set of training graphs against learning them on individual test graphs. Despite the computational expense, without much engineering and heuristic designing, Neural Combinatorial Optimization achieves close to optimal results on 2D Euclidean graphs with up to 100 nodes. Applied to the KnapSack, another NP-hard problem, the same method obtains optimal solutions for instances with up to 200 items.}
}  
@inproceedings{BelongieetalECCV-02,
     title = {Spectral Partitioning with Indefinite Kernels Using the Nystr{\"o}m Extension},
    author = {Serge Belongie and Charless Fowlkes and Fan R. K. Chung and Jitendra Malik},
 booktitle = {7th European Conference on Computer Vision}, 
      year = 2002,
     pages = {531-542},
}  
@article{BelongieetalPAMI-02,
       author = {Serge Belongie and Jitendra Malik and Jan Puzicha},
        title = {Shape Matching and Object Recognition Using Shape Contexts},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
       volume = 24,
        issue = 4,
        pages = {509-522},
         year = 2002,
}
@article{BeltramoetalNATURE-NEUROSCIENCE-13,
       author = {Beltramo, Riccardo and D'Urso, Giulia and Dal Maschio, Marco and Farisello, Pasqualina and Bovetti, Serena and Clovis, Yoanne and Lassi, Glenda and Tucci, Valter and De Pietri Tonelli, Davide and Fellin, Tommaso},
        title = {Layer-specific excitatory circuits differentially control recurrent network dynamics in the neocortex},
      journal = {Nature Neuroscience},
         year = 2013,
    publisher = {Nature Publishing Group},
       volume = 16,
        issue = 2,
        pages = {227-234},
     abstract = {In the absence of external stimuli, the mammalian neocortex shows intrinsic network oscillations. These dynamics are characterized by translaminar assemblies of neurons whose activity synchronizes rhythmically in space and time. How different cortical layers influence the formation of these spontaneous cellular assemblies is poorly understood. We found that excitatory neurons in supragranular and infragranular layers have distinct roles in the regulation of intrinsic low-frequency oscillations in mice in vivo. Optogenetic activation of infragranular neurons generated network activity that resembled spontaneous events, whereas photoinhibition of these same neurons substantially attenuated slow ongoing dynamics. In contrast, light activation and inhibition of supragranular cells had modest effects on spontaneous slow activity. This study represents, to the best of our knowledge, the first causal demonstration that excitatory circuits located in distinct cortical layers differentially control spontaneous low-frequency dynamics.}
}
@article{Ben-BassatIEEE-78,
       author = {Ben-Bassat, M.},
        title = {Myopic Policies in Sequential Classification},
      journal = {{IEEE} Transactions on Computers},
       volume = 27,
        pages = {170-174},
         year = 1978,
}
@article{Ben-ShaharandZuckerNC-04,
      title = {Geometrical computations explain projection patterns of long-range horizontal connections in visual cortex},
     author = {O. Ben-Shahar and S. Zucker}, 
      month = {March},
    journal = {Neural Computation},
     volume = 16, 
      issue = 3, 
      pages = {445-476},
       year = 2004, 
}
@article{BenAbdallahetalEN-11,
       author = {Nada M.-B. Ben Abdallah and Johannes Fuss and Massimo Trusel and Michael J. Galsworthy and Kristin Bobsin and Giovanni Colacicco and Robert M.J. Deacon and Marco A. Riva and Christoph Kellendonk and Rolf Sprengel and Hans-Peter Lipp and Peter Gass},
        title = {The puzzle box as a simple and efficient behavioral test for exploring impairments of general cognition and executive functions in mouse models of schizophrenia},
      journal = {Experimental Neurology},
       volume = {227},
       number = {1},
         year = {2011},
        pages = {42-52},
     abstract = {Deficits in executive functions are key features of schizophrenia. Rodent behavioral paradigms used so far to find animal correlates of such deficits require extensive effort and time. The puzzle box is a problem-solving test in which mice are required to complete escape tasks of increasing difficulty within a limited amount of time. Previous data have indicated that it is a quick but highly reliable test of higher-order cognitive functioning. We evaluated the use of the puzzle box to explore executive functioning in five different mouse models of schizophrenia: mice with prefrontal cortex and hippocampus lesions, mice treated sub-chronically with the NMDA-receptor antagonist MK-801, mice constitutively lacking the GluA1 subunit of AMPA-receptors, and mice over-expressing dopamine \{D2\} receptors in the striatum. All mice displayed altered executive functions in the puzzle box, although the nature and extent of the deficits varied between the different models. Deficits were strongest in hippocampus-lesioned and GluA1 knockout mice, while more subtle deficits but specific to problem solving were found in the medial prefrontal-lesioned mice, MK-801-treated mice, and in mice with striatal overexpression of \{D2\} receptors. Data from this study demonstrate the utility of the puzzle box as an effective screening tool for executive functions in general and for schizophrenia mouse models in particular.}
}
@article{BendorandWilsonNATURE-NEUROSCIENCE-12,
        title = {Biasing the content of hippocampal replay during sleep},
       author = {Bendor, Daniel and Wilson, Matthew A},
      journal = {Nature neuroscience},
       volume = {15},
       number = {10},
        pages = {1439},
         year = {2012},
    publisher = {Nature Publishing Group}
}
@article{BenedettietalSCHIZOPHRENIA-09,
       author = {Francesco Benedetti and Alessandro Bernasconi and Marta Bosia and Roberto Cavallaro and Sara Dallaspezia and Andrea Falini and Sara Poletti and Daniele Radaelli and Roberta Riccaboni and Giuseppe Scotti and Enrico Smeraldi},
        title = {Functional and structural brain correlates of theory of mind and empathy deficits in schizophrenia},
      journal = {Schizophrenia Research},
       volume = 114,
        issue = {1-3},
         year = 2009,
        pages = {154-160},
     abstract = {Background Patients affected by schizophrenia show deficits in social cognition, with abnormal performance on tasks targeting theory of mind (ToM) and empathy (Emp). Brain imaging studies suggested that ToM and Emp depend on the activation of brain networks mainly localized at the superior temporal lobe and temporo-parietal junction. Methods Participants included 24 schizophrenia patients and 20 control subjects. We used brain blood oxygen level dependent fMRI to study the neural responses to tasks targeting ToM and Emp. We then studied voxel-based morphometry of grey matter in areas where diagnosis influenced functional activation to both tasks. Outcomes were analyzed in the context of the general linear model, with global grey matter volume as nuisance covariate for structural MRI. Results Patients showed worse performance on both tasks. We found significant effects of diagnosis on neural responses to the tasks in a wide cluster in right posterior superior temporal lobe (encompassing \{BA\} 22-42), in smaller clusters in left temporo-parietal junction and temporal pole (BA 38 and 39), and in a white matter region adjacent to medial prefrontal cortex (BA 10). A pattern of double dissociation of the effects of diagnosis and task on neural responses emerged. Among these areas, grey matter volume was found to be reduced in right superior temporal lobe regions of patients. Conclusions Functional and structural abnormalities were observed in areas affected by the schizophrenic process early in the illness course, and known to be crucial for social cognition, suggesting a biological basis for social cognition deficits in schizophrenia.}
}
@book{Benford2008beyond,
        title = {Beyond Human: Living with Robots and Cyborgs},
       author = {Elisabeth Malartre and Gregory Benford},
         year = {2008},
    publisher = {Tom Doherty Associates}
}
@inproceedings{BengioCGEC-14,
       author = {Bengio, Yoshua},
        title = {Deep learning and cultural evolution},
    booktitle = {Proceedings of the Companion Publication of the 2014 Annual Conference on Genetic and Evolutionary Computation},
    publisher = {ACM},
     location = {New York, NY, USA},
         year = {2014},
     abstract = {We propose a theory and its first experimental tests, relating difficulty of learning in deep architectures to culture and language. The theory is articulated around the following hypotheses: learning in an individual human brain is hampered by the presence of effective local minima, particularly when it comes to learning higher-level abstractions, which are represented by the composition of many levels of representation, i.e., by deep architectures; a human brain can learn such high-level abstractions if guided by the signals produced by other humans, which act as hints for intermediate and higher-level abstractions; language and the recombination and optimization of mental concepts provide an efficient evolutionary recombination operator for this purpose. The theory is grounded in experimental observations of the difficulties of training deep artificial neural networks and an empirical test of the hypothesis regarding the need for guidance of intermediate concepts is demonstrated. This is done through a learning task on which all the tested machine learning algorithms failed, unless provided with hints about intermediate-level abstractions.}
}
@inproceedings{BengioCOGSCI-14,
       author = {Bengio, Yoshua},
        title = {{Deep learning, Brains and the Evolution of Culture}},
    booktitle = {Proceedings of the 36th Annual Conference of the Cognitive Science Society Workshop on Deep Learning and the Brain},
    publisher = {Cognitive Science Society},
     location = {Quebec City, Quebec, Canada},
         year = {2014},
}
@article{BengioCoRR-12,
       author = {Yoshua Bengio},
        title = {Evolving Culture vs Local Minima},
      journal = {CoRR},
       volume = {arXiv:1203.2990},
         year = {2012},
     abstract = {We propose a theory that relates difficulty of learning in deep architectures to culture and language. It is articulated around the following hypotheses: (1) learning in an individual human brain is hampered by the presence of effective local minima; (2) this optimization difficulty is particularly important when it comes to learning higher-level abstractions, i.e., concepts that cover a vast and highly-nonlinear span of sensory configurations; (3) such high-level abstractions are best represented in brains by the composition of many levels of representation, i.e., by deep architectures; (4) a human brain can learn such high-level abstractions if guided by the signals produced by other humans, which act as hints or indirect supervision for these high-level abstractions; and (5), language and the recombination and optimization of mental concepts provide an efficient evolutionary recombination operator, and this gives rise to rapid search in the space of communicable ideas that help humans build up better high-level internal representations of their world. These hypotheses put together imply that human culture and the evolution of ideas have been crucial to counter an optimization difficulty: this optimization difficulty would otherwise make it very difficult for human brains to capture high-level knowledge of the world. The theory is grounded in experimental observations of the difficulties of training deep artificial neural networks. Plausible consequences of this theory for the efficiency of cultural evolutions are sketched.}
}
@article{BengioCoRR-17,
       author = {Yoshua Bengio},
        title = {The Consciousness Prior},
      journal = {CoRR},
       volume = {arXiv:1709.08568},
         year = {2017},
     abstract = {A new prior is proposed for representation learning, which can be combined with other priors in order to help disentangling abstract factors from each other. It is inspired by the phenomenon of consciousness seen as the formation of a low-dimensional combination of a few concepts constituting a conscious thought, i.e., consciousness as awareness at a particular time instant. This provides a powerful constraint on the representation in that such low-dimensional thought vectors can correspond to statements about reality which are either true, highly probable, or very useful for taking decisions. The fact that a few elements of the current state can be combined into such a predictive or useful statement is a strong constraint and deviates considerably from the maximum likelihood approaches to modeling data and how states unfold in the future based on an agent's actions. Instead of making predictions in the sensory (e.g. pixel) space, the consciousness prior allow the agent to make predictions in the abstract space, with only a few dimensions of that space being involved in each of these predictions. The consciousness prior also makes it natural to map conscious states to natural language utterances or to express classical AI knowledge in the form of facts and rules, although the conscious states may be richer than what can be expressed easily in the form of a sentence, a fact or a rule.}
}
@article{BengioFTML-09,
        title = {Learning Deep Architectures for AI},
       author = {Yoshua Bengio},
      journal = {Foundations and Trends in Machine Learning},
       volume = 2,
        issue = 1,
         year = 2009,
        pages = {1-127},
}
@incollection{BengioandFrasconiNIPS-94a,
       author = {Bengio, Yoshua and Frasconi, P.},
        title = {An Input Output {HMM} Architecture},
       editor = {J. D. Cowan, et al.},
    booktitle = {Advances in Neural Information Processing 6},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1994,
}
@incollection{BengioandFrasconiNIPS-94b,
       author = {Bengio, Yoshua and Frasconi, P.},
        title = {Diffusion of Credit in {M}arkovian Models},
       editor = {J. D. Cowan, et al.},
    booktitle = {Advances in Neural Information Processing 6},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1994,
}
@incollection{BengioandLeCun07,
       author = {Yoshua Bengio and Yann LeCun},
        title = {Scaling learning algorithms towards {AI}},
    booktitle = {Large-Scale Kernel Machines},
       editor = {L. Bottou and O. Chapelle and D. DeCoste and J. Weston},
    publisher = {{MIT} Press},
          url = {http://yann.lecun.com/exdb/publis/pdf/bengio-lecun-07.pdf},
         year = 2007,
}
@article{BengioetalCoRR-13,
       author = {Yoshua Bengio and Nicholas L{\'{e}}onard and Aaron C. Courville},
        title = {Estimating or Propagating Gradients Through Stochastic Neurons for Conditional Computation},
      journal = {CoRR},
       volume = {arXiv:1308.3432},
         year = {2013},
     abstract = {Stochastic neurons and hard non-linearities can be useful for a number of reasons in deep learning models, but in many cases they pose a challenging problem: how to estimate the gradient of a loss function with respect to the input of such stochastic or non-smooth neurons? I.e., can we "back-propagate" through these stochastic neurons? We examine this question, existing approaches, and compare four families of solutions, applicable in different settings. One of them is the minimum variance unbiased gradient estimator for stochatic binary neurons (a special case of the REINFORCE algorithm). A second approach, introduced here, decomposes the operation of a binary stochastic neuron into a stochastic binary part and a smooth differentiable part, which approximates the expected effect of the pure stochatic binary neuron to first order. A third approach involves the injection of additive or multiplicative noise in a computational graph that is otherwise differentiable. A fourth approach heuristically copies the gradient with respect to the stochastic output directly as an estimator of the gradient with respect to the sigmoid argument (we call this the straight-through estimator). To explore a context where these estimators are useful, we consider a small-scale version of {\em conditional computation}, where sparse stochastic units form a distributed representation of gaters that can turn off in combinatorially many ways large chunks of the computation performed in the rest of the neural network. In this case, it is important that the gating units produce an actual 0 most of the time. The resulting sparsity can be potentially be exploited to greatly reduce the computational cost of large deep networks for which conditional computation would be useful.}
}
@article{BengioetalCoRR-14,
       author = {Ian J. Goodfellow and Dumitru Erhan and Pierre Luc Carrier and Aaron Courville and Mehdi Mirza and Ben Hamner and Will Cukierski and Yichuan Tang and David Thaler and Dong-Hyun Lee and Yingbo Zhou and Chetan Ramaiah and Fangxiang Feng and Ruifan Li and Xiaojie Wang and Dimitris Athanasakis and John Shawe-Taylor and Maxim Milakov and John Park and Radu Ionescu and Marius Popescu and Cristian Grozea and James Bergstra and Jingjing Xie and Lukasz Romaszko and Bing Xu and Zhang Chuang and Yoshua Bengio},
        title = {Challenges in Representation Learning: A report on three machine learning contests},
      journal = {CoRR},
       volume = {arXiv:1307.1414},
         year = 2013,
     abstract = {The ICML 2013 Workshop on Challenges in Representation Learning focused on three challenges: the black box learning challenge, the facial expression recognition challenge, and the multimodal learning challenge. We describe the datasets created for these challenges and summarize the results of the competitions. We provide suggestions for organizers of future challenges and some comments on what kind of knowledge can be gained from machine learning competitions.}
}
@article{BengioetalCoRR-15,
        title = {Scheduled Sampling for Sequence Prediction with Recurrent Neural Networks},
       author = {Samy Bengio and Oriol Vinyals and Navdeep Jaitly and Noam Shazeer},
      journal = {CoRR},
       volume = {arXiv:1506.03099},
         year = 2015,
     abstract = {Recurrent Neural Networks can be trained to produce sequences of tokens given some input, as exemplified by recent results in machine translation and image captioning. The current approach to training them consists in maximizing the likelihood of each token in the sequence given the current (recurrent) state and the previous token. At inference, the unknown previous token is then replaced by a token generated by the model itself. This discrepancy between training and inference can yield errors that can accumulate quickly along the generated sequence. We propose a curriculum learning strategy to gently change the training process from a fully guided scheme using the true previous token, towards a less guided scheme which mostly uses the generated token instead. Experiments on several sequence prediction tasks show that this approach yields significant improvements.}
}
@article{BengioetalCoRR-16,
       author = {Yoshua Bengio and Dong{-}Hyun Lee and J{\"{o}}rg Bornschein and Zhouhan Lin},
        title = {Towards Biologically Plausible Deep Learning},
      journal = {CoRR},
       volume = {arXiv:1502.04156},
         year = {2016},
     abstract = {Neuroscientists have long criticised deep learning algorithms as incompatible with current knowledge of neurobiology. We explore more biologically plausible versions of deep representation learning, focusing here mostly on unsupervised learning but developing a learning mechanism that could account for supervised, unsupervised and reinforcement learning. The starting point is that the basic learning rule believed to govern synaptic weight updates (Spike-Timing-Dependent Plasticity) arises out of a simple update rule that makes a lot of sense from a machine learning point of view and can be interpreted as gradient descent on some objective function so long as the neuronal dynamics push firing rates towards better values of the objective function (be it supervised, unsupervised, or reward-driven). The second main idea is that this corresponds to a form of the variational EM algorithm, i.e., with approximate rather than exact posteriors, implemented by neural dynamics. Another contribution of this paper is that the gradients required for updating the hidden states in the above variational interpretation can be estimated using an approximation that only requires propagating activations forward and backward, with pairs of layers learning to form a denoising auto-encoder. Finally, we extend the theory about the probabilistic interpretation of auto-encoders to justify improved sampling schemes based on the generative interpretation of denoising auto-encoders, and we validate all these ideas on generative learning tasks.},
}
@inproceedings{BengioetalICML-09,
       author = {Bengio, Yoshua and Louradour, J{\'e}r\^{o}me and Collobert, Ronan and Weston, Jason},
        title = {Curriculum Learning},
    booktitle = {Proceedings of the 26th Annual International Conference on Machine Learning},
    publisher = {ACM},
      address = {New York, NY, USA},
     location = {Montreal, Quebec, Canada},
         year = {2009},
        pages = {41-48},
     abstract = {Humans and animals learn much better when the examples are not randomly presented but organized in a meaningful order which illustrates gradually more concepts, and gradually more complex ones. Here, we formalize such training strategies in the context of machine learning, and call them "curriculum learning". In the context of recent research studying the difficulty of training in the presence of non-convex training criteria (for deep deterministic and stochastic neural networks), we explore curriculum learning in various set-ups. The experiments show that significant improvements in generalization can be achieved. We hypothesize that curriculum learning has both an effect on the speed of convergence of the training process to a minimum and, in the case of non-convex criteria, on the quality of the local minima obtained: curriculum learning can be seen as a particular form of continuation method (a general strategy for global optimization of non-convex functions).},
}
@article{BengioetalJMLR-03,
    author = {Bengio, Yoshua and Ducharme, R{\'e}jean and Vincent, Pascal and Janvin, Christian},
     title = {A neural probabilistic language model},
   journal = {Journal of Machine Learning Research},
    volume = 3,
      year = 2003,
     pages = {1137-1155},
}
@incollection{BengioetalNIPS-07,
       author = {Yoshua Bengio and Pascal Lamblin and Dan Popovici and Hugo Larochelle},
        title = {Greedy Layer-Wise Training of Deep Networks},
    booktitle = {Advances in Neural Information Processing Systems 19},
      address = {Cambridge, MA},
    publisher = {MIT Press},
        pages = {153-160},
         year = 2007,
     abstract = {Complexity theory of circuits strongly suggests that deep architectures can be much more efficient (sometimes exponentially) than shallow architectures, in terms of computational elements required to represent some functions. Deep multi-layer neural networks have many levels of non-linearities allowing them to compactly represent highly non-linear and highly-varying functions. However, until recently it was not clear how to train such deep networks, since gradient-based optimization starting from random initialization appears to often get stuck in poor solutions. Hinton et al. recently introduced a greedy layer-wise unsupervised learning algorithm for Deep Belief Networks (DBN), a generative model with many layers of hidden causal variables. In the context of the above optimization problem, we study this algorithm empirically and explore variants to better understand its success and extend it to cases where the inputs are continuous or where the structure of the input distribution is not revealing enough about the variable to be predicted in a supervised task. Our experiments also confirm the hypothesis that the greedy layer-wise unsupervised training strategy mostly helps the optimization, by initializing weights in a region near a good local minimum, giving rise to internal distributed representations that are high-level abstractions of the input, bringing better generalization.}
}
@article{BengioetalPAMI-13,
       author = {Bengio, Yoshua and Courville, Aaron and Vincent, Pascal},
        title = {Representation Learning: A Review and New Perspectives},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
    publisher = {IEEE Computer Society},
       volume = {35},
       number = {8},
         year = {2013},
        pages = {1798-1828},
     abstract = {The success of machine learning algorithms generally depends on data representation, and we hypothesize that this is because different representations can entangle and hide more or less the different explanatory factors of variation behind the data. Although specific domain knowledge can be used to help design representations, learning with generic priors can also be used, and the quest for AI is motivating the design of more powerful representation-learning algorithms implementing such priors. This paper reviews recent work in the area of unsupervised feature learning and deep learning, covering advances in probabilistic models, autoencoders, manifold learning, and deep networks. This motivates longer term unanswered questions about the appropriate objectives for learning good representations, for computing representations (i.e., inference), and the geometrical connections between representation learning, density estimation, and manifold learning.}
}
@article{BenjaminetalIEEE-14,
       author = {Benjamin, Ben Varkey and Gao, Peiran and McQuinn, Emmett and Choudhary, Swadesh and Chandrasekaran, Anand and Bussat, Jean-Marie and Alvarez-Icaza, Rodrigo and Arthur, John V. and Merolla, Paul and Boahen, Kwabena},
        title = {Neurogrid: A Mixed-Analog-Digital Multichip System for Large-Scale Neural Simulations},
      journal = {Proceedings of the {IEEE}},
       volume = 102,
       number = 5,
         year = 2014,
        pages = {699-716},
     abstract = {In this paper, we describe the design of Neurogrid, a neuromorphic system for simulating large-scale neural models in real time. Neuromorphic systems realize the function of biological neural systems by emulating their structure. Designers of such systems face three major design choices: (1) whether to emulate the four neural elementsVaxonal arbor, synapse, dendritic tree, and somaVwith dedicated or shared electronic circuits; (2) whether to implement these electronic circuits in an analog or digital manner; and (3) whether to interconnect arrays of these silicon neurons with a mesh or a tree network. The choices we made were: (1) we emulated all neural elements except the soma with shared electronic circuits; this choice maximized the number of synaptic connections; (2) we realized all electronic circuits except those for axonal arbors in an analog manner; this choice maximized energy efficiency; and (3) we interconnected neural arrays in a tree network; this choice maximized throughput. These three choices made it possible to simulate a million neurons with billions of synaptic connections in real time for the first time using 16 Neurocores integrated on a board that consumes three watts.},
}
@article{Bennett89,
        title = {Time/Space Trade-Offs for Reversible Computation},
       author = {Charles H. Bennett},
        pages = {766-776},
      journal = {{SIAM} Journal on Computing},
       volume = 18,
        issue = 4,
         year = 1989,
}
@article{BerantandLiangEMNLP-14,
        title = {Semantic Parsing via Paraphrasing},
       author = {Jonathan Berant and Percy Liang},
      journal = {Proceedings of the Conference on Empirical Methods in Natural Language Processing},
         year = 2014,
     abstract = {A central challenge in semantic parsing is handling the myriad ways in which knowledge base predicates can be expressed. Traditionally, semantic parsers are trained primarily from text paired with knowledge base information. Our goal is to exploit the much larger amounts of raw text not tied to any knowledge base. In this paper, we turn semantic parsing on its head. Given an input utterance, we first use a simple method to deterministically generate a set of candidate logical forms with a canonical realization in natural language for each. Then, we use a paraphrase model to choose the realization that best paraphrases the input, and output the corresponding logical form. We present two simple paraphrase models, an association model and a vector space model, and train them jointly from question-answer pairs. Our system PARASEMPRE improves state-of-the-art accuracies on two recently released question-answering datasets.}
}
@article{BerantetalEMNLP-14,
        title = {Modeling Biological Processes for Reading Comprehension},
       author = {Jonathan Berant and Vivek Srikumar and Pei-Chun Chen and Brad Huang and Christopher D. Manning and Abby Vander Linden and Brittany Harding and Peter Clark},
      journal = {Proceedings of the Conference on Empirical Methods in Natural Language Processing},
         year = 2014,
     abstract = {Machine reading calls for programs that read and understand text, but most current work only attempts to extract facts from redundant web-scale corpora. In this paper, we focus on a new reading comprehension task that requires complex reasoning over a single document. The input is a paragraph describing a biological process, and the goal is to answer questions that require an understanding of the relations between entities and events in the process. To answer the questions, we first predict a rich structure representing the process in the paragraph. Then, we map the question to a formal query, which is executed against the predicted structure. We demonstrate that answering questions via predicted structures substantially improves accuracy over baselines that use shallower representations.}
} 
@inproceedings{BergandMalikCVPR-01,
    author = {Alex Berg and Jitendra Malik},
     title = {Geometric blur for template matching},
 booktitle = {Proceedings of the 2001 {IEEE} Conference on Computer Vision and Pattern Recognition},
 publisher = {IEEE Computer Society},
    volume = 1,
     pages = {607-614},
      year = 2001,
}
@article{BerglesandSteinhauserGLIA-12,
       author = {Bergles, Dwight E. and Jabs, Ronald and Steinhauser, Christian},
        title = {Neuron-glia synapses in the brain},
      journal = {Brain Research Reviews},
         year = 2010,
       volume = 63,
        issue = {1-2},
        pages = {130-137},
     abstract = {The ability to investigate the electrophysiological properties of individual cells in acute brain tissue led to the discovery that many glial cells have the capacity to respond rapidly to neuronal activity. In particular, a distinct class of neuroglial cells known as NG2 cells, which exhibit many of the properties that have been described for glial subtypes such as complex cells, polydendrocytes, synantocytes and GluR cells, express ionotropic receptors for glutamate and GABA. In both gray and white matter, NG2 cells form direct synaptic junctions with axons, which enable transient activation of these receptors. Electrophysiological analyses have shown that these neuron-glia synapses exhibit all the hallmarks of classical neuron-neuron synapses, including rapid activation, quantized responses, facilitation and depression, and presynaptic inhibition. Electron microscopy indicates that axons form morphologically distinct junctions at discrete sites along processes of NG2 cells, suggesting that NG2 cells are an overt target of axonal projections. AMPA receptors expressed by NG2 cells exhibit varying degrees of Ca(2+) permeability, depending on the brain region and stage of development, and in white matter NG2 cells have also been shown to express functional NMDA receptors. Ca(2+) influx through AMPA receptors following repetitive stimulation can trigger long term potentiation of synaptic currents in NG2 cells. The expression of receptors with significant Ca(2+) permeability may increase the susceptibility of NG2 cells to excitotoxic injury. Future studies using transgenic mice in which expression of receptors can be manipulated selectively in NG2 cells have to define the functions of this enigmatic neuron-glia signaling in the normal and diseased CNS.},
}
@article{BergstraetalICML-13,
    author = {James Bergstra and Daniel Yamins and David Cox},
     title = {Making a Science of Model Search: Hyperparameter Optimization in Hundreds of Dimensions for Vision Architectures},
   journal = {Journal of Machine Learning Research},
     voume = 28,
     issue = 1,
      year = 2013,
     pages = {115-123},
  abstract = {Many computer vision algorithms depend on configuration settings that are typically hand-tuned in the course of evaluating the algorithm for a particular data set. While such parameter tuning is often presented as being incidental to the algorithm, correctly setting these parameter choices is frequently critical to realizing a method's full potential. Compounding matters, these parameters often must be re-tuned when the algorithm is applied to a new problem domain, and the tuning process itself often depends on personal experience and intuition in ways that are hard to quantify or describe. Since the performance of a given technique depends on both the fundamental quality of the algorithm and the details of its tuning, it is sometimes difficult to know whether a given technique is genuinely better, or simply better tuned. In this work, we propose a meta-modeling approach to support automated hyperparameter optimization, with the goal of providing practical tools that replace hand-tuning with a reproducible and unbiased optimization process. Our approach is to expose the underlying expression graph of how a performance metric (e.g. classification accuracy on validation examples) is computed from hyperparameters that govern not only how individual processing steps are applied, but even which processing steps are included. A hyperparameter optimization algorithm transforms this graph into a program for optimizing that performance metric. Our approach yields state of the art results on three disparate computer vision problems: a face-matching verification task (LFW), a face identification task (PubFig83) and an object recognition task (CIFAR-10), using a single broad class of feed-forward vision architectures.},
}
@article{BerkesandWiskott05,
       author = {Pietro Berkes and Laurenz Wiskott},
        title = {Slow feature analysis yields a rich repertoire of complex cell properties},
      journal = {Journal of Vision},
       volume = 5,
        issue = 6,
        pages = {579-602},
         year = 2005,
}
@article{BerkovitsPRE-91,
       author = {Richard Berkovits},
        title = {Sensitivity of the multiple-scattering speckle pattern to the motion of a single scatterer},
      journal = {Physical Review E},
       volume = 43,
        issue = 19,
         year = 1991,
        pages = {8638-8640},
}
@article{BermanetalJOTRS-14,
        title = {Mapping the stereotyped behaviour of freely moving fruit flies},
       author = {Berman, Gordon J and Choi, Daniel M and Bialek, William and Shaevitz, Joshua W},
      journal = {Journal of The Royal Society Interface},
    publisher = {The Royal Society},
       volume = {11},
       number = {99},
         year = {2014},
        pages = {20140672},
}
@book{Bernstein2004,
        title = {The Birth of Plenty: How the Prosperity of the Modern World was Created},
       author = {William Bernstein},
    publisher = {McGraw-Hill},
         year = 2004,
}
@book{Bernstein2008,
        title = {A Splendid Exchange: How Trade Shaped the World},
       author = {William Bernstein},
    publisher = {Atlantic Monthly Press},
         year = 2008,
}
 @article{NambuFiNA-11,
       author = {Nambu, A},
        title = {{S}omatotopic organization of the primate {B}asal {G}anglia},
      journal = {Frontiers in Neuroanatomy},
         year = {2011},
       volume = {5},
        pages = {26},
     abstract = {Somatotopic organization is a fundamental and key concept to understand how the cortico-basal ganglia loop works. It is also indispensable knowledge to perform stereotaxic surgery for movement disorders. Here I would like to describe the somatotopic organization of the basal ganglia, which consist of the striatum, subthalamic nucleus, globus pallidus, and substantia nigra. Projections from motor cortical regions representing different body parts terminate in different regions of these nuclei. Basal ganglia neurons respond not only to the stimulation of the corresponding regions of the motor cortices, but also to active and passive movements of the corresponding body parts. On the basis of these anatomical and physiological findings, somatotopic organization can be identified in the motor territories of these nuclei in the basal ganglia. In addition, projections from functionally interrelated cortical areas partially converge through the cortico-basal ganglia loop, but nevertheless the somatotopy is still preserved. Disorganized somatotopy may explain, at least in part, the pathophysiology of movement disorders, such as Parkinson's disease and dystonia.}
}
@book{BernsteinAGAINST_THE_GODS-96,
        title = {Against the Gods: The Remarkable Story of Risk},
       author = {Bernstein, Peter L.},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = {1996},
}
@book{BernsteinCAPITAL_IDEAS-05,
        title = {Capital Ideas: The Improbable Origins of Modern Wall Street},
       author = {Bernstein, Peter L.},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = {2005},
}
@book{BernsteinMASTERS_OF_THE_WORD-14,
        title = {Masters of the Word: How Media Shaped History from the Alphabet to the Internet},
       author = {Bernstein, William J.},
    publisher = {Grove / Atlantic, Incorporated},
         year = {2014},
}
@article{BernsteinandBoydenTCS-11,
        title = {Optogenetic tools for analyzing the neural circuits of behavior},
       author = {Jacob G. Bernstein and Edward S. Boyden},
      journal = {Trends in Cognitive Sciences},
       volume = 15,
        issue = 12,
         year = 2011,
        pages = {592-600},
      comment = {http://www.ncbi.nlm.nih.gov/pubmed/22055387},
}
@inproceedings{Berrouetal93,
       author = {G. Berrou and A. Glavieux and P. Thitimajshima},
        title = {Near Shannon limit error-correcting coding: Turbo codes},
    booktitle = {Proceedings International Conference on Communications},
     location = {Geneva, Switzerland},
        pages = {1064-1070},
         year = 1993,
}
@article{BerryetalNATURE-13,
       author = {Significant performance enhancement in photoconductive terahertz optoelectronics by incorporating plasmonic contact electrodes},
        title = {Berry, C.W. and Wang, N. and Hashemi, M.R. and Unlu, M. and Jarrahi, M.},
      journal = {Nature Communications},
    publisher = {Nature Publishing Group},
       volume = 4,
         year = 2013,
        pages = 1622,
}
@inproceedings{BertalmioetalSIGGRAPH-00,
       author = {M. Bertalm\'{i}o and G. Sapiro and V. Caselles and C. Ballester},
        title = {Image Inpainting},
    booktitle = {Proceedings {ACM} {SIGGRAPH}},
         year = 2000,
        pages = {417-424}
}
@book{BertotandCasteranCOQ-04,
       author = {Bertot, Yves and Cast\'{e}ran, Pierre},
        title = {Interactive Theorem Proving and Program Development: Coq'Art The Calculus of Inductive Constructions},
    publisher = {Springer},
      address = {New York},
         year = {2010},
}
@article{BertrandandCamosCOGNITIVE-DEVELOPMENT-15,
        title = {The role of attention in preschoolers working memory},
       author = {Rapha{\"{e}}lle Bertrand and Val{\`{e}}rie Camos},
      journal = {Cognitive Development},
       volume = {33},
         year = {2015},
        pages = {14-27},
     abstract = {Recent studies showed that introducing an unfilled delay or a secondary task between encoding and recall reduces recall performance in preschoolers, whereas increasing the attentional demand of this secondary task does not. Based on these results, different authors drew opposing conclusions regarding the use of attention in preschoolers' memory maintenance. This study aimed to bring further evidence to bear on the issue. In two experiments, recall was reduced when an unfilled delay was introduced before recall, but also when the demand of the concurrent task increased. These effects did not interact with age, although performance improved from 4 to 6 years of age. A third experiment ruled out an alternative interpretation. These findings show the role of attention in preschoolers' working memory and raise the question of how motor activity may help memory maintenance.}
}

@inproceedings{BertrandetalEMBC-14,
       author = {Bertrand, A. and Dongjin Seo and Maksimovic, F. and Carmena, J.M. and Maharbiz, M.M. and Alon, E. and Rabaey, J.M.},
    booktitle = {Proceedings of the {IEEE} International Conference of the Engineering in Medicine and Biology Society},
        title = {Beamforming approaches for untethered, ultrasonic neural dust motes for cortical recording: A simulation study},
         year = 2014,
        pages = {2625-2628},
     abstract = {In this paper, we examine the use of beamforming techniques to interrogate a multitude of neural implants in a distributed, ultrasound-based intra-cortical recording platform known as Neural Dust [1]. We propose a general framework to analyze system design tradeoffs in the ultrasonic beamformer that extracts neural signals from modulated ultrasound waves that are backscattered by free-floating neural dust (ND) motes. Simulations indicate that high-resolution linearly-constrained minimum variance beamforming sufficiently suppresses interference from unselected ND motes and can be incorporated into the ND-based cortical recording system.},
}
@book{Bertsekas87,
       author = {Bertsekas, Dimitri P.},
        title = {Dynamic Programming: Deterministic and Stochastic Models},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, N.J.},
         year = 1987,
}
@book{Bertsekas89,
       author = {Dimitri P. Bertsekas and John N. Tsitsiklis},
        title = {Parallel and Distributed Computation: Numerical Methods},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, NJ},
         year = 1989,
}
@article{BertsekasandCastanonIEEE-89,
       author = {Bertsekas, D. P. and Casta\~{n}on, D. A.},
        title = {Adaptive Aggregation for Infinite Horizon Dynamic Programming},
      journal = {{IEEE} Transactions on Automatic Control},
       volume = 34,
        issue = 6,
        pages = {589-598},
         year = 1989,
}
@book{BertsekasandShreve78,
       author = {Bertsekas, Dimitri P. and Shreve, Steven E.},
        title = {Stochastic Optimal Control: The Discrete Time Case},
    publisher = {Academic Press},
         year = 1978,
       volume = 139,
 series       = {Mathematics in Science and Engineering},
      address = {New York},
}
@book{BertsekasandTsitsiklis96,
       author = {Bertsekas, Dimitris P. and Tsitsiklis, John N.},
        title = {Neuro-Dynamic Programming},
    publisher = {Athena Scientific},
      address = {Belmont, Massachusetts},
         year = 1996,
}
@article{BertsekasetalHEURISTICS-97,
       author = {Bertsekas, D. P. and Tsitsiklis, John N. and Wu, C.},
        title = {Rollout Algorithms for Combinatorial Optimization},
      journal = {Heuristics},
       volume = 3,
        pages = {245-262},
         year = 1997,
}
@article{BerzuiniAIJ-91,
       author = {Berzuini, Carlo},
        title = {A Probabilistic Framework for Temporal Reasoning},
      journal = AIJ,
         year = {to appear},
}
@inproceedings{BerzuiniUAI-89,
       author = {Berzuini, Carlo and Bellazzi, Riccardo and Quaglini, Silvana},
        title = {Temporal reasoning with probabilities},
    booktitle = UAI-89,
         year = 1989,
     location = {Windsor, Ontario},
        pages = {14-21},
}
@inproceedings{BesshoetalACL-12,
        title = {Dialog System Using Real-Time Crowdsourcing and Twitter Large-Scale Corpus},
       author = {Bessho, Fumihiro and Harada, Tatsuya and Kuniyoshi, Yasuo},
    booktitle = {Proceedings of the 13th Annual Meeting of the Special Interest Group on Discourse and Dialogue},
    publisher = {Association for Computational Linguistics},
         year = 2012,
        pages = {227-231},
}
@article{BetzeletalCoRR-16,
       author = {Richard F. Betzel, Theodore D. Satterthwaite, Joshua I. Gold, Danielle S. Bassett},
        title = {A positive mood, a flexible brain},
      journal = {CoRR},
       volume = {arXiv:1601.07881},
         year = {2016},
     abstract = {Flexible reconfiguration of human brain networks supports cognitive flexibility and learning. However, modulating flexibility to enhance learning requires an understanding of the relationship between flexibility and brain state. In an unprecedented longitudinal data set, we investigate the relationship between flexibility and mood, demonstrating that flexibility is positively correlated with emotional state. Our results inform the modulation of brain state to enhance response to training in health and injury.},
}

@inproceedings{BharatandBroderWWW-98,
    author = {K. Bharat and A. Broder},
     title = {A technique for measuring the relative size and overlap of public {W}eb search engines},
 booktitle = {Proceedings of the 7th World Wide Web Conference},
     pages = {379-388},
      year = 1998,
}
@inproceedings{Bharatetal98,
    author = {K. Bharat and A. Broder and M. Henzinger and P. Kumar and S. Venkatasubramanian},
     title = {The Connectivity Server: fast access to linkage information on the Web},
 booktitle = {Proceedings of the 7th World Wide Web Conference},
     pages = {},
      year = 1998,
       url = {http://www7.scu.edu.au/programme/fullpapers/1938/com1938.htm},
}
@inproceedings{BharathetalECTC-09,
       author = {Bharath, K. and Jae Young Choi and Swaminathan, M.},
    booktitle = {Electronic Components and Technology Conference},
        title = {Use of the finite element method for the modeling of multi-layered power/ground planes with small features},
         year = {2009},
        pages = {1630-1635},
     abstract = {In this paper, a novel formulation for the modeling of multi-layer package power/ground planes has been proposed. The formulation is developed by first applying the finite element method (FEM) for a single plane-pair geometry, from which a SPICE-compatible equivalent circuit is extracted. Next, the admittance matrices obtained for individual plane-pairs are coupled together, to extend the technique, and to enable the modeling of multiple plane-pairs. This method, the mutli-layer finite element method (MFEM), uses an adaptive triangular mesh. This enables MFEM to capture the effects of small geometrical features that can affect the frequency response, with only a modest increase in computational cost as compared to methods that use regular square meshes. MFEM is able to correctly model the effect of apertures or voids in the planes, which cause vertical coupling of energy. Several examples have been shown to illustrate the efficacy of the method for both single and multiple plane-pair geometries.},
}
@article{BhatiaAandIngber2014,
       author = {Bhatia, Sangeeta N. and Ingber, Donald E.},
        title = {Microfluidic organs-on-chips},
      journal = {Nature Biotechnology},
         year = {2014},
    publisher = {Nature Publishing Group},
       volume = {32},
        issue = {8},
        pages = {760-772},
     abstract = {An organ-on-a-chip is a microfluidic cell culture device created with microchip manufacturing methods that contains continuously perfused chambers inhabited by living cells arranged to simulate tissue- and organ-level physiology. By recapitulating the multicellular architectures, tissue-tissue interfaces, physicochemical microenvironments and vascular perfusion of the body, these devices produce levels of tissue and organ functionality not possible with conventional 2D or 3D culture systems. They also enable high-resolution, real-time imaging and in vitro analysis of biochemical, genetic and metabolic activities of living cells in a functional tissue and organ context. This technology has great potential to advance the study of tissue development, organ physiology and disease etiology. In the context of drug discovery and development, it should be especially valuable for the study of molecular mechanisms of action, prioritization of lead candidates, toxicity testing and biomarker identification.},
}
@misc{BhattREINFORCEMENT_LEARNING-18,
        title = {Five Things You Need to Know about Reinforcement Learning},
       author = {Shweta Bhatt},
 howpublished = {\url{https://www.kdnuggets.com/2018/03/5-things-reinforcement-learning.htmlhttp://www.diee.unica.it/DRIEI/tesi/19_setzu.pdf}},
         year = {2018},
}
@article{BhoopchandetalCoRR-16,
       author = {Avishkar Bhoopchand and Tim Rockt{\"{a}}schel and Earl T. Barr and Sebastian Riedel},
        title = {Learning Python Code Suggestion with a Sparse Pointer Network},
      journal = {CoRR},
       volume = {arXiv:1611.08307},
         year = {2016},
     abstract = {To enhance developer productivity, all modern integrated development environments (IDEs) include code suggestion functionality that proposes likely next tokens at the cursor. While current IDEs work well for statically-typed languages, their reliance on type annotations means that they do not provide the same level of support for dynamic programming languages as for statically-typed languages. Moreover, suggestion engines in modern IDEs do not propose expressions or multi-statement idiomatic code. Recent work has shown that language models can improve code suggestion systems by learning from software repositories. This paper introduces a neural language model with a sparse pointer network aimed at capturing very long-range dependencies. We release a large-scale code suggestion corpus of 41M lines of Python code crawled from GitHub. On this corpus, we found standard neural language models to perform well at suggesting local phenomena, but struggle to refer to identifiers that are introduced many tokens in the past. By augmenting a neural language model with a pointer network specialized in referring to predefined classes of identifiers, we obtain a much lower perplexity and a 5 percentage points increase in accuracy for code suggestion compared to an LSTM baseline. In fact, this increase in code suggestion accuracy is due to a 13 times more accurate prediction of identifiers. Furthermore, a qualitative analysis shows this model indeed captures interesting long-range dependencies, like referring to a class member defined over 60 tokens in the past.}
}
@inproceedings{BhoopchandetalICLR-17,
       author = {Avishkar Bhoopchand and Tim Rockt\"{a}schel and Earl T. Barr and Sebastian Riedel},
        title = {Learning Python Code Suggestion with a Sparse Pointer Network},
    booktitle = {International Conference on Learning Representations},
       volume = {arXiv:/1611.08307},
         year = {2017},
          url = {https://github.com/uclmr/pycodesuggest},
     abstract = {To enhance developer productivity, all modern integrated development environments (IDEs) include code suggestion functionality that proposes likely next tokens at the cursor. While current IDEs work well for statically-typed languages, their reliance on type annotations means that they do not provide the same level of support for dynamic programming languages as for statically-typed languages. Moreover, suggestion engines in modern IDEs do not propose expressions or multi-statement idiomatic code. Recent work has shown that language models can improve code suggestion systems by learning from software repositories. This paper introduces a neural language model with a sparse pointer network aimed at capturing very long-range dependencies. We release a large-scale code suggestion corpus of 41M lines of Python code crawled from GitHub. On this corpus, we found standard neural language models to perform well at suggesting local phenomena, but struggle to refer to identifiers that are introduced many tokens in the past. By augmenting a neural language model with a pointer network specialized in referring to predefined classes of identifiers, we obtain a much lower perplexity and a 5 percentage points increase in accuracy for code suggestion compared to an LSTM baseline. In fact, this increase in code suggestion accuracy is due to a 13 times more accurate prediction of identifiers. Furthermore, a qualitative analysis shows this model indeed captures interesting long-range dependencies, like referring to a class member defined over 60 tokens in the past.}
}
@book{Bialek2012,
        title = {Biophysics: Searching for Principles},
       author = {William Bialek},
    publisher = {Princeton University Press},
      address = {Princeton, New Jersey},
         year = {2012},
}
@article{BialeketalSCIENCE-91,
       author = {Bialek, William and Rieke, Fred and de Ruyter van Steveninck, R.R. and Warland, D.},
        title = {Reading a Neural Code},
      journal = {Science},
       volume = 252,
        issue = 5014,
         year = 1991,
        pages = {1854-1857},
     abstract = {Traditional approaches to neural coding characterize the encoding of known stimuli in average neural responses. Organisms face nearly the opposite task--extracting information about an unknown time-dependent stimulus from short segments of a spike train. Here the neural code was characterized from the point of view of the organism, culminating in algorithms for real-time stimulus estimation based on a single example of the spike train. These methods were applied to an identified movement-sensitive neuron in the fly visual system. Such decoding experiments determined the effective noise level and fault tolerance of neural computation, and the structure of the decoding algorithms suggested a simple model for real-time analog signal processing with spiking neurons.}
}
@article{BicanskiandBurgessCURRENT-BIOLOGY-19,
       author = {Bicanski, Andrej and Burgess, Neil},
        title = {A Computational Model of Visual Recognition Memory via Grid Cells},
      journal = {Current Biology},
    publisher = {Cell Press},
       volume = {29},
        issue = {6},
         year = {2019},
        pages = {979-990},
     abstract = {Models of face, object, and scene recognition traditionally focus on massively parallel processing of low-level features, with higher-order representations emerging at later processing stages. However, visual perception is tightly coupled to eye movements, which are necessarily sequential. Recently, neurons in entorhinal cortex have been reported with grid cell-like firing in response to eye movements, i.e., in visual space. Following the presumed role of grid cells in vector navigation, we propose a model of recognition memory for familiar faces, objects, and scenes, in which grid cells encode translation vectors between salient stimulus features. A sequence of saccadic eye-movement vectors, moving from one salient feature to the expected location of the next, potentially confirms an initial hypothesis (accumulating evidence toward a threshold) about stimulus identity, based on the relative feature layout (i.e., going beyond recognition of individual features). The model provides an explicit neural mechanism for the long-held view that directed saccades support hypothesis-driven, constructive perception and recognition; is compatible with holistic face processing; and constitutes the first quantitative proposal for a role of grid cells in visual recognition. The variance of grid cell activity along saccade trajectories exhibits 6-fold symmetry across 360 degrees akin to recently reported fMRI data. The model suggests that disconnecting grid cells from occipitotemporal inputs may yield prosopagnosia-like symptoms. The mechanism is robust with regard to partial visual occlusion, can accommodate size and position invariance, and suggests a functional explanation for medial temporal lobe involvement in visual memory for relational information and memory-guided attention.},
}
@incollection{Biederman07,
       author = {Biederman, Irving},
        title = {Recent psychophysical and neural research in shape recognition},
    booktitle = {Object Recognition, Attention and Action},
    publisher = {Springer},
       editor = {N. Osaka and I. Rentschler and I. Biederman},
        pages = {71-88},
         year = 2007,
}
@incollection{Biederman1988,
        title = {Aspects and extensions of a theory of human image understanding},
       author = {Irving Biederman},
    booktitle = {Computational processes in human vision: An interdisciplinary perspective},
       editor = {Zenon W. Pylyshyn},
    publisher = {Ablex},
      address = {Norwood, NJ},
         year = 1988,
      comment = {See page 399 for Biederman's estimate of the number of readily-distinguishable categories acquired by humans},
        pages = {370-428}
}
@article{BiedermanCVGIP-1985,
       author = {I. Biederman},
        title = {Human image understanding: Recent research and a theory},
      journal = {Computer Vision Graphics Image Processing},
         year = 1985,
       volume = 32,
        pages = {29-73},
}
@article{BiedermanetalCP-82,
       author = {Biederman, I. and Mezzanotte, R. J. and Rabinowitz, J. C.},
        title = {Scene perception: Detecting and judging objects undergoing relational violations},
      journal = {Cognitive Psychology},
       volume = 14,
        pages = {143-177},
         year = 1982,
}
@article{BiedermanetalJEP-74,
       author = {Biederman, I. and Rabinowitz, J. and Glass, A. and Stacy, W.E.},
        title = {On the information extracted from a glance at a scene},
      journal = {Journal of Experimental Psychology},
       volume = 103,
        issue = 3,
         year = 1974,
        pages = {597-600},
}
@article{BieliketalCoRR-16,
       author = {Pavol Bielik and Veselin Raychev and Martin T. Vechev},
        title = {Learning a Static Analyzer from Data},
      journal = {CoRR},
       volume = {arXiv:1611.01752},
         year = {2016},
     abstract = {To be practically useful, modern static analyzers must precisely model the effect of both, statements in the programming language as well as frameworks used by the program under analysis. While important, manually addressing these challenges is difficult for at least two reasons: (i) the effects on the overall analysis can be non-trivial, and (ii) as the size and complexity of modern libraries increase, so is the number of cases the analysis must handle. In this paper we present a new, automated approach for creating static analyzers: instead of manually providing the various inference rules of the analyzer, the key idea is to learn these rules from a dataset of programs. Our method consists of two ingredients: (i) a synthesis algorithm capable of learning a candidate analyzer from a given dataset, and (ii) a counter-example guided learning procedure which generates new programs beyond those in the initial dataset, critical for discovering corner cases and ensuring the learned analysis generalizes to unseen programs. We implemented and instantiated our approach to the task of learning JavaScript static analysis rules for a subset of points-to analysis and for allocation sites analysis. These are challenging yet important problems that have received significant research attention. We show that our approach is effective: our system automatically discovered practical and useful inference rules for many cases that are tricky to manually identify and are missed by state-of-the-art, manually tuned analyzers.}
}
@inproceedings{BieliketalICML-16,
        title = {{PHOG}: {P}robabilistic Model for Code},
       author = {Pavol Bielik and Veselin Raychev and Martin Vechev},
    booktitle = {Proceedings of The 33rd International Conference on Machine Learning},
       editor = {Maria Florina Balcan and Kilian Q. Weinberger},
       volume = {48},
       series = {Proceedings of Machine Learning Research},
    publisher = {PMLR},
         year = {2016},
        pages = {2933-2942},
     abstract = {We introduce a new generative model for code called probabilistic higher order grammar (PHOG). PHOG generalizes probabilistic context free grammars (PCFGs) by allowing conditioning of a production rule beyond the parent non-terminal, thus capturing rich contexts relevant to programs. Even though PHOG is more powerful than a PCFG, it can be learned from data just as efficiently. We trained a PHOG model on a large JavaScript code corpus and show that it is more precise than existing models, while similarly fast. As a result, PHOG can immediately benefit existing programming tools based on probabilistic models of code.}
}
@article{Bienenstock95,
       author = {Elie Bienenstock},
        title = {Model of Neocortex},
      journal = {Network: Computation in Neural Systems},
       volume = 6,
        pages = {179-224},
         year = 1995,
}
@incollection{BienenstockGeman,
       author = {Elie Bienenstock and Stuart Geman},
        title = {Compositionality in Neural Systems},
    booktitle = {The Handbook of Brain Theory and Neural Networks},
       editor = {M. Arbib},
    publisher = {Bradford Books/MIT Press},
        pages = {223-226},
         year = 1995,
}
@incollection{Bienenstocketal98,
       author = {Bienenstock, Elie and Geman, Stuart and Potter, Daniel},
        title = {Compositionality, {MDL} priors and object recognition},
    booktitle = {Advances in Neural Information Processing Systems 9},
       editor = {M.C. Mozer and M.I. Jordan and T. Petsche},
    publisher = {MIT Press},
      address = {Cambridge, MA},
        pages = {838-844},
         year = 1998,
}
@article{BilboNLM-10,
       author = {Bilbo, Staci D.},
        title = {Early-life infection is a vulnerability factor for aging-related glial alterations and cognitive decline},
      journal = {Neurobiology of Learning and Memory},
         year = {2010},
       volume = {94},
       number = {1},
        pages = {57-64},
     abstract = {There is significant individual variability in cognitive decline during aging, suggesting the existence of "vulnerability factors" for eventual deficits. Neuroinflammation may be one such factor; increased glial reactivity is a common outcome of aging, which in turn is associated with numerous neurodegenerative conditions. Early-life infection leads to cognitive impairment in conjunction with an inflammatory challenge in young adulthood, which led us to explore whether it might also accelerate the cognitive decline associated with aging. Rats were treated on postnatal day 4 with PBS or Escherichia coli, and then tested for learning and memory at 2 or 16months of age, using two fear-conditioning tasks (context pre-exposure and ambiguous cue), and a spatial water maze task. Neonatally-infected rats exhibited memory impairments in both the ambiguous cue fear-conditioning task and in the water maze, but only at 16months. There were no differences in anxiety between groups. Neonatally-infected rats also exhibited greater aging-induced increases in glial markers (CD11b and MHCII on microglia, and GFAP on astrocytes), as well as selective changes in NMDA receptor subunit expression within the hippocampus, but not in amygdala or parietal cortex compared to controls. Taken together, these data suggest that early-life infection leads to less successful cognitive aging, which may be linked to changes in glial reactivity.}
}
@article{BilboandSchwarzFBN-09,
       author = {Bilbo, Staci D. and Schwarz, J. M.},
        title = {Early-life programming of later-life brain and behavior: a critical role for the immune system},
      journal = {Frontiers in Behavioral Neuroscience},
       volume = {3},
        pages = {14},
         year = {2009},
     abstract = {The immune system is well characterized for its critical role in host defense. Far beyond this limited role however, there is mounting evidence for the vital role the immune system plays within the brain, in both normal, "homeostatic" processes (e.g., sleep, metabolism, memory), as well as in pathology, when the dysregulation of immune molecules may occur. This recognition is especially critical in the area of brain development. Microglia and astrocytes, the primary immunocompetent cells of the CNS, are involved in every major aspect of brain development and function, including synaptogenesis, apoptosis, and angiogenesis. Cytokines such as tumor necrosis factor (TNF)alpha, interleukin [IL]-1beta, and IL-6 are produced by glia within the CNS, and are implicated in synaptic formation and scaling, long-term potentiation, and neurogenesis. Importantly, cytokines are involved in both injury and repair, and the conditions underlying these distinct outcomes are under intense investigation and debate. Evidence from both animal and human studies implicates the immune system in a number of disorders with known or suspected developmental origins, including schizophrenia, anxiety/depression, and cognitive dysfunction. We review the evidence that infection during the perinatal period of life acts as a vulnerability factor for later-life alterations in cytokine production, and marked changes in cognitive and affective behaviors throughout the remainder of the lifespan. We also discuss the hypothesis that long-term changes in brain glial cell function underlie this vulnerability.}
}
@article{BilboandSchwarzFNE-12,
       author = {Bilbo, Staci D. and Schwarz, J.M.},
        title = {The immune system and developmental programming of brain and behavior},
      journal = {Frontiers in Neuroendocrinology},
       volume = {33},
       number = {3},
         year = {2012},
        pages = {267-286},
     abstract = {The brain, endocrine, and immune systems are inextricably linked. Immune molecules have a powerful impact on neuroendocrine function, including hormone-behavior interactions, during health as well as sickness. Similarly, alterations in hormones, such as during stress, can powerfully impact immune function or reactivity. These functional shifts are evolved, adaptive responses that organize changes in behavior and mobilize immune resources, but can also lead to pathology or exacerbate disease if prolonged or exaggerated. The developing brain in particular is exquisitely sensitive to both endogenous and exogenous signals, and increasing evidence suggests the immune system has a critical role in brain development and associated behavioral outcomes for the life of the individual. Indeed, there are associations between many neuropsychiatric disorders and immune dysfunction, with a distinct etiology in neurodevelopment. The goal of this review is to describe the important role of the immune system during brain development, and to discuss some of the many ways in which immune activation during early brain development can affect the later-life outcomes of neural function, immune function, mood and cognition.}
}
@article{BinderandDesaiTiCS-11,
       author = {Binder, J. R. and Desai, R. H.},
        title = {The neurobiology of semantic memory},
      journal = {Trends in Cognitive Science},
         year = {2011},
       volume = {15},
       number = {11},
        pages = {527-536},
     abstract = {Semantic memory includes all acquired knowledge about the world and is the basis for nearly all human activity, yet its neurobiological foundation is only now becoming clear. Recent neuroimaging studies demonstrate two striking results: the participation of modality-specific sensory, motor, and emotion systems in language comprehension, and the existence of large brain regions that participate in comprehension tasks but are not modality-specific. These latter regions, which include the inferior parietal lobe and much of the temporal lobe, lie at convergences of multiple perceptual processing streams. These convergences enable increasingly abstract, supramodal representations of perceptual experience that support a variety of conceptual functions including object recognition, social cognition, language, and the remarkable human capacity to remember the past and imagine the future.}
}
@article{BinzandEndresCoRR-19,
       author = {Marcel Binz and Dominik Endres},
        title = {Emulating Human Developmental Stages with Bayesian Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1902.07579},
         year = {2019},
     abstract = {We compare the acquisition of knowledge in humans and machines. Research from the field of developmental psychology indicates, that human-employed hypothesis are initially guided by simple rules, before evolving into more complex theories. This observation is shared across many tasks and domains. We investigate whether stages of development in artificial learning systems are based on the same characteristics. We operationalize developmental stages as the size of the data-set, on which the artificial system is trained. For our analysis we look at the developmental progress of Bayesian Neural Networks on three different data-sets, including occlusion, support and quantity comparison tasks. We compare the results with prior research from developmental psychology and find agreement between the family of optimized models and pattern of development observed in infants and children on all three tasks, indicating common principles for the acquisition of knowledge.},
}
@article{BioleketalCoRR-14,
       author = {D. Biolek and M. Di Ventra and Y. V. Pershin},
        title = {Reliable {SPICE} Simulations of Memristors, Memcapacitors and Meminductors},
      journal = {CoRR},
       volume = {arXiv:1307.2717},
         year = {2014},
     abstract = {Memory circuit elements, namely memristive, memcapacitive and meminductive systems, are gaining considerable attention due to their ubiquity and use in diverse areas of science and technology. Their modeling within the most widely used environment, SPICE, is thus critical to make substantial progress in the design and analysis of complex circuits. Here, we present a collection of models of different memory circuit elements and provide a methodology for their accurate and reliable modeling in the SPICE environment. We also provide codes of these models written in the most popular SPICE versions (PSpice, LTspice, HSPICE) for the benefit of the reader. We expect this to be of great value to the growing community of scientists interested in the wide range of applications of memory circuit elements.},
}
@article{BirdBisbyandBurgessFiHN-12,
       author = {Bird, Chris and Bisby, James and Burgess, Neil},
        title = {The hippocampus and spatial constraints on mental imagery},
      journal = {Frontiers in Human Neuroscience},
       volume = {6},
        pages = {142},
         year = {2012},
     abstract = {We review a model of imagery and memory retrieval based on allocentric spatial representation by place cells and boundary vector cells in the medial temporal lobe, and their translation into egocentric images in retrosplenial and parietal areas. In this model the activity of place cells constrain the contents of imagery and retrieval to be coherent and consistent with the subject occupying a single location, while the activity of head-direction cells along Papez's circuit determine the viewpoint direction for which the egocentric image is generated. An extension of this model is discussed in which a role for grid cells in dynamic updating of representations (mental navigation) is included. We also discuss the extension of this model to implement a version of the dual representation theory of post-traumatic stress disorder (PTSD) in which PTSD arises from an imbalance between weak allocentric hippocampal-mediated contextual representations and strong affective/sensory representations. The implications of these models for behavioural, neuropsychological and neuroimaging data in humans are explored.}
}
@techreport{BirnbaumTR-86,
       author = {Birnbaum, Larry},
        title = {Integrated Processing in Planning and Understanding},
        issue = 489,
  institution = {Yale University Computer Science Department},
         year = 1986,
}
@article{BirneyMONIST-05,
        title = {Did the monks preserve the latin classics?},
       author = {William Birney},
      journal = {The Monist},
    publisher = {Oxford University Press},
       volume = {15},
       number = {1},
         year = {1905},
        pages = {87-108},
}
@book{Bishop06,
       author = {Christopher M. Bishop},
        title = {Pattern Recognition and Machine Learning},
    publisher = {Springer},
      address = {New York},
         year = 2006,
}
@inproceedings{BishopetalICANN-97,
       author = {Bishop, C. M. and G. E. Hinton and I. G. D. Strachan},
        title = {{GTM} through time},
    booktitle = {Proceedings IEEE Fifth International Conference on Artificial Neural Networks},
     location = {Cambridge, UK},
         year = 1997,
}
@incollection{BishopetalNIPS-97,
       author = {Christopher M. Bishop and Markus Svens{\'e}n and Christopher K. I. Williams},
        title = {{GTM}: {A} Principled Alternative to the Self-Organizing Map},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 9,
    publisher = {MIT Press},
      comment = {Generalized Topographic Mapping},
      address = {Cambridge, MA},
       editor = {Michael C. Mozer and Michael I. Jordan and Thomas Petsche},
        pages = {354-360},
         year = 1997,
}
@article{BlackandFleetIJCV-00,
       author = {Michael Black and David Fleet},
        title = {Probabilistic Detection and Tracking of Motion Boundaries},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 38,
        issue = 3,
        pages = {231-245},
         year = 2000,
}
@article{BlackandJepsonIJCV-98,
       author = {Michael Black and Allan Jepson},
        title = {EigenTracking: Robust matching and tracking of articulated objects using a view-based representation},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 26,
        issue = 1,
        pages = {63-84},
         year = 1998,
}
@article{BlackandJepsonPAMI-96,
       author = {Michael Black and Allan Jepson},
        title = {Estimating Optical Flow in Segmented Images Using Variable-Order Parametric Models With Local Deformations},
      journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
    publisher = {IEEE Computer Society},
       volume = 18,
        issue = 10,
        pages = {972-986},
         year = 1996,
}
@article{BlackandScholes-73,
        title = {The Pricing of Options and Corporate Liabilities},
       author = {Black, Fischer and Scholes, Myron},
      journal = {Journal of Political Economy},
    publisher = {University of Chicago Press},
       volume = {81},
       number = {3},
         year = {1973},
     abstract = {If options are correctly priced in the market, it should not be possible to make sure profits by creating portfolios of long and short positions in options and their underlying stocks. Using this principle, a theoretical valuation formula for options is derived. Since almost all corporate liabilities can be viewed as combinations of options, the formula and the analysis that led to it are also applicable to corporate liabilities such as common stock, corporate bonds, and warrants. In particular, the formula can be used to derive the discount that should be applied to a corporate bond because of the possibility of default.},
}
@article{BlacketalTIP-88,
       author = {M.J. Black and G. Sapiro and D. Marimont and D. Heeger},
        title = {Robust anisotropic diffusion},
      journal = {{IEEE} Transactions Image Processing},
         year = 1998,
       volume = 7,
        issue = 3,
        pages = {421-432},
}
@book{BlakeandIsard98,
       author = {Blake, A. and Isard, M.},
    publisher = {Springer-Verlag},
        title = {Active Contours},
         year = 1998,
}
@article{BlakemoreandCampbellJoP-69,
       author = {Blakemore, C. and Campbell, F. W.},
        title = {On the existence of neurons in the human visual system selectively sensitive to the orientation and size of retinal images},
      journal = {Journal of Physiology},
       volume = 203,
        pages = {237-260},
         year = 1969,
}
@book{Blakeslee2007,
        title = {The Body Has a Mind of Its Own},
       author = {Sandra Blakeslee and Matthew Blakeslee},
    publisher = {Random House},
         year = 2007,
}
@article{BlandMOR-77,
       author = {R. G. Bland},
        title = {New finite pivoting rules for the simplex method},
      journal = {Mathematics of Operations Research},
       volume = 2,
        pages = {103-107},
         year = 1977,
}
@misc{BlandfordetalKAVLI-10,
       author = {Roger Blandford and Michael Roukes and Larry Abbott and Terry Sejnowski},
        title = {Growing High Performance Computing in a Green Environment},
         year = {2011},
 howpublished = {Report on the Third Kavli Futures Symposium},
     abstract = {September 9-11, 2010, Tromo, Norway},
}
@article{BlanhceetalJoN-05,
       author = {Blanche, T.J. and Spacek, M.A. and Hetke, J.F. and Swindale, N.V.},
        title = {Polytrodes: high-density silicon electrode arrays for large-scale multiunit recording},
      journal = {Journal of Neurophysiology. 2005 May;93(5):. Epub 2004 Nov 17.},
       volume = {93},
        issue = {5},
         year = {2005},
        pages = {2987-3000},
}
@article{BlankertzFiN-10,
       author = {Blankertz, B. and Tangermann, M. and Vidaurre, C. and Fazli, S. and Sannelli, C. and Haufe, S. and Maeder, C. and Ramsey, L. and Sturm, I. and Curio, G. and Muller, K. R.},
        title = {The Berlin Brain-Computer Interface: Non-Medical Uses of {BCI} Technology},
      journal = {Frontiers in Neuroscience},
         year = {2010},
       volume = {4},
        pages = {198},
     abstract = {Brain-computer interfacing (BCI) is a steadily growing area of research. While initially BCI research was focused on applications for paralyzed patients, increasingly more alternative applications in healthy human subjects are proposed and investigated. In particular, monitoring of mental states and decoding of covert user states have seen a strong rise of interest. Here, we present some examples of such novel applications which provide evidence for the promising potential of BCI technology for non-medical uses. Furthermore, we discuss distinct methodological improvements required to bring non-medical applications of BCI technology to a diversity of layperson target groups, e.g., ease of use, minimal training, general usability, short control latencies.}
}
@article{BleiandJordanBA-06,
       author = {Blei, D. and Jordan, M.},
        title = {Variational Inference for Dirichlet Process Mixtures},
      journal = {Bayesian Analysis},
       volume = 3,
        issue = 1,
        pages = {121-144},
         year = 2006,
}
@inproceedings{BleiandMoreno01,
       author = {David M. Blei and Pedro J. Moreno},
        title = {Topic Segmentation with an Aspect Hidden {M}arkov Model},
    booktitle = {Research and Development in Information Retrieval},
          url = {citeseer.ist.psu.edu/blei01topic.html},
        pages = {343-348},
         year = 2001,
}
@inproceedings{BleiandMorenoSIGIR-01,
        title = {Topic segmentation with an aspect hidden {M}arkov model},
       author = {David M. Blei and Pedro J. Moreno},
    booktitle = {Proceedings of the 24th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval},
    publisher = {ACM Press},
        pages = {343-348},
         year = 2001,
}
@article{BleietalCoRR-17,
        title = {Variational Inference: A Review for Statisticians},
       author = {David M. Blei and Alp Kucukelbir and Jon D. McAuliffe},
      journal = {CoRR},
         year = {2017},
       volume = {arXiv:1601.00670},
     abstract = {One of the core problems of modern statistics is to approximate difficult-to-compute probability densities. This problem is especially important in Bayesian statistics, which frames all inference about unknown quantities as a calculation involving the posterior density. In this paper, we review variational inference (VI), a method from machine learning that approximates probability densities through optimization. VI has been used in many applications and tends to be faster than classical methods, such as Markov chain Monte Carlo sampling. The idea behind VI is to first posit a family of densities and then to find the member of that family which is close to the target. Closeness is measured by Kullback-Leibler divergence. We review the ideas behind mean-field variational inference, discuss the special case of VI applied to exponential family models, present a full example with a Bayesian mixture of Gaussians, and derive a variant that uses stochastic optimization to scale up to massive data. We discuss modern research in VI and highlight important open problems. VI is powerful, but it is not yet well understood. Our hope in writing this paper is to catalyze statistical research on this class of algorithms.}
}
@article{BleietalJMLR-03,
       author = {Blei, David and Ng, Andrew and Jordan, Michael},
        title = {Latent {D}irichlet Allocation},
      journal = {Journal of Machine Learning Research},
       volume = 3,
         year = 2003,
        pages = {993-1022},
          url = {http://www.cs.berkeley.edu/~blei/papers/blei03a.ps.gz},
}
@inproceedings{BleietalNIPS-02,
       author = {David Blei and Andrew Ng and Michael Jordan},
        title = {Latent {D}irichlet Allocation},
    booktitle = {Advances in Neural Information Processing Systems (NIPS 14)},
         year = 2002,
}
@inproceedings{BleietalNIPS-04,
       author = {D. Blei and T. Griffiths and M. Jordan and J. Tenenbaum},
        title = {Hierarchical topic models and the nested {C}hinese restaurant process},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 16,
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2004,
}
@article{BlellochCACM-96,
       author = {Guy E. Blelloch},
        title = {Programming Parallel Algorithms},
      journal = {Communications of the ACM},
       volume = 39,
        pages = {85-97},
         year = 1996,
}
@article{BlellochetalJPDC-94,
       author = {Guy E. Blelloch and Siddhartha Chatterjee and Jonathan C. Hardwick and Jay Sipelstein and Marco Zagha},
        title = {Implementation of a Portable Nested Data-Parallel Language},
      journal = {Journal of Parallel and Distributed Computing},
       volume = 21,
        issue = 1,
        pages = {4-14},
         year = 1994
}
@inproceedings{BlouwandEliasmithCSS-13,
        title = {A Neurally Plausible Encoding of Word Order Information into a Semantic Vector Space},
       author = {Peter Blouw and Chris Eliasmith},
    booktitle = {35th Annual Conference of the Cognitive Science Society},
         year = 2013,
        pages = {1905-1910},
     abstract = {Distributed models of lexical semantics increasingly incorporate information about word order. One influential method for encoding this information into high-dimensional spaces uses convolution to bind together vectors to form representations of numerous n-grams that a target word is a part of. The computational complexity of this method has led to the development of an alternative that uses random permutation to perform order-sensitive vector combinations. We describe a simplified form of order encoding with convolution that yields comparable performance to earlier models, and we discuss considerations of neural implementation that favor the use of the proposed encoding. We conclude that this new encoding method is a more neurally plausible alternative than its predecessors.},
}
@inproceedings{BlumandChalasaniCOLT-92,
       author = {Avrim Blum and Prasad Chalasani},
        title = {Learning Switching Concepts},
    booktitle = {{COLT}: Proceedings of the Workshop on Computational Learning Theory},
    publisher = {Morgan Kaufmann Publishers},
         year = 1992,
}
@inproceedings{BlumandFurstIJCAI-95,
        title = {Fast Planning Through Planning Graph Analysis},
       author = {Blum, Avrim L. and Furst, Merrick L.},
    booktitle = {Proceedings {IJCAI} 14},
     location = {Montreal, Canada},
 organization = {IJCAII},
        pages = {1636-1642},
         year = 1995,
}
@inproceedings{BlumandKozenFOCS-78,
       author = {Blum, Manuel and Kozen, Dexter},
        title = {On the Power of the Compass (or, Why Mazes Are Easier to Search than Graphs)},
    booktitle = {Proceedings of the 19th Symposium on the Foundations of Computer Science},
         year = 1978,
        pages = {132-142},
}
@article{BlumeretalIPL-87,
       author = {Blumer, Anselm and Ehrenfeucht, Andrzej and Haussler, David and Warmuth, Manfred K.},
        title = {Occam's Razor},
      journal = {Information Processing Letters},
    publisher = {North Holland},
       volume = 24,
         year = 1987,
        pages = {377-380},
}
@article{BlumeretalJACM-89,
       author = {Blumer, Anselm and Ehrenfeucht, Andrzej and Haussler, David and Warmuth, Manfred K.},
        title = {Learnability and the Vapnik-Chervonenkis Dimension},
      journal = {Journal of the {ACM}},
         year = 1989,
}
@article{BlumofeJPDC-96,
     author = {Robert D. Blumofe and Christopher F. Joerg and Bradley C. Kuszmaul and 
               Charles E. Leiserson and Keith H. Randall and Yuli Zhou},
      title = {{C}ilk: An Efficient Multithreaded Runtime System},
    journal = {Journal of Parallel and Distributed Computing},
     volume = 37,
      issue = 1,
      pages = {55-69},
       year = 1996,
}
@article{BlundelletalCoRR-16,
       author = {Charles Blundell, Benigno Uria, Alexander Pritzel, Yazhe Li, Avraham Ruderman, Joel Z Leibo, Jack Rae, Daan Wierstra, Demis Hassabis},
        title = {Model-Free Episodic Control},
      journal = {CoRR},
       volume = {rXiv:1606.04460},
         year = {2016},
     abstract = {State of the art deep reinforcement learning algorithms take many millions of interactions to attain human-level performance. Humans, on the other hand, can very quickly exploit highly rewarding nuances of an environment upon first discovery. In the brain, such rapid learning is thought to depend on the hippocampus and its capacity for episodic memory. Here we investigate whether a simple model of hippocampal episodic control can learn to solve difficult sequential decision-making tasks. We demonstrate that it not only attains a highly rewarding strategy significantly faster than state-of-the-art deep reinforcement learning algorithms, but also achieves a higher overall reward on some of the more challenging domains.}
}
@article{BoahenSA-05,
      author = {Boahen, Kwabena},
       title = {Neuromorphic Microchips},
     journal = {Scientific American},
      volume = 292,
       issue = 5,
       pages = {55-63},
        year = 2005,
         url = {http://search.epnet.com/login.aspx?direct=true&db=aph&an=16729948},
}
@article{BocketalNATURE-11,
        title = {Network anatomy and in vivo physiology of visual cortical neurons},
       author = {Bock, Davi D. and Lee, Wei-Chung Allen and Kerlin, Aaron M. and Andermann, Mark L. and Hood, Greg and Wetzel, Arthur W. and Yurgenson, Sergey and Soucy, Edward R. and Kim, Hyon Suk and Reid, R. Clay},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = 471,
        issue = 7337,
         year = 2011,
        pages = {177-182},
     abstract = {In the cerebral cortex, local circuits consist of tens of thousands of neurons, each of which makes thousands of synaptic connections. Perhaps the biggest impediment to understanding these networks is that we have no wiring diagrams of their interconnections. Even if we had a partial or complete wiring diagram, however, understanding the network would also require information about each neuron's function. Here we show that the relationship between structure and function can be studied in the cortex with a combination of in vivo physiology and network anatomy. We used two-photon calcium imaging to characterize a functional property--the preferred stimulus orientation--of a group of neurons in the mouse primary visual cortex. Large-scale electron microscopy of serial thin sections was then used to trace a portion of these neurons' local network. Consistent with a prediction from recent physiological experiments, inhibitory interneurons received convergent anatomical input from nearby excitatory neurons with a broad range of preferred orientations, although weak biases could not be rejected.}
}
@inproceedings{BoddyAAAI-91,
      author = {Boddy, Mark},
       title = {Anytime Problem Solving Using Dynamic Programming},
   booktitle = {Proceedings {AAAI}-91},
    location = {Anaheim, California},
   publisher = {MIT Press},
     address = {Cambridge, Massachusetts},
organization = {AAAI},
       pages = {738-743},
        year = 1991,
}
@techreport{BoddyPhD,
       author = {Boddy, Mark},
        title = {Solving Time-Dependent Problems: A Decision-Theoretic Approach to Planning in Dynamic Environments},
  institution = {Brown University Department of Computer Science},
         year = 1991,
        issue = {CS-91-06},
      address = {Providence, RI},
}
@article{BoddyandDeanAIJ-94,
       author = {Boddy, Mark and Dean, Thomas},
        title = {Decision-Theoretic Deliberation Scheduling for Problem Solving in Time-Constrained Environments},
      journal = AIJ,
       volume = 67,
        issue = 2,
        pages = {245-286},
         year = 1994,
}
@inproceedings{BoddyandDeanIJCAI-89,
      author = {Boddy, Mark and Dean, Thomas},
       title = {Solving Time-Dependent Planning Problems},
   booktitle = {Proceedings {IJCAI}-89},
    location = {Detroit, Michigan},
   publisher = {Morgan Kaufmann Publishers},
     address = {San Francisco, California},
organization = {IJCAI},
       pages = {979-984},
        year = 1989,
}
@inproceedings{BoddyandDeanJPL-89,
      author = {Boddy, Mark and Dean, Thomas},
       title = {Approximation Algorithms for Planning and Control},
   booktitle = {Proceedings of the {NASA} Conference on Telerobotics},
        year = 1989,
}
@incollection{BodhesakoandWilliams1987,
        title = {Consciousness ({V}i\~{n}\~{n}\={a}\d{n}a)},
       author = {\~{N}\={a}\d{n}av\={i}ra Thera},
    booktitle = {Clearing the {P}ath: {W}ritings of {\~N}{\=a}ṇav{\=\i}ra Thera (1960-1965)},
       editor = {Samonera Bodhesako and Forrest Williams},
         year = {1987},
    publisher = {Path Press}
}
@article{BodinandGoldenNETWORKS-81,
       author = {Bodin, Lawrence and Golden, Bruce},
        title = {Classification in Vehicle Routing and Scheduling},
      journal = {Networks},
       volume = 11,
         year = 1981,
        pages = {97-108},
},
      journal = {Journal Neuroscience},
       volume = {19},
       number = {10},
         year = {1999},
        pages = {4090-4101},
     abstract = {During slow wave sleep (SWS), traces of neuronal activity patterns from preceding behavior can be observed in rat hippocampus and neocortex. The spontaneous reactivation of these patterns is manifested as the reinstatement of the distribution of pairwise firing-rate correlations within a population of simultaneously recorded neurons. The effects of behavioral state [quiet wakefulness, SWS, and rapid eye movement (REM)], interactions between two successive spatial experiences, and global modulation during 200 Hz electroencephalographic (EEG) "ripples" on pattern reinstatement were studied in CA1 pyramidal cell population recordings. Pairwise firing-rate correlations during often repeated experiences accounted for a significant proportion of the variance in these interactions in subsequent SWS or quiet wakefulness and, to a lesser degree, during SWS before the experience on a given day. The latter effect was absent for novel experiences, suggesting that a persistent memory trace develops with experience. Pattern reinstatement was strongest during sharp wave-ripple oscillations, suggesting that these events may reflect system convergence onto attractor states corresponding to previous experiences. When two different experiences occurred in succession, the statistically independent effects of both were evident in subsequent SWS. Thus, the patterns of neural activity reemerge spontaneously, and in an interleaved manner, and do not necessarily reflect persistence of an active memory (i.e., reverberation). Firing-rate correlations during REM sleep were not related to the preceding familiar experience, possibly as a consequence of trace decay during the intervening SWS. REM episodes also did not detectably influence the correlation structure in subsequent SWS, suggesting a lack of strengthening of memory traces during REM sleep, at least in the case of familiar experiences.}
}
@phdthesis{BohusPhD-07,
       author = {Dan Bohus},
        title = {Error Awareness and Recovery in Conversational Spoken Language Interfaces},
       school = {Carnegie Mellon University},
         year = 2007,
}
@article{BohusandRudnickyCS-09,
       author = {Bohus, Dan and Rudnicky, Alexander I.},
        title = {{The RavenClaw Dialogue Management Framework: Architecture and Systems}},
      journal = {Computer Speech and Language},
    publisher = {Academic Press Ltd.},
       volume = 23,
        issue = 3,
         year = 2009,
        pages = {332-361},
}
@article{BohusandRudnickyCSL-09,
        title = {{The {RavenClaw} dialogue management framework: architecture and systems}},
       author = {Dan Bohus and Alexander I. Rudnicky},
      journal = {Computer Speech \& Language},
       volume = 23,
        issue = 3,
         year = 2009,
        pages = {332-361},
     abstract = {In this paper, we describe RavenClaw, a plan-based, task-independent dialog management framework. RavenClaw isolates the domain-specific aspects of the dialog control logic from domain-independent conversational skills, and in the process facilitates rapid development of mixed-initiative systems operating in complex, task-oriented domains. System developers can focus exclusively on describing the dialog task control logic, while a large number of domain-independent conversational skills such as error handling, timing and turn-taking are transparently supported and enforced by the RavenClaw dialog engine. To date, RavenClaw has been used to construct and deploy a large number of systems, spanning different domains and interaction styles, such as information access, guidance through procedures, command-and-control, medical diagnosis, etc. The framework has easily adapted to all of these domains, indicating a high degree of versatility and scalability.},
}
@inproceedings{BohusandRudnickyHLTEMNLP-05,
       author = {Dan Bohus and Alexander I. Rudnicky},
        title = {Error handling in the RavenClaw dialog management architecture},
    booktitle = {In Proceedings of the Human Language Technology Conference Conference on Empirical Methods in Natural Language Processing},
         year = 2005,
}
@inproceedings{BojnordiandIpekHPCA-16,
       author = {M.N. Bojnordi and E. Ipek},
        title = {Memristive Boltzmann Machine: A Hardware Accelerator for Combinatorial Optimization and Deep Learning},
    booktitle = {Proceedings of the International Symposium on High Performance Computer Architecture},
         year = {2016},
     abstract = {The Boltzmann machine is a massively parallel computational model capable of solving a broad class of combinatorial optimization problems. In recent years, it has been successfully applied to training deep machine learning models on massive datasets. High performance implementations of the Boltzmann machine using GPUs, MPI-based HPC clusters, and FPGAs have been proposed in the literature. Regrettably, the required all-to-all communication among the processing units limits the performance of these efforts. This paper examines a new class of hardware accelerators for large-scale combinatorial optimization and deep learning based on memristive Boltzmann machines. A massively parallel, memory-centric hardware accelerator is proposed based on recently developed resistive RAM (RRAM) technology. The proposed accelerator exploits the electrical properties of RRAM to realize in situ, fine-grained parallel computation within memory arrays, thereby eliminating the need for exchanging data between the memory cells and the computational units. Two classical optimization problems, graph partitioning and boolean satisfiability, and a deep belief network application are mapped onto the proposed hardware. As compared to a multicore system, the proposed accelerator achieves 57\% higher performance and 25\% lower energy with virtually no loss in the quality of the solution to the optimization problems. The memristive accelerator is also compared against an RRAM based processing-in-memory (PIM) system, with respective performance and energy improvements of 6.89\% and 5.2\%.},
}
@article{BolanosetalCOiN-18,
       author = {Noelia Ant{\'{o}}n-Bola{\~{o}}os and Ana Espinosa and Guillermina L{\'{o}}pez-Bendito},
        title = {Developmental interactions between thalamus and cortex: A true love reciprocal story},
      journal = {Current Opinion in Neurobiology},
       volume = {40},
         year = {2018},
        pages = {33-41},
     abstract = {The developmental programs that control the specification of cortical and thalamic territories are maintained largely as independent processes. However, bulk of evidence demonstrates the requirement of the reciprocal interactions between cortical and thalamic neurons as key for the correct development of functional thalamocortical circuits. This reciprocal loop of connections is essential for sensory processing as well as for the execution of complex sensory-motor tasks. Here, we review recent advances in our understanding of how mutual collaborations between both brain regions define area patterning and cell differentiation in the thalamus and cortex.}
}
@article{BoldogetalNATURE-NEUROSCIENCE-18,
       author = {Boldog, Eszter and Bakken, Trygve E. and Hodge, Rebecca D. and Novotny, Mark and Aevermann, Brian D. and Baka, Judith and Bord{\~A}{\textcopyright}, S{\~A}{\textexclamdown}ndor and Close, Jennie L. and Diez-Fuertes, Francisco and Ding, Song-Lin and Farag{\~A}{\textthreesuperior}, N{\~A}{\textthreesuperior}ra and Kocsis, {\~A}?gnes K. and Kov{\~A}{\textexclamdown}cs, Bal{\~A}{\textexclamdown}zs and Maltzer, Zoe and McCorrison, Jamison M. and Miller, Jeremy A. and Moln{\~A}{\textexclamdown}r, G{\~A}{\textexclamdown}bor and Ol{\~A}{\textexclamdown}h, G{\~A}{\textexclamdown}sp{\~A}{\textexclamdown}r and Ozsv{\~A}{\textexclamdown}r, Attila and R{\~A}{\textthreesuperior}zsa, M{\~A}{\textexclamdown}rton and Shehata, Soraya I. and Smith, Kimberly A. and Sunkin, Susan M. and Tran, Danny N. and Venepally, Pratap and Wall, Abby and Pusk{\~A}{\textexclamdown}s, L{\~A}{\textexclamdown}szl{\~A}{\textthreesuperior} G. and Barz{\~A}{\textthreesuperior}, P{\~A}{\textexclamdown}l and Steemers, Frank J. and Schork, Nicholas J. and Scheuermann, Richard H. and Lasken, Roger S. and Lein, Ed S. and Tam{\~A}{\textexclamdown}s, G{\~A}{\textexclamdown}bor},
        title = {Transcriptomic and morphophysiological evidence for a specialized human cortical GABAergic cell type},
      journal = {Nature Neuroscience},
         year = {2018},
       volume = {21},
        issue = {9},
        pages = {1185-1195},
     abstract = {We describe convergent evidence from transcriptomics, morphology, and physiology for a specialized GABAergic neuron subtype in human cortex. Using unbiased single-nucleus RNA sequencing, we identify ten GABAergic interneuron subtypes with combinatorial gene signatures in human cortical layer 1 and characterize a group of human interneurons with anatomical features never described in rodents, having large 'rosehip'-like axonal boutons and compact arborization. These rosehip cells show an immunohistochemical profile (GAD1+CCK+CNR1-SST+CALB2+PVALB) matching a single transcriptomically defined cell type whose specific molecular marker signature is not seen in mouse cortex. Rosehip cells in layer 1 make homotypic gap junctions, predominantly target apical dendritic shafts of layer 3 pyramidal neurons, and inhibit backpropagating pyramidal action potentials in microdomains of the dendritic tuft. These cells are therefore positioned for potent local control of distal dendritic computation in cortical pyramidal neurons.},
}
@article{BolhuisetalPLoS-14,
       author = {Bolhuis, Johan J. and Tattersall, Ian and Chomsky, Noam and Berwick, Robert C.},
      journal = {{PLoS} Biology},
    publisher = {Public Library of Science},
        title = {How Could Language Have Evolved?},
       volume = 12,
        issue = 8,
         year = 2014,
        pages = {e1001934},
     abstract = {How could language have evolved? What is the key innovation underlying the evolution of human language? This Essay argues that the ability to "merge" two syntactic elements uniquely explains the recentness and stability of language.},
}
@article{BolhuisetalRNN-10,
       author = {Bolhuis, Johan J. and Okanoya, Kazuo and Scharff, Constance},
        title = {Twitter evolution: converging mechanisms in birdsong and human speech},
      journal = {Nature Reviews Neuroscience},
         year = 2010,
    publisher = {Nature Publishing Group},
       volume = 11,
        issue = 11,
        pages = {747-759},
     abstract = {Vocal imitation in human infants and in some orders of birds relies on auditory-guided motor learning during a sensitive period of development. It proceeds from 'babbling' (in humans) and 'subsong' (in birds) through distinct phases towards the full-fledged communication system. Language development and birdsong learning have parallels at the behavioural, neural and genetic levels. Different orders of birds have evolved networks of brain regions for song learning and production that have a surprisingly similar gross anatomy, with analogies to human cortical regions and basal ganglia. Comparisons between different songbird species and humans point towards both general and species-specific principles of vocal learning and have identified common neural and molecular substrates, including the forkhead box P2 (FOXP2) gene.}
}
@inproceedings{BollackeretalSIGMOD-08,
       author = {Bollacker, Kurt and Evans, Colin and Paritosh, Praveen and Sturge, Tim and Taylor, Jamie},
        title = {Freebase: a collaboratively created graph database for structuring human knowledge},
    booktitle = {{ACM} {SIGMOD} International Conference on Management of Data},
    publisher = {ACM},
      address = {New York, NY},
     location = {Vancouver, Canada},
         year = 2008,
        pages = {1247-1250},
}
@article{BollePAMI-86,
       author = {Bolle, R. M. and Cooper, D. B.},
        title = {On Optimally Combining Pieces of Information, with Application to Estimating 3-{D} Complex-Object Position From Range Data},
      journal = PAMI,
       volume = 8,
         year = 1986,
        pages = {619-638},
}
@book{Bollinger88,
       author = {Bollinger, John G. and Duffie, Neil A.},
        title = {Computer Control of Machines and Processes},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1988,
}
@book{Bollobas85,
       author = {Bollob\'{a}s, B\'{e}la},
        title = {Random Graphs},
    publisher = {Academic Press},
      address = {London},
         year = 1985,
}
@article{BolouretalSIGART-82,
       author = {Bolour, A. and Anderson, T. L. and Dekeyser, L. J. and Wong, H. K. T.},
        title = {The Role of Time in Information Processing: A Survey},
      journal = {SIGART Newsletter},
         year = 1982,
        pages = {28-48},
}
@article{BonanomietalCELL-12,
       author = {Dario Bonanomi and Onanong Chivatakarn and Ge Bai and Houari Abdesselem and Karen Lettieri and Till Marquardt and Brian A. Pierchala and Samuel L. Pfaff},
        title = {Ret Is a Multifunctional Coreceptor that Integrates Diffusible- and Contact-Axon Guidance Signals},
      journal = {Cell},
       volume = 148,
        issue = 3,
         year = 2012,
        pages = {568-582},
     abstract = {Growing axons encounter multiple guidance cues, but it is unclear how separate signals are resolved and integrated into coherent instructions for growth cone navigation. We report that glycosylphosphatidylinositol (GPI)-anchored ephrin-As function as "reverse" signaling receptors for motor axons when contacted by transmembrane EphAs present in the dorsal limb. Ephrin-A receptors are thought to depend on transmembrane coreceptors for transmitting signals intracellularly. We show that the receptor tyrosine kinase Ret is required for motor axon attraction mediated by ephrin-A reverse signaling. Ret also mediates GPI-anchored GFR$\alpha$1 signaling in response to GDNF, a diffusible chemoattractant in the limb, indicating that Ret is a multifunctional coreceptor for guidance molecules. Axons respond synergistically to coactivation by GDNF and EphA ligands, and these cooperative interactions are gated by GFR$\alpha$1 levels. Our studies uncover a hierarchical GPI-receptor signaling network that is constructed from combinatorial components and integrated through Ret using ligand coincidence detection.}
}
@article{BonassoandDeanAIM-97,
       author = {Dean, Thomas and Bonasso, R. Peter},
        title = {A Retrospective on the  {AAAI} Robot Competitions},
      journal = {{AI} Magazine},
       volume = 18,
        issue = 1,
        pages = {11-23},
         year = 1997,
}
@article{BooneandDoeDEVELOPMENTAL-NEUROBIOLOGY-08,
       author = {Boone, Jason Q. and Doe, Chris Q.},
        title = {Identification of Drosophila type {II} neuroblast lineages containing transit amplifying ganglion mother cells},
      journal = {Devopmental Neurobiology},
         year = {2008},
       volume = {68},
        issue = {9},
        pages = {1185-1195},
     abstract = {Mammalian neural stem cells generate transit amplifying progenitors that expand the neuronal population, but these type of progenitors have not been studied in Drosophila. The Drosophila larval brain contains {\textasciitilde}100 neural stem cells (neuroblasts) per brain lobe, which are thought to bud off smaller ganglion mother cells (GMCs) that each produce two post-mitotic neurons. Here we use molecular markers and clonal analysis to identify a novel neuroblast cell lineage containing "transit amplifying GMCs" (TA-GMCs). TA-GMCs differ from canonical GMCs in several ways: each TA-GMC has nuclear Deadpan, cytoplasmic Prospero, forms Prospero crescents at mitosis, and generates up to 10 neurons; canonical GMCs lack Deadpan, have nuclear Prospero, lack Prospero crescents at mitosis, and generate two neurons. We conclude that there are at least two types of neuroblast lineages: a type I lineage where GMCs generate two neurons, and a type II lineage where TA-GMCs have longer lineages. Type II lineages allow more neurons to be produced faster than type I lineages, which may be advantageous in a rapidly developing organism like Drosophila.},
}
@book{Boorstin03,
       author = {Jon Boorstin},
        title = {The Newsboys' Lodging House: Or The confessions of Williams James},
    publisher = {Viking},
      address = {New York, NY},
         year = 2003,
}
@book{Booth1967,
        title = {The Knowledge Most Worth Having},
       author = {Booth, Wayne C.},
         year = {1967},
    publisher = {University of Chicago Press},
}
@article{BordensteiandBordensteinNATURE_COMMUNICATIONS-16,
       author = {Bordenstein, Sarah R. and Bordenstein, Seth R.},
        title = {Eukaryotic association module in phage {WO} genomes from {W}olbachia},
      journal = {Nature Communications},
         year = {2016},
    publisher = {Nature Publishing Group},
       volume = {7},
        pages = {13155 EP},
     abstract = {Viruses are trifurcated into eukaryotic, archaeal and bacterial categories. This domain-specific ecology underscores why eukaryotic viruses typically co-opt eukaryotic genes and bacteriophages commonly harbour bacterial genes. However, the presence of bacteriophages in obligate intracellular bacteria of eukaryotes may promote DNA transfers between eukaryotes and bacteriophages. Here we report a metagenomic analysis of purified bacteriophage WO particles of Wolbachia and uncover a eukaryotic association module in the complete WO genome. It harbours predicted domains, such as the black widow latrotoxin C-terminal domain, that are uninterrupted in bacteriophage genomes, enriched with eukaryotic protease cleavage sites and combined with additional domains to forge one of the largest bacteriophage genes to date (14,256 bp). To the best of our knowledge, these eukaryotic-like domains have never before been reported in packaged bacteriophages and their phylogeny, distribution and sequence diversity imply lateral transfers between bacteriophage/prophage and animal genomes. Finally, the WO genome sequences and identification of attachment sites will potentially advance genetic manipulation of Wolbachia.}
}
@article{BordesandWestonCoRR-16,
       author = {Antoine Bordes and Jason Weston},
        title = {Learning End-to-End Goal-Oriented Dialog},
      journal = {CoRR},
       volume = {arXiv:1605.07683},
         year = {2016},
     abstract = {Traditional dialog systems used in goal-oriented applications require a lot of domain-specific handcrafting, which hinders scaling up to new domains. End-to-end dialog systems, in which all components are trained from the dialogs themselves, escape this limitation. But the encouraging success recently obtained in chit-chat dialog may not carry over to goal-oriented settings. This paper proposes a testbed to break down the strengths and shortcomings of end-to-end dialog systems in goal-oriented applications. Set in the context of restaurant reservation, our tasks require manipulating sentences and symbols, so as to properly conduct conversations, issue API calls and use the outputs of such calls. We show that an end-to-end dialog system based on Memory Networks can reach promising, yet imperfect, performance and learn to perform non-trivial operations. We confirm those results by comparing our system to a hand-crafted slot-filling baseline on data from the second Dialog State Tracking Challenge (Henderson et al., 2014a). We show similar result patterns on data extracted from an online concierge service.}
}
@inproceedings{BordesetalAAAI-11,
       author = {Antoine Bordes and Jason Weston and Ronan Collobert and Yoshua Bengio},
        title = {Learning Structured Embeddings of Knowledge Bases},
    booktitle = {Proceedings of the Twenty-Sixth {AAAI} Conference on Artificial Intelligence},
         year = 2012,
}
@inproceedings{BordesetalAISTATS-12,
       author = {Bordes, Antoine and Glorot, Xavier and Weston, Jason and Bengio, Yoshua},
        title = {Joint Learning of Words and Meaning Representations for Open-Text Semantic Parsing},
    booktitle = {Proceedings of the 15th International Conference on Artificial Intelligence and Statistics},
         year = 2012
}
@inproceedings{BordesetalICLR-17,
       author = {Antoine Bordes and Y-Lan Boureau and Jason Weston},
        title = {Learning End-to-End Goal-Oriented Dialog},
    booktitle = {International Conference on Learning Representations},
         year = 2017,
}
@inproceedings{BorensteinandMalikCVPR-06,
       author = {Eran Borenstein and Jitendra Malik},
        title = {Shape Guided Object Segmentation},
    booktitle = {Proceedings of the 2006 {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
        pages = {969-976},
         year = 2006,
}
@inproceedings{BorensteinandUllmanECCV-02,
       author = {Eran Borenstein and Shimon Ullman},
        title = {Class-Specific, Top-Down Segmentation},
    booktitle = {Proceedings of the 7th European Conference on Computer Vision},
    publisher = {Springer-Verlag},
      address = {London, UK},
         year = 2002,
        pages = {109-124},
}
@inproceedings{BorensteinetalCVPRW-04,
       author = {Eran Borenstein and Eitan Sharon and Shimon Ullman},
        title = {Combining Top-Down and Bottom-Up Segmentation},
    booktitle = {Proceedings of the 2004 {IEEE} Conference on Computer Vision and Pattern Recognition 
              Workshop on Perceptual Organization in Computer Vision},
    publisher = {IEEE Computer Society},
       volume = 4,
        pages = 46,
         year = 2004,
}
@article{BorgersetalPLoS-12,
        title = {Minimal size of cell assemblies coordinated by gamma oscillations},
       author = {Christoph B\"{o}rgers and Giovanni Talei Franzesi and Fiona E. N. LeBeau and Edward S. Boyden and Nancy J. Kopell},
      journal = {{PLoS} Biology},
    publisher = {Public Library of Science},
       volume = 8,
        issue = 2,
         year = 2012,
        pages = {e1002362},
     comment = {http://www.ncbi.nlm.nih.gov/pubmed/22346741},
}
@article{BorghuisetalJoN-11,
       author = {Borghuis, Bart G and Tian, Lin and Xu, Ying and Nikonov, Sergei S and Vardi, Noga and Zemelman, Boris V and Looger, Loren L},
      journal = {The Journal of Neuroscience},
        title = {Imaging light responses of targeted neuron populations in the rodent retina},
       volume = 31,
        issue = 8,
         year = 2011,
        pages = {2855-2867},
     abstract = {Decoding the wiring diagram of the retina requires simultaneous observation of activity in identified neuron populations. Available recording methods are limited in their scope: electrodes can access only a small fraction of neurons at once, whereas synthetic fluorescent indicator dyes label tissue indiscriminately. Here, we describe a method for studying retinal circuitry at cellular and subcellular levels combining two-photon microscopy and a genetically encoded calcium indicator. Using specific viral and promoter constructs to drive expression of GCaMP3, we labeled all five major neuron classes in the adult mouse retina. Stimulus-evoked GCaMP3 responses as imaged by two-photon microscopy permitted functional cell type annotation. Fluorescence responses were similar to those measured with the small molecule dye OGB-1. Fluorescence intensity correlated linearly with spike rates >10 spikes/s, and a significant change in fluorescence always reflected a significant change in spike firing rate. GCaMP3 expression had no apparent effect on neuronal function. Imaging at subcellular resolution showed compartment-specific calcium dynamics in multiple identified cell types.},
}
@article{BornscheuerCOCB-01,
        title = {Improved biocatalysts by directed evolution and rational protein design},
       author = {Bornscheuer, U.T. and Pohl, M.},
      journal = {Current Opinion Chemical Biology},
       volume = 5,
        issue = 2,
        pages = {137-43},
         year = 2001,
     abstract = {The efficient application of biocatalysts requires the availability of suitable enzymes with high activity and stability under process conditions, desired substrate selectivity and high enantioselectivity. However, wild-type enzymes often need to be optimized to fulfill these requirements. Two rather contradictory tools can be used on a molecular level to create tailor-made biocatalysts: directed evolution and rational protein design.}
}
@book{Borrie86,
       author = {Borrie, John A.},
        title = {Modern Control Systems: A Manual of Design Methods},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1986,
}
@article{BorstNATURE-14,
       author = {Borst, Alexander},
        title = {Fly visual course control: behaviour, algorithms and circuits},
      journal = {Nature Reviews Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {15},
        issue = {9},
        pages = {590-599},
     abstract = {Understanding how the brain controls behaviour is undisputedly one of the grand goals of neuroscience research, and the pursuit of this goal has a long tradition in insect neuroscience. However, appropriate techniques were lacking for a long time. Recent advances in genetic and recording techniques now allow the participation of identified neurons in the execution of specific behaviours to be interrogated. By focusing on fly visual course control, I highlight what has been learned about the neuronal circuit modules that control visual guidance in Drosophila melanogaster through the use of these techniques.},
}
@article{BorstandEulerNEURON-11,
        title = {Seeing Things in Motion: Models, Circuits, and Mechanisms},
      journal = {Neuron},
       author = {Alexander Borst and Thomas Euler},
       volume = {71},
       number = {6},
        pages = {974-994},
         year = {2011},
     abstract = {Motion vision provides essential cues for navigation and course control as well as for mate, prey, or predator detection. Consequently, neurons responding to visual motion in a direction-selective way are found in almost all species that see. However, directional information is not explicitly encoded at the level of a single photoreceptor. Rather, it has to be computed from the spatio-temporal excitation level of at least two photoreceptors. How this computation is done and how this computation is implemented in terms of neural circuitry and membrane biophysics have remained the focus of intense research over many decades. Here, we review recent progress made in this area with an emphasis on insects and the vertebrate retina.},
}
@article{BorstandHelmstaedterNATURE-NEUROSCIENCE-15,
       author = {Borst, Alexander and Helmstaedter, Moritz},
        title = {Common circuit design in fly and mammalian motion vision},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {18},
        issue = {8},
         year = {2015},
        pages = {1067-1076},
     abstract = {Motion-sensitive neurons have long been studied in both the mammalian retina and the insect optic lobe, yet striking similarities have become obvious only recently. Detailed studies at the circuit level revealed that, in both systems, (i) motion information is extracted from primary visual information in parallel ON and OFF pathways; (ii) in each pathway, the process of elementary motion detection involves the correlation of signals with different temporal dynamics; and (iii) primary motion information from both pathways converges at the next synapse, resulting in four groups of ON-OFF neurons, selective for the four cardinal directions. Given that the last common ancestor of insects and mammals lived about 550 million years ago, this general strategy seems to be a robust solution for how to compute the direction of visual motion with neural hardware.},
}
@article{BosmanetalNEURON-12,
       author = {Conrado A. Bosman and Jan-Mathijs Schoffelen and Nicolas Brunet and Robert Oostenveld and Andre M. Bastos and Thilo Womelsdorf and Birthe Rubehn and Thomas Stieglitz and Peter De Weerd and Pascal Fries},
        title = {Attentional Stimulus Selection through Selective Synchronization between Monkey Visual Areas},
      journal = {Neuron},
       volume = {75},
        issue = {5},
         year = {2012},
        pages = {875-888},
}
@article{BossuandSiegelAIJ-85,
       author = {Bossu, Genevieve and Siegel, Pierre},
        title = {Saturation, Nonmonotonic Reasoning, and the Closed-World Assumption},
      journal = AIJ,
       volume = 25,
         year = 1985,
        pages = {13-63},
}
@article{BostanandStrickNATURE-REVIEWS-NEUROSCIENCE-18,
       author = {Bostan, Andreea C. and Strick, Peter L.},
        title = {The basal ganglia and the cerebellum: nodes in an integrated network},
      journal = {Nature Reviews Neuroscience},
         year = {2018},
       volume = {19},
        issue = {6},
        pages = {338-350},
     abstract = {The basal ganglia and the cerebellum are considered to be distinct subcortical systems that perform unique functional operations. The outputs of the basal ganglia and the cerebellum influence many of the same cortical areas but do so by projecting to distinct thalamic nuclei. As a consequence, the two subcortical systems were thought to be independent and to communicate only at the level of the cerebral cortex. Here, we review recent data showing that the basal ganglia and the cerebellum are interconnected at the subcortical level. The subthalamic nucleus in the basal ganglia is the source of a dense disynaptic projection to the cerebellar cortex. Similarly, the dentate nucleus in the cerebellum is the source of a dense disynaptic projection to the striatum. These observations lead to a new functional perspective that the basal ganglia, the cerebellum and the cerebral cortex form an integrated network. This network is topographically organized so that the motor, cognitive and affective territories of each node in the network are interconnected. This perspective explains how synaptic modifications or abnormal activity at one node can have network-wide effects. A future challenge is to define how the unique learning mechanisms at each network node interact to improve performance.},
}
@article{BostanetalPNAS-10,
       author = {Bostan, Andreea C. and Dum, Richard P. and Strick, Peter L.},
        title = {The basal ganglia communicate with the cerebellum},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {107},
       number = {18},
         year = {2010},
        pages = {8452-8456},
     abstract = {The basal ganglia and cerebellum are major subcortical structures that influence not only movement, but putatively also cognition and affect. Both structures receive input from and send output to the cerebral cortex. Thus, the basal ganglia and cerebellum form multisynaptic loops with the cerebral cortex. Basal ganglia and cerebellar loops have been assumed to be anatomically separate and to perform distinct functional operations. We investigated whether there is any direct route for basal ganglia output to influence cerebellar function that is independent of the cerebral cortex. We injected rabies virus (RV) into selected regions of the cerebellar cortex in cebus monkeys and used retrograde transneuronal transport of the virus to determine the origin of multisynaptic inputs to the injection sites. We found that the subthalamic nucleus of the basal ganglia has a substantial disynaptic projection to the cerebellar cortex. This pathway provides a means for both normal and abnormal signals from the basal ganglia to influence cerebellar function. We previously showed that the dentate nucleus of the cerebellum has a disynaptic projection to an input stage of basal ganglia processing, the striatum. Taken together these results provide the anatomical substrate for substantial two-way communication between the basal ganglia and cerebellum. Thus, the two subcortical structures may be linked together to form an integrated functional network.},
}
@article{BotoetalNATURE-18,
       author = {Boto, Elena and Holmes, Niall and Leggett, James and Roberts, Gillian and Shah, Vishal and Meyer, Sofie S. and Mu{\~A}{\textpm}oz, Leonardo Duque and Mullinger, Karen J.and Tierney, Tim M. and Bestmann, Sven and Barnes, Gareth R. and Bowtell, Richard and Brookes, Matthew J.},
        title = {Moving magnetoencephalography towards real-world applications with a wearable system},
      journal = {Nature},
         year = {2018},
    publisher = {Macmillan Publishers Limited},
       volume = {555},
        pages = {657-661},
     abstract = {Imaging human brain function with techniques such as magnetoencephalography1 typically requires a subject to perform tasks while their head remains still within a restrictive scanner. This artificial environment makes the technique inaccessible to many people, and limits the experimental questions that can be addressed. For example, it has been difficult to apply neuroimaging to investigation of the neural substrates of cognitive development in babies and children, or to study processes in adults that require unconstrained head movement (such as spatial navigation). Here we describe a magnetoencephalography system that can be worn like a helmet, allowing free and natural movement during scanning. This is possible owing to the integration of quantum sensors, which do not rely on superconducting technology, with a system for nulling background magnetic fields. We demonstrate human electrophysiological measurement at millisecond resolution while subjects make natural movements, including head nodding, stretching, drinking and playing a ball game. Our results compare well to those of the current state-of-the-art, even when subjects make large head movements. The system opens up new possibilities for scanning any subject or patient group, with myriad applications such as characterization of the neurodevelopmental connectome, imaging subjects moving naturally in a virtual environment and investigating the pathophysiology of movement disorders.}
}
@article{BotvinickCOiN-12,
       author = {Matthew Michael Botvinick},
        title = {Hierarchical reinforcement learning and decision making},
      journal = {Current Opinion in Neurobiology},
       volume = {22},
       number = {6},
        pages = {956-962},
         year = {2012},
     abstract = {The hierarchical structure of human and animal behavior has been of critical interest in neuroscience for many years. Yet understanding the neural processes that give rise to such structure remains an open challenge. In recent research, a new perspective on hierarchical behavior has begun to take shape, inspired by ideas from machine learning, and in particular the framework of hierarchical reinforcement learning. Hierarchical reinforcement learning builds on traditional reinforcement learning mechanisms, extending them to accommodate temporally extended behaviors or subroutines. The resulting computational paradigm has begun to influence both theoretical and empirical work in neuroscience, conceptually aligning the study of hierarchical behavior with research on other aspects of learning and decision making, and giving rise to some thought-provoking new findings.}
}
@article{BotvinickPTRS_B-07,
       author = {Botvinick, Matthew M.},
        title = {Multilevel structure in behaviour and in the brain: a model of Fuster's hierarchy},
      journal = {Philosophical transactions of the Royal Society of London. Series B, Biological sciences},
    publisher = {The Royal Society},
       volume = {362},
        issue = {1485},
         year = {2007},
        pages = {1615-1626},
     abstract = {A basic question, intimately tied to the problem of action selection, is that of how actions are assembled into organized sequences. Theories of routine sequential behaviour have long acknowledged that it must rely not only on environmental cues but also on some internal representation of temporal or task context. It is assumed, in most theories, that such internal representations must be organized into a strict hierarchy, mirroring the hierarchical structure of naturalistic sequential behaviour. This article reviews an alternative computational account, which asserts that the representations underlying naturalistic sequential behaviour need not, and arguably cannot, assume a strictly hierarchical form. One apparent liability of this theory is that it seems to contradict neuroscientific evidence indicating that different levels of sequential structure in behaviour are represented at different levels in a hierarchy of cortical areas. New simulations, reported here, show not only that the original computational account can be reconciled with this alignment between behavioural and neural organization, but also that it gives rise to a novel explanation for how this alignment might develop through learning.},
}
@article{BotvinickTiCS-08,
       author = {Botvinick, M. Matthew},
        title = {Hierarchical models of behavior and prefrontal function},
      journal = {Trends Cognitive Science},
       volume = {12},
       number = {5},
         year = {2008},
        pages = {201-208},
     abstract = {The recognition of hierarchical structure in human behavior was one of the founding insights of the cognitive revolution. Despite decades of research, however, the computational mechanisms underlying hierarchically organized behavior are still not fully understood. Recent findings from behavioral and neuroscientific research have fueled a resurgence of interest in the problem, inspiring a new generation of computational models. In addition to developing some classic proposals, these models also break fresh ground, teasing apart different forms of hierarchical structure, placing a new focus on the issue of learning and addressing recent findings concerning the representation of behavioral hierarchies within the prefrontal cortex. In addition to offering explanations for some key aspects of behavior and functional neuroanatomy, the latest models also pose new questions for empirical research.}
}
@article{BotvinickandAnANIPS-09,
       author = {Botvinick, Matthew and An, James},
        title = {Goal-directed decision making in prefrontal cortex: A computational framework},
      journal = {Advances in Neural Information Processing Systems},
       volume = {21},
         year = {2009},
        pages = {169-176},
     abstract = {Research in animal learning and behavioral neuroscience has distinguished between two forms of action control: a habit-based form, which relies on stored action values, and a goal-directed form, which forecasts and compares action outcomes based on a model of the environment. While habit-based control has been the subject of extensive computational research, the computational principles underlying goal-directed control in animals have so far received less attention. In the present paper, we advance a computational framework for goal-directed control in animals and humans. We take three empirically motivated points as founding premises: (1) Neurons in dorsolateral prefrontal cortex represent action policies, (2) Neurons in orbitofrontal cortex represent rewards, and (3) Neural computation, across domains, can be appropriately understood as performing structured probabilistic inference. On a purely computational level, the resulting account relates closely to previous work using Bayesian inference to solve Markov decision problems, but extends this work by introducing a new algorithm, which provably converges on optimal plans. On a cognitive and neuroscientific level, the theory provides a unifying framework for several different forms of goal-directed action selection, placing emphasis on a novel form, within which orbitofrontal reward representations directly drive policy selection.},
}
@article{BotvinickandBartoCOGNITION-09,
       author = {Botvinick, M. M. and Niv, Y. and Barto, A. C.},
        title = {Hierarchically organized behavior and its neural foundations: a reinforcement learning perspective},
      journal = {Cognition},
       volume = {113},
       number = {3},
         year = {2009},
        pages = {262-280},
     abstract = {Research on human and animal behavior has long emphasized its hierarchical structure-the divisibility of ongoing behavior into discrete tasks, which are comprised of subtask sequences, which in turn are built of simple actions. The hierarchical structure of behavior has also been of enduring interest within neuroscience, where it has been widely considered to reflect prefrontal cortical functions. In this paper, we reexamine behavioral hierarchy and its neural substrates from the point of view of recent developments in computational reinforcement learning. Specifically, we consider a set of approaches known collectively as hierarchical reinforcement learning, which extend the reinforcement learning paradigm by allowing the learning agent to aggregate actions into reusable subroutines or skills. A close look at the components of hierarchical reinforcement learning suggests how they might map onto neural structures, in particular regions within the dorsolateral and orbital prefrontal cortex. It also suggests specific ways in which hierarchical reinforcement learning might provide a complement to existing psychological models of hierarchically structured behavior. A particularly important question that hierarchical reinforcement learning brings to the fore is that of how learning identifies new action routines that are likely to provide useful building blocks in solving a wide range of future problems. Here and at many other points, hierarchical reinforcement learning offers an appealing framework for investigating the computational and neural underpinnings of hierarchically structured behavior.}
}
@article{BotvinickandPlautPSYCHOLOGICAL-RESEARCH-02,
       author = {Botvinick, Matthew and Plaut, David C.},
        title = {Representing task context: proposals based on a connectionist model of action},
      journal = {Psychological Research},
       volume = {66},
       number = {4},
         year = {2002},
        pages = {298-311},
     abstract = {Representations of task context play a crucial role in shaping human behavior. While the nature of these representations remains poorly understood, existing theories share a number of basic assumptions. One of these is that task representations are discrete, independent, and non-overlapping. We present here an alternative view, according to which task representations are instead viewed as graded, distributed patterns occupying a shared, continuous representational space. In recent work, we have implemented this view in a computational model of routine sequential action. In the present article, we focus specifically on this model's implications for understanding task representation, considering the implications of the account for two influential concepts: (1) cognitive underspecification, the idea that task representations may be imprecise or vague, especially in contexts where errors occur, and (2) information-sharing, the idea that closely related operations rely on common sets of internal representations.},
}
@article{BotvinickandPlautPSYCHOLOGICAL-REVIEW-04,
       author = {Botvinick, Matthew and Plaut, David C.},
        title = {Doing Without Schema Hierarchies: A Recurrent Connectionist Approach to Normal and Impaired Routine Sequential Action},
      journal = {Psychological Review},
       volume = {111},
       number = {2},
         year = {2004},
        pages = {395-429},
     abstract = {In everyday tasks, selecting actions in the proper sequence requires a continuously updated representation of temporal context. Previous models have addressed this problem by positing a hierarchy of processing units, mirroring the roughly hierarchical structure of naturalistic tasks themselves. The present study considers an alternative framework, in which the representation of context depends on recurrent connections within a network mapping from environmental inputs to actions. The ability of this approach to account for human performance was evaluated by applying it, through simulation, to a specific everyday task. The resulting model learned to deal flexibly with a complex set of sequencing constraints, encoding contextual information at multiple time scales within a single, distributed internal representation. Degrading this representation led to errors resembling those observed both in everyday behavior and in apraxia. Analysis of the model's function yielded numerous predictions relevant to both normal and apraxic performance.},
}
@article{BotvinickandWeinsteinPTRS-14,
       author = {Botvinick, M. and Weinstein, A.},
        title = {{M}odel-based hierarchical reinforcement learning and human action control},
      journal = {Philosophical Transactions Royal Society, London, B, Biological Science},
         year = {2014},
       volume = {369},
       number = {1655},
     abstract = {Recent work has reawakened interest in goal-directed or 'model-based' choice, where decisions are based on prospective evaluation of potential action outcomes. Concurrently, there has been growing attention to the role of hierarchy in decision-making and action control. We focus here on the intersection between these two areas of interest, considering the topic of hierarchical model-based control. To characterize this form of action control, we draw on the computational framework of hierarchical reinforcement learning, using this to interpret recent empirical findings. The resulting picture reveals how hierarchical model-based mechanisms might play a special and pivotal role in human decision-making, dramatically extending the scope and complexity of human behaviour.}
}
@article{BotvinicketalCOGNITION-09,
       author = {Matthew M. Botvinick and Yael Niv and Andrew C. Barto},
        title = {Hierarchically organized behavior and its neural foundations: A reinforcement learning perspective},
      journal = {Cognition},
       volume = {113},
       number = {3},
         year = {2009},
        pages = {262-280},
     abstract = {Research on human and animal behavior has long emphasized its hierarchical structure-the divisibility of ongoing behavior into discrete tasks, which are comprised of subtask sequences, which in turn are built of simple actions. The hierarchical structure of behavior has also been of enduring interest within neuroscience, where it has been widely considered to reflect prefrontal cortical functions. In this paper, we reexamine behavioral hierarchy and its neural substrates from the point of view of recent developments in computational reinforcement learning. Specifically, we consider a set of approaches known collectively as hierarchical reinforcement learning, which extend the reinforcement learning paradigm by allowing the learning agent to aggregate actions into reusable subroutines or skills. A close look at the components of hierarchical reinforcement learning suggests how they might map onto neural structures, in particular regions within the dorsolateral and orbital prefrontal cortex. It also suggests specific ways in which hierarchical reinforcement learning might provide a complement to existing psychological models of hierarchically structured behavior. A particularly important question that hierarchical reinforcement learning brings to the fore is that of how learning identifies new action routines that are likely to provide useful building blocks in solving a wide range of future problems. Here and at many other points, hierarchical reinforcement learning offers an appealing framework for investigating the computational and neural underpinnings of hierarchically structured behavior.}
}
@article{BotvinicketalCoRR-19,
       author = {Matthew Botvinick and David G. T. Barrett and Peter W. Battaglia and Nando de Freitas and Dharshan Kumaran and Joel Z. Leibo and Tim Lillicrap and Joseph Modayil and S. Mohamed and Neil C. Rabinowitz and Danilo Jimenez Rezende and Adam Santoro and Tom Schaul and Christopher Summerfield and Greg Wayne and Theophane Weber and Daan Wierstra and Shane Legg and Demis Hassabis},
        title = {Building Machines that Learn and Think for Themselves: Commentary on Lake et al., Behavioral and Brain Sciences, 2017},
      journal = {CoRR},
       volume = {arXiv:1711.08378},
         year = {2017},
     abstract = {Brette contends that the neural coding metaphor is an invalid basis for theories of what the brain does (Brette, 2019). Here, we argue that it is an insufficient guide for building an artificial intelligence (AI) that learns to accomplish short- and long-term goals in a complex, changing environment.}
}
@article{BotvinicketalNEUROIMAGE-05,
       author = {Botvinick, M.  and Jha, A. P.  and Bylsma, L. M.  and Fabian, S. A.  and Solomon, P. E.  and Prkachin, K. M.},
        title = {Viewing facial expressions of pain engages cortical areas involved in the direct experience of pain},
      journal = {Neuroimage},
         year = {2005},
       volume = {25},
       number = {1},
        pages = {312-319},
     abstract = {Recent neuroimaging and neuropsychological work has begun to shed light on how the brain responds to the viewing of facial expressions of emotion. However, one important category of facial expression that has not been studied on this level is the facial expression of pain. We investigated the neural response to pain expressions by performing functional magnetic resonance imaging (fMRI) as subjects viewed short video sequences showing faces expressing either moderate pain or, for comparison, no pain. In alternate blocks, the same subjects received both painful and non-painful thermal stimulation. Facial expressions of pain were found to engage cortical areas also engaged by the first-hand experience of pain, including anterior cingulate cortex and insula. The reported findings corroborate other work in which the neural response to witnessed pain has been examined from other perspectives. In addition, they lend support to the idea that common neural substrates are involved in representing one's own and others' affective states.}
}
@article{BotvinicketalPSYCHOLOGICAL_REVIEW-01,
        title = {Conflict monitoring and cognitive control},
       author = {Botvinick, Matthew M. and Carter, Cameron S. and Braver, Todd S. and Carter, Cameron S. and Cohen, Jonathan D.},
      journal = {Psychological Review},
    publisher = {American Psychological Association Inc.},
         year = {2001},
        pages = {624-652},
       volume = {108},
       number = {3},
     abstract = {A neglected question regarding cognitive control is how control processes might detect situations calling for their involvement. The authors propose here that the demand for control may be evaluated in part by monitoring for conflicts in information processing. This hypothesis is supported by data concerning the anterior cingulate cortex, a brain area involved in cognitive control, which also appears to respond to the occurrence of conflict. The present article reports two computational modeling studies, serving to articulate the conflict monitoring hypothesis and examine its implications. The first study tests the sufficiency of the hypothesis to account for brain activation data, applying a measure of conflict to existing models of tasks shown to engage the anterior cingulate. The second study implements a feedback loop connecting conflict monitoring to cognitive control, using this to simulate a number of important behavioral phenomena.},
}
@article{BotvinicketalTiCS-19,
       author = {Botvinick, Matthew and Ritter, Sam and Wang, Jane X. and Kurth-Nelson, Zeb and Blundell, Charles and Hassabis, Demis},
        title = {Reinforcement Learning, Fast and Slow},
      journal = {Trends in Cognitive Sciences},
    publisher = {Elsevier},
         year = {2019},
       volume = {23},
        issue = {5},
        pages = {408-422},
     abstract = {Deep reinforcement learning (RL) methods have driven impressive advances in artificial intelligence in recent years, exceeding human performance in domains ranging from Atari to Go to no-limit poker. This progress has drawn the attention of cognitive scientists interested in understanding human learning. However, the concern has been raised that deep RL may be too sample-inefficient ? that is, it may simply be too slow ? to provide a plausible model of how humans learn. In the present review, we counter this critique by describing recently developed techniques that allow deep RL to operate more nimbly, solving problems much more quickly than previous methods. Although these techniques were developed in an AI context, we propose that they may have rich implications for psychology and neuroscience. A key insight, arising from these AI methods, concerns the fundamental connection between fast RL and slower, more incremental forms of learning.},
}
@article{BouajjanietalSCP-92,
       author = {A. Bouajjani and J.-C. Fernandez and N. Halbwachs and P. Raymond and C. Ratel},
        title = {Minimal state graph generation},
      journal = {Science of Computer Programming},
       volume = 18,
        pages = {247-269},
         year = 1992,
}
@article{BouchardetalCOMPUTER-18,
       author = {Bouchard, Kristofer and Aimone, James and Chun, Miyoung and Dean, Thomas and Donofrio, David and Frank, Loren and Koch, Christof and R{\"{u}}bel, Oliver and Simon, Horst D. and Sommer, Friedrich and Prabhat, Mr},
        title = {International Neuroscience Initiatives Through the Lens of High-Performance Computing},
      journal = {{IEEE} {C}omputer},
       volume = {51},
        issue = {4},
         year = {2018},
        pages = {50-59},
}
@article{BouchardetalNATURE-PHOTONICS-15,
       author = {Bouchard, Matthew B. and Voleti, Venkatakaushik and Mendes, C\'{e}sar S. and Lacefield, Clay and Grueber, Wesley B. and Mann, Richard S. and Bruno, Randy M. and Hillman, Elizabeth M. C.},
        title = {Swept confocally-aligned planar excitation (SCAPE) microscopy for high-speed volumetric imaging of behaving organisms},
      journal = {Nature Photonics},
    publisher = {Nature Publishing Group},
         year = {2015},
       volume = {9},
        issue = {2},
        pages = {113-119},
     abstract = {We report a three-dimensional microscopy technique\emdash{}swept, confocally-aligned planar excitation (SCAPE) microscopy\emdash{}that allows volumetric imaging of living samples at ultrahigh speeds. Although confocal and two-photon microscopy have revolutionized biomedical research, current implementations are costly, complex and limited in their ability to image three-dimensional volumes at high speeds. Light-sheet microscopy techniques using two-objective, orthogonal illumination and detection require a highly constrained sample geometry and either physical sample translation or complex synchronization of illumination and detection planes. In contrast, SCAPE microscopy acquires images using an angled, swept light sheet in a single-objective, en face geometry. Unique confocal descanning and image rotation optics map this moving plane onto a stationary high-speed camera, permitting completely translationless three-dimensional imaging of intact samples at rates exceeding 20 volumes per second. We demonstrate SCAPE microscopy by imaging spontaneous neuronal firing in the intact brain of awake behaving mice, as well as freely moving transgenic Drosophila larvae.},
}
@article{BouchardetalNEURON-16,
       author = {Kristofer E. Bouchard and James B. Aimone and Miyoung Chun and Thomas Dean and Michael Denker and Markus Diesmann and David D. Donofrio and Loren M. Frank and Narayanan Kasthuri and Chirstof Koch and Oliver Ruebel and Horst D. Simon and Friedrich T. Sommer and Prabhat},
        title = {High-Performance Computing in Neuroscience for Data-Driven Discovery, Integration, and Dissemination},
      journal = {Neuron},
       volume = {92},
        pages = {628-631},
         year = {2016},
     abstract = {Opportunities offered by new neuro-technologies are threatened by lack of coherent plans to analyze, manage, and understand the data. High-performance computing will allow exploratory analysis of massive datasets stored in standardized formats, hosted in open repositories, and integrated with simulations.}
}
@article{BoudewijnsetalCIB-11,
        title = {Semi-automated three-dimensional reconstructions of individual neurons reveal cell type-specific circuits in cortex},
       author = {Boudewijns, Zimbo SRM and Kleele, Tatjana and Mansvelder, Huibert D. and Sakmann, Bert and de Kock, Christiaan PJ and Oberlaender, Marcel},
      journal = {Communications Integrative Biology},
    publisher = {Landes Bioscience},
       volume = 4,
        issue = 4,
         year = 2011,
        pages = {486-488},
     abstract = {Despite a long history of anatomical mapping of neuronal networks, we are only beginning to understand the detailed three-dimensional (3D) organization of the cortical micro-circuitry. This is in part due to the lack of complete reconstructions of individual cortical neurons. Morphological studies are either performed on incomplete cells in vitro, or when performed in vivo, lack the necessary cellular resolution. We recently reconstructed the in vivo axonal and dendritic morphology of two types of L(ayer) 5 neurons from vibrissal cortex. The {3D} profiles of short-range as well as longrange projections indicate that L5 slender-tufted and L5 thick-tufted neurons represent very different building blocks of the cortical circuitry. In this addendum to Oberlaender et al. (PNAS 2011), we motivate our technical approach and the advancements this may give in reconstructing the cortical micro-circuitry.},
}
@article{BoulangeretalPAMI-07,
        title = {Space-Time Adaptation for Patch-Based Image Sequence Restoration},
       author = {Boulanger, Jerome and Kervrann, Charles and Bouthemy, Patrick},
      journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
    publisher = {IEEE Computer Society},
       volume = 29,
        issue = 6,
         year = 2007,
        pages = {1096-1102},
}
@inproceedings{BoureauetalCVPR-10,
        title = {Learning Mid-Level Features for Recognition},
       author = {Boureau, {Y-Lan} and Bach, Francis and LeCun, Yann and Ponce, Jean},
    booktitle = {Proceedings of International Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE},
         year = 2010,
}
@inproceedings{BoureauetalICML-10,
        title = {A theoretical analysis of feature pooling in vision algorithms},
       author = {Boureau, {Y-Lan} and Ponce, Jean and LeCun, Yann},
    booktitle = {Proceedings of International Conference on Machine learning},
         year = 2010,
}
@inproceedings{BoutilierandDeardenAAAI-94,
       author = {Boutilier, Craig and Dearden, Richard},
        title = {Using Abstractions for Decision Theoretic Planning with Time Constraints},
    booktitle = {Proceedings AAAI-94},
     location = {Seattle, Washington},
         year = 1994,
        pages = {1016-1022},
}
@inproceedings{BoutilierandDeardenICML-96,
       author = {Boutilier, Craig and Dearden, Richard},
        title = {Approximating Value Trees in Structured Dynamic Programming},
    booktitle = {Proceedings Thirteenth International Conference on Machine Learning},
     location = {Bari, Italy},
        pages = {54-62},
         year = 1996,
}
@inproceedings{BoutilierandPooleAAAI-96,
       author = {Boutilier, Craig and Poole, David},
        title = {Computing Optimal Policies for Partially Observable Decision Processes using Compact Representations},
    booktitle = {Proceedings AAAI-96},
     location = {Portland, Oregon},
 organization = {AAAI},
        pages = {1169-1175},
         year = 1996,
}
@inproceedings{BoutilieretalEWSP-95,
        title = {Planning Under Uncertainty: Structural Assumptions and Computational Leverage},
       author = {Boutilier, Craig and Thomas Dean and Steve Hanks},
    booktitle = {Proceedings of the 3rd European Workshop on Planning},
     location = {Assisi, Italy},
         year = 1995,
}
@inproceedings{BoutilieretalIJCAI-95,
        title = {Exploiting Structure in Policy Construction},
       author = {Boutilier, Craig and Dearden, Richard and Goldszmidt, Moises},
    booktitle = {Proceedings {IJCAI} 14},
     location = {Montreal, Canada},
 organization = {IJCAII},
        pages = {1104-1111},
         year = 1995,
}
@inproceedings{BoutilieretalIJCAI-97,
       author = {Craig Boutilier and Ronen I. Brafman and Christopher Geib},
        title = {Prioritized Goal Decomposition of Markov Decision Processes: Toward a Synthesis of Classical and Decision Theoretic Planning},
    booktitle = {Proceedings {IJCAI} 14},
     location = {Montreal, Canada},
 organization = {IJCAII},
         year = 1997,
        pages = {1156-1162},
}
@article{BoutilieretalJAIR-99,
       author = {Boutilier, Craig and Dean, Thomas and Hanks, Steve},
        title = {Decision Theoretic Planning: Structural Assumptions and Computational Leverage},
      journal = JAIR,
       volume = 11,
        pages = {1-94},
         year = 1999,
}
@inproceedings{BoutilieretalUAI-96,
       author = {Craig Boutilier and Nir Friedman and Moises Goldszmidt and Daphne Koller},
        title = {Context-Specific Independence in {Bayesian} Networks},
        pages = {115-123},
    booktitle = {Proceedings of the Conference on Uncertainty in Artificial Intelligence},
         year = 1996,
}
@article{BoutonetalNATURE-16,
       author = {Bouton, Chad E. and Shaikhouni, Ammar and Annetta, Nicholas V. and Bockbrader, Marcia A. and Friedenberg, David A. and Nielson, Dylan M. and Sharma, Gaurav and Sederberg, Per B. and Glenn, Bradley C. and Mysiw, W. Jerry and Morgan, Austin G. and Deogaonkar, Milind and Rezai, Ali R.},
        title = {Restoring cortical control of functional movement in a human with quadriplegia},
      journal = {Nature},
         year = {2016},
    publisher = {Nature Publishing Group},
       volume = {533},
        issue = {7602},
        pages = {247-250},
     abstract = {Millions of people worldwide suffer from diseases that lead to paralysis through disruption of signal pathways between the brain and the muscles. Neuroprosthetic devices are designed to restore lost function and could be used to form an electronic 'neural bypass' to circumvent disconnected pathways in the nervous system. It has previously been shown that intracortically recorded signals can be decoded to extract information related to motion, allowing non-human primates and paralysed humans to control computers and robotic arms through imagined movements. In non-human primates, these types of signal have also been used to drive activation of chemically paralysed arm muscles. Here we show that intracortically recorded signals can be linked in real-time to muscle activation to restore movement in a paralysed human. We used a chronically implanted intracortical microelectrode array to record multiunit activity from the motor cortex in a study participant with quadriplegia from cervical spinal cord injury. We applied machine- learning algorithms to decode the neuronal activity and control activation of the participant's forearm muscles through a custom- built high-resolution neuromuscular electrical stimulation system. The system provided isolated finger movements and the participant achieved continuous cortical control of six different wrist and hand motions. Furthermore, he was able to use the system to complete functional tasks relevant to daily living. Clinical assessment showed that, when using the system, his motor impairment improved from the fifth to the sixth cervical (C5-C6) to the seventh cervical to first thoracic (C7-T1) level unilaterally, conferring on him the critical abilities to grasp, manipulate, and release objects. This is the first demonstration to our knowledge of successful control of muscle activation using intracortically recorded signals in a paralysed human. These results have significant implications in advancing neuroprosthetic technology for people worldwide living with the effects of paralysis.}
}
@techreport{Bowen81,
       author = {Bowen, D. L. and Byrd, L. and Pereira, L. M. and Pereira, F. C. N. and Warren, D. H. D.},
        title = {PROLOG on the DEC System-10 User's Manual},
  institution = {University of Edinburgh, Department of Artificial Intelligence},
         year = 1981,
}
@article{BowmanCoRR-14,
       author = {Sam Bowman},
        title = {Can recursive neural tensor networks learn logical reasoning?},
      journal = {CoRR},
         year = 2014,
          url = {http://arxiv.org/abs/1312.6192},
     abstract = {Recursive neural network models and their accompanying vector representations for words have seen success in an array of increasingly semantically sophisticated tasks, but almost nothing is known about their ability to accurately capture the aspects of linguistic meaning that are necessary for interpretation or reasoning. To evaluate this, I train a recursive model on a new corpus of constructed examples of logical reasoning in short sentences, like the inference of "some animal walks" from "some dog walks" or "some cat walks," given that dogs and cats are animals. This model learns representations that generalize well to new types of reasoning pattern in all but a few cases, a result which is promising for the ability of learned representation models to capture logical reasoning.}
}
@article{BowmanetalCoRR-14,
       author = {Samuel R. Bowman and Christopher Potts and Christopher D. Manning},
        title = {Recursive Neural Networks for Learning Logical Semantics},
      journal = {CoRR},
       volume = {arXiv:1406.1827},
         year = {2014},
          url = {http://arxiv.org/abs/1406.1827},
     abstract = {Recursive neural networks (RNNs) for sentence meaning have been successful for many tasks, but it remains an open question whether they can learn compositional semantic representations that support logical deduction. We pursue this question by evaluating whether two such models-plain RNNs and recursive neural tensor networks (RNTNs)-can correctly learn to identify logical relationships such as entailment and contradiction. In our first set of experiments, we generate artificial data from a logical grammar and use it to evaluate the models' ability to learn to handle basic relational reasoning, recursive structures, and quantification. We then evaluate the models on the more natural SICK challenge data. Both models perform competitively on the SICK data and generalize well in all three experiments on simulated data, suggesting that they can learn suitable representations for logical inference in natural language.},
}
@article{BowmanetalDS-12,
       author = {Bowman, L. C. and Liu, D. and Meltzoff, A. N. and Wellman, H. M.},
        title = {{{N}eural correlates of belief- and desire-reasoning in 7- and 8-year-old children: an event-related potential study}},
      journal = {Development Science},
       volume = {15},
       number = {5},
         year = {2012},
        pages = {618-632},
     abstract = {Theory of mind requires belief- and desire-understanding. Event-related brain potential (ERP) research on belief- and desire-reasoning in adults found mid-frontal activations for both desires and beliefs, and selective right-posterior activations only for beliefs. Developmentally, children understand desires before beliefs; thus, a critical question concerns whether neural specialization for belief-reasoning exists in childhood or develops later. Neural activity was recorded as 7- and 8-year-olds (N = 18) performed the same diverse-desires, diverse-beliefs, and physical control tasks used in a previous adult ERP study. Like adults, mid-frontal scalp activations were found for belief- and desire-reasoning. Moreover, analyses using correct trials alone yielded selective right-posterior activations for belief-reasoning. Results suggest developmental links between increasingly accurate understanding of complex mental states and neural specialization supporting this understanding.}
}
@article{BoydenBIOLOGY-11,
        title = {A history of optogenetics: the development of tools for controlling brain circuits with light},
       author = {Boyden, Edward},
      journal = {F1000 Biology Reports},
       volume = 3,
         year = 2011
}
@article{BoydenetalNATURE-05,
        title = {Millisecond-timescale, genetically targeted optical control of neural activity},
       author = {Boyden, Edward S and Zhang, Feng and Bamberg, Ernst and Nagel, Georg and Deisseroth, Karl},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = 8,
        issue = 9,
         year = 2005,
        pages = {1263-1268},
}
@inproceedings{BoyenandKollerAAAI-99,
       author = {Boyen, Xavier and Koller, Daphne},
        title = {Exploiting the Architecture of Dynamic Systems},
    booktitle = {Proceedings AAAI-99},
     location = {Orlando, Florida},
 organization = {AAAI},
         year = 1999,
}
@inproceedings{BoyenetalUAI-99,
       author = {Boyen, Xavier and Friedman, Nir and Koller, Daphne},
        title = {Discovering the Hidden Structure of Complex Dynamic Systems},
    booktitle = {Proceedings of the Fifteenth Conference on Uncertainty in Artificial Intelligence},
      address = {Stockholm, Sweden},
         year = 1999,
}
@article{BrachmanAIM-85,
       author = {Brachman, Ronald J.},
        title = {"I lied about the trees" or, Defaults and Definitions in Knowledge Representation},
      journal = {AI Magazine},
       volume = 6,
        issue = 3,
        pages = {80-93},
         year = 1985,
}
@book{Brachmanetal2004,
        title = {Knowledge Representation and Reasoning},
       author = {Brachman, Ronald and Levesque, Hector},
       series = {The Morgan Kaufmann Series in Artificial Intelligence},
         year = {2004},
    publisher = {Elsevier Science}
}
@article{BradskiITJ-98,
       author = {Gary Bradski},
        title = {Computer Vision Face Tracking For Use in a Perceptual User Interface},
      journal = {Intel Technology Journal},
       volume = {Q2},
        pages = 15,
         year = 1998,
}
@article{BradyandFieldPERCEPTION-00,
       author = {Nuala Brady and David Field},
        title = {Local contrast in natural images: {N}ormalisation and coding efficiency},
      journal = {Perception},
       volume = 29,
        issue = 9,
        pages = {1041-1055},
         year = 2000,
}
@article{BradyetalJoV-11,
       author = {Brady, Timothy F. and Konkle, Talia and Alvarez, George A.},
        title = {A review of visual memory capacity: {B}eyond individual items and toward structured representations},
      journal = {Journal of Vision},
       volume = {11},
       number = {5},
        pages = {4-4},
         year = {2011},
     abstract = {bTraditional memory research has focused on identifying separate memory systems and exploring different stages of memory processing. This approach has been valuable for establishing a taxonomy of memory systems and characterizing their function but has been less informative about the nature of stored memory representations. Recent research on visual memory has shifted toward a representation-based emphasis, focusing on the contents of memory and attempting to determine the format and structure of remembered information. The main thesis of this review will be that one cannot fully understand memory systems or memory processes without also determining the nature of memory representations. Nowhere is this connection more obvious than in research that attempts to measure the capacity of visual memory. We will review research on the capacity of visual working memory and visual long-term memory, highlighting recent work that emphasizes the contents of memory. This focus impacts not only how we estimate the capacity of the system-going beyond quantifying how many items can be remembered and moving toward structured representations-but how we model memory systems and memory processes.},
}
@article{BradyetalPNAS-08,
       author = {Brady, Timothy F. and Konkle, Talia and Alvarez, George A. and Oliva, Aude},
        title = {Visual long-term memory has a massive storage capacity for object details},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {105},
       number = {38},
        pages = {14325-14329},
         year = {2008},
     abstract = {One of the major lessons of memory research has been that human memory is fallible, imprecise, and subject to interference. Thus, although observers can remember thousands of images, it is widely assumed that these memories lack detail. Contrary to this assumption, here we show that long-term memory is capable of storing a massive number of objects with details from the image. Participants viewed pictures of 2,500 objects over the course of 5.5 h. Afterward, they were shown pairs of images and indicated which of the two they had seen. The previously viewed item could be paired with either an object from a novel category, an object of the same basic-level category, or the same object in a different state or pose. Performance in each of these conditions was remarkably high (92\%, 88\%, and 87\%, respectively), suggesting that participants successfully maintained detailed representations of thousands of images. These results have implications for cognitive models, in which capacity limitations impose a primary computational constraint (e.g., models of object recognition), and pose a challenge to neural models of memory storage and retrieval, which must be able to account for such a large and detailed storage capacity.},
}
@article{BrafmanandTennenholtzJMLR-02,
       author = {Ronen I. Brafman and Moshe Tennenholtz},
        title = {{R-MAX}: A General Polynomial Time Algorithm for Near-Optimal Reinforcement Learning},
      journal = {Journal of Machine Learning Research},
       volume = 3,
        issue = {Oct},
        pages = {213-231},
         year = 2002,
}
@article{BrainardandDoupeARN-13,
       author = {Brainard, M. S. and Doupe, A. J.},
        title = {Translating birdsong: songbirds as a model for basic and applied medical research},
      journal = {Annual Review of Neuroscience},
       volume = {36},
         year = {2013},
        pages = {489-517},
     abstract = {Songbirds, long of interest to basic neuroscience, have great potential as a model system for translational neuroscience. Songbirds learn their complex vocal behavior in a manner that exemplifies general processes of perceptual and motor skill learning and, more specifically, resembles human speech learning. Song is subserved by circuitry that is specialized for vocal learning and production but that has strong similarities to mammalian brain pathways. The combination of highly quantifiable behavior and discrete neural substrates facilitates understanding links between brain and behavior, both in normal states and in disease. Here we highlight (a) behavioral and mechanistic parallels between birdsong and aspects of speech and social communication, including insights into mirror neurons, the function of auditory feedback, and genes underlying social communication disorders, and (b) contributions of songbirds to understanding cortical-basal ganglia circuit function and dysfunction, including the possibility of harnessing adult neurogenesis for brain repair.}
}
@article{BrainardandDoupeNATURE-02,
       author = {Brainard, M. S. and Doupe, A. J.},
        title = {What songbirds teach us about learning},
      journal = {Nature},
       volume = {417},
       number = {6886},
         year = {2002},
        pages = {351-358},
     abstract = {Bird fanciers have known for centuries that songbirds learn their songs. This learning has striking parallels to speech acquisition: like humans, birds must hear the sounds of adults during a sensitive period, and must hear their own voice while learning to vocalize. With the discovery and investigation of discrete brain structures required for singing, songbirds are now providing insights into neural mechanisms of learning. Aided by a wealth of behavioural observations and species diversity, studies in songbirds are addressing such basic issues in neuroscience as perceptual and sensorimotor learning, developmental regulation of plasticity, and the control and function of adult neurogenesis.}
}
@article{BrainardandFitchCOIN-14,
       author = {Michael S. Brainard and W. Tecumseh Fitch},
        title = {Editorial overview: Communication and language: Animal communication and human language },
      journal = {Current Opinion in Neurobiology},
       volume = {28},
         year = {2014},
        pages = {5-8},
      comment = {A broad overview of acoustic communication in animals and spoken language in humans. Including the perspective of an ethnologist, Niko Tinbergen, spanning questions from the mechanistic to the evolutionary. For example the authors review in detail the capacity for complex vocal learning (the ability to reproduce sounds heard in the environment that are not part of an inborn species repertoire). The capacity for vocal learning is a fundamental prerequisite for our ability to acquire a large and flexible vocabulary shared among a speech community, and although highly developed in humans, appears to be absent or extremely limited in nonhuman primates. Nonetheless this capacity has evolved convergently in many other vertebrates including a number of mammalian clades along with songbirds (where it is best understood from a neuroscientific viewpoint).},
}
@inproceedings{Braitenberg1978cellassemblies,
       author = {Braitenberg, Valentino},
       editor = {Heim, Roland and Palm, G{\"{u}}nther},
        title = {Cell Assemblies in the Cerebral Cortex},
    booktitle = {Theoretical Approaches to Complex Systems},
         year = {1978},
    publisher = {Springer Berlin Heidelberg},
      address = {Berlin, Heidelberg},
        pages = {171-188},
     abstract = {To say that an animal responds to sensory stimuli may not be the most natural and efficient way to describe behaviour. Rather, it appears that animals most of the time react to situations, to opponents or things which they actively isolate from their environment, Situations, things, partners or opponents are, in a way, the terms of behaviour. It is legitimate, therefore, to ask what phenomena correspond to them in the internal activity of the brain, or, in other words: how are the meaningful chunks of experience "represented" in the brain?},
}
@book{Braitenberg1986,
       author = {Valentino Braitenberg},
        title = {Vehicles: Experiments in Synthetic Psychology},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1986,
}
@book{BraitenbergandSchuz1991,
       author = {Braitenberg, V. and Schuz, A.},
        title = {Anatomy of the Cortex},
      address = {Berlin},
    publisher = {Springer-Verlag},
         year = 1991,
}
@book{BraitenbergandSchuz1998,
   author = {Braitenberg, V. and Schuz, A.},
    title = {Cortex: Statistics and Geometry of Neuronal Connections},
  address = {Berlin},
publisher = {Springer-Verlag},
     year = 1991,
}
@article{BrakkeandSavage-RumbaughLC-95,
        title = {The development of language skills in bonobo and chimpanzee-{I. Comprehension}},
       author = {Karen E. Brakke and E. Sue Savage-Rumbaugh},
      journal = {Language \& Communication},
       volume = {15},
       number = {2},
         year = {1995},
        pages = {121-148},
}
@article{BrakkeandSavage-RumbaughLC-96,
        title = {The development of language skills in Pan-{II. Production}},
       author = {Karen E. Brakke and E. Sue Savage-Rumbaugh},
      journal = {Language \& Communication},
       volume = {16},
       number = {4},
         year = {1996},
        pages = {361-380},
}
@book{BrammerandSiffling89,
       author = {Brammer, Karl and Siffling, Gerhard},
        title = {Kalman-Bucy Filters},
    publisher = {Artech House},
      address = {Norwood, Massachusetts},
         year = 1989,
}
@article{BrancoandStarasNATURE-09,
       author = {Branco, Tiago and Staras, Kevin},
        title = {The probability of neurotransmitter release: variability and feedback control at single synapses},
      journal = {Nature Reviews Neuroscience},
         year = 2009,
    publisher = {Nature Publishing Group},
       volume = 10,
        issue = 5,
        pages = {373-383},
     abstract = {Information transfer at chemical synapses occurs when vesicles fuse with the plasma membrane and release neurotransmitter. This process is stochastic and its likelihood of occurrence is a crucial factor in the regulation of signal propagation in neuronal networks. The reliability of neurotransmitter release can be highly variable: experimental data from electrophysiological, molecular and imaging studies have demonstrated that synaptic terminals can individually set their neurotransmitter release probability dynamically through local feedback regulation. This local tuning of transmission has important implications for current models of single-neuron computation.}
}
@incollection{BrandandMarkowitschHBN-08,
       author = {Matthias Brand and Hans J. Markowitsch},
       editor = {Dere, E. and Easton, A. and Nadel, L. and Huston, J.P.},
        title = {Chapter 3.5 The role of the prefrontal cortex in episodic memory},
    booktitle = {Handbook of Behavioral Neuroscience},
    publisher = {Elsevier Science},
         year = {2008},
        pages = {332-356}
}
@inproceedings{BrandetalCVPR-96,
       author = {M. Brand and N. Oliver and A. Pentland},
        title = {Coupled hidden {M}arkov models for complex action recognition},
    booktitle = {Proceedings of {IEEE} Computer Vision and Pattern Recognition},
        pages = {994-999},
         year = 1996,
}
@article{BrandetalPsyArXiv20,
        title = {Analogy as a catalyst for cumulative cultural evolution},
       author = {Brand, Charlotte O. and Mesoudi, Alex and Smaldino, Paul E.},
          url = {psyarxiv.com/ynkqf},
      journal = {PsyArXiv},
         year = {2020},
     abstract = {'Analogies', broadly defined, map novel concepts onto familiar concepts, making them essential for perception, reasoning, and communication. We argue that analogy-building served a critical role in the evolution of cumulative culture, by allowing humans to learn and transmit complex behavioural sequences that would otherwise be too cognitively demanding or opaque to acquire. The emergence of a protolanguage consisting of simple labels would have provided early humans with the cognitive tools to build explicit analogies and to communicate them to others. This focus on analogy-building can shed new light on the coevolution of cognition and culture, and addresses recent calls for better integration of the field of cultural evolution with cognitive science.}
}
@article{BransonandFreemanCELL-16,
       author = {Branson, Kristin and Freeman, Jeremy},
        title = {Imaging the Neural Basis of Locomotion},
      journal = {Cell},
         year = {2016},
    publisher = {Elsevier},
       volume = {163},
        issue = {3},
        pages = {541-542},
     abstract = {To investigate the fundamental question of how nervous systems encode, organize, and sequence behaviors, Kato et al. imaged neural activity with cellular resolution across the brain of the worm Caenorhabditis elegans. Locomotion behavior seems to be continuously represented by cyclical patterns of distributed neural activity that are present even in immobilized animals.},
}
@article{BrashPHOTOBIOLOGY-15,
       author = {Brash, Douglas E.},
        title = {{UV} Signature Mutations},
      journal = {Photochemical Photobiology},
         year = {2015},
       volume = {91},
        issue = {1},
        pages = {15-26},
     abstract = {Sequencing complete tumor genomes and exomes has sparked the cancer field's interest in mutation signatures for identifying the tumor's carcinogen. This review and meta-analysis discusses signatures and their proper use. We first distinguish between a mutagen's canonical mutations - deviations from a random distribution of base changes to create a pattern typical of that mutagen - and the subset of signature mutations, which are unique to that mutagen and permit inference backward from mutations to mutagen. To verify UV signature mutations, we assembled literature datasets on cells exposed to UVC, UVB, UVA, or solar simulator light (SSL) and tested canonical UV mutation features as criteria for clustering datasets. A confirmed UV signature was: 60\% of mutations are C-T at a dipyrimidine site, with 5\% CC-TT. Other canonical features such as a bias for mutations on the non-transcribed strand or at the 3' pyrimidine had limited application. The most robust classifier combined these features with criteria for the rarity of non-UV canonical mutations. In addition, several signatures proposed for specific UV wavelengths were limited to specific genes or species; non-signature mutations induced by UV may cause melanoma BRAF mutations; and the mutagen for sunlight-related skin neoplasms may vary between continents.},
}
@book{Bratko86,
       author = {Bratko, Ivan},
        title = {Prolog Programming for Artificial Intelligence},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1986,
}
@article{Bratman84,
       author = {Bratman, M.},
        title = {Two Faces of Intention},
      journal = {The Philosophical Review},
       volume = 93,
         year = 1984,
        pages = {375-405},
}
@incollection{BrayandMartinezNIPS-02,
       author = {A. Bray and D. Martinez},
        title = {Kernel-based extraction of slow features: Complex cells learn disparity and translation invariance from natural images},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 14,
      address = {Cambridge, MA},
    publisher = {MIT Press},
         year = 2002,
}
@article{BrebissonandMontanaCoRR-15,
       author = {Alexandre de Br{\'{e}}bisson and Giovanni Montana},
        title = {Deep Neural Networks for Anatomical Brain Segmentation},
      journal = {CoRR},
       volume = {arXiv:1502.02445},
         year = {2015},
          url = {http://arxiv.org/abs/1502.02445},
     abstract = {We present a novel approach to automatically segment magnetic resonance (MR) images of the human brain into anatomical regions. Our methodology is based on a deep artificial neural network that assigns each voxel in an MR image of the brain to its corresponding anatomical region. The inputs of the network capture information at different scales around the voxel of interest: 3D and orthogonal 2D intensity patches capture the local spatial context while large, compressed 2D orthogonal patches and distances to the regional centroids enforce global spatial consistency. Contrary to commonly used segmentation methods, our technique does not require any non-linear registration of the MR images. To benchmark our model, we used the dataset provided for the MICCAI 2012 challenge on multi-atlas labelling, which consists of 35 manually segmented MR images of the brain. We obtained competitive results (mean dice coefficient 0.725, error rate 0.163) showing the potential of our approach. To our knowledge, our technique is the first to tackle the anatomical segmentation of the whole brain using deep neural networks.}
}
@techreport{BreeseTR-87,
       author = {Breese, John S.},
        title = {Knowledge Representation and Inference in Intelligent Decision Systems},
        issue = 2,
  institution = {Rockwell International Science Center},
         year = 1987,
}
@inproceedings{BreeseandFehlingUAI-88,
       author = {Breese, John S. and Fehling, Michael R.},
        title = {Decision-Theoretic Control of Problem Solving: Principles and Architecture},
    booktitle = UAI-88,
     location = {Minneapolis, Minnesota},
        pages = {30-37},
         year = 1988,
}
@inproceedings{BreeseetalUAI-90,
       author = {Breese, John S. and Fertig, Kenneth W.},
        title = {Decision making with interval influence diagrams},
    booktitle = {Sixth Conference on Uncertainty in Artificial Intelligence},
     location = {Cambridge, Massachusetts},
        pages = {122-129},
         year = 1990,
}
@inproceedings{BregonzioetalCVPR-09,
       author = {Matteo Bregonzio and Shaogang Gong and Tao Xiang},
        title = {Recognising Action as Clouds of Space-Time Interest Points},
    booktitle = {{IEEE} Computer Society Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 2009,
}
@book{Breimanetal84,
       author = {Breiman, L. and Friedman, J. H. and Olshen, R. A. and Stone, C. J.},
        title = {Classification and Regression Trees},
    publisher = {Wadsworth and Brooks},
      address = {Monterey, California},
         year = 1984,
}
@article{BreuleuxetalJMLR-10,
     author = {Breuleux, O. and Bergstra, J. and Turian, J. and Bastien, F. and Lamblin, P. and Desjardins, G. and Pascanu, R. and Delalleau, O. and Bengio, Y.},
      title = {Theano: A package for efficient computation in Python},
    journal = {Journal of Machine Learning Research},
      pages = {(submitted)},
       year = 2010,
}
@article{BrezinaPTRS-B-10,
       author = {Brezina, Vladimir},
        title = {Beyond the wiring diagram: signalling through complex neuromodulator networks},
      journal = {Philosophical transactions of the Royal Society of London. Series B, Biological sciences},
    publisher = {The Royal Society},
       volume = {365},
        issue = {1551},
         year = {2010},
        pages = {2363-2374},
     abstract = {During the computations performed by the nervous system, its 'wiring diagram'--the map of its neurons and synaptic connections--is dynamically modified and supplemented by multiple actions of neuromodulators that can be so complex that they can be thought of as constituting a biochemical network that combines with the neuronal network to perform the computation. Thus, the neuronal wiring diagram alone is not sufficient to specify, and permit us to understand, the computation that underlies behaviour. Here I review how such modulatory networks operate, the problems that their existence poses for the experimental study and conceptual understanding of the computations performed by the nervous system, and how these problems may perhaps be solved and the computations understood by considering the structural and functional 'logic' of the modulatory networks.},
}
@article{BridgeandClarePTRS-B-06,
       author = {Bridge, Holly and Clare, Stuart},
        title = {High-resolution {MRI}: {\em{in vivo}} histology?},
      journal = {Philosophical transactions of the Royal Society of London. Series B, Biological sciences},
    publisher = {The Royal Society},
       volume = {361},
        issue = {1465},
         year = {2006},
        pages = {137-146},
     abstract = {For centuries scientists have been fascinated with the question of how the brain works. Investigators have looked at both where different functions are localized and how the anatomical microstructure varies across the brain surface. Here we discuss how advances in magnetic resonance imaging (MRI) have allowed in vivo visualization of the fine structure of the brain that was previously only visible in post-mortem brains. We present data showing the correspondence between definitions of the primary visual cortex defined anatomically using very high-resolution MRI and functionally using functional MRI. We consider how this technology can be applied to allow the investigation of brains that differ from normal, and what this ever-evolving technology may be able to reveal about in vivo brain structure in the next few years.},
}
@inproceedings{BridsonSIGGRAPH-07,
       author = {Bridson, Robert},
        title = {Fast Poisson Disk Sampling in Arbitrary Dimensions},
    booktitle = {{ACM SIGGRAPH}},
     location = {San Diego, California},
    publisher = {ACM},
      address = {New York, NY, USA},
         year = 2007,
}
@incollection{BrigardDENNETT-17,
        title = {Memory and the Intentional Stance},
       author = {Felipe De Brigard},
    booktitle = {The Philosophy of Daniel Dennett},
       editor = {Huebner, Bryce},
    publisher = {Oxford University Press},
      address = {New York, NY},
         year = {2017},
        pages = {92-94}
}
@incollection{BrigardMTT-16,
        title = {Time Is Not of the Essence: {U}nderstanding the Neural Correlates of Mental Time Travel},
       author = {Felipe De Brigard and Bryce S. Gessell},
    booktitle = {Seeing the Future: {T}heoretical Perspectives on Future-Oriented Mental Time Travel},
       editor = {Kourken Michaelian and Stanley B. Klein and Karl K. Szpunar},
    publisher = {Oxford Scholarship Online},
      address = {New York, NY},
         year = {2016},
        pages = {133-175},
     abstract = {Recent research in neuropsychology and cognitive neuroscience has shown that episodic memory and future thinking engage a common core brain network. Some have interpreted these results as showing that episodic memory and future thinking are two operations of a single cognitive system for mental time travel. However, some recent results cast doubt upon this tensed-content view, as they strongly suggest that a temporal content may not be essential for the engagement of this core brain network. This chapter reviews evidence both in favor of and against the tensed-content view, and suggests instead a different approach that can accommodate these discrepant results: the dynamic-structure view. According to this view, what explains the engagement of the core brain network is not that the mental representations are about time but rather that they unfold in time, i.e. that they have a dynamic structure.}
}
@article{BriggmanandBockCON-12,
       author = {Briggman, K.L. and Bock, D.D.},
        title = {Current Opinion Neurobiology},
      journal = {Volume electron microscopy for neuronal circuit reconstruction},
       volume = 22,
        issue = 1,
         year = 2012,
        pages = {154-61},
}
@article{BriggmanetalNATURE-11,
       author = {Briggman, K.L. and Helmstaedter, M. and Denk, W.},
        title = {Wiring specificity in the direction-selectivity circuit of the retina},
      journal = {Nature},
       volume = 471,
        issue = 7337,
         year = 2011,
        pages = {183-188},
}
@inproceedings{BrinandPageWWW-98,
       author = {Brin, Sergey and Page, Larry},
        title = {The anatomy of a large-scale hypertextual {W}eb search engine},
    booktitle = {Proceedings of the 7th World Wide Web Conference},
        pages = {},
         year = 1998,
          url = {http://www7.scu.edu.au/programme/fullpapers/1921/com1921.htm},
}
@article{BrockmanetalCoRR-16,
       author = {Greg Brockman and Vicki Cheung and Ludwig Pettersson and Jonas Schneider and John Schulman and Jie Tang and Wojciech Zaremba},
        title = {{OpenAI Gym}},
      journal = {CoRR},
       volume = {arXiv:1606.01540},
         year = {2016},
     abstract = {OpenAI Gym is a toolkit for reinforcement learning research. It includes a growing collection of benchmark problems that expose a common interface, and a website where people can share their results and compare the performance of algorithms. This whitepaper discusses the components of OpenAI Gym and the design decisions that went into the software.}
}
@inproceedings{Broderetal00,
    author = {A. Broder and R. Krauthgamer and M. Mitzenmacher},
     title = {Improved classification via connectivity information},
 booktitle = {Proceedings of the 11th ACM-SIAM Symposium on Discrete Algorithms},
     pages = {576-585},
      year = 2000,
}
@inproceedings{Broderetal00b,
    author = {A. Broder and R. Kumar and F. Maghoul and P. Raghavan and S. Rajagopalan and R. Stata and A. Tomkins and J. Wiener},
     title = {Graph structure in the {W}eb: experiments and models},
 booktitle = {Proceedings of the 9th World Wide Web Conference},
      year = 2000,
       url = {http://www9.org/w9cdrom/index.html},
}
@article{Broderetal98,
   author = {A.Z. Broder and A.M. Frieze and S. Suen and E. Upfal},
    title = {Optimal construction of edge-disjoint paths in random graphs},
  journal = {SIAM Journal on Computing},
   volume = 28,
    pages = {541-573},
     year = 1998,
}
@inproceedings{Broderetal99,
     author = {A. Z. Broder and S. R. Kumar and F. Maghoul and P. Raghavan and S. Rajagopalan and R. Stata and A. Tomkins and J. Wiener},
      title = {Graph structure of the {W}eb: experiments and models},
  booktitle = {Proceedings of the Ninth International
               World Wide Web Conference},
       year = 1999,
}
@article{BroderetalCN-00,
    author = {Broder, Andrei and Kumar, Ravi and Maghoul, Farzin and Raghavan, Prabhakar and Rajagopalan, Sridhar and Stata, Raymie and Tomkins, Andrew and Wiener, Janet},
     title = {Graph Structure in the Web},
   journal = {Computer Networks},
    volume = 33,
     issue = {1-6},
      year = 2000,
     pages = {309-320},
}
@inproceedings{BroderetalSTOC-98,
       author = {Broder, Andrei Z. and Charikar, Moses and Frieze, Alan M. and Mitzenmacher, Michael},
        title = {Min-wise Independent Permutations (Extended Abstract)},
    booktitle = {Proceedings of the Thirtieth Annual ACM Symposium on Theory of Computing},
    publisher = {ACM},
      address = {New York, NY, USA},
         year = {1998},
        pages = {327-336},
}
@incollection{BrodieetalICAC-05,
       author = {M. Brodie and Sheng Ma and G. Lohman and L. Mignet and N. Modani and M. Wilding and J. Champlin and P. Sohn},
        title = {Quickly Finding Known Software Problems via Automated Symptom Matching},
    booktitle = {Proceedings of the Second International Conference on Autonomic Computing},
    publisher = {ACM},
      address = {New York, NY, USA},
         year = {2005},
     abstract = {We present an architecture for and prototype of a system for quickly detecting software problem recurrences. Re-discovery of the same problem is very common in many large software products and is a major cost component of product support. At run-time, when a problem occurs, the system collects the problem symptoms, including the program call-stack, and compares it against a database of symptoms to find the closest matches. The database is populated off-line using solved cases and indexed to allow for efficient matching. Thus problems that occur repeatedly can be easily and automatically resolved without requiring any human problem-solving expertise. We describe a prototype implementation of the system, including the matching algorithm, and present some experimental results demonstrating the value of automatically detecting re-occurrence of the same problem for a popular sofware product}
}
@book{Brodmann1909,
        title = {{Vergleichende Lokalisationslehre der Grosshirnrinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues}},
       author = {Brodmann, Korbinian},
    publisher = {Johann Ambrosius Barth Verlag},
      address = {Leipzig},
         year = 1909,
}
@misc{BronsteinSCALABLE-GRAPH-NETWORKS-20,
        title = {Simple scalable graph neural networks},
       author = {Bronstein, Michael and Fabrizo Frasca and Emanuele Rossi},
 howpublished = {\url{https://towardsdatascience.com/simple-scalable-graph-neural-networks-7eb04f366d07}},
         year = {2020},
}
@book{Brooks02,
       author = {Rodney Brooks},
        title = {Flesh and Machines: How Robots Will Change Us},
    publisher = {Pantheon Books},
      address = {New York, NY},
         year = 2002,
}
@inproceedings{Brooks84,
       author = {Brooks, Rodney A.},
        title = {Aspects of Mobile Robot Visual Map Making},
    booktitle = {Second International Symposium on Robotics Research},
       editor = {H. Hanafusa and H. Inoue},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1984,
        pages = {325-331},
}
@book{Brooks95,
       author = {Frederick Brooks},
        title = {The Mythical Man-Month: Essays on Software Engineering},
    publisher = {Addison Wesley},
      address = {Reading, MA},
         year = 1995,
}
@article{BrooksAIJ-81,
       author = {Brooks, Rodney A.},
        title = {Symbolic Reasoning Among 3-{D} Models and 2-{D} Images},
      journal = {Artificial Intelligence},
      comment = {ACRONYM},
       volume = 17,
         year = 1981,
        pages = {285-348},
}
@inproceedings{BrooksIJCAI-91,
       author = {Brooks, Rodney A.},
        title = {Intelligence Without Reason},
    booktitle = {Proceedings of the 12th International Joint Conference on Artificial Intelligence - Volume 1},
    publisher = {Morgan Kaufmann Publishers Inc.},
      address = {San Francisco, CA, USA},
     location = {Sydney, New South Wales, Australia},
         year = {1991},
        pages = {569-595},
     abstract = {Computers and Thought are the two categories that together define Artificial Intelligence as a discipline. It is generally accepted that work in Artificial Intelligence over the last thirty years has had a strong influence on aspects of computer architectures. In this paper we also make the converse claim; that the state of computer architecture has been a strong influence on our models of thought. The Von Neumann model of computation has lead Artificial Intelligence in particular directions. Intelligence in biological systems is completely different. Recent work in behavior-based Artificial Intelligence has produced new models of intelligence that are much closer in spirit to biological systems. The non-Von Neumann computational models they use share many characteristics with biological computation.}
}
@article{BrooksRA-86,
       author = {Brooks, Rodney A.},
        title = {A Robust Layered Control System for a Mobile Robot},
      journal = {{IEEE} Journal of Robotics and Automation},
       volume = 2,
         year = 1986,
        pages = {14-23},
}
@book{BrownandDaguid,
       author = {John Seely Brown and Paul Daguid},
        title = {The Social Life of Information},
    publisher = {Harvard Business School},
         year = 2000,
}
@inproceedings{BrownandLoweBMVC-02,
       author = {Matthew Brown and David Lowe},
        title = {Invariant features from interest point groups},
    booktitle = {Proceedings of the {B}ritish Machine Vision Conference, ({BMVC})},
        pages = {253-262},
         year = 2002,
}
@article{BrownandMilnerNATURE-03,
       author = {Brown, Richard E. and Milner, Peter M.},
        title = {The legacy of Donald O. Hebb: more than the Hebb Synapse},
      journal = {Nature Reviews Neuroscience},
         year = {2003},
       volume = {4},
        issue = {12},
        pages = {1013-1019},
     abstract = {Neuroscientists associate the name of Donald O. Hebb with the Hebbian synapse and the Hebbian learning rule, which underlie connectionist theories and synaptic plasticity, but Hebb's work has also influenced developmental psychology, neuropsychology, perception and the study of emotions, as well as learning and memory. Here, we review the work of Hebb and its lasting influence on neuroscience in honour of the 2004 centenary of his birth.},
}
@inproceedings{BrownandNajorkACM-93,
       author = {Brown, Marc H. and Najork, Marc A.},
        title = {Algorithm Animation Using 3D Interactive Graphics},
    publisher = {Association for Computing Machinery},
    booktitle = {Proceedings of the 6th Annual ACM Symposium on User Interface Software and Technology},
         year = {1993},
        pages = {93-100},
}
@article{BrownandNeherNATURE-REVIEWS-NEUROSCIENCE-14,
       author = {Brown, Guy C. and Neher, Jonas J.},
        title = {Microglial phagocytosis of live neurons},
      journal = {Nature Reviews Neuroscience},
         year = {2014},
    publisher = {Nature Publishing Group},
       volume = {15},
        issue = {4},
        pages = {209-216},
     abstract = {Microglia, the brain's professional phagocytes, can remove dead and dying neurons as well as synapses and the processes of live neurons. However, we and others have recently shown that microglia can also execute neuronal death by phagocytosing stressed-but-viable neurons [mdash] a process that we have termed phagoptosis. In this Progress article, we discuss evidence suggesting that phagoptosis may contribute to neuronal loss during brain development, inflammation, ischaemia and neurodegeneration.},
}
@article{BrownandSedgewick84,
       author = {Brown, Marc H. and Robert Sedgewick},
        title = {A System for Algorithm Animation},
      journal = {Computer Graphics},
       volume = 18,
         year = 1984,
        pages = {177-186},
}
@article{BrownetalIET-15,
       author = {A. D. Brown and R. Mills and K. J. Dugan and J. S. Reeve and S. B. Furber},
      journal = {{IET} Computers Digital Techniques},
        title = {Reliable computation with unreliable computers},
         year = {2015},
       volume = {9},
       number = {4},
        pages = {230-237},
     abstract = {As computing systems continue their unquenchable rise towards and through million core architectures, two considerations that used to be unimportant become more and more dominant: power consumption (be it FLOPS/W or W/mm2) and reliability. This study is concerned with the latter: in a system of a million cores, it is unrealistic to expect 100\% functionality on power-up; equally, operational availability degrades with time. Monitoring and maintaining the health of such a system using traditional techniques is costly, and most rely on the concept of some sort of central overseer or monitor to make a final judgement about system availability, giving a single point of failure. Large systems of the future will consist of hardware and software that work synergistically to cope with isolated points of failure, allowing the gross behaviour of the system to degrade gracefully and in a meaningful way in the face of faults. This study describes one such system: spiking neural network architecture is a million-core machine with layered fault-tolerance built in at many levels. The authors show how the system may be used to solve the canonical distributed heat diffusion equation, and how the quality of solution is modulated by the effects of partial system failure.},
}
@inproceedings{BroxandMalikECCV-10,
      title = {Object segmentation by long term analysis of point trajectories},
     author = {Brox, Thomas and Malik, Jitendra},
  booktitle = {Proceedings of the 11th European Conference on Computer Vision},
   location = {Heraklion, Crete, Greece},
  publisher = {Springer-Verlag},
    address = {Berlin, Heidelberg},
       year = 2010,
      pages = {282-295},
}
@article{BruceAIJ-72,
       author = {Bruce, B.},
        title = {A model for temporal reference and its application in a question answering program},
      journal = {Artificial Intelligence},
       volume = 3,
         year = 1972,
        pages = {1-25},
}
@article{BubenikJMLR-15,
       author = {Peter Bubenik},
        title = {Statistical topological data analysis using persistence landscapes},
      journal = {Journal of Machine Learning Research},
       volume = {11},
       number = {1},
         year = {2015},
        pages = {77-102},
     abstract = {We define a new topological summary for data that we call the persistence landscape. Since this summary lies in a vector space, it is easy to combine with tools from statistics and machine learning, in contrast to the standard topological summaries. Viewed as a random variable with values in a Banach space, this summary obeys a strong law of large numbers and a central limit theorem. We show how a number of standard statistical tests can be used for statistical inference using this summary. We also prove that this summary is stable and that it can be used to provide lower bounds for the bottleneck and Wasserstein distances.}
}
@article{BubenikandDlotkoCoRR-15,
       author = {Peter Bubenik and Pawel Dlotko},
        title = {A persistence landscapes toolbox for topological statistics},
      journal = {CoRR},
       volume = {arXiv:1501.00179},
         year = {2015},
     abstract = {Topological data analysis provides a multiscale description of the geometry and topology of quantitative data. The persistence landscape is a topological summary that can be easily combined with tools from statistics and machine learning. We give efficient algorithms for calculating persistence landscapes, their averages, and distances between such averages. We discuss an implementation of these algorithms and some related procedures. These are intended to facilitate the combination of statistics and machine learning with topological data analysis. We present an experiment showing that the low-dimensional persistence landscapes of points sampled from spheres (and boxes) of varying dimensions differ.}
}
@misc{BubenikandKim2003,
       author = {Peter Bubenik and Peter T. Kim},
        title = {A Statistical Approach to Persistent Homology},
         year = {2006},
     abstract = {Assume that a finite set of points is randomly sampled from a subspace of a metric space. Recent advances in computational topology have provided several approaches to recovering the geometric and topological properties of the underlying space. In this paper we take a statistical approach to this problem. We assume that the data is randomly sampled from an unknown probability distribution. We define two filtered complexes with which we can calculate the persistent homology of a probability distribution. Using statistical estimators for samples from certain families of distributions, we show that we can recover the persistent homology of the underlying distribution.}
}
@article{BucketalICLR-18,
       author = {Christian Buck and Jannis Bulian and Massimiliano Ciaramita and Andrea Gesmundo and Neil Houlsby and Wojciech Gajewski and Wei Wang},
        title = {Ask the Right Questions: Active Question Reformulation with Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1705.07830},
         year = {2018},
     abstract = {We frame Question Answering (QA) as a Reinforcement Learning task, an approach that we call Active Question Answering. We propose an agent that sits between the user and a black box QA system and learns to reformulate questions to elicit the best possible answers. The agent probes the system with, potentially many, natural language reformulations of an initial question and aggregates the returned evidence to yield the best answer. The reformulation system is trained end-to-end to maximize answer quality using policy gradient. We evaluate on SearchQA, a dataset of complex questions extracted from Jeopardy!. The agent outperforms a state-of-the-art base model, playing the role of the environment, and other benchmarks. We also analyze the language that the agent has learned while interacting with the question answering system. We find that successful question reformulations look quite different from natural language paraphrases. The agent is able to discover non-trivial reformulation strategies that resemble classic information retrieval techniques such as term re-weighting (tf-idf) and stemming.}
}
@article{BucknerNEURON-13,
       author = {Buckner, R. L.},
        title = {The cerebellum and cognitive function: 25 years of insight from anatomy and neuroimaging},
      journal = {Neuron},
       volume = {80},
       number = {3},
         year = {2013},
        pages = {807-815},
     abstract = {Twenty-five years ago the first human functional neuroimaging studies of cognition discovered a surprising response in the cerebellum that could not be attributed to motor demands. This controversial observation challenged the well-entrenched view that the cerebellum solely contributes to the planning and execution of movement. Recurring neuroimaging findings combined with key insights from anatomy and case studies of neurological patients motivated a reconsideration of the traditional model of cerebellar organization and function. The majority of the human cerebellum maps to cerebral association networks in an orderly manner that includes a mirroring of the prominent cerebral asymmetries for language and attention. These findings inspire exploration of the cerebellum's contributions to a diverse array of functional domains and neuropsychiatric disorders.}
}

@incollection{BucknerandGarsonCONNECTIONISM-19,
       author = {Buckner, Cameron and Garson, James},
        title = {Connectionism},
    booktitle = {The Stanford Encyclopedia of Philosophy},
       editor = {Edward N. Zalta },
    publisher = {Stanford University},
         year = {2019},
          url = {https://plato.stanford.edu/archives/fall2019/entries/connectionism/},
}
@article{BucknerandKrienenTiCS-13,
     title = {The evolution of distributed association networks in the human brain},
   journal = {Trends in Cognitive Sciences},
    volume = 17,
     issue = 12,
      year = 2013,
     pages = {648-665},
    author = {Randy L. Buckner and Fenna M. Krienen},
  abstract = {The human cerebral cortex is vastly expanded relative to other primates and disproportionately occupied by distributed association regions. Here we offer a hypothesis about how association networks evolved their prominence and came to possess circuit properties vital to human cognition. The rapid expansion of the cortical mantle may have untethered large portions of the cortex from strong constraints of molecular gradients and early activity cascades that lead to sensory hierarchies. What fill the gaps between these hierarchies are densely interconnected networks that widely span the cortex and mature late into development. Limitations of the tethering hypothesis are discussed as well as its broad implications for understanding critical features of the human brain as a byproduct of size scaling.}
}
@book{Budd00,
      author = {Timothy Budd},
       title = {Understanding Object-Oriented Programming with Java},
   publisher = {Addison-Wesley},
     address = {Reading, MA},
        year = 2000,
}
@book{Budd01,
      author = {Timothy Budd},
       title = {Introduction to Object-Oriented Programming},
   publisher = {Third Edition, Addison-Wesley},
     address = {Reading, MA},
        year = 2001,
}
@book{Budd87,
      author = {Timothy Budd},
       title = {A Little Smalltalk},
   publisher = {Addison-Wesley},
     address = {Reading, MA},
        year = 1987,
}
@article{BuddPTRS_B-98,
       author = {Julian M.L. Budd},
        title = {Extrastriate feedback to primary visual cortex in primates: {A} quantitative analysis of connectivity},
      journal = {Proceedings of the Royal Society of London. Series B: Biological Sciences},
       volume = {265},
       number = {1400},
         year = {1998},
        pages = {1037-1044},
     abstract = {Knowledge-based or top-down influences on primary visual cortex (area V1) are believed to originate from information conveyed by extrastriate feedback axon connections. Understanding how this information is communicated to area V1 neurons relies in part on elucidating the quantitative as well as the qualitative nature of extrastriate pathway connectivity. A quantitative analysis of the connectivity based on anatomical data regarding the feedback pathway from extrastriate area V2 to area V1 in macaque monkey suggests: (i) a total of around ten million or more area V2 axons project to area V1; (ii) the mean number of synaptic inputs from area V2 per upper-layer pyramidal cell in area V1 is less than 6\% of all excitatory inputs; and (iii) the mean degree of convergence of area V2 afferents may be high, perhaps more than a hundred afferent axons per cell. These results are consistent with empirical observations of the density of radial myelinated axons present in the upper layers in macaque area V1 and the proportion of excitatory extrastriate feedback synaptic inputs onto upper-layer neurons in rat visual cortex. Thus in primate area V1, extrastriate feedback synapses onto upper-layer cells may, like geniculocortical afferent synapses onto layer IVC neurons, form only a small percentage of the total excitatory synaptic input. }
}
@article{BudzianowskiCoRR-17,
       author = {Pawel Budzianowski and Stefan Ultes and Pei-Hao Su and Nikola Mrksic and Tsung-Hsien Wen and I\~{n}igo Casanueva and Lina Maria Rojas-Barahona and Milica Gasic},
        title = {Sub-domain Modelling for Dialogue Management with Hierarchical Reinforcement Learning},
      journal = {CoRR},
       volume = {abs/1706.06210},
         year = {2017},
          url = {http://arxiv.org/abs/1706.06210},
     abstract = {Human conversation is inherently complex, often spanning many different topics/domains. This makes policy learning for dialogue systems very challenging. Standard flat reinforcement learning methods do not provide an efficient framework for modelling such dialogues. In this paper, we focus on the under-explored problem of multi-domain dialogue management. First, we propose a new method for hierarchical reinforcement learning using the option framework. Next, we show that the proposed architecture learns faster and arrives at a better policy than the existing flat ones do. Moreover, we show how pretrained policies can be adapted to more complex systems with an additional set of new actions. In doing that, we show that our approach has the potential to facilitate policy optimisation for more sophisticated multi-domain dialogue systems.}
}
@article{BuesingetalNETWORK-12,
       author = {Buesing, L. and Macke, J. H. and Sahani, M.},
        title = {Learning stable, regularised latent models of neural population dynamics},
      journal = {Network},
         year = {2012},
       volume = {23},
       number = {1-2},
        pages = {24-47},
     abstract = {Ongoing advances in experimental technique are making commonplace simultaneous recordings of the activity of tens to hundreds of cortical neurons at high temporal resolution. Latent population models, including Gaussian-process factor analysis and hidden linear dynamical system (LDS) models, have proven effective at capturing the statistical structure of such data sets. They can be estimated efficiently, yield useful visualisations of population activity, and are also integral building-blocks of decoding algorithms for brain-machine interfaces (BMI). One practical challenge, particularly to LDS models, is that when parameters are learned using realistic volumes of data the resulting models often fail to reflect the true temporal continuity of the dynamics; and indeed may describe a biologically-implausible unstable population dynamic that is, it may predict neural activity that grows without bound. We propose a method for learning LDS models based on expectation maximisation that constrains parameters to yield stable systems and at the same time promotes capture of temporal structure by appropriate regularisation. We show that when only little training data is available our method yields LDS parameter estimates which provide a substantially better statistical description of the data than alternatives, whilst guaranteeing stable dynamics. We demonstrate our methods using both synthetic data and extracellular multi-electrode recordings from motor cortex.}
}
@article{BuiandJiangCoRR-18,
       author = {Nghi D. Q. Bui and Lingxiao Jiang},
        title = {Hierarchical Learning of Cross-Language Mappings through Distributed Vector Representations for Code},
      journal = {CoRR},
       volume = {arXiv:1803.04715},
         year = {2018},
     abstract = {Translating a program written in one programming language to another can be useful for software development tasks that need functionality implementations in different languages. Although past studies have considered this problem, they may be either specific to the language grammars, or specific to certain kinds of code elements (e.g., tokens, phrases, API uses). This paper proposes a new approach to automatically learn cross-language representations for various kinds of structural code elements that may be used for program translation. Our key idea is two folded: First, we normalize and enrich code token streams with additional structural and semantic information, and train cross-language vector representations for the tokens (a.k.a. shared embeddings based on word2vec, a neural-network-based technique for producing word embeddings; Second, hierarchically from bottom up, we construct shared embeddings for code elements of higher levels of granularity (e.g., expressions, statements, methods) from the embeddings for their constituents, and then build mappings among code elements across languages based on similarities among embeddings. Our preliminary evaluations on about 40,000 Java and C# source files from 9 software projects show that our approach can automatically learn shared embeddings for various code elements in different languages and identify their cross-language mappings with reasonable Mean Average Precision scores. When compared with an existing tool for mapping library API methods, our approach identifies many more mappings accurately. The mapping results and code can be accessed at this https URL We believe that our idea for learning cross-language vector representations with code structural information can be a useful step towards automated program translation.}
}
@article{BuibasandSilvaCoRR-15,
       author = {Gabriel Silva},
        title = {Algebraic identification of the effective connectivity of constrained geometric network models of neural signaling},
      journal = {CoRR},
       volume = {arXiv:1505.03964},
         year = {2015},
     abstract = {Cellular neural circuit and networks consisting of interconnected neurons and glia are ultimately responsible for the information processing associated with information processing in the brain. While there are major efforts aimed at mapping the structural and (electro)physiological connectivity of brain networks, such as the White House BRAIN Initiative aimed at the development of neurotechnologies capable of high density neural recordings, theoretical and computational methods for analyzing and making sense of all this data seem to be further behind. Here, we propose and provide a summary of an approach for calculating effective connectivity from experimental observations of neuronal network activity. The proposed method operates on network-level data, makes use of all relevant prior knowledge, such as dynamical models of individual cells in the network and the physical structural connectivity of the network, and is broadly applicable to large classes of biological and non-biological networks.} 
}
@article{BulthoffetalNATURE-89,
       author = {{B\"{u}lthoff}, H. and Little, J. and Poggio, T},
        title = {A Parallel Algorithm for Real-time Computation of Optical Flow},
      journal = {Nature},
       volume = {337(6207)},
         year = 1989,
        pages = {549-553},
}
@incollection{BundyUAI-88,
       author = {Bundy, Alan},
        title = {Incidence Calculus: A Mechanism for Probabilistic Reasoning},
     crossref = UAI-88,
        pages = {177-184},
}
@article{BungeetalCEREBRAL-CORTEX-05,
       author = {Bunge, Silvia and Wendelken, Carter and Badre, David and Wagner, Anthony},
        title = {Analogical Reasoning and Prefrontal Cortex: Evidence for Separable Retrieval and Integration Mechanisms},
      journal = {Cerebral Cortex},
       volume = {15},
         year = {2005},
        pages = {239-49},
     abstract = {The present study examined the contributions of prefrontal cortex (PFC) subregions to two component processes underlying verbal analogical reasoning: semantic retrieval and integration. Event- related functional magnetic resonance imaging data were acquired while subjects performed propositional analogy and semantic decision tasks. On each trial, subjects viewed a pair of words (pair 1), followed by an instructional cue and a second word pair (pair 2). On analogy trials, subjects evaluated whether pair 2 was semantically analogous to pair 1. On semantic trials, subjects indicated whether the pair 2 words were semantically related to each other. Thus, analogy -- but not semantic -- trials required integration across multiple retrieved relations. To identify regions involved in semantic retrieval, we manipulated the associative strength of pair 1 words in both tasks. Anterior left inferior PFC (aLIPC) was modulated by associative strength, consistent with a role in controlled semantic retrieval. Left frontopolar cortex was insensitive to associative strength, but was more sensitive to integration demands than was aLIPC, consistent with a role in integrating the products of semantic retrieval to evaluate whether distinct representations are analogous. Right dorsolateral PFC exhibited a profile consistent with a role in response selection rather than retrieval or integration. These findings indicate that verbal analogical reasoning depends on multiple, PFC-mediated computations.}
}


@article{BuntineJAIR-94,
       author = {Buntine, Wray L.},
        title = {Operations for Learning with Graphical Models},
      journal = jair,
       volume = 2,
        pages = {159-225},
         year = 1994,
}
@phdthesis{BuntinePhD,
       author = {Buntine, W.L.},
        title = {A Theory of Learning Classification Rules},
       school = {University of Technology, Sydney, Austrailia},
         year = 1990,
}
@article{BurchetalIEEE-94,
       author = {Burch, Jerry and Clarke, Edmund M. and Long, David and McMillan, Kenneth L. and Dill, David L.},
        title = {Symbolic Model Checking for Sequential Circuit Verification},
      journal = {{IEEE} Transactions on Computer Aided Design},
       volume = 13,
        issue = 4,
        pages = {401-424},
         year = 1994,
}

@article{BuretteetalFiN-15,
       author = {Alain Burette and Forrest Collman and Kristina D. Micheva and Stephen J. Smith and Richard J. Weinberg},
        title = {Knowing a synapse when you see one},
      journal = {Frontiers in Neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = 9,
        issue = 100,
         year = 2015,
     abstract = {Recent years have seen a rapidly growing recognition of the complexity and diversity of the myriad individual synaptic connections that define brain synaptic networks. It has also become increasingly apparent that the synapses themselves are a major key to understanding the development, function and adaptability of those synaptic networks. In spite of this growing appreciation, the molecular, structural and functional characteristics of individual synapses and the patterning of their diverse characteristics across functional networks have largely eluded quantitative study with available imaging technologies. Here we offer an overview of new computational imaging methods that promise to bring single-synapse analysis of synaptic networks to the fore. We focus especially on the challenges and opportunities associated with quantitative detection of individual synapses and with measuring individual synapses across network scale populations in mammalian brain.},
}
@article{BurgerandLangNCJB-01,
    author = {Burger, T. and Lang, E. W.},
     title = {{S}elf-organization of local cortical circuits and cortical orientation maps: a nonlinear {H}ebbian model of the visual cortex with adaptive lateral couplings},
   journal = {Z. Naturforsch., C, J. Bioscience},
      year = 2001,
    volume = 56,
     issue = {5-6},
     pages = {464-478},
  abstract = {A nonlinear, recurrent neural network model of the visual cortex is presented. Orientation maps emerge from adaptable afferent as well as plastic local intracortical circuits driven by random input stimuli. Lateral coupling structures self-organize into DOG profiles under the influence of pronounced emerging cortical activity blobs. The model's simplified architecture and features are modeled to largely mimik neurobiological findings.}
}
@article{Burges98,
      author = {Christopher J. C. Burges},
      title = {A Tutorial on Support Vector Machines for Pattern Recognition},
      journal = {Data Mining and Knowledge Discovery},
      volume = 2,
       issue = 2,
      pages = {121-167},
      year = 1998,
}
@article{BurgessANYAS-08,
       author = {Burgess, Neil},
        title = {Spatial Cognition and the Brain},
      journal = {Annals of the New York Academy of Sciences},
       volume = {1124},
       number = {1},
         year = {2008},
        pages = {77-97},
     abstract = {Recent advances in the understanding of spatial cognition are reviewed, focusing on memory for locations in large-scale space and on those advances inspired by single-unit recording and lesion studies in animals. Spatial memory appears to be supported by multiple parallel representations, including egocentric and allocentric representations, and those updated to accommodate self-motion. The effects of these representations can be dissociated behaviorally, developmentally, and in terms of their neural bases. It is now becoming possible to construct a mechanistic neural-level model of at least some aspects of spatial memory and imagery, with the hippocampus and medial temporal lobe providing allocentric environmental representations, the parietal lobe egocentric representations, and the retrosplenial cortex and parieto-occipital sulcus allowing both types of representation to interact. Insights from this model include a common mechanism for the construction of spatial scenes in the service of both imagery and episodic retrieval and a role for the remainder of Papez's circuit in orienting the viewpoint used. In addition, it appears that hippocampal and striatal systems process different aspects of environmental layout (boundaries and local landmarks, respectively) and do so using different learning rules (incidental learning and associative reinforcement, respectively).},
}
@article{BurgessandHitchPR-99,
       author = {Burgess, Neil and Hitch, Graham},
        title = {Memory for Serial Order: A Network Model of the Phonological Loop and its Timing},
      journal = {Psychological Review},
         year = {1999},
        pages = {},
       volume = {106},
       number = {3},
     abstract = {A connectionist model of human short-term memory is presented that extends the 'phonological loop' (A.D. Baddeley, 1986) to encompass serial order and learning. Psychological and neuropsychological data motivate separate layers of lexical, timing and input and output phonemic information. Connection weights between layers show Hebbian learning and decay over short and long time scales. At recall, the timing signal is rerun, phonemic information feeds back from output to input and lexical nodes compete to be selected. The selected node then receives decaying inhibition. The model provides an explanatory mechanism for the phonological loop, and for the effects of serial position, presentation modality, lexicality, grouping and Hebb repetition. It makes new psychological and neuropsychological predictions and is a starting point for understanding the role of the phonological loop in vocabulary acquisition and for interpreting data from functional neuroimaging.}
}
@incollection{BurgessetalGATEWAY-05,
     title = {The gateway hypothesis of rostral prefrontal cortex (area 10) function},
    author = {Paul W. Burgess and Jon S. Simons and Iroise Dumontheil and Sam J. Gilbert},
 booktitle = {Measuring the Mind: Speed, Control, and Age},
    editor = {John Duncan and Louise Phillips and Peter McLeod},
 publisher = {Oxford University Press},
   address = {Oxford, UK},
      year = 2005,
     pages = {217-248}
}
@article{BurgessetalNEURON-02,
        title = {The Human Hippocampus and Spatial and Episodic Memory},
       author = {Neil Burgess and Eleanor A Maguire and John O'Keefe},
      journal = {Neuron},
       volume = {35},
       number = {4},
        pages = {625-641},
         year = {2002},
     abstract = {Finding one's way around an environment and remembering the events that occur within it are crucial cognitive abilities that have been linked to the hippocampus and medial temporal lobes. Our review of neuropsychological, behavioral, and neuroimaging studies of human hippocampal involvement in spatial memory concentrates on three important concepts in this field: spatial frameworks, dimensionality, and orientation and self-motion. We also compare variation in hippocampal structure and function across and within species. We discuss how its spatial role relates to its accepted role in episodic memory. Five related studies use virtual reality to examine these two types of memory in ecologically valid situations. While processing of spatial scenes involves the parahippocampus, the right hippocampus appears particularly involved in memory for locations within an environment, with the left hippocampus more involved in context-dependent episodic or autobiographical memory.}
}
@article{BurghardtetalHIPPOCAMPUS-12,
       author = {Burghardt, N. S. and Park, E. H. and Hen, R. and Fenton, A. A.},
        title = {Adult-born hippocampal neurons promote cognitive flexibility in mice},
      journal = {Hippocampus},
         year = {2012},
       volume = {22},
       number = {9},
        pages = {1795-1808},
     abstract = {The hippocampus is involved in segregating memories, an ability that utilizes the neural process of pattern separation and allows for cognitive flexibility. We evaluated a proposed role for adult hippocampal neurogenesis in cognitive flexibility using variants of the active place avoidance task and two independent methods of ablating adult-born neurons, focal X-irradiation of the hippocampus, and genetic ablation of glial fibrillary acidic protein positive neural progenitor cells, in mice. We found that ablation of adult neurogenesis did not impair the ability to learn the initial location of a shock zone. However, when conflict was introduced by switching the location of the shock zone to the opposite side of the room, irradiated and transgenic mice entered the new shock zone location significantly more than their respective controls. This impairment was associated with increased upregulation of the immediate early gene Arc in the dorsal dentate gyrus, suggesting a role for adult neurogenesis in modulating network excitability and/or synaptic plasticity. Additional experiments revealed that irradiated mice were also impaired in learning to avoid a rotating shock zone when it was added to an initially learned stationary shock zone, but were unimpaired in learning the identical simultaneous task variant if it was their initial experience with place avoidance. Impaired avoidance could not be attributed to a deficit in extinction or an inability to learn a new shock zone location in a different environment. Together these results demonstrate that adult neurogenesis contributes to cognitive flexibility when it requires changing a learned response to a stimulus-evoked memory.}
}
@book{Burgin2010,
        title = {Theory of Information: Fundamentality, Diversity and Unification},
       author = {Burgin, Mark},
       series = {World Scientific Series in Information Studies},
    publisher = {World Scientific Publishing Company},
         year = {2010},
}
@article{BurginENTROPY-03,
       author = {Burgin, Mark},
        title = {{Information Theory: a Multifaceted Model of Information}},
      journal = {Entropy},
       volume = {5},
        issue = {2},
         year = {2003},
        pages = {146-160},
}
@article{BurkeBRAIN-07,
     title = {{Sir Charles Sherrington's The integrative action of the nervous system: a centenary appreciation}}, 
    author = {Burke, Robert E.}, 
    volume = 130, 
     issue = 4, 
     pages = {887-894}, 
      year = 2007, 
   journal = {Brain}, 
  abstract = {In 1906 Sir Charles Sherrington published The Integrative
  Action of the Nervous System, which was a collection of ten lectures
  delivered two years before at Yale University in the United
  States. In this monograph Sherrington summarized two decades of
  painstaking experimental observations and his incisive
  interpretation of them. It settled the then-current debate between
  the "Reticular Theory" versus "Neuron Doctrine" ideas about the
  fundamental nature of the nervous system in mammals in favor of the
  latter, and it changed forever the way in which subsequent
  generations have viewed the organization of the central nervous
  system. Sherrington's magnum opus contains basic concepts and even
  terminology that are now second nature to every student of the
  subject. This brief article reviews the historical context in which
  the book was written, summarizes its content, and considers its
  impact on Neurology and Neuroscience.}, 
}
@book{Burns1981,
        title = {Feeling good: the new mood therapy},
       author = {Burns, David},
         year = {1981},
    publisher = {Penguin Books}
}
@book{BurnsDISTORTIONS-80,
       author = {Burns, David D.},
        title = {Feeling Good: The New Mood Therapy},
         year = {William Morrow \& Company},
    publisher = {New York},
      address = 1980
}
@inproceedings{BurnsetalSS-94,
     author = {Burns, G. and Daoud, R. and Vaigl, J.},
      title = {{LAM}: An Open Cluster Environment for {MPI}},
  booktitle = {Proceedings of Supercomputing Symposium},
      pages = {379-386},
       year = 1994,
}
@article{BurroneandMurthyCOIN-03,
       author = {Juan Burrone and Venkatesh N Murthy},
        title = {Synaptic gain control and homeostasis},
      journal = {Current Opinion in Neurobiology},
       volume = {13},
       number = {5},
         year = {2003},
        pages = {560-567},
     abstract = {Chronic changes in activity can induce neurons to alter the strength of all their synapses in unison. Although the specific changes that occur appear to vary depending on the experimental preparation, their net effect is to counter the experimentally induced modification of activity. Such adaptive, cell-wide changes in synaptic strength serve to stabilize neuronal activity and are collectively referred to as homeostatic synaptic plasticity. Recent studies have shed light on what triggers homeostatic synaptic plasticity, whether or not it is distinct from other forms of synaptic plasticity and whether or not it occurs in the intact brain.}
}
@article{BurtandAdelsonCOM-83,
    author = {Peter J. Burt and Edward H. Adelson},
     title = {The {L}aplacian Pyramid as a compact image code},
   journal = {IEEE Transactions on Communications},
    volume = 31,
     issue = 4,
 publisher = {IEEE},
     pages = {532-540},
      year = 1983,
}
@book{Burton2008,
     author = {Robert A. Burton},
      title = {On Being Certain: Believing You Are Right Even When You're Not},
  publisher = {St. Martin's Griffin},
    address = {New York, NY},
       year = 2008,
    comment = {\urlh{http://en.wikipedia.org/wiki/Robert_A._Burton}{Robert Burton}},
}
@article{Bush1945,
      author = {Vannevar Bush},
       title = {As We May Think},
     journal = {Atlantic Monthly},
      volume = 176,
       issue = 1,
       pages = {101-108},
        year = 1945,
}
@book{Bush2010,
        title = {Nanoscale Communication Networks},
       author = {Bush, Stephen F.},
       series = {Nanoscale Science and Engineering},
    publisher = {Artech House},
         year = 2010,
}
@article{Bush45,
      author = {Vannevar Bush},
       title = {As We May Think},
     journal = {Atlantic Monthly},
      volume = 176,
       issue = 1,
       pages = {101-108},
        year = 1945,
}
@article{BushFICSC-2011,
       author = {Bush, Stephen F.},
        title = {Toward {\it{in vivo}} nanoscale communication networks: utilizing an active network architecture},
      journal = {Frontiers in Computer Science China},
    publisher = {Springer-Verlag},
       volume = 5,
        issue = 3,
         year = 2011,
        pages = {316-326},
}
@article{ButcheretalJoN-17,
       author = {Butcher, Peter A. and Ivry, Richard B. and Kuo, Sheng-Han and Rydz, David and Krakauer, John W. and Taylor, Jordan A.},
        title = {The cerebellum does more than sensory prediction error-based learning in sensorimotor adaptation tasks},
      journal = {Journal of Neurophysiology},
       volume = {118},
       number = {3},
         year = {2017},
        pages = {1622-1636},
     abstract = { Individuals with damage to the cerebellum perform poorly in sensorimotor adaptation paradigms. This deficit has been attributed to impairment in sensory prediction error-based updating of an internal forward model, a form of implicit learning. These individuals can, however, successfully counter a perturbation when instructed with an explicit aiming strategy. This successful use of an instructed aiming strategy presents a paradox: In adaptation tasks, why do individuals with cerebellar damage not come up with an aiming solution on their own to compensate for their implicit learning deficit? To explore this question, we employed a variant of a visuomotor rotation task in which, before executing a movement on each trial, the participants verbally reported their intended aiming location. Compared with healthy control participants, participants with spinocerebellar ataxia displayed impairments in both implicit learning and aiming. This was observed when the visuomotor rotation was introduced abruptly (experiment 1) or gradually (experiment 2). This dual deficit does not appear to be related to the increased movement variance associated with ataxia: Healthy undergraduates showed little change in implicit learning or aiming when their movement feedback was artificially manipulated to produce similar levels of variability (experiment 3). Taken together the results indicate that a consequence of cerebellar dysfunction is not only impaired sensory prediction error-based learning but also a difficulty in developing and/or maintaining an aiming solution in response to a visuomotor perturbation. We suggest that this dual deficit can be explained by the cerebellum forming part of a network that learns and maintains action-outcome associations across trials. NEW \& NOTEWORTHY Individuals with cerebellar pathology are impaired in sensorimotor adaptation. This deficit has been attributed to an impairment in error-based learning, specifically, from a deficit in using sensory prediction errors to update an internal model. Here we show that these individuals also have difficulty in discovering an aiming solution to overcome their adaptation deficit, suggesting a new role for the cerebellum in sensorimotor adaptation tasks. }
}
@incollection{Butterworth2003,
       author = {George Butterworth},
        title = {Pointing is the royal road to language},
    booktitle = {Pointing: Where language, culture, and cognition meet},
       editor = {Sotaro Kita},
    publisher = {Lawrence Erlbaum Associates, Inc.},
      address = {Mahwah, NJ, USA},
         year = 2003
},
      journal = {Neuroscience \& Biobehavioral Reviews},
         year = {2015},
       volume = {48},
        pages = {92-147},
     abstract = {The aims of the paper are to update Rolls' quantitative computational theory of hippocampal function and the predictions it makes about the different subregions (dentate gyrus, CA3 and CA1), and to examine behavioral and electrophysiological data that address the functions of the hippocampus and particularly its subregions. Based on the computational proposal that the dentate gyrus produces sparse representations by competitive learning and via the mossy fiber pathway forces new representations on the CA3 during learning (encoding), it has been shown behaviorally that the dentate gyrus supports spatial pattern separation during learning. Based on the computational proposal that CA3-CA3 autoassociative networks are important for episodic memory, it has been shown behaviorally that the CA3 supports spatial rapid one-trial learning, learning of arbitrary associations where space is a component, pattern completion, spatial short-term memory, and spatial sequence learning by associations formed between successive items. The concept that the CA1 recodes information from CA3 and sets up associatively learned backprojections to neocortex to allow subsequent retrieval of information to neocortex, is consistent with findings on consolidation. Behaviorally, the CA1 is implicated in processing temporal information as shown by investigations requiring temporal order pattern separation and associations across time; and computationally this could involve associations in CA1 between object and timing information that have their origins in the lateral and medial entorhinal cortex respectively. The perforant path input from the entorhinal cortex to DG is implicated in learning, to CA3 in retrieval from CA3, and to CA1 in retrieval after longer time intervals ("intermediate-term memory") and in the temporal sequence memory for objects.}
}
@article{BuxhoevedenandCasanovaBRAIN-02,
       author = {Buxhoeveden, Daniel P. and Casanova, Manuel F.},
        title = {The minicolumn hypothesis in neuroscience},
      journal = {Brain},
       volume = 125,
        issue = 5,
        pages = {935-951},
         year = 2002,
     abstract = {The minicolumn is a continuing source of research and debate more than half a century after it was identified as a component of brain organization. The minicolumn is a sophisticated local network that contains within it the elements for redundancy and plasticity. Although it is sometimes compared to subcortical nuclei, the design of the minicolumn is a distinctive form of module that has evolved specifically in the neocortex. It unites the horizontal and vertical components of cortex within the same cortical space. Minicolumns are often considered highly repetitive, even clone‐like, units. However, they display considerable heterogeneity between areas and species, perhaps even within a given macrocolumn. Despite a growing recognition of the anatomical basis of the cortical minicolumn, as well as its physiological properties, the potential of the minicolumn has not been exploited in fields such as comparative neuroanatomy, abnormalities of the brain and mind, and evolution.},
}
@article{BuxtonetalNEUROIMAGING-04,
        title = {Modeling the hemodynamic response to brain activation},
       author = {Buxton, R.B. and Uludag, K. and Dubowitz, D.J. and Liu, T.T.},
      journal = {Neuroimaging},
       volume = 23,
        issue = 1,
         year = 2004,
        pages = {220-233},
     abstract = {Neural activity in the brain is accompanied by changes in cerebral blood flow (CBF) and blood oxygenation that are detectable with functional magnetic resonance imaging (fMRI) techniques. In this paper, recent mathematical models of this hemodynamic response are reviewed and integrated. Models are described for: (1) the blood oxygenation level dependent (BOLD) signal as a function of changes in cerebral oxygen extraction fraction (E) and cerebral blood volume (CBV); (2) the balloon model, proposed to describe the transient dynamics of CBV and deoxy-hemoglobin (Hb) and how they affect the BOLD signal; (3) neurovascular coupling, relating the responses in CBF and cerebral metabolic rate of oxygen (CMRO(2)) to the neural activity response; and (4) a simple model for the temporal nonlinearity of the neural response itself. These models are integrated into a mathematical framework describing the steps linking a stimulus to the measured BOLD and CBF responses. Experimental results examining transient features of the BOLD response (post-stimulus undershoot and initial dip), nonlinearities of the hemodynamic response, and the role of the physiologic baseline state in altering the BOLD signal are discussed in the context of the proposed models. Quantitative modeling of the hemodynamic response, when combined with experimental data measuring both the BOLD and CBF responses, makes possible a more specific and quantitative assessment of brain physiology than is possible with standard BOLD imaging alone. This approach has the potential to enhance numerous studies of brain function in development, health, and disease.}
}
@book{Buzsaki2006,
        title = {{Rhythms of the Brain}},
       author = {Buzs\'{a}ki, Gy\"{o}rgy},
    publisher = {Oxford University Press},
         year = 2006,
     abstract = {Studies of mechanisms in the brain that allow complicated things to happen in a coordinated fashion have produced some of the most spectacular discoveries in neuroscience. This book provides eloquent support for the idea that spontaneous neuron activity, far from being mere noise, is actually the source of our cognitive abilities. It takes a fresh look at the co-evolution of structure and function in the mammalian brain, illustrating how self-emerged oscillatory timing is the brains fundamental organizer of neuronal information. The small world-like connectivity of the cerebral cortex allows for global computation on multiple spatial and temporal scales. The perpetual interactions among the multiple network oscillators keep cortical systems in a highly sensitive metastable state and provide energy-efficient synchronizing mechanisms via weak links. In a sequence of cycles, Gyorgy Buzsaki guides the reader from the physics of oscillations through neuronal assembly organization to complex cognitive processing and memory storage. His clear, fluid writing  accessible to any reader with some scientific knowledge  is supplemented by extensive footnotes and references that make it just as gratifying and instructive a read for the specialist. The coherent view of a single author who has been at the forefront of research in this exciting field, this volume is essential reading for anyone interested in our rapidly evolving understanding of the brain.},
}
@article{BuzsakiHIPPOCAMPUS-15,
       author = {Gy\"{o}rgy Buzs\`{a}ki},
        title = {Hippocampal sharp wave-ripple: A cognitive biomarker for episodic memory and planning},
      journal = {Hippocampus},
    publisher = {John Wiley and Sons Inc.},
         year = {2015},
       volume = {25},
        issue = {10},
        pages = {1073-1188},
     abstract = {Sharp wave ripples (SPW-Rs) represent the most synchronous population pattern in the mammalian brain. Their excitatory output affects a wide area of the cortex and several subcortical nuclei. SPW-Rs occur during ``off-line'' states of the brain, associated with consummatory behaviors and non-REM sleep, and are influenced by numerous neurotransmitters and neuromodulators. They arise from the excitatory recurrent system of the CA3 region and the SPW-induced excitation brings about a fast network oscillation (ripple) in CA1. The spike content of SPW-Rs is temporally and spatially coordinated by a consortium of interneurons to replay fragments of waking neuronal sequences in a compressed format. SPW-Rs assist in transferring this compressed hippocampal representation to distributed circuits to support memory consolidation; selective disruption of SPW-Rs interferes with memory. Recently acquired and pre-existing information are combined during SPW-R replay to influence decisions, plan actions and, potentially, allow for creative thoughts. In addition to the widely studied contribution to memory, SPW-Rs may also affect endocrine function via activation of hypothalamic circuits. Alteration of the physiological mechanisms supporting SPW-Rs leads to their pathological conversion, ``p-ripples,'' which are a marker of epileptogenic tissue and can be observed in rodent models of schizophrenia and Alzheimer's Disease. Mechanisms for SPW-R genesis and function are discussed in this review.},
}
@article{BuzsakiNEURON-10,
        title = {Neural Syntax: Cell Assemblies, Synapsembles, and Readers},
       author = {Gy{\"{o}}rgy Buzs{\'{a}}ki},
      journal = {Neuron},
       volume = {68},
       number = {3},
         year = {2010},
        pages = {362-385},
     abstract = {A widely discussed hypothesis in neuroscience is that transiently active ensembles of neurons, known as "cell assemblies," underlie numerous operations of the brain, from encoding memories to reasoning. However, the mechanisms responsible for the formation and disbanding of cell assemblies and temporal evolution of cell assembly sequences are not well understood. I introduce and review three interconnected topics, which could facilitate progress in defining cell assemblies, identifying their neuronal organization, and revealing causal relationships between assembly organization and behavior. First, I hypothesize that cell assemblies are best understood in light of their output product, as detected by "reader-actuator" mechanisms. Second, I suggest that the hierarchical organization of cell assemblies may be regarded as a neural syntax. Third, constituents of the neural syntax are linked together by dynamically changing constellations of synaptic weights ("synapsembles"). The existing support for this tripartite framework is reviewed and strategies for experimental testing of its predictions are discussed.}
}
@book{BuzsakiandChristen2016,
       author = {Buzs\`{a}ki, Gy\"{o}rgy and Christen, Yves},
        title = {Micro-, Meso- and Macro-Dynamics of the Brain},
    publisher = {Springer},
         year = 2016
}
@article{BuzsakiandDraguhnSCIENCE-04,
       author = {Buzs\'{a}ki, Gy\"{o}rgy and Draguhn, Andreas},
        title = {Neuronal Oscillations in Cortical Networks},
      journal = {Science},
       volume = 304,
        issue = 5679,
         year = 2004,
        pages = {1926-1929},
     abstract = {Clocks tick, bridges and skyscrapers vibrate, neuronal networks oscillate. Are neuronal oscillations an inevitable by-product, similar to bridge vibrations, or an essential part of the brain's design? Mammalian cortical neurons form behavior-dependent oscillating networks of various sizes, which span five orders of magnitude in frequency. These oscillations are phylogenetically preserved, suggesting that they are functionally relevant. Recent findings indicate that network oscillations bias input selection, temporally link neurons into assemblies, and facilitate synaptic plasticity, mechanisms that cooperatively support temporal representation and long-term consolidation of information.},
}
@article{BuzsakiandTingleyTiCN-18,
        title = {Space and Time: {T}he Hippocampus as a Sequence Generator},
       author = {Gy\"{o}rgy Buzs\`{a}ki and David Tingley},
      journal = {Trends in Cognitive Sciences},
       volume = {22},
       number = {10},
         year = {2018},
        pages = {853-869},
     abstract = {Neural computations are often compared to instrument-measured distance or duration, and such relationships are interpreted by a human observer. However, neural circuits do not depend on human-made instruments but perform computations relative to an internally defined rate-of-change. While neuronal correlations with external measures, such as distance or duration, can be observed in spike rates or other measures of neuronal activity, what matters for the brain is how such activity patterns are utilized by downstream neural observers. We suggest that hippocampal operations can be described by the sequential activity of neuronal assemblies and their internally defined rate of change without resorting to the concept of space or time.}
}
@article{BuzsakiandWatsonDNC-12,
       author = {Buzsaki, G. and Watson, B. O.},
        title = {Brain rhythms and neural syntax: implications for efficient coding of cognitive content and neuropsychiatric disease},
      journal = {Dialogues Clinical Neuroscience},
         year = {2012},
       volume = {14},
        issue = {4},
        pages = {345-367},
     abstract = {The perpetual activity of the cerebral cortex is largely supported by the variety of oscillations the brain generates, spanning a number of frequencies and anatomical locations, as well as behavioral correlates. First, we review findings from animal studies showing that most forms of brain rhythms are inhibition-based, producing rhythmic volleys of inhibitory inputs to principal cell populations, thereby providing alternating temporal windows of relatively reduced and enhanced excitability in neuronal networks. These inhibition-based mechanisms offer natural temporal frames to group or "chunk" neuronal activity into cell assemblies and sequences of assemblies, with more complex multi-oscillation interactions creating syntactical rules for the effective exchange of information among cortical networks. We then review recent studies in human psychiatric patients demonstrating a variety alterations in neural oscillations across all major psychiatric diseases, and suggest possible future research directions and treatment approaches based on the fundamental properties of brain rhythms.}
}
@article{ByersetalVR-12,
        title = {Exploring the relationship between perceptual learning and top-down attentional control},
      journal = {Vision Research},
       volume = 74,
        issue = 0,
         year = 2012,
        pages = 30039,
     abstract = {Here, we review the role of top-down attention in both the acquisition and the expression of perceptual learning, as well as the role of learning in more efficiently guiding attentional modulations. Although attention often mediates learning at the outset of training, many of the characteristic behavioral and neural changes associated with learning can be observed even when stimuli are task irrelevant and ignored. However, depending on task demands, attention can override the effects of perceptual learning, suggesting that even if top-down factors are not strictly necessary to observe learning, they play a critical role in determining how learning-related changes in behavior and neural activity are ultimately expressed. In turn, training may also act to optimize the effectiveness of top-down attentional control by improving the efficiency of sensory gain modulations, regulating intrinsic noise, and altering the read-out of sensory information.}
}
@article{BylanderAIJ-94,
       author = {Bylander, Tom},
        title = {The Computational Complexity of Propositional {STRIPS} Planning},
      journal = AIJ,
       volume = 69,
        pages = {165-204},
         year = 1994,
}
@article{BylanderAIJ-96,
       author = {Bylander, Tom},
        title = {A Probabilistic Analysis of {STRIPS} Planning},
      journal = AIJ,
         year = 1996,
}
@inproceedings{BylanderIJCAI-91,
       author = {Bylander, Thomas},
        title = {Complexity Results for Planning},
    booktitle = {Proceedings {IJCAI} 12},
     location = {Sydney, Australia},
 organization = {IJCAII},
         year = 1991,
        pages = {274-279},
}
@article{ByrneJOTA-13,
       author = {Byrne, Charles L.},
        title = {Alternating Minimization as Sequential Unconstrained Minimization: {A} Survey},
      journal = {Journal of Optimization Theory and Applications},
       volume = {156},
       number = {3},
         year = {2013},
        pages = {554-566},
     abstract = {Sequential unconstrained minimization is a general iterative method for minimizing a function over a given set. At each step of the iteration we minimize the sum of the objective function and an auxiliary function. The aim is to select the auxiliary functions so that, at least, we get convergence in function value to the constrained minimum. The SUMMA is a broad class of these methods for which such convergence holds. Included in the SUMMA class are the barrier-function methods, entropic and other proximal minimization algorithms, the simultaneous multiplicative algebraic reconstruction technique, and, after some reformulation, penalty-function methods. The alternating minimization method of Csisz\'{a}r and Tusn\'{a}dy also falls within the SUMMA class, whenever their five-point property holds. Therefore, the expectation maximization maximum likelihood algorithm for the Poisson case is also in the SUMMA class.},
}
@article{ByrneetalNEURO-10,
    author = {Byrne, M. J. and Waxham, M. N. and Kubota, Y.},
     title = {{C}ellular dynamic simulator: an event driven molecular simulation environment for cellular physiology},
   journal = {Neuroinformatics},
      year = 2010,
    volume = 8,
     issue = 2,
     pages = {63-82},
}
@article{BystronetalNATURE-06,
     title = {The first neurons of the human cerebral cortex},
    author = {Bystron, Irina and Rakic, Pasko and Molnar, Zoltan and Blakemore, Colin},
   journal = {Nature Neuroscience},
    volume = 9,
     issue = 7,
      year = 2006,
     pages = {880-886},
abstract = {We describe a distinctive, widespread population of neurons situated beneath the pial surface of the human embryonic forebrain even before complete closure of the neural tube. These 'predecessor' cells include the first neurons seen in the primordium of the cerebral cortex, before the onset of local neurogenesis.  Morphological analysis, combined with the study of centrosome  location, regional transcription factors and patterns of mitosis and  neurogenesis, indicates that predecessor cells invade the cortical  primordium by tangential migration from the subpallium. These  eurons, described here for the first time, precede all other known cell types of the developing cortex.},
}
@misc{CS379C_Class_Discussion_Notes_07-04-20,
        title = {CS379C Class Discussion Notes for 2020},
       author = {Thomas Dean},
 howpublished = {https://web.stanford.edu/class/cs379c/archive/2020/class_messages_listing/index.html},
         year = {2019},
}
@misc{CS379C_Class_Discussion_Notes_08-09-19,
        title = {CS379C Class Discussion Notes for 2019},
       author = {Thomas Dean},
 howpublished = {https://web.stanford.edu/class/cs379c/archive/2019/class_messages_listing/index.html},
         year = {2019},
}
@misc{CS379C_Final_Project_Gomezetal-18,
        title = {Imitation Learning for Code Generation via Recurrent State Space Embeddings},
       author = {Marcus Gomez and Nate Gruver and Michelle Lam and Rohun Saxena and Lucy Wang},
 howpublished = {Representative Final Project Archive for Stanford CS379C in Spring 2018},
         year = {2018},
}
@misc{CS379C_Final_Project_Luetal-18,
        title = {Integrating Reinforcement Learning Agents for Error Correction in Abstract Syntax Trees},
       author = {Peter Lu and Sophia Sanchez and Yousef Hindy and Maurice Chiang and Michael Smith},
 howpublished = {Representative Final Project Archive for Stanford CS379C in Spring 2018},
         year = {2018},
}
@misc{CS379C_Final_Project_Wong-18,
        title = {Reconstructive Memory for Abstract Selective Recall},
       author = {Catherine Wong},
 howpublished = {Representative Final Project Archive for Stanford CS379C in Spring 2018},
         year = {2018},
}
@article{CaceresetalPNAS-03,
     title = {Elevated gene expression levels distinguish human from non-human primate brains},
    author = {C\'{a}ceres, Mario and Lachuer, Joel and Zapala, Matthew A. and Redmond, John C. and Kudo, Lili and Geschwind, Daniel H. and Lockhart, David J. and Preuss, Todd M. and Barlow, Carrolee},
   journal = {Proceedings of the National Academy of Sciences of the United States of America},
     issue = 22,
     pages = {13030-13035},
    volume = 100,
      year = 2003,
  abstract = {Little is known about how the human brain differs from that of our closest relatives. To investigate the genetic basis of human specializations in brain organization and cognition, we compared gene expression profiles for the cerebral cortex of humans  chimpanzees, and rhesus macaques by using several independent techniques. We identified 169 genes that exhibited expression differences between human and chimpanzee cortex, and 91 were ascribed to the human lineage by using macaques as an outgroup. Surprisingly, most differences between the brains of humans and non-human primates involved up-regulation, with 90\% of the genes being more highly expressed in humans. By  contrast, in the comparison of human and chimpanzee heart and liver   the numbers of up- and down-regulated genes were nearly identical   Our results indicate that the human brain displays a distinctive  pattern of gene expression relative to non-human primates, with  higher expression levels for many genes belonging to a wide variety  of functional classes. The increased expression of these genes could  provide the basis for extensive modifications of cerebral physiology  and function in humans and suggests that the human brain is characterized by elevated levels of neuronal activity.},
}
@incollection{CadieuandOlshausenNIPS-08,
     title = {Learning Transformational Invariants from Time-Varying Natural Images},
    author = {Charles F. Cadieu and Bruno A. Olshausen},
 booktitle = {Advances in Neural Information Processing Systems 21},
    editor = {Dale Schuurmans and Yoshua Bengio},
 publisher = {MIT Press},
   address = {Cambridge, MA},
      year = 2008
}
@article{CadieuandOshausenNC-12,
     title = {Learning Intermediate-Level Representations of Form and Motion from Natural Movies},
    author = {Charles F. Cadieu and Bruno A. Olshausen},
   journal = {Neural Computation},
    volume = 24,
     issue = 4,
      year = 2012,
     pages = {827-866},
}
@inproceedings{CadieuetalCoRR-13,
       author = {Cadieu, C. F. and Hong, H. and Yamins, D. L. and Nicolas Pinto and N. J. Majaj and J. J. {DiCarlo}},
        title = {The Neural Representation Benchmark and its Evaluation on Brain and Machine},
    booktitle = {International Conference on Learning Representations (ICLR)},
         year = {2013},
      address = {Scottsdale, AZ},
}
@article {CadieuetalPLoS-14,
       author = {Cadieu, Charles F. and Hong, Ha and Yamins, Daniel L. K. and Nicolas Pinto and Ardila, Diego and Solomon, Ethan A. and Majaj, Najib J. and James J. {DiCarlo}},
        title = {Deep Neural Networks Rival the Representation of Primate {IT} Cortex for Core Visual Object Recognition.},
      journal = {PLoS Computational Biology},
       volume = {10},
         year = {2014},
        pages = {e1003963},
    abstract = {The primate visual system achieves remarkable visual object recognition performance even in brief presentations, and under changes to object exemplar, geometric transformations, and background variation (a.k.a. core visual object recognition). This remarkable performance is mediated by the representation formed in inferior temporal (IT) cortex. In parallel, recent advances in machine learning have led to ever higher performing models of object recognition using artificial deep neural networks (DNNs). It remains unclear, however, whether the representational performance of DNNs rivals that of the brain. To accurately produce such a comparison, a major difficulty has been a unifying metric that accounts for experimental limitations, such as the amount of noise, the number of neural recording sites, and the number of trials, and computational limitations, such as the complexity of the decoding classifier and the number of classifier training examples. In this work, we perform a direct comparison that corrects for these experimental limitations and computational considerations. As part of our methodology, we propose an extension of "kernel analysis" that measures the generalization accuracy as a function of representational complexity. Our evaluations show that, unlike previous bio-inspired models, the latest DNNs rival the representational performance of IT cortex on this visual object recognition task. Furthermore, we show that models that perform well on measures of representational performance also perform well on measures of representational similarity to IT, and on measures of predicting individual IT multi-unit responses. Whether these DNNs rely on computational mechanisms similar to the primate visual system is yet to be determined, but, unlike all previous bio-inspired models, that possibility cannot be ruled out merely on representational performance grounds.},
}
@article{CaietalCoRR-17,
       author = {Jonathon Cai and Richard Shin and Dawn Song},
        title = {A Comprehensive Survey of Graph Embedding: Problems, Techniques and Applications},
      journal = {CoRR},
       volume = {arXiv:1704.06611},
         year = {2017},
     abstract = {Graph is an important data representation which appears in a wide diversity of real-world scenarios. Effective graph analytics provides users a deeper understanding of what is behind the data, and thus can benefit a lot of useful applications such as node classification, node recommendation, link prediction, etc. However, most graph analytics methods suffer the high computation and space cost. Graph embedding is an effective yet efficient way to solve the graph analytics problem. It converts the graph data into a low dimensional space in which the graph structural information and graph properties are maximumly preserved. In this survey, we conduct a comprehensive review of the literature in graph embedding. We first introduce the formal definition of graph embedding as well as the related concepts. After that, we propose two taxonomies of graph embedding which correspond to what challenges exist in different graph embedding problem settings and how the existing work address these challenges in their solutions. Finally, we summarize the applications that graph embedding enables and suggest four promising future research directions in terms of computation efficiency, problem settings, techniques and application scenarios.}
}
@article{CaietalICLR-17,
       author = {Jonathon Cai and Richard Shin and Dawn Song},
        title = {Making Neural Programming Architectures Generalize via Recursion},
      journal = {CoRR},
       volume = {arXiv:1704.06611},
         year = {2017},
     abstract = {Empirically, neural networks that attempt to learn programs from data have exhibited poor generalizability. Moreover, it has traditionally been difficult to reason about the behavior of these models beyond a certain level of input complexity. In order to address these issues, we propose augmenting neural architectures with a key abstraction: recursion. As an application, we implement recursion in the Neural Programmer-Interpreter framework on four tasks: grade-school addition, bubble sort, topological sort, and quicksort. We demonstrate superior generalizability and interpretability with small amounts of training data. Recursion divides the problem into smaller pieces and drastically reduces the domain of each neural network component, making it tractable to prove guarantees about the overall system's behavior. Our experience suggests that in order for neural architectures to robustly learn program semantics, it is necessary to incorporate a concept like recursion.},
}
@article{CaietalNANOSCALE-07,
     title = {Are quantum dots ready for in vivo imaging in human subjects?},
    author = {Cai, W. and Hsu, A.R. and Li, Z. and Chen, X.},
   journal = {Nanoscale Research Letters},
    volume = 2,
     issue = 6,
     pages = {265-281},
      year = 2007,
  abstract = {Nanotechnology has the potential to profoundly transform the nature of cancer diagnosis and cancer patient management in the future. Over the past decade, quantum dots (QDs) have become one of the fastest growing areas of research in nanotechnology. QDs are fluorescent semiconductor nanoparticles suitable for multiplexed in vitro and in vivo imaging. Numerous studies on QDs have resulted in major advancements in QD surface modification, coating, biocompatibility, sensitivity, multiplexing, targeting specificity, as well as important findings regarding toxicity and applicability. For in vitro applications, QDs can be used in place of traditional organic fluorescent dyes in virtually any system, outperforming organic dyes in the majority of cases. In vivo targeted tumor imaging with biocompatible QDs has recently become possible in mouse models. With new advances in QD technology such as bioluminescence resonance energy transfer, synthesis of smaller size non-Cd based QDs, improved surface coating and conjugation, and multifunctional probes for multimodality imaging, it is likely that human applications of QDs will soon be possible in a clinical setting.}
}
@article{CaietalNATURE-COMMUNICATIONS-19,
       author = {Cai, Weidong and Duberg, Katherine and Padmanabhan, Aarthi and Rehert, Rachel and Bradley, Travis and Carrion, Victor and Menon, Vinod},
        title = {Hyperdirect insula-basal-ganglia pathway and adult-like maturity of global brain responses predict inhibitory control in children},
      journal = {Nature Communications},
       volume = {10},
        issue = {1},
         year = {2019},
        pages = {4798},
     abstract = {Inhibitory control is fundamental to children's self-regulation and cognitive development. Here we investigate cortical-basal ganglia pathways underlying inhibitory control in children and their adult-like maturity. We first conduct a comprehensive meta-analysis of extant neurodevelopmental studies of inhibitory control and highlight important gaps in the literature. Second, we examine cortical-basal ganglia activation during inhibitory control in children ages 9-12 and demonstrate the formation of an adult-like inhibitory control network by late childhood. Third, we develop a neural maturation index (NMI), which assesses the similarity of brain activation patterns between children and adults, and demonstrate that higher NMI in children predicts better inhibitory control. Fourth, we show that activity in the subthalamic nucleus and its effective connectivity with the right anterior insula predicts children's inhibitory control. Fifth, we replicate our findings across multiple cohorts. Our findings provide insights into cortical-basal ganglia circuits and global brain organization underlying the development of inhibitory control.},
}
@article{CaietalNATURE-METHODS-13,
       author = {Cai, Dawen and Cohen, Kimberly B. and Luo, Tuanlian and Lichtman, Jeff W. and Sanes, Joshua R.},
        title = {Improved tools for the Brainbow toolbox},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = {10},
        issue = {6},
         year = {2013},
        pages = {540-547},
     abstract = {In the transgenic multicolor labeling strategy called 'Brainbow', Cre-loxP recombination is used to create a stochastic choice of expression among fluorescent proteins, resulting in the indelible marking of mouse neurons with multiple distinct colors. This method has been adapted to non-neuronal cells in mice and to neurons in fish and flies, but its full potential has yet to be realized in the mouse brain. Here we present several lines of mice that overcome limitations of the initial lines, and we report an adaptation of the method for use in adeno-associated viral vectors. We also provide technical advice about how best to image Brainbow-expressing tissue.},
}
@article {CaietalPNAS-09,
       author = {Cai, Denise J. and Mednick, Sarnoff A. and Harrison, Elizabeth M. and Kanady, Jennifer C. and Mednick, Sara C.},
        title = {{REM}, not incubation, improves creativity by priming associative networks},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {106},
       number = {25},
         year = {2009},
        pages = {10130-10134},
     abstract = {The hypothesized role of rapid eye movement (REM) sleep, which is rich in dreams, in the formation of new associations, has remained anecdotal. We examined the role of REM on creative problem solving, with the Remote Associates Test (RAT). Using a nap paradigm, we manipulated various conditions of prior exposure to elements of a creative problem. Compared with quiet rest and non-REM sleep, REM enhanced the formation of associative networks and the integration of unassociated information. Furthermore, these REM sleep benefits were not the result of an improved memory for the primed items. This study shows that compared with quiet rest and non-REM sleep, REM enhances the integration of unassociated information for creative problem solving, a process, we hypothesize, that is facilitated by cholinergic and noradrenergic neuromodulation during REM sleep.},
}
@article{CainandShea-BrownCOIN-12,
       author = {Nicholas Cain and Eric Shea-Brown},
        title = {Computational models of decision making: integration, stability, and noise},
      journal = {Current Opinion in Neurobiology},
       volume = {22},
       number = {6},
         year = {2012},
        pages = {1047-1053},
     abstract = {Decision making demands the accumulation of sensory evidence over time. Questions remain about how this occurs, but recent years have seen progress on several fronts. The first concerns when optimal accumulation of evidence coincides with the simplest method of accumulating neural activity: summation over time. The second involves what computations the brain might perform when summation is difficult due to imprecision in neural circuits or is suboptimal due to uncertainty or variability in how evidence arrives. Finally, the third concerns sources of noise in decision circuits. Empirical studies have better constrained the extent of this noise, and modeling work is helping to clarify its possible origins.}
}
@inproceedings{CainesandWangDC-90,
       author = {Peter E. Caines and S. Wang},
        title = {{COCOLOG}: A Conditional Observer and Controller Logic for Finite Machines},
    booktitle = {Proceedings of the 29th {IEEE} Conference on Decision and Control},
      address = {Hawaii},
         year = 1990,
}
@article {CalabreseetalPHYSIOLOGY-06,
       author = {Calabrese, Barbara and Wilson, Margaret S. and Halpain, Shelley},
        title = {Development and Regulation of Dendritic Spine Synapses},
      journal = {Physiology},
    publisher = {American Physiological Society},
       volume = {21},
       number = {1},
         year = {2006},
        pages = {38-47},
     abstract = {Dendritic spines are small protrusions from neuronal dendrites that form the postsynaptic component of most excitatory synapses in the brain. They play critical roles in synaptic transmission and plasticity. Recent advances in imaging and molecular technologies reveal that spines are complex, dynamic structures that contain a dense array of cytoskeletal, transmembrane, and scaffolding molecules. Several neurological and psychiatric disorders exhibit dendritic spine abnormalities.},
}
@article{CalhounetalLRE-10,
       author = {Calhoun, Sasha and Carletta, Jean and Brenier, Jason M. and Mayo, Neil and Jurafsky, Dan and Steedman, Mark and Beaver, David},
        title = {The NXT-format Switchboard Corpus: A Rich Resource for Investigating the Syntax, Semantics, Pragmatics and Prosody of Dialogue},
      journal = {Language Resources and Evaluation},
       volume = 44,
        issue = 4,
         year = 2010,
        pages = {387-419},
    publisher = {Springer-Verlag New York, Inc.},
      address = {Secaucus, NJ, USA},
}
@article{CaligioreetalCEREBELLUM-17,
       author = {Caligiore, D. and Pezzulo, G. and Baldassarre, G. and Bostan, A. C. and Strick, P. L. and Doya, K. and Helmich, R. C. and Dirkx, M. and Houk, J. and Jorntell, H. and Lago-Rodriguez, A. and Galea, J. M. and Miall, R. C. and Popa, T. and Kishore, A. and Verschure, P. F. and Zucca, R. and Herreros, I.},
        title = {{C}onsensus Paper: {T}owards a Systems-Level View of Cerebellar Function: the Interplay Between Cerebellum, Basal Ganglia, and Cortex},
      journal = {Cerebellum},
       volume = {16},
       number = {1},
         year = {2017},
        pages = {203-229},
     abstract = {Despite increasing evidence suggesting the cerebellum works in concert with the cortex and basal ganglia, the nature of the reciprocal interactions between these three brain regions remains unclear. This consensus paper gathers diverse recent views on a variety of important roles played by the cerebellum within the cerebello-basal ganglia-thalamo-cortical system across a range of motor and cognitive functions. The paper includes theoretical and empirical contributions, which cover the following topics: recent evidence supporting the dynamical interplay between cerebellum, basal ganglia, and cortical areas in humans and other animals; theoretical neuroscience perspectives and empirical evidence on the reciprocal influences between cerebellum, basal ganglia, and cortex in learning and control processes; and data suggesting possible roles of the cerebellum in basal ganglia movement disorders. Although starting from different backgrounds and dealing with different topics, all the contributors agree that viewing the cerebellum, basal ganglia, and cortex as an integrated system enables us to understand the function of these areas in radically different ways. In addition, there is unanimous consensus between the authors that future experimental and computational work is needed to understand the function of cerebellar-basal ganglia circuitry in both motor and non-motor functions. The paper reports the most advanced perspectives on the role of the cerebellum within the cerebello-basal ganglia-thalamo-cortical system and illustrates other elements of consensus as well as disagreements and open questions in the field.}
}
@article{CallaertsetalARN-97,
        title = {{PAX-6} in development and evolution},
       author = {Callaerts, P. and Halder, G. and Gehring, W. J.},
      journal = {Annual Review of Neuroscience},
        pages = {483-532},
       volume = 20,
         year = 1997,
     abstract = {Pax-6 is a member of the Pax gene class and encodes  a protein containing a paired domain and a homeodomain. The molecular characterization of Pax-6 genes from species of different animal phyla and the analysis of Pax-6 function in the developing eyes and central nervous system of vertebrates  Drosophila melanogaster, and Caenorhabditis elegans suggest that Pax-6 homologues share conserved functions. In this review, we present recent data on the structural and functional characterization of Pax-6 homologues from species of different animal phyla. We discuss the implications of these findings for our understanding of the development and evolution of eyes and nervous systems.},
}
@article{CallanetalCEREBELLUM-07,
       author = {Callan, Daniel E. and Kawato, Mitsuo and Parsons, Lawrence and Turner, Robert},
        title = {Speech and song: The role of the cerebellum},
      journal = {The Cerebellum},
         year = {2007},
       volume = {6},
       number = {4},
        pages = {321-327},
     abstract = {An exploration into cerebellar activity during the perception and production of speech and song may elucidate general underlying cerebellar functions. Recently, the cerebellum has been hypothesized to be involved with sharpening sensory input, temporal coordination and processing of motor articulation and perception, as well as instantiation of internal models that simulate the input-output characteristics of a specific system. Sung language and spoken language share many common features (physiology for articulation and perception as well as phonology, phonotactics, syntax, and semantics of the underlying language), although they differ in certain vocal and prosodic aspects. A review of the literature on perception and production of singing and speech reveals considerable overlap in the lateral aspect of the VI lobule of the posterior cerebellum, a region known to somatotopically represent the lips and tongue. This region may instantiate internal models of vocal tract articulation that simulate well learned phonological and/or segmental articulatory---auditory/orosensory mappings utilized for both speech and singing. Recent results show tendencies for left cerebellar hemispheric specialization for processing of singing and right specialization for processing of speech, both in the VI lobule of the cerebellum, inferior to that found for representing both speech and singing. Given the crossed pattern of cerebellar-cortical anatomical connectivity the findings are consistent with the hypothesis that the right cerebellum differentially processes high pass filtered information (segmental properties) and the left cerebellum differentially processes low pass filtered information (prosodic, melodic properties). Further research is necessary to examine these hypotheses and their alternatives directly.},
}
@article{CallanetalJAMH-15,
       author = {Judith A. Callan and Greg J. Siegle and Kaleab Abebe and Beverly Black and Lynn Martire and Richard Schulz and Charles Reynolds and Martica H. Hall},
        title = {Feasibility of a pocket-PC based cognitive control intervention in dementia spousal caregivers},
      journal = {Aging and Mental Health},
       volume = 19,
        issue = 8,
         year = 2015,
        pages = {1364-6915},
     abstract = {Spousal caregivers of patients with dementia are in need of interventions to bolster their quality of life. Computer-based, self-administered cognitive training is an innovative approach to target spousal caregiver distress and coping. We tested the feasibility of administering one such intervention with minimal clinician intervention. Twenty-seven elderly adults (>64 years old), who each were the primary caregiver for a spouse with dementia, were recruited through the Memory Disorders Clinic of the Alzheimer Disease Research Center in Pittsburgh, PA. Spousal caregivers were instructed to use a handheld computer version of the Adaptive Paced Visual Serial Attention Task (APVSAT) at least three times per week for four weeks as part of a larger caregiver intervention trial. Feasibility was explored by examining the frequency of APVSAT usage. Results suggest that self-directed cognitive training is feasible for spousal caregivers of dementia patients. The mean usage of the APVSAT was 42 (SD = 28.58). Performance increased from the beginning to the end of the trial, and usage was not affected by stress, worry, or poor sleep quality. Findings suggest the potential utility of cognitive training via handheld computer for spousal caregivers of dementia patients to improve problem solving, coping and adaptation, planning, and persevering with goal-directed tasks.}
}
@article{CallawayCOIN-08,
        title = {Transneuronal circuit tracing with neurotropic viruses},
       author = {Callaway, E.M.},
      journal = {Current Opinion in Neurobiology},
       volume = 18,
        issue = 6,
         year = 2008,
        pages = {617-23},
     abstract = {Because neurotropic viruses naturally traverse neural pathways, they are extremely valuable for elucidating neural circuits. Naturally occurring herpes and rabies viruses have been used for transneuronal circuit tracing for decades. Depending on the type of virus and strain, virus can travel preferentially in the anterograde or the retrograde direction. More recently, genetic modifications have allowed for many improvements. These include: reduced pathogenicity; addition of marker genes; control of synaptic spread; pseudotyping for infection of selected cells; addition of ancillary genetic elements for combining circuit tracing with manipulation of activity or functional assays. These modifications, along with the likelihood of future developments, suggest that neurotropic viruses will be increasingly important and effective tools for future studies of neural circuits.}
}
@article{CallawayCURRENT-OPINION-08,
       author = {Callaway, Edward M.},
        title = {Transneuronal circuit tracing with neurotropic viruses},
      journal = {Current opinion in neurobiology},
       volume = {18},
        issue = {6},
         year = {2008},
        pages = {617-623},
     abstract = {Because neurotropic viruses naturally traverse neural pathways, they are extremely valuable for elucidating neural circuits. Naturally occurring herpes and rabies viruses have been used for transneuronal circuit tracing for decades. Depending on the type of virus and strain, virus can travel preferentially in the anterograde or the retrograde direction. More recently, genetic modifications have allowed for many improvements. These include: reduced pathogenicity; addition of marker genes; control of synaptic spread; pseudotyping for infection of selected cells; addition of ancillary genetic elements for combining circuit tracing with manipulation of activity or functional assays. These modifications, along with the likelihood of future developments, suggest that neurotropic viruses will be increasingly important and effective tools for future studies of neural circuits.},
}
@article{CallawayandYusteNEUROBIOLOGY-02,
        title = {Stimulating neurons with light},
       author = {Callaway, E.M. and Yuste, R.},
      journal = {Current Opinion in Neurobiology},
       volume = 12,
        issue = 5,
         year = 2002,
        pages = {587-92},
     abstract = {Recent technological advances have enabled the use of different optical methods to activate neurons, including 'caged' glutamate, photoactivation of genetically engineered cascades, and direct two-photon excitation. The ability to use light as a stimulation tool provides, in principle, a non-invasive method for the temporally and spatially precise activation of any neuron or any part of a neuron. When combined with two-photon excitation, excellent spatial control can be achieved even in complex and highly scattering preparations, such as living nervous tissue. Different methods that have been developed in the last several decades have been used to probe neuronal sensitivity, mimic synaptic input, and elucidate patterns of neural connectivity.}
}
@book{CalvoandSymonsTACO-14,
       editor = {Paco Calvo and John Symons},
        title = {The Architecture of Cognition: Rethinking Fodor and Pylyshyn's Systematicity Challenge},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 2014,
     abstract = {In 1988, Jerry Fodor and Zenon Pylyshyn challenged connectionist theorists to explain the systematicity of cognition. In a highly influential critical analysis of connectionism, they argued that connectionist explanations, at best, can only inform us about details of the neural substrate; explanations at the cognitive level must be classical insofar as adult human cognition is essentially systematic. More than twenty-five years later, however, conflicting explanations of cognition do not divide along classicist-connectionist lines, but oppose cognitivism (both classicist and connectionist) with a range of other methodologies, including distributed and embodied cognition, ecological psychology, enactivism, adaptive behavior, and biologically based neural network theory. This volume reassesses Fodor and Pylyshyn's "systematicity challenge" for a post-connectionist era.}
}
@techreport{CameronandDurrant-WhyteTR-88,
       author = {Cameron, Alec and Durrant-Whyte, Hugh F.},
        title = {A {B}ayesian Approach to Optimal Sensor Placement},
  institution = {Oxford University Robotics Research Group},
         year = 1988,
}
@article{CamillaandWarrenBRAIN-15,
       author = {Clark, Camilla N. and Warren, Jason D.},
        title = {Music, memory and mechanisms in Alzheimer's disease},
      journal = {Brain},
       volume = {138},
       number = {8},
         year = {2015},
        pages = {2122-2125},
     abstract = {This scientific commentary refers to 'Why musical memory can be preserved in advanced Alzheimer's disease', by Jacobsen et al. (doi:10.1093/brain/awv135).The power of music to unlock memories and other cognitive capacities in Alzheimer's disease is a cherished tenet of clinical neurology, and music is unquestionably a welcome source of comfort to many people with this devastating illness. Scientific validation of clinical instinct, however, is rarely straightforward. Though ubiquitous, music is hard to study scientifically; until comparatively recently, we lacked the conceptual framework to analyse it as a neuropsychological phenomenon and the practical tools to examine brain structures and processes that allow us to experience it. Thanks in part to the current ascendancy of functional neuroimaging, these tools are now to hand and the time is ripe to take music neuroscience to the clinic. The work of Jacobsen and colleagues reported in this issue of Brain (Jacobsen et al., 2015) contributes an interesting fresh perspective on why musical memories might survive the predations of dementia.},
}
@misc{Campbell-Kelly-Programming_the_EDSAC,
        title = {Programming the EDSAC, 1946-1949},
       author = {Martin Campbell-Kelly},
  institution = {Warwick University},
         year = {2017},
 howpublished = {http://chiphack.org/talks/edsac-part-2.pdf}
}
@misc{Campbell-Kelly_Building_the_EDSAC,
        title = {Building the EDSAC, 1946-1949},
       author = {Martin Campbell-Kelly},
  institution = {Warwick University},
         year = {2017},
 howpublished = {http://chiphack.org/talks/edsac-part-1.pdf}
}
@book{Campbell2016,
        title = {The China Study Solution: The Simple Way to Lose Weight and Reverse Illness, Using a Whole-Food, Plant-Based Diet},
       author = {Campbell, T.},
         year = {2016},
    publisher = {Potter/Ten Speed/Harmony/Rodale}
}
@article{CampbellandTurnerCURRENT-BIOLOGY-10,
       author = {Robert A.A. Campbell and Glenn C. Turner},
        title = {The mushroom body},
      journal = {Current Biology},
       volume = {20},
       number = {1},
        pages = {R11-R12},
         year = {2010},
     abstract = {The mushroom body is a prominent and striking structure in the brain of several invertebrates, mainly arthropods. It is found in insects, scorpions, spiders, and even segmented worms. With its long stalk crowned with a cap of cell bodies, a GFP-labeled mushroom body certainly lives up to its name (Figure 1). The mushroom body is composed of small neurons known as Kenyon cells, named after Frederick Kenyon, who first applied the Golgi staining technique to the insect brain. The honey bee brain, for instance, contains roughly 175,000 neurons per mushroom body while the brain of the smaller fruit fly Drosophila melanogaster only possesses about 2,500. Kenyon cells thus make up 20\% and 2\%, respectively, of the total number of neurons in each insect's brain. Kenyon cell bodies sit atop the calyx, a tangled zone of synapses representing the site of sensory input. Projecting away from the calyx is the stalk comprised of Kenyon cell axons carrying information away to the output lobes. }
}
@incollection{Campos-OrtegaandStrausfeldIPVSA-72,
       author = {Campos-Ortega, J. A. and Strausfeld, N. J.},
       editor = {Wehner, R\"{u}diger},
        title = {Columns and Layers in the Second Synaptic Region of the Fly's Visual System: The Case for Two Superimposed Neuronal Architectures},
    booktitle = {Information Processing in the Visual Systems of Anthropods},
     location = {Symposium at Department of Zoology, University of Zurich},
    publisher = {Springer Berlin Heidelberg},
      address = {Berlin, Heidelberg},
         year = {1972},
        pages = {31-36},
}
@techreport{CamusTR-90,
       author = {Camus, Theodore},
        title = {Applications of Pyramid Structures to Multiscale Optical Flow},
        issue = {CS-90-09},
  institution = {Brown University Department of Computer Science},
         year = 1990,
}
@inproceedings{CamusandBulthoffSPIE-91,
       author = {Camus, Theodore and B\"{u}lthoff, Heinrich},
        title = {Space-Time Tradeoffs for Adaptive Real-Time Tracking},
    booktitle = {Proceedings of the {SPIE} Conference on Advances in Intelligent Robotic Systems},
 organization = {SPIE},
        pages = {268-276},
         year = 1991,
}
@inproceedings{CamusetalIUW-93,
       author = {Camus, Theodore and Monsarrat, Jonathan and Dean, Thomas},
        title = {Planning and Selective Perception for Mobile Robot Object Retrieval Tasks},
    booktitle = {Proceedings of the DARPA Image Understanding Workshop},
     location = {Washington, DC},
 organization = {DARPA},
         year = 1993,
}
@article{CandyandBanksVR-99,
     author = {Candy, T.R. and Banks, M.S.}, 
      title = {Use of an early nonlinearity to measure optical and receptor resolution in the human infant},
    journal = {Vision Research}, 
     volume = 39, 
      pages = {3386-3398},
       year = 1999, 
    comment = {Candy, T.R. and Banks, M.S. (1999). Use of an early nonlinearity to measure optical 
               and receptor resolution in the human infant. Vision Research, 39, 3386-3398.},
}
@article{CangandWeiCoRR-17,
       author = {Zixuan Cang, Guo-Wei Wei},
        title = {TopologyNet: Topology based deep convolutional neural networks for biomolecular property predictions},
      journal = {CoRR},
       volume = {arXiv:1704.00063},
         year = {2017},
          url = {https://arxiv.org/abs/1704.00063},
     abstract = {Although deep learning approaches have had tremendous success in image, video and audio processing, computer vision, and speech recognition, their applications to three-dimensional (3D) biomolecular structural data sets have been hindered by the entangled geometric complexity and biological complexity. We introduce topology, i.e., element specific persistent homology (ESPH), to untangle geometric complexity and biological complexity. ESPH represents 3D complex geometry by one-dimensional (1D) topological invariants and retains crucial biological information via a multichannel image representation. It is able to reveal hidden structure-function relationships in biomolecules. We further integrate ESPH and convolutional neural networks to construct a multichannel topological neural network (TopologyNet) for the predictions of protein-ligand binding affinities and protein stability changes upon mutation. To overcome the limitations to deep learning arising from small and noisy training sets, we present a multitask topological convolutional neural network (MT-TCNN). We demonstrate that the present TopologyNet architectures outperform other state-of-the-art methods in the predictions of protein-ligand binding affinities, globular protein mutation impacts, and membrane protein mutation impacts.},
}
@book{Cannon67,
   author =    {Cannon, Robert H.},
   title =     {Dynamics of Physical Systems},
   publisher = {McGraw-Hill},
   address =   {New York},
   year =      1967,
}
@article{CannyPAMI-86,
    author = {John Canny},
     title = {A Computational Approach to Edge Detection}, 
   journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence}, 
    volume = 8, 
     issue = 6, 
     month = {November}, 
     pages = {679-698}, 
      year = 1986, 
}
@article{CanoltyetalPNAS-10,
     title = {Oscillatory phase coupling coordinates anatomically dispersed functional cell assemblies},
    author = {Canolty, Ryan T. and Ganguly, Karunesh and Kennerley, Steven W. and Cadieu, Charles F. and Koepsell, Kilian and Wallis, Jonathan D. and Carmena, Jose M.},
   journal = {Proceedings of the National Academy of Sciences},
    volume = 107,
     issue = 40,
      year = 2010,
     pages = {17356-17361},
  abstract = {Hebb proposed that neuronal cell assemblies are critical for effective perception, cognition, and action. However, evidence for brain mechanisms that coordinate multiple coactive assemblies remains lacking. Neuronal oscillations have been suggested as one possible mechanism for cell assembly coordination. Prior studies have shown that spike timing depends upon local field potential (LFP) phase proximal to the cell body, but few studies have examined the dependence of spiking on distal LFP phases in other brain areas far from the neuron or the influence of LFP--LFP phase coupling between distal areas on spiking. We investigated these interactions by recording LFPs and single-unit activity using multiple microelectrode arrays in several brain areas and then used a unique probabilistic multivariate phase distribution to model the dependence of spike timing on the full pattern of proximal LFP phases, distal LFP phases, and LFP--LFP phase coupling between electrodes. Here we show that spiking activity in single neurons and neuronal ensembles depends on dynamic patterns of oscillatory phase coupling between multiple brain areas, in addition to the effects of proximal LFP phase. Neurons that prefer similar patterns of phase coupling exhibit similar changes in spike rates, whereas neurons with different preferences show divergent responses, providing a basic mechanism to bind different neurons together into coordinated cell assemblies. Surprisingly, phase-coupling-based rate correlations are independent of interneuron distance. Phase-coupling preferences correlate with behavior and neural function and remain stable over multiple days. These findings suggest that neuronal oscillations enable selective and dynamic control of distributedfunctional cell assemblies.},
}
@article{CantleyetalACS-18,
       author = {Cantley, William L. and Du, Chuang and Lomoio, Selene and DePalma, Thomas and Peirent, Emily and Kleinknecht, Dominic and Hunter, Martin and Tang-Schomer, Min D. and Tesco, Giuseppina and Kaplan, David L.},
        title = {Functional and Sustainable 3D Human Neural Network Models from Pluripotent Stem Cells},
      journal = {ACS Biomaterials Science \& Engineering},
    publisher = {American Chemical Society},
         year = {2018},
     abstract = {Three-dimensional (3D) in vitro cell and tissue culture models, particularly for the central nervous system, allow for the exploration of mechanisms of organ development, cellular interactions, and disease progression within defined environments. Here, we describe the development and characterization of human 3D tissue models that promote the differentiation and long-term survival of functional neural networks. This work builds upon previous work where primary rodent neurons were successfully grown in a similar 3D system.(1) The model was adapted to human induced pluripotent stem cells, allowing for a more direct exploration of the human condition. These tissue cultures show diverse cell populations, including neurons and astroglial cells, interacting in 3D and exhibit spontaneous neural activity confirmed through electrophysiological recordings and calcium imaging over at least nine months. This approach allows for the direct integration of pluripotent stem cells into the 3D construct, bypassing early neural differentiation steps (embryoid bodies and neural rosettes). The streamlined process, in combination with the longevity of the cultures, provides a system that can be manipulated to support a variety of experimental applications, including the study of network development, maturation, plasticity, and/or degeneration. This tissue model was tested with stem cells derived from healthy individuals as well as Alzheimer's and Parkinson's disease patients. We observed similar growth and gene expression, which indicates the feasibility of generating patient-derived brain tissue models. These could be used to uncover early stage biomarkers of the disease state, in turn supporting earlier diagnosis and improving understanding of disease progression. With additional model development, this approach would have potential use for investigating drug targets in neurodegenerative diseases.}
}
@article{CaoetalCELL-13,
     title = {Genetically Targeted Optical Electrophysiology in Intact Neural Circuits},
    author = {Cao, Guan and Platisa, Jelena and Pieribone, Vincent A. and Raccuglia, Davide and Kunst, Michael and Nitabach, Michael N.},
   journal = {Cell},
    volume = 154,
     issue = 4,
      year = 2013,
     pages = {904-913},
  abstract = {Nervous systems process information by integrating the electrical activity of neurons in complex networks. This motivates the long-standing interest in using optical methods to simultaneously monitor the membrane potential of multiple genetically targeted neurons via expression of genetically encoded fluorescent voltage indicators (GEVIs) in intact neural circuits. No currently available GEVIs have demonstrated robust signals in intact brain tissue that enable reliable recording of individual electrical events simultaneously in multiple neurons. Here, we show that the recently developed "ArcLight" GEVI robustly reports both subthreshold events and action potentials in genetically targeted neurons in the intact Drosophila fruit fly brain and reveals electrical signals in neurite branches. In the same way that genetically encoded fluorescent sensors have revolutionized the study of intracellular Ca2+ signals, ArcLight now enables optical measurement in intact neural circuits of membrane potential, the key cellular parameter that underlies neuronal information processing.}
}
@article{CaporaleandDanARN-08,
       author = {Natalia Caporale and Yang Dan},
        title = {Spike Timing-Dependent Plasticity: A {Hebbian} Learning Rule},
      journal = {Annual Review of Neuroscience},
       volume = {31},
        issue = {1},
         year = {2008},
        pages = {25-46},
     abstract = {Spike timing-dependent plasticity (STDP) as a Hebbian synaptic learning rule has been demonstrated in various neural circuits over a wide spectrum of species, from insects to humans. The dependence of synaptic modification on the order of pre- and postsynaptic spiking within a critical window of tens of milliseconds has profound functional implications. Over the past decade, significant progress has been made in understanding the cellular mechanisms of STDP at both excitatory and inhibitory synapses and of the associated changes in neuronal excitability and synaptic integration. Beyond the basic asymmetric window, recent studies have also revealed several layers of complexity in STDP, including its dependence on dendritic location, the nonlinear integration of synaptic modification induced by complex spike trains, and the modulation of STDP by inhibitory and neuromodulatory inputs. Finally, the functional consequences of STDP have been examined directly in an increasing number of neural circuits in vivo.}
}
@article{CarandiniNATURE-12,
     title = {From circuits to behavior: a bridge too far?},
    author = {Matteo Carandini},
   journal = {Nature Neuroscience},
    volume = 15,
     issue = 4,
     pages = {507-509},
      year = 2012,
  abstract = {Neuroscience seeks to understand how neural circuits lead to behavior. However, the gap between circuits and behavior is too wide. An intermediate level is one of neural computations, which occur in individual neurons and populations of neurons. Some computations seem to be canonical: repeated and combined in different ways across the brain. To understand neural computations, we must record from a myriad of neurons in multiple brain regions. Understanding computation guides research in the underlying circuits and provides a language for theories of behavior.}
}
@article{CarandiniandHeegerNATURE-12,
       author = {Matteo Carandini and David J. Heeger},
        title = {Normalization as a canonical neural computation},
      journal = {Nature Reviews Neuroscience},
       volume = 13,
         year = 2012,
        pages = {51-62},
     abstract = {There is increasing evidence that the brain relies on a set of canonical neural computations, repeating them across brain regions and modalities to apply similar operations to different problems. A promising candidate for such a computation is normalization, in which the responses of neurons are divided by a common factor that typically includes the summed activity of a pool of neurons. Normalization was developed to explain responses in the primary visual cortex and is now thought to operate throughout the visual system, and in many other sensory modalities and brain regions. Normalization may underlie operations such as the representation of odours, the modulatory effects of visual attention, the encoding of value and the integration of multisensory information. Its presence in such a diversity of neural systems in multiple species, from invertebrates to mammals, suggests that it serves as a canonical neural computation.},
}
@article{CarandinietalJN-97,
       author = {Matteo Carandini and David J. Heeger and J. Anthony Movshon},
        title = {Linearity and Normalization in Simple Cells of the Macaque Primary Visual Cortex},
      journal = {Journal of Neuroscience},
       volume = 17,
         year = 1997,
        pages = {8621-8644},
}
@article{CarlenSCIENCE-17,
       author = {Carl\`{e}n, Marie},
        title = {What constitutes the prefrontal cortex?},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {358},
       number = {6362},
         year = {2017},
        pages = {478-482},
     abstract = {During evolution, the prefrontal region grew in size relative to the rest of the cortex. It reached its largest extent in the human brain, where it constitutes 30\% of the total cortical area. This growth was accompanied by phylogenetic differentiation of the cortical areas. It has been argued that the human brain holds prefrontal regions that are both qualitatively and functionally unique. Present-day neuroscientists studying the prefrontal cortex increasingly use mice. An important goal is to reveal how the prefrontal cortex enables complex behavior. However, the prefrontal cortex still lacks a conclusive definition. The structure and function of this brain area across species remain unresolved. This state of affairs is often overlooked, warranting renewed focus on what the prefrontal cortex is and does.},
}
@article {CarleoandTroyerSCIENCE-17,
       author = {Carleo, Giuseppe and Troyer, Matthias},
        title = {Solving the quantum many-body problem with artificial neural networks},
       volume = {355},
      journal = {Science},
       number = {6325},
    publisher = {American Association for the Advancement of Science},
         year = {2017},
        pages = {602-606},
     abstract = {Elucidating the behavior of quantum interacting systems of many particles remains one of the biggest challenges in physics. Traditional numerical methods often work well, but some of the most interesting problems leave them stumped. Carleo and Troyer harnessed the power of machine learning to develop a variational approach to the quantum many-body problem (see the Perspective by Hush). The method performed at least as well as state-of-the-art approaches, setting a benchmark for a prototypical two-dimensional problem. With further development, it may well prove a valuable piece in the quantum toolbox.Science, this issue p. 602; see also p. 580The challenge posed by the many-body problem in quantum physics originates from the difficulty of describing the nontrivial correlations encoded in the exponential complexity of the many-body wave function. Here we demonstrate that systematic machine learning of the wave function can reduce this complexity to a tractable computational form for some notable cases of physical interest. We introduce a variational representation of quantum states based on artificial neural networks with a variable number of hidden neurons. A reinforcement-learning scheme we demonstrate is capable of both finding the ground state and describing the unitary time evolution of complex interacting quantum systems. Our approach achieves high accuracy in describing prototypical interacting spins models in one and two dimensions.},
}
@article{CarlsonBC-90,
       author = {A. Carlson},
        title = {Anti-Hebbian learning in a non-linear neural network},
      journal = {Biological Cybernetics},
       volume = 64,
        pages = {171-176},
         year = 1990,
}
@article{CarlssonBAMS-09,
       author = {Gunnar Carlsson},
        title = {Topology and Data},
      journal = {Bulletin of the American Mathematical Society},
       volume = {46},
       number = {2},
         year = {2009},
        pages = {255-308},
}
@book{Carpenter1986turing,
        title = {A.M. Turing's {ACE} report of 1946 and other papers},
       author = {Carpenter, B.E. and Doran, R.W. and Turing, A.M. and Woodger, M.},
       series = {Charles Babbage Institute reprint series for the history of computing},
         year = {1986},
    publisher = {MIT Press}
}
@incollection{CarpenterART-17,
       author = {Carpenter, Gail},
        title = {Adaptive Resonance Theory},
       editor = {C. Sammut and G.I. Webb},
    booktitle = {Encyclopedia of Machine Learning and Data Mining},
    publisher = {Springer Science+Business Media New York},
         year = {2017},
     abstract = {Computational models based on cognitive and neural systems are now deeply embedded in the standard repertoire of machine learning and data mining methods, with intelligent learning systems enhancing performance in nearly every existing application area. Beyond data mining, this article shows how models based on adaptive resonance theory (ART) may provide entirely new questions and practical solutions for technological applications. ART models carry out hypothesis testing, search, and incremental fast or slow, self-stabilizing learning, recognition, and prediction in response to large nonstationary databases (big data). Three computational examples, each based on the distributed ART neural network, frame questions and illustrate how a learning system (each with no free parameters) may enhance the analysis of large-scale data. Performance of each task is simulated on a common mapping platform, a remote sensing dataset called the Boston Testbed, available online along with open-source system code. Key design elements of ART models and links to software for each system are included. The article further points to future applications for integrative ART-based systems that have already been computationally specified and simulated. New application directions include autonomous robotics, general-purpose machine vision, audition, speech recognition, language acquisition, eye movement control, visual search, figure-ground separation, invariant object recognition, social cognition, object and spatial attention, scene understanding, spacetime integration, episodic memory, navigation, object tracking, system-level analysis of mental disorders, and machine consciousness.}
}
@inproceedings{CarpenteretalRTSS-94,
       author = {T. Carpenter and K. Driscoll and K. Hoyme and J. Carciofini},
        title = {{ARINC} 659 Scheduling: Problem Definition},
    booktitle = {Proceedings of the 15th {IEEE} Real-Time Systems Symposium},
     location = {San Juan, Puerto Rico},
         year = 1994,
}
@article{CarretalNATURE-NEUROSCIENCE-11,
        title = {Hippocampal replay in the awake state: a potential substrate for memory consolidation and retrieval},
       author = {Carr, Margaret F and Jadhav, Shantanu P and Frank, Loren M},
      journal = {Nature neuroscience},
       volume = {14},
       number = {2},
        pages = {147},
         year = {2011},
    publisher = {Nature Publishing Group}
}
@article {Carrillo-ReidetalSCIENCE-16,
       author = {Carrillo-Reid, Luis and Yang, Weijian and Bando, Yuki and Peterka, Darcy S. and Yuste, Rafael},
        title = {Imprinting and recalling cortical ensembles},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {353},
       number = {6300},
         year = {2016},
        pages = {691-694},
     abstract = {Donald Hebb's hypothesis that coactivation of neurons leads to the formation of ensembles of neurons has inspired neuroscientists for decades. The experimental creation of such ensembles has been technically challenging. Using two-photon optogenetic stimulation with single-cell resolution, Carrillo-Reid et al. discovered that recurrent activation of a group of neurons creates an ensemble that is imprinted in the brain circuitry. Activation of a single neuron can lead to recall of the entire ensemble in a phenomenon called pattern completion. The artificial ensemble persists over days and can be reactivated at later time points without interfering with endogenous circuitry. [Science, this issue p. 691] Neuronal ensembles are coactive groups of neurons that may represent building blocks of cortical circuits. These ensembles could be formed by Hebbian plasticity, whereby synapses between coactive neurons are strengthened. Here we report that repetitive activation with two-photon optogenetics of neuronal populations from ensembles in the visual cortex of awake mice builds neuronal ensembles that recur spontaneously after being imprinted and do not disrupt preexisting ones. Moreover, imprinted ensembles can be recalled by single-cell stimulation and remain coactive on consecutive days. Our results demonstrate the persistent reconfiguration of cortical circuits by two-photon optogenetics into neuronal ensembles that can perform pattern completion.},
}
@book{Carroll2005,
        title = {Endless forms most beautiful: the new science of evo devo and the making of the animal kingdom},
       author = {Carroll, Sean B.},
    publisher = {W.W. Norton \& Co.},
         year = 2005,
}
@book{Carroll2016,
        title = {The Big Picture: On the Origins of Life, Meaning, and the Universe Itself},
       author = {Carroll, Sean},
    publisher = {Penguin Publishing Group},
         year = {2016},
}
@book{CarrollLIFE_MEANING_UNIVERSE-16,
        title = {The Big Picture: On the Origins of Life, Meaning, and the Universe Itself},
       author = {Carroll, Sean},
    publisher = {Penguin Publishing Group},
         year = {2016},
}
@article{CarstensenandFrankCOiBS-20,
       author = {Carstensen, Alexandra and Frank, Michael C},
        title = {Do graded representations support abstract thought?},
      journal = {Current Opinion in Behavioral Sciences},
       volume = {37},
        pages = {90-97},
         year = {2020},
     abstract = {Relational reasoning requires the reasoner to go beyond her/his specific experience, abstracting from items to make inferences about categories and kinds on the basis of structural or analogical similarities. Reasoning about the relations same and different is one of the best-studied cases of relational reasoning, both across species and over human development, and has become a paradigm case study for abstract representation. However, decades of careful study have nonetheless produced seemingly contradictory findings - with surprising successes and puzzling failures - in both the comparative and developmental literatures. In this article, we review these literatures and suggest first steps toward a reconciliation of these contradictions by suggesting that samedifferent reasoning is supported by graded representations.}
}
@article {CartaetalSCIENCE-19,
       author = {Carta, Ilaria and Chen, Christopher H. and Schott, Amanda L. and Dorizan, Schnaude and Khodakhah, Kamran},
        title = {Cerebellar modulation of the reward circuitry and social behavior},
    publisher = {American Association for the Advancement of Science},
      journal = {Science},
       volume = {363},
         year = {2019},
       number = {6424},
     abstract = {Although the cerebellum has long been considered to be a purely motor structure, recent studies have revealed that it also has critical nonmotor functions. Cerebellar dysfunction is implicated in addictive behavior and in mental disorders such as autism spectrum disorder (ASD), cognitive affective syndrome, and schizophrenia. The cerebellum is well poised to contribute to behavior because it receives a wide array of cortical and sensory information and is subject to control by a number of neuromodulators. To perform its function, the cerebellum is believed to integrate these diverse inputs to provide the rest of the brain with predictions required for optimal behavior. Although there are many pathways for this to occur in the motor domain, fewer exist for the nonmotor domain. In mice, we found that monosynaptic excitatory projections from the cerebellar nuclei to the VTA powerfully activate the reward circuitry and contribute to social behavior. Using anatomical tracing, we showed that axonal projections from the cerebellar nuclei form synapses with both dopaminergic and nondopaminergic neurons in the VTA. The cerebello-VTA (Cb-VTA) projections were powerful and their optogenetic stimulation robustly increased the activity of VTA neurons both in vivo and in vitro. Behavioral tests to examine reward processing showed that stimulation of the Cb-VTA projections was sufficient to cause short-term and long-term place preference, thereby demonstrating that the pathway was rewarding. Although optogenetic inhibition of Cb-VTA projections was not aversive, it completely abolished social preference in the three-chamber test for sociability, which suggests that the cerebellar input to the VTA is required for normal social behavior.}
}
@incollection{CasagrandeandPurushothaman2009,
       author = {Casagrande, V.A. and G. Purushothaman},
    booktitle = {{Encyclopedia of Perception, Volume 1}},
       editor = {E. Bruce Goldstein},
    publisher = {Sage Publications},
      address = {Thousand Oaks, CA},
        title = {Modularity},
         year = 2009,
        pages = {561-566},
}
@article{CasellatoetalPLoS-ONE-14,
       author = {Casellato, Claudia and Antonietti, Alberto and Garrido, Jesus A. and Carrillo, Richard R. and Luque, Niceto R. and Ros, Eduardo and Pedrocchi, Alessandra and D'Angelo, Egidio},
      journal = {PLOS ONE},
    publisher = {Public Library of Science},
        title = {Adaptive Robotic Control Driven by a Versatile Spiking Cerebellar Network},
       volume = {9},
       number = {11},
         year = {2014},
        pages = {1-17},
     abstract = {The cerebellum is involved in a large number of different neural processes, especially in associative learning and in fine motor control. To develop a comprehensive theory of sensorimotor learning and control, it is crucial to determine the neural basis of coding and plasticity embedded into the cerebellar neural circuit and how they are translated into behavioral outcomes in learning paradigms. Learning has to be inferred from the interaction of an embodied system with its real environment, and the same cerebellar principles derived from cell physiology have to be able to drive a variety of tasks of different nature, calling for complex timing and movement patterns. We have coupled a realistic cerebellar spiking neural network (SNN) with a real robot and challenged it in multiple diverse sensorimotor tasks. Encoding and decoding strategies based on neuronal firing rates were applied. Adaptive motor control protocols with acquisition and extinction phases have been designed and tested, including an associative Pavlovian task (Eye blinking classical conditioning), a vestibulo-ocular task and a perturbed arm reaching task operating in closed-loop. The SNN processed in real-time mossy fiber inputs as arbitrary contextual signals, irrespective of whether they conveyed a tone, a vestibular stimulus or the position of a limb. A bidirectional long-term plasticity rule implemented at parallel fibers-Purkinje cell synapses modulated the output activity in the deep cerebellar nuclei. In all tasks, the neurorobot learned to adjust timing and gain of the motor responses by tuning its output discharge. It succeeded in reproducing how human biological systems acquire, extinguish and express knowledge of a noisy and changing world. By varying stimuli and perturbations patterns, real-time control robustness and generalizability were validated. The implicit spiking dynamics of the cerebellar model fulfill timing, prediction and learning functions.},
}
@techreport{CassandraTR-94,
       author = {Anthony R. Cassandra},
        title = {Algorithms for Partially Observable {M}arkov Decision Processes},
  institution = {Brown University},
         year = 1994,
        issue = {CS-94-14},
      address = {Providence, Rhode Island},
} 
@inproceedings{CassandraetalAAAI-94,
       author = {Cassandra, Anthony R. and Kaelbling, Leslie and Littman, Michael},
        title = {Acting Optimally in Partially Observable Stochastic Domains},
    booktitle = {Proceedings AAAI-94},
     location = {Seattle, Washington},
 organization = {AAAI},
        pages = {1023-1028},
         year = 1994,
}
@inproceedings{CassandraetalCUAI-97,
       author = {Cassandra, Anthony and Littman, Michael L. and Zhang, Nevin L.},
        title = {Incremental Pruning: A Simple, Fast, Exact Algorithm for
                Partially Observable {M}arkov Decision Processes},
    booktitle = {Thirteenth Conference on Uncertainty in Artificial Intelligence},
     location = {Providence, Rhode Island},
         year = 1997,
}
@book{Cassandras93,
       author = {Cassandras, Christos G.},
        title = {Discrete Event Systems: Modeling and Performance Analysis},
    publisher = {Aksen Associates, Inc.},
      address = {Boston, Massachusetts},
         year = 1993,
}
@inproceedings{Castanon97,
       author = {David A. Casta\~{n}on},
        title = {Approximate Dynamic Programming for Sensor Management},
    booktitle = {Proceedings of the Conference on Decision and Control},
      address = {San Diego, California},
         year = 1997,
}
@article {CastelloetalENEURO-18,
       author = {Visconti di Oleggio Castello, Matteo and Taylor, Morgan and Cavanagh, Patrick and Gobbini, M. Ida},
        title = {Idiosyncratic, Retinotopic Bias in Face Identification Modulated by Familiarity},
      journal = {eNeuro},
       volume = {5},
       number = {5},
         year = {2018},
    publisher = {Society for Neuroscience},
     abstract = {The perception of gender and age of unfamiliar faces is reported to vary idiosyncratically across retinal locations such that, for example, the same androgynous face may appear to be male at one location but female at another. Here, we test spatial heterogeneity for the recognition of the identity of personally familiar faces in human participants. We found idiosyncratic biases that were stable within participants and that varied more across locations for low as compared to high familiar faces. These data suggest that like face gender and age, face identity is processed, in part, by independent populations of neurons monitoring restricted spatial regions and that the recognition responses vary for the same face across these different locations. Moreover, repeated and varied social interactions appear to lead to adjustments of these independent face recognition neurons so that the same familiar face is eventually more likely to elicit the same recognition response across widely separated visual field locations. We provide a mechanistic account of this reduced retinotopic bias based on computational simulations.},
}
@book{Casti90,
       author = {John L. Casti},
        title = {Searching for Certainty},
    publisher = {William Morrow and Company},
      address = {New York},
         year = 1990,
}
@article{CataniBRAIN-17,
       author = {Catani, Marco},
        title = {{A little man of some importance}},
      journal = {Brain},
       volume = {140},
       number = {11},
         year = {2017},
        pages = {3055-3061},
     abstract = {The homunculus made his first appearance in the field of neurology on 1 December 1937, when Wilder Penfield and Edwin Boldrey (Fig. 1) published in Brain a 55-page article entitled Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation (Penfield and Boldrey, 1937). The article is filled with painstakingly crafted summaries of data sourced from the cortical stimulation of 126 patients, who were operated under local anaesthesia by Penfield between 1928 and 1936. Compared to previous publications in animals, the authors had the advantage of operating on awake patients and relying on their verbal report of elicited movements and tactile sensation. All patient recordings were collated to obtain the first comprehensive map of motor and somatosensory localization in the human brain. This map was visualized as a distorted human-like figure - the homunculus - whose form indicates the amount of cortical area dedicated to motor or somatosensory functions of each body part. },
}
@article{CataniaJCN-02,
       author = {Catania, Kenneth C.},
        title = {Barrels, stripes, and fingerprints in the brain---Implications for theories of cortical organization},
      journal = {Journal of Neurocytology},
       number = {3},
       volume = {31},
         year = {2002},
        pages = {347-358},
     abstract = {In the last decade improvements in the histological processing of cortical tissue in conjunction with the investigation of additional mammalian species in comparative brain studies has expanded the information available to guide theories of cortical organization. Here I review some of these recent findings in the somatosensory system with an emphasis on modules related to specializations of the peripheral sensory surface. The diversity of modular representations, or cortical ``isomorphs'' suggest that information from the sensory sheet guides many of the features of cortical maps and suggest that cortex is not constrained to form circular units in the form of a traditional cortical column.},
}
@article{CataniandffytcheBRAIN-05,
       author = {Catani, Marco and ffytche, Dominic H.},
        title = {The rises and falls of disconnection syndromes},
      journal = {Brain},
       volume = {128},
       number = {10},
         year = {2005},
        pages = {2224-2239},
    publisher = {Oxford University Press},
     abstract = {In a brain composed of localized but connected specialized areas, disconnection leads to dysfunction. This simple formulation underlay a range of 19th century neurological disorders, referred to collectively as disconnection syndromes. Although disconnectionism fell out of favour with the move against localized brain theories in the early 20th century, in 1965, an American neurologist brought disconnection to the fore once more in a paper entitled, {\textquoteleft}Disconnexion syndromes in animals and man{\textquoteright}. In what was to become the manifesto of behavioural neurology, Norman Geschwind outlined a pure disconnectionist framework which revolutionized both clinical neurology and the neurosciences in general. For him, disconnection syndromes were higher function deficits that resulted from white matter lesions or lesions of the association cortices, the latter acting as relay stations between primary motor, sensory and limbic areas. From a clinical perspective, the work reawakened interest in single case studies by providing a useful framework for correlating lesion locations with clinical deficits. In the neurosciences, it helped develop contemporary distributed network and connectionist theories of brain function. Geschwind{\textquoteright}s general disconnectionist paradigm ruled clinical neurology for 20 years but in the late 1980s, with the re-emergence of specialized functional roles for association cortex, the orbit of its remit began to diminish and it became incorporated into more general models of higher dysfunction. By the 1990s, textbooks of neurology were devoting only a few pages to classical disconnection theory. Today, new techniques to study connections in the living human brain allow us, for the first time, to test the classical formulation directly and broaden it beyond disconnections to include disorders of hyperconnectivity. In this review, on the 40th anniversary of Geschwind{\textquoteright}s publication, we describe the changing fortunes of disconnection theory and adapt the general framework that evolved from it to encompass the entire spectrum of higher function disorders in neurology and psychiatry. {\textcopyright} The Author (2005). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org},
}
@article{CepelewiczQUANTA-19,
       author = {Jordana Cepelewicz},
        title = {What a New Theory of Attention Says About Consciousness},
      journal = {Quanta Magazine},
    publisher = {The Atlantic Magazine},
         year = {2019},
     abstract = {Filtering out distractions might be more important for the brain than highlighting important information.}
}
@incollection{CerandReillyCOGNITIVE-SCIENCE-98,
       author = {Cer, Daniel and O'Reilly, Randall},
        title = {Neural mechanisms of binding in the hippocampus and neocortex: {I}nsights from computational models},
    booktitle = {Handbook of Binding and Memory: Perspectives from Cognitive Neuroscience},
       editor = {Hubert Zimmer, Axel Mecklinger and Ulman Lindenberger},
    publisher = {Oxford University Press},
         year = {1998},
     abstract = {The development of accurate models of the neural mechanisms underlying binding represents a critical step in the understanding of the mechanisms that give rise to most cognitive processes. This chapter presents a range of computational models based on the biological specializations associated with different brain areas that support a range of different contributions to binding. The posterior cortex can learn coarse-coded distributed representations (CCDRs) of low-order conjunctions which can efficiently and systematically bind information in the service of many different forms of cortical information processing. However, these representations are learned slowly over experience. In contrast, the hippocampus is specialized for rapidly binding novel information into high-order conjunctive representations (e.g., episodes or locations). The prefrontal cortex can actively maintain dynamic bindings in working memory and, through more abstract rule-like representations, support more flexible generalization of behaviour across novel task contexts. Taken together, this overall biologically based cognitive architecture represents a more plausible framework for understanding binding than that provided by temporal synchrony approaches.}
}
@article{CerquidesandMantarasJML-05,
       author = {Jes\'{u}s Cerquides and Ramon L\'{o}pez M\`{a}ntaras},
        title = {{TAN} Classifiers Based on Decomposable Distributions},
      journal = {Machine Learning},
       volume = 59,
        issue = 3,
         year = 2005,
        pages = {323-354},
    publisher = {Kluwer Academic Publishers},
      address = {Hingham, MA},
}
@article{Chafeeetal07,
       author = {Chafee, M. V. and Averbeck, B. B. and Crowe, D. A.},
        title = {Representing Spatial Relationships in Posterior Parietal Cortex: Single Neurons Code Object-Referenced Position},
      journal = {Cerebral Cortex},
         year = 2007,
}
@article{ChaikeeratisaketalSCIENCE-17,
       author = {Vorrapon Chaikeeratisak and Katrina Nguyen and Kanika Khanna and Axel F. Brilot and Marcella L. Erb and Joanna K. C. Coker and Anastasia Vavilina and Gerald L. Newton and Robert Buschauer and Kit Pogliano and Elizabeth Villa and David A. Agard and Joe Pogliano},
        title = {Assembly of a Nucleus-Like Structure During Viral Replication in Bacteria},
      journal = {Science},
       volume = {355},
        issue = {6321},
         year = {2017},
        pages = {194-197},
     abstract = {We observed the assembly of a nucleus-like structure in bacteria during viral infection. Using fluorescence microscopy and cryo-electron tomography, we showed that Pseudomonas chlororaphis phage 201f2-1 assembled a compartment that separated viral DNA from the cytoplasm (shown in blue). The phage compartment was centered by a bipolar tubulin-based spindle (shown in green), and it segregated phage and bacterial proteins according to function. Proteins involved in DNA replication and transcription localized inside the compartment, whereas proteins involved in translation (shown in yellow) and nucleotide synthesis localized outside. Later during infection, viral capsids assembled on the cytoplasmic membrane (shown in red) and moved to the surface of the compartment for DNA packaging (shown in purple). Ultimately, viral particles were released from the compartment and the cell lysed. These results demonstrate that phages have evolved a specialized structure to compartmentalize viral replication.},
}
@book{Chakrabarti02,
       author = {Soumen Chakrabarti},
        title = {Mining the Web: Discovering Knowledge from Hypertext Data},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, CA},
         year = 2002,
}
@inproceedings{Chakrabartietal02,
       author = {S. Chakrabarti and K. Punera and M. Subramanyam},
        title = {Accelerated focused crawling through online relevance feedback},
    booktitle = {Proceedings of the 11th World Wide Web Conference},
    publisher = {ACM},
         year = 2002,
}
@inproceedings{Chakrabartietal98,
       author = {S. Chakrabarti and B. Dom and P. Indyk},
        title = {Enhanced hypertext categorization using hyperlinks},
    booktitle = {Proceedings of the ACM SIGMOD International Conference on Management of Data},
        pages = {307-318},
         year = 1998,
}
@inproceedings{Chakrabartietal99,
       author = {Soumen Chakrabarti and Martin van den Berg and Byron Dom},
        title = {Focused crawling: a new approach to topic-specific {W}eb resource discovery},
    booktitle = {Proceeding of the 8th international conference on World Wide Web},
        pages = {1623-1640},
    publisher = {Elsevier North-Holland, Inc.},
         year = 1999,
}
@article{ChakrabartietalTA-10,
       author = {Chakrabarti, Amit and Cormode, Graham and Mcgregor, Andrew},
        title = {A near-optimal algorithm for estimating the entropy of a stream},
      journal = {{ACM} Transactions on Algorithms},
       volume = 6,
        issue = 3,
         year = 2010,
        pages = {51:1-51:21},
    publisher = {ACM},
      address = {New York, NY},
}
@article{ChakrabartyandMartinJoN-00,
       author = {Chakrabarty, Samit and Martin, John},
        title = {Postnatal Development of the Motor Representation in Primary Motor Cortex},
      journal = {Journal of neurophysiology},
       volume = {84},
         year = {2000},
        pages = {2582-94},
}
@inproceedings{ChalfounetalITS-06,
        title = {Predicting the emotional reaction of the learner with a machine learning technique},
       author = {Chalfoun, Pierre and Chaffar, Soumaya and Frasson, Claude},
    booktitle = {Workshop on Motivaional and Affective Issues, International Conference on Intelligent Tutoring Systems},
         year = 2006,
     abstract = {Emotions play an important role in cognitive processes and specially in learning tasks. Online learning is no exception. Detecting a learner's emotional reaction for a given situation is an essential element for every Distant Learning Environment. Nevertheless, inferring a learner's emotional reaction in those environments is not a trivial task. In this paper, we present an agent capable of predicting a learner's emotional reaction in a distant learning environment based on the learner's personal and non-personal traits using a machine learning technique, namely the ID3 algorithm. We then describe the agent's method for predicting the learner's emotional reaction and discuss the obtained results.}
}
@article{ChalkPNAS-18,
       author = {Chalk, Matthew and Marre, Olivier and Tka{\v c}ik, Ga{\v s}per},
        title = {Toward a unified theory of efficient, predictive, and sparse coding},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {115},
       number = {1},
         year = {2018},
        pages = {186-191},
     abstract = {Sensory neural circuits are thought to efficiently encode incoming signals. Several mathematical theories of neural coding formalize this notion, but it is unclear how these theories relate to each other and whether they are even fully consistent. Here we develop a unified framework that encompasses and extends previous proposals. We highlight key tradeoffs faced by sensory neurons; we show that trading off future prediction against efficiently encoding past inputs generates qualitatively different predictions for neural responses to natural visual stimulation. Our approach is a promising first step toward theoretically explaining the observed diversity of neural responses. A central goal in theoretical neuroscience is to predict the response properties of sensory neurons from first principles. To this end, "efficient coding" posits that sensory neurons encode maximal information about their inputs given internal constraints. There exist, however, many variants of efficient coding (e.g., redundancy reduction, different formulations of predictive coding, robust coding, sparse coding, etc.), differing in their regimes of applicability, in the relevance of signals to be encoded, and in the choice of constraints. It is unclear how these types of efficient coding relate or what is expected when different coding objectives are combined. Here we present a unified framework that encompasses previously proposed efficient coding models and extends to unique regimes. We show that optimizing neural responses to encode predictive information can lead them to either correlate or decorrelate their inputs, depending on the stimulus statistics; in contrast, at low noise, efficiently encoding the past always predicts decorrelation. Later, we investigate coding of naturalistic movies and show that qualitatively different types of visual motion tuning and levels of response sparsity are predicted, depending on whether the objective is to recover the past or predict the future. Our approach promises a way to explain the observed diversity of sensory neural responses, as due to multiple functional goals and constraints fulfilled by different cell types and/or circuits.},
}
@article {ChalketalPNAS-18,
       author = {Chalk, Matthew and Marre, Olivier and Tka\v{c}ik, Ga\v{s}per},
        title = {Toward a unified theory of efficient, predictive, and sparse coding},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {115},
       number = {1},
         year = {2018},
        pages = {186-191},
     abstract = {Sensory neural circuits are thought to efficiently encode incoming signals. Several mathematical theories of neural coding formalize this notion, but it is unclear how these theories relate to each other and whether they are even fully consistent. Here we develop a unified framework that encompasses and extends previous proposals. We highlight key tradeoffs faced by sensory neurons; we show that trading off future prediction against efficiently encoding past inputs generates qualitatively different predictions for neural responses to natural visual stimulation. Our approach is a promising first step toward theoretically explaining the observed diversity of neural responses.A central goal in theoretical neuroscience is to predict the response properties of sensory neurons from first principles. To this end, {\textquotedblleft}efficient coding{\textquotedblright} posits that sensory neurons encode maximal information about their inputs given internal constraints. There exist, however, many variants of efficient coding (e.g., redundancy reduction, different formulations of predictive coding, robust coding, sparse coding, etc.), differing in their regimes of applicability, in the relevance of signals to be encoded, and in the choice of constraints. It is unclear how these types of efficient coding relate or what is expected when different coding objectives are combined. Here we present a unified framework that encompasses previously proposed efficient coding models and extends to unique regimes. We show that optimizing neural responses to encode predictive information can lead them to either correlate or decorrelate their inputs, depending on the stimulus statistics; in contrast, at low noise, efficiently encoding the past always predicts decorrelation. Later, we investigate coding of naturalistic movies and show that qualitatively different types of visual motion tuning and levels of response sparsity are predicted, depending on whether the objective is to recover the past or predict the future. Our approach promises a way to explain the observed diversity of sensory neural responses, as due to multiple functional goals and constraints fulfilled by different cell types and/or circuits.},
}
@article{ChalmersJoCS-18,
       author = {David J. Chalmers},
        title = {The Meta-Problem of Consciousness},
      journal = {Journal of Consciousness Studies},
       volume = {25},
        issue = {9-10},
         year = {2018},
        pages = {6-61},
}
@article{ChalmersJoCS-95,
    publisher = {Imprint Academic},
       author = {David J. Chalmers},
        title = {Facing Up to the Problem of Consciousness},
      journal = {Journal of Consciousness Studies},
         year = {1995},
       number = {3},
       volume = {2},
        pages = {200-219},
}
@article{ChamanzaretalNATURE-COMMUNICATION-19,
       author = {Chamanzar, Maysamreza and Scopelliti, Matteo Giuseppe and Bloch, Julien and Do, Ninh and Huh, Minyoung and Seo, Dongjin and Iafrati, Jillian and Sohal, Vikaas S. and Alam, Mohammad-Reza and Maharbiz, Michel M.},
        title = {Ultrasonic sculpting of virtual optical waveguides in tissue},
      journal = {Nature Communications},
         year = {2019},
       volume = {10},
        issue = {1},
        pages = {92},
     abstract = {Optical imaging and stimulation are widely used to study biological events. However, scattering processes limit the depth to which externally focused light can penetrate tissue. Optical fibers and waveguides are commonly inserted into tissue when delivering light deeper than a few millimeters. This approach, however, introduces complications arising from tissue damage. In addition, it makes it difficult to steer light. Here, we demonstrate that ultrasound can be used to define and steer the trajectory of light within scattering media by exploiting local pressure differences created by acoustic waves that result in refractive index contrasts. We show that virtual light pipes can be created deep into the tissue (>18 scattering mean free paths). We demonstrate the application of this technology in confining light through mouse brain tissue. This technology is likely extendable to form arbitrary light patterns within tissue, extending both the reach and the flexibility of light-based methods.},
}
@inproceedings{ChanandShachterUAI-92,
       author = {Chan, Brian Y. and Shachter, Ross},
        title = {Structural Controllability and Observability in Influence Diagrams},
    booktitle = {Eighth Conference on Uncertainty in Artificial Intelligence},
     location = {Los Angeles, California},
        pages = {25-32},
         year = 1992,
}
@article{ChanceandNortonCOiP-15,
        title = {The what and why of self-deception},
      journal = {Current Opinion in Psychology},
       volume = {6},
       number = {Supplement C},
        pages = {104-107},
         year = {2015},
       author = {Zo\"{e} Chance and Michael I. Norton},
     abstract = {Scholars from many disciplines have investigated self-deception, but defining self-deception and establishing its possible benefits have been a matter of heated debate--a debate impoverished by a relative lack of empirical research. Drawing on recent research, we first classify three distinct definitions of self-deception, ranging from a view that self-deception is synonymous with positive illusions to a more stringent view that self-deception requires the presence of simultaneous conflicting beliefs. We then review recent research on the possible benefits of self-deception, identifying three adaptive functions: deceiving others, social status, and psychological benefits. We suggest potential directions for future research.}
}
@article{ChanesandBarrettTCS-16,
        title = {Redefining the role of limbic areas in cortical processing},
       author = {Chanes, Lorena and Barrett, Lisa Feldman},
      journal = {Trends in Cognitive Sciences},
    publisher = {Elsevier Current Trends},
       volume = {20},
       number = {2},
         year = {2016},
        pages = {96-106},
     abstract = {There is increasing evidence that the brain actively constructs action and perception using past experience. In this paper, we propose that the direction of information flow along gradients of laminar differentiation provides important insight on the role of limbic cortices in cortical processing. Cortical limbic areas, with a simple laminar structure (e.g., no or rudimentary layer IV), send 'feedback' projections to lower level better laminated areas. We propose that this 'feed- back' functions as predictions that drive processing throughout the cerebral cortex. This hypothesis has the potential to provide a unifying framework for an increasing number of proposals that use predictive coding to explain a myriad of neural processes and disorders, and has important implications for hypotheses about consciousness.}
}
@article{ChanetalPHYSICS-07,
       author = {Wai Lam Chan and Jason Deibel and Daniel M. Mittleman},
        title = {Imaging with terahertz radiation},
      journal = {Reports on Progress in Physics},
       volume = 70,
        issue = 8,
         year = 2007,
        pages = 1325,
     abstract = {Within the last several years, the field of terahertz science and technology has changed dramatically. Many new advances in the technology for generation, manipulation, and detection of terahertz radiation have revolutionized the field. Much of this interest has been inspired by the promise of valuable new applications for terahertz imaging and sensing. Among a long list of proposed uses, one finds compelling needs such as security screening and quality control, as well as whimsical notions such as counting the almonds in a bar of chocolate. This list has grown in parallel with the development of new technologies and new paradigms for imaging and sensing. Many of these proposed applications exploit the unique capabilities of terahertz radiation to penetrate common packaging materials and provide spectroscopic information about the materials within. Several of the techniques used for terahertz imaging have been borrowed from other, more well established fields such as x-ray computed tomography and synthetic aperture radar. Others have been developed exclusively for the terahertz field, and have no analogies in other portions of the spectrum. This review provides a comprehensive description of the various techniques which have been employed for terahertz image formation, as well as discussing numerous examples which illustrate the many exciting potential uses for these emerging technologies.}
}
@article {ChanetalPNAS-17,
       author = {Chan, Russell W. and Leong, Alex T. L. and Ho, Leon C. and Gao, Patrick P. and Wong, Eddie C. and Dong, Celia M. and Wang, Xunda and He, Jufang and Chan, Ying-Shing and Lim, Lee Wei and Wu, Ed X.},
        title = {Low-frequency hippocampal-cortical activity drives brain-wide resting-state functional MRI connectivity},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {114},
       number = {33},
         year = {2017},
        pages = {E6972-E6981},
     abstract = {The hippocampus with its dense reciprocal axonal projections to and from cortex is widely believed to mediate numerous cognitive functions. However, it is unknown whether and how specific hippocampal{\textendash}cortical activity contributes to the brain-wide functional connectivity. Here, we use optogenetics and fMRI to examine how excitatory neural activity initiated in the dorsal dentate gyrus of the hippocampus propagates and modulates resting-state fMRI (rsfMRI) connectivity. We discover its robust propagation brain-wide at low frequency (1 Hz), which enhances interhemispheric rsfMRI connectivity and cortical and subcortical visual responses. Our findings highlight the important role of slow hippocampal{\textendash}cortical oscillatory activity in driving brain-wide rsfMRI connectivity and mediating sensory processing.The hippocampus, including the dorsal dentate gyrus (dDG), and cortex engage in bidirectional communication. We propose that low-frequency activity in hippocampal{\textendash}cortical pathways contributes to brain-wide resting-state connectivity to integrate sensory information. Using optogenetic stimulation and brain-wide fMRI and resting-state fMRI (rsfMRI), we determined the large-scale effects of spatiotemporal-specific downstream propagation of hippocampal activity. Low-frequency (1 Hz), but not high-frequency (40 Hz), stimulation of dDG excitatory neurons evoked robust cortical and subcortical brain-wide fMRI responses. More importantly, it enhanced interhemispheric rsfMRI connectivity in various cortices and hippocampus. Subsequent local field potential recordings revealed an increase in slow oscillations in dorsal hippocampus and visual cortex, interhemispheric visual cortical connectivity, and hippocampal{\textendash}cortical connectivity. Meanwhile, pharmacological inactivation of dDG neurons decreased interhemispheric rsfMRI connectivity. Functionally, visually evoked fMRI responses in visual regions also increased during and after low-frequency dDG stimulation. Together, our results indicate that low-frequency activity robustly propagates in the dorsal hippocampal{\textendash}cortical pathway, drives interhemispheric cortical rsfMRI connectivity, and mediates visual processing.},
}
@book{ChangandLee73,
       author = {Chang, Chin-Liang and Lee, Richard Char-Tung},
        title = {Symbolic Logic and Mechanical Theorem Proving},
    publisher = {Academic Press},
      address = {New York},
         year = 1973,
}
@inproceedings{ChangandTanIJCAI-17,
       author = {Poo-Hee Chang, Ah-Hwee Tan},
        title = {Encoding and Recall of Spatio-Temporal Episodic Memory in Real Time},
    booktitle = {Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence, {IJCAI-17}},
        pages = {1490-1496},
         year = {2017},
     abstract = {Episodic memory enables a cognitive system to improve its performance by reflecting upon past events. In this paper, we propose a computational model called STEM for encoding and recall of episodic events together with the associated con- textual information in real time. Based on a class of self-organizing neural networks, STEM is designed to learn memory chunks or cognitive nodes, each encoding a set of co-occurring multi-modal activity patterns across multiple pattern channels. We present algorithms for recall of events based on partial and inexact input patterns. Our empirical results based on a public domain data set show that STEM displays a high level of efficiency and robustness in encoding and retrieval with both partial and noisy search cues when compared with a state-of-the-art associative memory model.},
}
@article{ChangetalCoRR-16,
       author = {Michael B. Chang and Tomer Ullman and Antonio Torralba and Joshua B. Tenenbaum},
        title = {A Compositional Object-Based Approach to Learning Physical Dynamics},
      journal = {CoRR},
       volume = {arXiv:1612.00341},
         year = {2016},
     abstract = {We present the Neural Physics Engine (NPE), a framework for learning simulators of intuitive physics that naturally generalize across variable object count and different scene configurations. We propose a factorization of a physical scene into composable object-based representations and a neural network architecture whose compositional structure factorizes object dynamics into pairwise interactions. Like a symbolic physics engine, the NPE is endowed with generic notions of objects and their interactions; realized as a neural network, it can be trained via stochastic gradient descent to adapt to specific object properties and dynamics of different worlds. We evaluate the efficacy of our approach on simple rigid body dynamics in two-dimensional worlds. By comparing to less structured architectures, we show that the NPE's compositional representation of the structure in physical interactions improves its ability to predict movement, generalize across variable object count and different scene configurations, and infer latent properties of objects such as mass.},
}
@article{ChangetalCoRR-18,
       author = {Michael Chang and Abhishek Gupta and Sergey Levine and Thomas L. Griffiths},
        title = {Automatically Composing Representation Transformations as a Means for Generalization},
      journal = {CoRR},
       volume = {arXiv:1807.04640},
         year = {2018},
      comment = {Published as a conference paper at ICLR 2019},
     abstract = {A generally intelligent learner should generalize to more complex tasks than it has previously encountered, but the two common paradigms in machine learning -- either training a separate learner per task or training a single learner for all tasks -- both have difficulty with such generalization because they do not leverage the compositional structure of the task distribution. This paper introduces the compositional problem graph as a broadly applicable formalism to relate tasks of different complexity in terms of problems with shared subproblems. We propose the compositional generalization problem for measuring how readily old knowledge can be reused and hence built upon. As a first step for tackling compositional generalization, we introduce the compositional recursive learner, a domain-general framework for learning algorithmic procedures for composing representation transformations, producing a learner that reasons about what computation to execute by making analogies to previously seen problems. We show on a symbolic and a high-dimensional domain that our compositional approach can generalize to more complex problems than the learner has previously encountered, whereas baselines that are not explicitly compositional do not.}
}
@inproceedings{ChangetalICLR-19,
       author = {Michael Chang and Abhishek Gupta and Sergey Levine and Thomas L. Griffiths},
        title = {Automatically Composing Representation Transformations as a Means for Generalization},
    booktitle = {International Conference on Learning Representations},
         year = {2019},
}
@article{ChangetalTCS-08,
       author = {Fay Chang and Jeffrey Dean and Sanjay Ghemawat and Wilson C. Hsieh and Deborah A. Wallach and Michael Burrows and Tushar Chandra and Andrew Fikes and Robert E. Gruber},
        title = {{BigTable}: A Distributed Storage System for Structured Data},
      journal = {{ACM} Transactions on Computing Systems},
       volume = 26,
        issue = 2,
         year = 2008,
}
@article{ChapmanAIJ-87,
       author = {Chapman, David},
        title = {Planning for Conjunctive Goals},
      journal = {Artificial Intelligence},
       volume = 32,
         year = 1987,
        pages = {333-377},
}
@techreport{ChapmanPhD,
       author = {Chapman, David},
        title = {Vision, Instruction, and Action},
        issue = {AI-TR-1204},
  institution = {{MIT} {AI} Laboratory},
         year = 1990,
}
@techreport{ChapmanTR-85,
       author = {Chapman, David},
        title = {Planning for Conjunctive Goals},
        issue = {AI-TR-802},
  institution = {{MIT} {AI} Laboratory},
         year = 1985,
}
@incollection{ChapmanandAgrePW-86,
       author = {Chapman, David and Agre, Philip E.},
        title = {Abstract Reasoning as Emergent from Concrete Activity},
     crossref = {GeorgeffandLanskyPW-86},
        pages = {411-424},
}
@inproceedings{ChapmanandKaelblingIJCAI-91,
       author = {Chapman, David and Kaelbling, Leslie},
        title = {Input Generalization in Delayed Reinforcement Learning: An Algorithm and Performance Comparisons},
    booktitle = {Proceedings {IJCAI} 12},
     location = {Sydney, Australia},
 organization = {IJCAII},
         year = 1991,
}
@article{CharmanetalCOGNITIVE-DEVELOPMENT-00,
       author = {Tony Charman and Simon Baron-Cohen and John Swettenham and Gillian Baird and Antony Cox and Auriol Drew},
        title = {Testing joint attention, imitation, and play as infancy precursors to language and theory of mind},
      journal = {Cognitive Development},
       volume = {15},
       number = {4},
        pages = {481-498},
         year = {2000},
     abstract = {Various theoretical accounts propose that an important developmental relation exists between joint attention, play, and imitation abilities, and later theory of mind ability. However, very little direct empirical evidence supports these claims for putative "precursor" theory of mind status. A small sample (N=13) of infants, for whom measures of play, joint attention, and imitation had been collected at 20 months of age, was followed-up longitudinally at 44 months and a battery of theory of mind measures was conducted. Language and IQ were measured at both timepoints. Imitation ability at 20 months was longitudinally associated with expressive, but not receptive, language ability at 44 months. In contrast, only the joint attention behaviours of gaze switches between an adult and an active toy and looking to an adult during an ambiguous goal detection task at 20 months were longitudinally associated with theory of mind ability at 44 months. It is argued that joint attention, play, and imitation, and language and theory of mind, might form part of a shared social-communicative representational system in infancy that becomes increasingly specialised and differentiated as development progresses.}
}
@book{Charniak2019deeplearning,
        title = {Introduction to Deep Learning},
       author = {Charniak, Eugene},
       series = {Mit Press},
         year = {2019},
    publisher = {MIT Press}
}
@book{Charniak80,
       author = {Charniak, Eugene and Riesbeck, Christopher K. and McDermott, Drew V.},
        title = {Artificial Intelligence Programming},
    publisher = {Lawrence Erlbaum Associates},
      address = {Hillsdale, New Jersey},
         year = 1980,
}
@book{Charniak87,
       author = {Charniak, Eugene and Riesbeck, Christopher K. and McDermott, Drew V. and Meehan, James R.},                
        title = {Artificial Intelligence Programming, Second Edition},
    publisher = {Lawrence Erlbaum Associates},
      address = {Hillsdale, New Jersey},
         year = 1987,
}
@book{Charniak93,
       author = {Charniak, Eugene},
        title = {Statistical Language Learning},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1993,
}
@inproceedings{CharniakAAAI-83,
       author = {Charniak, Eugene},
        title = {The {B}ayesian Basis of Common Sense Medical Diagnosis},
    booktitle = {Proceedings of AAAI-83},
     location = {Washington, DC, USA},
 organization = {AAAI},
         year = 1983,
}
@article{CharniakAIJ-86,
       author = {Charniak, Eugene},
        title = {Motivation Analyisis, Abductive Unification, and Non-Monotonic Equality},
      journal = AIJ,
         year = 1986,
}
@article{CharniakAIM-91,
        title = {Bayesian Networks without Tears},
       author = {Charniak, Eugene},
      journal = {AI Magazine},
       volume = {Winter},
        pages = {51-63},
         year = 1991,
}
@inproceedings{CharniakandGoldmanAAAI-91,
        title = {A Probabilistic Model for Plan Recognition},
       author = {Charniak, Eugene and Goldman, Robert},
    booktitle = {Proceedings AAAI-91},
     location = {Anaheim, California},
 organization = {AAAI},
        pages = {160-165},
         year = 1991,
}
@book{CharniakandMcDermott85,
       author = {Charniak, Eugene and McDermott, Drew V.},
        title = {Introduction to Artificial Intelligence},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1985,
}
@book{Chater2018,
        title = {The Mind is Flat: The Illusion of Mental Depth and The Improvised Mind},
       author = {Chater, Nick},
         year = {2018},
    publisher = {Penguin Books Limited}
}
@article{ChaterandChristiansenCOiBS-18,
        title = {Language acquisition as skill learning},
      journal = {Current Opinion in Behavioral Sciences},
         year = {2018},
       author = {Nick Chater and Morten H. Christiansen},
     abstract = {Language acquisition researchers have often viewed children as 'mini-linguists,' attempting to infer abstract knowledge of language from exposure to their native language. From this perspective, the challenge of acquisition can seem so formidable that meeting it would appear to require that much of this knowledge must be built-in, as a language instinct or universal grammar. From this viewpoint, language acquisition is also disconnected from language processing, and from the acquisition of other learned perceptuo-motor or cognitive abilities. This paper explores a recent alternative viewpoint, the 'language-as-skill' framework, according to which the child's challenge is practical, not theoretical: the child learns to understand and produce the language from practicing conversational interactions. Language acquisition can thus be seen as a type of skill acquisition, using similar mechanisms to those involved in learning to ride a bicycle, play a musical instrument, or draw a picture; and the need to acquire knowledge of the abstract structure of language is dissolved. This perspective takes the pressure off biological adaptation as the primary driver of language evolution, emphasizing instead the cultural evolution of linguistic structure.}
}
@incollection{ChaterandChristiansenHLB-11,
        title = {A solution to the logical problem of language evolution: language as an adaptation to the human brain},
       author = {Nick Chater and Morten H. Christiansen},
    booktitle = {The Oxford Handbook of Language Evolution},
       editor = {Kathleen R. Gibson and Maggie Tallerman},
    publisher = {Oxford University Press},
         year = {2011},
     abstract = {This article addresses the logical problem of language evolution that arises from a conventional universal grammar (UG) perspective and investigates the biological and cognitive constraints that are considered when explaining the cultural evolution of language. The UG prespective states that language acquisition should not be viewed as a process of learning at all but it should be viewed as a process of growth, analogous to the growth of the arm or the liver. UG is intended to characterize a set of universal grammatical principles that hold across all languages. Language has the same status as other cultural products, such as styles of dress, art, music, social structure, moral codes, or patterns of religious beliefs. Language may be particularly central to culture and act as the primary vehicle through which much other cultural information is transmitted. The biological and cognitive constraints helps to determine which types of linguistic structure tend to be learned, processed, and hence transmitted from person to person, and from generation to generation. The communicative function of language is likely to shape language structure in relation to the thoughts that are transmitted and regarding the processes of pragmatic interpretation that people use to understand each other's behavior. A source of constraints derives from the nature of cognitive architecture, including learning, processing, and memory. The language processing involves generating and decoding regularities from highly complex sequential input, indicating a connection between general-purpose cognitive mechanisms for learning and processing sequential material, and the structure of natural language.}
}
@article{ChaterandOaksfordCS-13,
       author = {Chater, Nick and Oaksford, Mike},
        title = {Programs as Causal Models: Speculations on Mental Programs and Mental Representation},
      journal = {Cognitive Science},
       volume = {37},
       number = {6},
         year = {2013},
        pages = {1171-1191},
     abstract = {Judea Pearl has argued that counterfactuals and causality are central to intelligence, whether natural or artificial, and has helped create a rich mathematical and computational framework for formally analyzing causality. Here, we draw out connections between these notions and various current issues in cognitive science, including the nature of mental 'programs' and mental representation. We argue that programs (consisting of algorithms and data structures) have a causal (counterfactual-supporting) structure; these counterfactuals can reveal the nature of mental representations. Programs can also provide a causal model of the external world. Such models are, we suggest, ubiquitous in perception, cognition, and language processing.},
}
@article{ChateretalJML-16,
        title = {Language as skill: Intertwining comprehension and production},
       author = {Nick Chater and Stewart M. McCauley and Morten H. Christiansen},
      journal = {Journal of Memory and Language},
       volume = {89},
        pages = {244-254},
         year = {2016},
     abstract = {Are comprehension and production a single, integrated skill, or are they separate processes drawing on a shared abstract knowledge of language? We argue that a fundamental constraint on memory, the Now-or-Never bottleneck, implies that language processing is incremental and that language learning occurs on-line. These properties are difficult to reconcile with the 'abstract knowledge' viewpoint, and crucially suggest that language comprehension and production are facets of a unitary skill. This viewpoint is exemplified in the Chunk-Based Learner, a computational acquisition model that processes incrementally and learns on-line. The model both parses and produces language; and implements the idea that language acquisition is nothing more than learning to process. We suggest that the Now-or-Never bottleneck also provides a strong motivation for unified perception-production models in other domains of communication and cognition.}
}
@book{Chatfield89,
       author = {Chatfield, C.},
        title = {The Analysis of Time Series: An Introduction},
    publisher = {Chapman \& Hall},
         year = 1989,
}
@article{ChathametalJCN-11,
       author = {Chatham, C.H. and Herd, S.A. and Brant, A.M. and Hazy, T.E. and Miyake, A. and O'Reilly, R. and Friedman N.P.},
        title = {From an executive network to executive control: a computational model of the n-back task},
      journal = {Journal of Cognitive Neuroscience},
       volume = {23},
        issue = {11},
    publisher = {MIT Press},
         year = {2011},
        pages = {3598-3619},
     abstract = {A paradigmatic test of executive control, the n-back task, is known to recruit a widely distributed parietal, frontal, and striatal "executive network," and is thought to require an equally wide array of executive functions. The mapping of functions onto substrates in such a complex task presents a significant challenge to any theoretical framework for executive control. To address this challenge, we developed a biologically constrained model of the n-back task that emergently develops the ability to appropriately gate, bind, and maintain information in working memory in the course of learning to perform the task. Furthermore, the model is sensitive to proactive interference in ways that match findings from neuroimaging and shows a U-shaped performance curve after manipulation of prefrontal dopaminergic mechanisms similar to that observed in studies of genetic polymorphisms and pharmacological manipulations. Our model represents a formal computational link between anatomical, functional neuroimaging, genetic, behavioral, and theoretical levels of analysis in the study of executive control. In addition, the model specifies one way in which the pFC, BG, parietal, and sensory cortices may learn to cooperate and give rise to executive control.}
}
@article{ChathametalJoCS-11,
       author = {Chatham, Christopher H. and Herd, Seth A. and Brant, Angela M. and Hazy, Thomas E. and Miyake, Akira and O'Reilly, Randy and Friedman, Naomi P.},
        title = {From an executive network to executive control: {A} computational model of the n-back task},
      journal = {Journal of Cognitive Neuroscience},
       volume = {23},
        issue = {11},
         year = {2011},
        pages = {3598-3619},
     abstract = {A paradigmatic test of executive control, the n-back task, is known to recruit a widely distributed parietal, frontal, and striatal ``executive network,'' and is thought to require an equally wide array of executive functions. The mapping of functions onto substrates in such a complex task presents a significant challenge to any theoretical framework for executive control. To address this challenge, we developed a biologically constrained model of the n-back task that emergently develops the ability to appropriately gate, bind, and maintain information in working memory in the course of learning to perform the task. Furthermore, the model is sensitive to proactive interference in ways that match findings from neuroimaging and shows a U-shaped performance curve after manipulation of prefrontal dopaminergic mechanisms similar to that observed in studies of genetic polymorphisms and pharmacological manipulations. Our model represents a formal computational link between anatomical, functional neuroimaging, genetic, behavioral, and theoretical levels of analysis in the study of executive control. In addition, the model specifies one way in which the pFC, BG, parietal, and sensory cortices may learn to cooperate and give rise to executive control.},
}
@inproceedings{ChatilaandLaumond85,
       author = {Chatila, R. and J.-P. Laumond},
        title = {Position Referencing and Consistent World Modeling on Mobile Robots},
    booktitle = {{IEEE} International Conference on Robotics and Automation},
     location = {St. Louis, MO},
         year = 1985,
        pages = {138-145},
}
@article{ChaudhurietalNEURON-15,
       author = {Rishidev Chaudhuri and Kenneth Knoblauch and Marie-Alice Gariel and Henry Kennedy and Xiao-Jing Wang},
        title = {A Large-Scale Circuit Mechanism for Hierarchical Dynamical Processing in the Primate Cortex},
      journal = {Neuron},
       volume = {88},
       number = {2},
         year = {2015},
        pages = {419-431},
     abstract = {Summary We developed a large-scale dynamical model of the macaque neocortex, which is based on recently acquired directed- and weighted-connectivity data from tract-tracing experiments, and which incorporates heterogeneity across areas. A hierarchy of timescales naturally emerges from this system: sensory areas show brief, transient responses to input (appropriate for sensory processing), whereas association areas integrate inputs over time and exhibit persistent activity (suitable for decision-making and working memory). The model displays multiple temporal hierarchies, as evidenced by contrasting responses to visual versus somatosensory stimulation. Moreover, slower prefrontal and temporal areas have a disproportionate impact on global brain dynamics. These findings establish a circuit mechanism for "temporal receptive windows" that are progressively enlarged along the cortical hierarchy, suggest an extension of time integration in decision making from local to large circuits, and should prompt a re-evaluation of the analysis of functional connectivity (measured by fMRI or electroencephalography/magnetoencephalography) by taking into account inter-areal heterogeneity.}
}
@inproceedings{Chauetal01,
     author = {Michael Chau and Daniel Zeng and Hinchun Chen},
      title = {Personalized spiders for {W}eb search and analysis},
  booktitle = {Proceedings of the first ACM/IEEE-CS joint conference on Digital libraries},
  publisher = {ACM Press},
      pages = {79-87},
       year = 2001,
}
@article{ChavezNETWORKS-90,
       author = {Chavez, R. Martin and Cooper, Gregory F.},
        title = {A Fully Polynomial Randomized Approximation Scheme for the {B}ayesian Inferencing Problem},
      journal = {Networks},
       volume = 20,
         year = 1990,
}
@techreport{ChavezTR-72,
       author = {R. Martin Chavez},
        title = {Fully Polynomial Randomized Approximation Schemes for the {B}ayesian Infererencing Problem},
  institution = {Section on Medical Informatics, Stanford University School of Medicine},
 type         = {Report},
        issue = {KSL-88-72},
         year = 1988,
}
@inproceedings{ChavezandHenrionUAI-94,
       author = {Chavez, Tom and Henrion, Max},
        title = {Focusing on What Matters in Plan Evaluation: Efficiently Estimating the Value of Information},
    booktitle = {Tenth Conference on Uncertainty in Artificial Intelligence},
     location = {Seattle, Washington},
         year = 1994,
}
@inproceedings{CheathametalHCSS-95,
       author = {Cheatham, T. and Fahmy, A. and Stefanescu, D. and Valiant, L.},
        title = {Bulk Synchronous Parallel Computing: {A} Paradigm for Transportable Software},
    booktitle = {Proceedings of the 28th Annual Hawaii Conference on System Sciences},
       volume = {II},
    publisher = {IEEE Computer Society Press},
        pages = {268-275},
         year = 1995,
}
@inproceedings{Cheeseman84,
       author = {Cheeseman, Peter},
        title = {A Representation of Time for Automatic Planning},
    booktitle = {Proceedings {IEEE} International Conference on Robotics},
 organization = {IEEE},
         year = 1984,
}
@incollection{Cheeseman86,
       author = {Cheeseman, Peter},
        title = {Probabilistic versus fuzzy reasoning},
     crossref = {UAI-86},
}
@inproceedings{CheesemanIJCAI-83,
       author = {Cheeseman, Peter},
        title = {A Method of Computing Generalized {B}ayesian Probability Values for Expert Systems},
    booktitle = {Proceedings {IJCAI} 8},
     location = {Karlsruhe, West Germany},
 organization = {IJCAII},
         year = 1983,
}
@incollection{CheesemanSFI-95,
       author = {Cheeseman, Peter},
        title = {On {B}ayesian Model Selection},
    booktitle = {The Mathematics of Generalization},
       editor = {Wolpert, David H.},
    publisher = {Addison-Wesley},
         year = 1995,
}
@inproceedings{CheesemanetalIJCAI-91,
       author = {Cheeseman, P. and Kanefsky, B. and Taylor, W. M.},
        title = {Where the really hard problems are},
    booktitle = {Proceedings {IJCAI} 12},
     location = {Sydney, Australia},
 organization = {IJCAII},
        pages = {331-337},
         year = 1991,
}
@inproceedings{CheesemanetalML-88,
       author = {Cheeseman, Peter and Kelly, James and Self, Matthew and Stutz, John and Taylor, Will and Freeman, Don},
        title = {AutoClass: A {B}ayesian Classification System},
    booktitle = {Proceedings Fifth International Conference on Machine Learning},
         year = 1988,
}
@techreport{ChekalukMSc,
       author = {Chekaluk, Robert A.},
        title = {Using Influence Diagrams in Recognizing Locally-Distinctive Places},
         type = {{M.Sc.} Thesis},
  institution = {Brown University},
         year = 1989,
}
@article{ChellaandPipitoneCSR-20,
       author = {Antonio Chella and Arianna Pipitone},
        title = {A cognitive architecture for inner speech},
      journal = {Cognitive Systems Research},
       volume = {59},
         year = {2020},
        pages = {287-292},
     abstract = {A cognitive architecture for inner speech is presented. It is based on the Standard Model of Mind, integrated with modules for self-talking. Briefly, the working memory of the proposed architecture includes the phonological loop as a component which manages the exchanging information between the phonological store and the articulatory control system. The inner dialogue is modeled as a loop where the phonological store hears the inner voice produced by the hidden articulator process. A central executive module drives the whole system, and contributes to the generation of conscious thoughts by retrieving information from long-term memory. The surface form of thoughts thus emerges by the phonological loop. Once a conscious thought is elicited by inner speech, the perception of new context takes place and then repeating the cognitive loop. A preliminary formalization of some of the described processes by event calculus, and early results of their implementation on the humanoid robot Pepper by SoftBank Robotics are discussed.}
}
@book{Chen70,
       author = {Chen, C. T.},
        title = {Introduction to Linear System Theory},
    publisher = {Holt, Rinehart, and Winston},
      address = {New York},
         year = 1970,
}
@article{ChenSCIAM-18,
       author = {Helen Chen},
        title = {Does the Adult Brain Really Grow New Neurons?},
      journal = {Scientific American},
       volume = {March 7},
         year = {2018}
}
@article{ChenandHongNEURON-18,
       author = {Patrick Chen and Weizhe Hong},
        title = {Neural Circuit Mechanisms of Social Behavior},
      journal = {Neuron},
       volume = {98},
       number = {1},
        pages = {16-30},
         year = {2018},
     abstract = {We live in a world that is largely socially constructed, and we are constantly involved in and fundamentally influenced by a broad array of complex social interactions. Social behaviors among conspecifics, either conflictive or cooperative, are exhibited by all sexually reproducing animal species and are essential for the health, survival, and reproduction of animals. Conversely, impairment in social function is a prominent feature of several neuropsychiatric disorders, such as autism spectrum disorders and schizophrenia. Despite the importance of social behaviors, many fundamental questions remain unanswered. How is social sensory information processed and integrated in the nervous system? How are different social behavioral decisions selected and modulated in brain circuits? Here we discuss conceptual issues and recent advances in our understanding of brain regions and neural circuit mechanisms underlying the regulation of social behaviors.}
}
@article{ChenandKnierimNEURON-18,
        title = {It's About Time: Temporal Dynamics of Dentate Gyrus Pattern Separation},
      journal = {Neuron},
       volume = {98},
       number = {4},
        pages = {681-683},
         year = {2018},
       author = {Xiaojing Chen and James J. Knierim},
     abstract = {The dentate gyrus (DG) is hypothesized to be a pattern separator crucial for distinguishing similar memories. In this issue of Neuron, a new study by van Dijk and Fenton (2018) revealed that transient changes of firing patterns of DG ensembles, but not the remapping of session-averaged place fields, contributed to a place discrimination task.}
}
@article{ChenetalAM-15,
       author = {Chen, Chien-Yu and Lee, Wei-Kai and Chen, Yi-Jiun and Lu, Chun-Yang and Lin, Hoang Yan and Wu, Chung-Chih},
        title = {Enhancing Optical Out-Coupling of Organic Light-Emitting Devices with Nanostructured Composite Electrodes Consisting of Indium Tin Oxide Nanomesh and Conducting Polymer},
      journal = {Advanced Materials},
         year = 2015,
     abstract = {A nanostructured composite electrode consisting of the high-index indium-tin-oxide nanomesh and low-index high-conductivity conducting polymer effectively enhances coupling of internal radiation of organic light-emitting devices into their substrates. When combining this internal extraction structure and the external extraction scheme, a very high external quantum efficiency of nearly 62\% was achieved with a green phosphorescent device.}
}
@article{ChenetalCoRR-18,
       author = {Tian Qi Chen and Yulia Rubanova and Jesse Bettencourt and David Duvenaud},
        title = {Neural Ordinary Differential Equations},
      journal = {CoRR},
       volume = {arXiv:1806.07366},
         year = {2018},
     abstract = {We introduce a new family of deep neural network models. Instead of specifying a discrete sequence of hidden layers, we parameterize the derivative of the hidden state using a neural network. The output of the network is computed using a black-box differential equation solver. These continuous-depth models have constant memory cost, adapt their evaluation strategy to each input, and can explicitly trade numerical precision for speed. We demonstrate these properties in continuous-depth residual networks and continuous-time latent variable models. We also construct continuous normalizing flows, a generative model that can train by maximum likelihood, without partitioning or ordering the data dimensions. For training, we show how to scalably backpropagate through any ODE solver, without access to its internal operations. This allows end-to-end training of ODEs within larger models.}
}
@article{ChenetalICLR-18a,
       author = {Xinyun Chen and Chang Liu and Dawn Song},
        title = {Towards Synthesizing Complex Programs from Input-Output Examples},
      journal = {CoRR},
       volume = {arXiv:1706.01284},
         year = {2018},
          url = {https://arxiv.org/abs/1706.01284},
     abstract = {In recent years, deep learning techniques have been developed to improve the performance of program synthesis from input-output examples. Albeit its significant progress, the programs that can be synthesized by state-of-the-art approaches are still simple in terms of their complexity. In this work, we move a significant step forward along this direction by proposing a new class of challenging tasks in the domain of program synthesis from input-output examples: learning a context-free parser from pairs of input programs and their parse trees. We show that this class of tasks are much more challenging than previously studied tasks, and the test accuracy of existing approaches is almost 0\%. We tackle the challenges by developing three novel techniques inspired by three novel observations, which reveal the key ingredients of using deep learning to synthesize a complex program. First, the use of a non-differentiable machine is the key to effectively restrict the search space. Thus our proposed approach learns a neural program operating a domain-specific non-differentiable machine. Second, recursion is the key to achieve generalizability. Thus, we bake-in the notion of recursion in the design of our non-differentiable machine. Third, reinforcement learning is the key to learn how to operate the non-differentiable machine, but it is also hard to train the model effectively with existing reinforcement learning algorithms from a cold boot. We develop a novel two-phase reinforcement learning-based search algorithm to overcome this issue. In our evaluation, we show that using our novel approach, neural parsing programs can be learned to achieve 100\% test accuracy on test inputs that are 500x longer than the training samples.}
}
@article{ChenetalICLR-18b,
       author = {Xinyun Chen and Chang Liu and Dawn Song},
        title = {Tree-to-tree Neural Networks for Program Translation},
      journal = {CoRR},
       volume = {arXiv:1802.03691},
         year = {2018},
          url = {https://arxiv.org/abs/1802.03691},
     abstract = {Program translation is an important tool to migrate legacy code in one language into an ecosystem built in a different language. In this work, we are the first to consider employing deep neural networks toward tackling this problem. We observe that program translation is a modular procedure, in which a sub-tree of the source tree is translated into the corresponding target sub-tree at each step. To capture this intuition, we design a tree-to-tree neural network as an encoder-decoder architecture to translate a source tree into a target one. Meanwhile, we develop an attention mechanism for the tree-to-tree model, so that when the decoder expands one non-terminal in the target tree, the attention mechanism locates the corresponding sub-tree in the source tree to guide the expansion of the decoder. We evaluate the program translation capability of our tree-to-tree model against several state-of-the-art approaches. Compared against other neural translation models, we observe that our approach is consistently better than the baselines with a margin of up to 15 points. Further, our approach can improve the previous state-of-the-art program translation approaches by a margin of 20 points on the translation of real-world projects.}
}
@article{ChenetalIEEE-12,
       author = {Quan Chen and Ho, S.L. and Fu, W.N.},
        title = {Numerical Investigation of Magnetic Resonant Coupling Technique in Inter-Chip Communication via Electromagnetics-{TCAD} Coupled Simulation},
      journal = {{IEEE Transactions on Magnetics}},
         year = 2012,
       volume = 48,
        issue = 11,
        pages = {4253-4256},
     abstract = {Wireless power transfer based on magnetic resonant coupling (MRC) has attracted extensive studies for its applications in macroscopic engineering. In this work, we explore one potential application of MRC in microelectronics, i.e., as an enhancement of inductive coupling for wireless inter-chip communication. Theoretical analysis and numerical simulation are conducted to study the advantages of MRC in power transfer efficiency. In particular, a new EM-semiconductor simulator is employed to provide a faithful description for the interactions between magnetic fields and semiconductor carrier transport.},
}
@article{ChenetalJACS-11,
    author = {Chen, Chia-Chun and Lin, Yen-Ping and Wang, Chih-Wei and Tzeng, Hsiao-Chien and Wu, 
              Chia-Hsuan and Chen, Yi-Cheng and Chen, Chin-Pei and Chen, Li-Chyong and Wu, Yi-Chun},
     title = {DNA−Gold Nanorod Conjugates for Remote Control of Localized Gene Expression by near Infrared Irradiation},
   journal = {Journal of the American Chemical Society},
    volume = {128},
     issue = {11},
      year = {2006},
     pages = {3709-3715},
}
@article{ChenetalMEMS-10,
    author = {Chen, Chang-Hsiao and Chuang, Shih-Chang and Lee, Yu-Tao and Chang, Yen-Chung and Yeh, Shih-Rung and Yao, Da-Jeng},
     title = {Three-dimensional flexible microprobe for recording the neural signal},
   journal = {Journal of Micro/Nanolithography, MEMS, and MOEMS},
    volume = 9,
     issue = 3,
      year = 2010,
     pages = 031007,
}
@article{ChenetalNATURE-13,
    author = {Chen, Tsai-Wen and Wardill, Trevor J. and Sun, Yi and Pulver, Stefan R. and Renninger, Sabine L. and Baohan, Amy and Schreiter, Eric R. and Kerr, Rex A. and Orger, Michael B. and Jayaraman, Vivek and Looger, Loren L. and Svoboda, Karel and Kim, Douglas S.},
     title = {Ultrasensitive fluorescent proteins for imaging neuronal activity},
   journal = {Nature},
 publisher = {Nature Publishing Group},
    volume = 499,
     issue = 7458,
      year = 2013,
     pages = {295-300},
  abstract = {Fluorescent calcium sensors are widely used to image neural activity. Using structure-based mutagenesis and neuron-based screening, we developed a family of ultrasensitive protein calcium sensors (GCaMP6) that outperformed other sensors in cultured neurons and in zebrafish, flies and mice in vivo. In layer 2/3 pyramidal neurons of the mouse visual cortex, GCaMP6 reliably detected single action potentials in neuronal somata and orientation-tuned synaptic calcium transients in individual dendritic spines. The orientation tuning of structurally persistent spines was largely stable over timescales of weeks. Orientation tuning averaged across spine populations predicted the tuning of their parent cell. Although the somata of GABAergic neurons showed little orientation tuning, their dendrites included highly tuned dendritic segments (5-40-mm long). GCaMP6 sensors thus provide new windows into the organization and dynamics of neural circuits over multiple spatial and temporal scales.},
}
@article{ChenetalNATURE-NEUROSCIENCE-14,
       author = {Chen, C. H. and Fremont, R.  and Arteaga-Bracho, E. E. and Khodakhah, K.},
        title = {Short latency cerebellar modulation of the basal ganglia},
      journal = {Nature Neuroscience},
       volume = {17},
       number = {12},
         year = {2014},
        pages = {1767-1775},
     abstract = {The graceful, purposeful motion of our body is an engineering feat that remains unparalleled in robotic devices using advanced artificial intelligence. Much of the information required for complex movements is generated by the cerebellum and the basal ganglia in conjunction with the cortex. Cerebellum and basal ganglia have been thought to communicate with each other only through slow, multi-synaptic cortical loops, begging the question as to how they coordinate their outputs in real time. We found that the cerebellum rapidly modulates the activity of the striatum via a disynaptic pathway in mice. Under physiological conditions, this short latency pathway was capable of facilitating optimal motor control by allowing the basal ganglia to incorporate time-sensitive cerebellar information and by guiding the sign of cortico-striatal plasticity. Conversely, under pathological condition, this pathway relayed aberrant cerebellar activity to the basal ganglia to cause dystonia.}
}
@article{ChenetalNEURON-11,
     title = {Genetic Influences on Cortical Regionalization in the Human Brain},
    author = {Chi-Hua Chen and Matthew S. Panizzon and Lisa T. Eyler and Terry L. Jernigan and Wes Thompson and Christine Fennema-Notestine and Amy J. Jak and Michael C. Neale and  Carol E. Franz and Samar Hamza and Michael J. Lyons and Michael D. Grant and Bruce Fischl and Larry J. Seidman and Ming T. Tsuang and William S. Kremen and Anders M. Dale},
   journal = {Neuron},
    volume = 72,
     issue = 4,
     pages = {537-544},
      year = 2011,
  abstract = {Animal data demonstrate that the development of distinct cortical areas is influenced by genes that exhibit highly regionalized expression patterns. In this paper, we show genetic patterning of cortical surface area derived from MRI data from 406 adult human twins. We mapped genetic correlations of areal expansion between selected seed regions and all other cortical locations, with the selection of seed points based on results from animal studies. "Marching seeds" and a data-driven, hypothesis-free  fuzzy-clustering approach provided convergent validation. The results reveal strong anterior-to-posterior graded, bilaterally symmetric patterns of regionalization, largely consistent with patterns previously reported in non-human mammalian models. Broad similarities in genetic patterning between rodents and humans might suggest a conservation of cortical patterning mechanisms, whereas dissimilarities might reflect the functionalities most essential to each species.},
  comment = {Highlights: Overall genetic regionalization in humans reflects general mammalian patterning. The patterns are largely bilaterally symmetric and correspond to lobar divisions. There are also regionalization differences consistent with species-specific features. The genetic patterns of the cortical surface area and thickness are different.},
}
@article{ChenetalNEURON-14,
       author = {Chen, Yi and Akin, Orkun and Nern, Aljoscha and Tsui, CY Kimberly and Pecot, Matthew Y. and Zipursky, S. Lawrence},
        title = {Cell-type Specific Labeling of Synapses in vivo through Synaptic Tagging with Recombination ({STaR})},
      journal = {Neuron},
         year = {2014},
       volume = {81},
        issue = {2},
        pages = {280-293},
     abstract = {The study of synaptic specificity and plasticity in the Central Nervous System (CNS) is limited by the inability to efficiently visualize synapses in identified neurons using light microscopy. Here we describe Synaptic Tagging with Recombination (STaR), a method for labeling endogenous presynaptic and postsynaptic proteins in a cell-type specific fashion. We modified genomic loci encoding synaptic proteins within Bacterial Artificial Chromosomes such that these proteins, expressed at endogenous levels and with normal spatiotemporal patterns, were labeled in an inducible fashion in specific neurons through targeted expression of site-specific recombinases. Within the Drosophila visual system, the number and distribution of synapses correlate with Electron Microscopy studies. Using two different recombination systems, presynaptic and postsynaptic specializations of synaptic pairs can be co-labeled. STaR also allows synapses within the CNS to be studied in live animals non-invasively. In principle, STaR can be adapted to the mammalian nervous system.},
}
@article{ChenetalNEURON-18,
       author = {Chen, X. and Mu, Y. and Hu, Y. and Kuan, A. T. and Nikitchenko, M. and Randlett, O. and Chen, A. B. and Gavornik, J. P. and Sompolinsky, H. and Engert, F. and Ahrens, M. B.},
        title = {{Brain-wide Organization of Neuronal Activity and Convergent Sensorimotor Transformations in Larval Zebrafish}},
      journal = {Neuron},
         year = {2018},
       volume = {100},
       number = {4},
        pages = {876-890},
     abstract = {Simultaneous recordings of large populations of neurons in behaving animals allow detailed observation of high-dimensional, complex brain activity. However, experimental approaches often focus on singular behavioral paradigms or brain areas. Here, we recorded whole-brain neuronal activity of larval zebrafish presented with a battery of visual stimuli while recording fictive motor output. We identified neurons tuned to each stimulus type and motor output and discovered groups of neurons in the anterior hindbrain that respond to different stimuli eliciting similar behavioral responses. These convergent sensorimotor representations were only weakly correlated to instantaneous motor activity, suggesting that they critically inform, but do not directly generate, behavioral choices. To catalog brain-wide activity beyond explicit sensorimotor processing, we developed an unsupervised clustering technique that organizes neurons into functional groups. These analyses enabled a broad overview of the functional organization of the brain and revealed numerous brain nuclei whose neurons exhibit concerted activity patterns.}
}
@article{ChenetalNIPS-16,
       author = {Xinyun Chen and Chang Liu and Richard Shin and Dawn Song and Mingcheng Chen},
        title = {Latent Attention for If-Then Program Synthesis},
      journal = {CoRR},
       volume = {arXiv:1611.01867},
         year = {2016},
          url = {https://arxiv.org/abs/1611.01867},
     abstract = {Automatic translation from natural language descriptions into programs is a longstanding challenging problem. In this work, we consider a simple yet important sub-problem: translation from textual descriptions to If-Then programs. We devise a novel neural network architecture for this task which we train end-to-end. Specifically, we introduce Latent Attention, which computes multiplicative weights for the words in the description in a two-stage process with the goal of better leveraging the natural language structures that indicate the relevant parts for predicting program elements. Our architecture reduces the error rate by 28.57\% compared to prior art. We also propose a one-shot learning scenario of If-Then program synthesis and simulate it with our existing dataset. We demonstrate a variation on the training procedure for this scenario that outperforms the original procedure, significantly closing the gap to the model trained with all data.}
}
@article{ChenetalPAMI-13,
       author = {Bo Chen and Polatkan, G. and Sapiro, G. and Blei, D. and Dunson, D. and Carin, L.},
      journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
        title = {Deep Learning with Hierarchical Convolutional Factor Analysis},
         year = 2013,
       volume = 35,
        issue = 8,
        pages = {1887-1901},
     abstract = {Unsupervised multilayered ("deep") models are considered for imagery. The model is represented using a hierarchical convolutional factor-analysis construction, with sparse factor loadings and scores. The computation of layer-dependent model parameters is implemented within a Bayesian setting, employing a Gibbs sampler and variational Bayesian (VB) analysis that explicitly exploit the convolutional nature of the expansion. To address large-scale and streaming data, an online version of VB is also developed. The number of dictionary elements at each layer is inferred from the data, based on a beta-Bernoulli implementation of the Indian buffet process. Example results are presented for several image-processing applications, with comparisons to related models in the literature.},
}
@article{ChenetalSCIENCE-15,
       author = {Chen, Fei and Tillberg, Paul W. and Boyden, Edward S.},
        title = {Expansion microscopy},
      journal = {Science},
       volume = 347,
        issue = 6221,
        pages = {543-548},
         year = 2015,
     abstract = {In optical microscopy, fine structural details are resolved by using refraction to magnify images of a specimen. We discovered that by synthesizing a swellable polymer network within a specimen, it can be physically expanded, resulting in physical magnification. By covalently anchoring specific labels located within the specimen directly to the polymer network, labels spaced closer than the optical diffraction limit can be isotropically separated and optically resolved, a process we call expansion microscopy (ExM). Thus, this process can be used to perform scalable superresolution microscopy with diffraction-limited microscopes. We demonstrate ExM with apparent \tilde{}70-nanometer lateral resolution in both cultured cells and brain tissue, performing three-color superresolution imaging of \tilde{}107 cubic micrometers of the mouse hippocampus with a conventional confocal microscope.},
}
@misc{ChenetalSCIENCE-15.supplement,
        title = {Expansion microscopy: Supplementary Materials},
       author = {Chen, Fei and Tillberg, Paul W. and Boyden, Edward S.},
         year = 2015,
 howpublished = {Available from the pages of Science Magazine},
}
@article{CheneyandSeyfarthBBS-11,
       author = {Cheney, Dorothy and Seyfarth, Robert},
         year = {2011},
        pages = {135-147},
        title = {Pr{\'{e}}cis of How monkeys see the world},
       volume = {15},
      journal = {Behavioral and Brain Sciences},
}
@article{ChengPAMI-95,
       author = {Yizong Cheng},
        title = {Mean Shift, Mode Seeking, and Clustering},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
       volume = 17,
        issue = 8,
         year = 1995,
        pages = {790-799},
}
@article{ChengetalNATURE_METHODS-11,
       author = {Cheng, Adrian and Goncalves, J. Tiago and Golshani, Peyman and Arisaka, Katsushi and Portera-Cailliau, Carlos},
        title = {Simultaneous two-photon calcium imaging at different depths with spatiotemporal multiplexing},
      journal = {Nature Methods},
         year = {2011},
    publisher = {Nature Publishing Group},
       volume = {8},
        issue = {2},
        pages = {139-142},
     abstract = {In vivo two-photon calcium imaging would benefit from the use of multiple excitation beams to increase scanning speed, signalto-noise ratio and field of view or to image different axial planes simultaneously. Using spatiotemporal multiplexing we circumvented light-scattering ambiguity inherent to deep-tissue multifocal two-photon microscopy. We demonstrate calcium imaging at multiple axial planes in the intact mouse brain to monitor network activity of ensembles of cortical neurons in three spatial dimensions.}
}
@book{ChernoffandMoses59,
       author = {Chernoff, Herman and Moses, Lincoln E.},
        title = {Elementary Decision Theory},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1959,
}
@article{CherubiniandMilesFiCN-15,
       author = {Cherubini, Enrico and Miles, Richard},
        title = {The {CA3} region of the hippocampus: {H}ow is it? {W}hat is it for? {H}ow does it do it?},
      journal = {Frontiers in Cellular Neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = {9},
         year = {2015},
        pages = {19-19},
     abstract = {The hippocampus, in the temporal lobe, is phylogenetically one of the oldest parts of the brain and forms part of the limbic system. The hippocampus proper is defined by the dentate gyrus and Cornu Ammonis (CA). While the dentate gyrus contains the fascia dentata and the hilus, the CA is anatomically and functionally differentiated into distinct subfields named CA1, CA2, CA3, and CA4. The CA3 region has attracted major attention in recent years for its specific role in memory processes, susceptibility to seizures and neuro-degeneration. Internal connectivity in the CA3 subfield is more rich than in other hippocampal regions. Recurrent axon collaterals of CA3 pyramidal cells ramify extensively making excitatory contacts with neighboring excitatory and inhibitory neurons [...]}
}
@article{ChiaandWuJBE-15,
       author = {Chia, Helena N. and Wu, Benjamin M.},
        title = {Recent advances in {3D} printing of biomaterials},
      journal = {Journal of Biological Engineering},
    publisher = {BioMed Central},
      address = {London},
       volume = {9},
        pages = {4},
         year = {2015},
     abstract = {3D Printing promises to produce complex biomedical devices according to computer design using patient-specific anatomical data. Since its initial use as pre-surgical visualization models and tooling molds, {3D} Printing has slowly evolved to create one-of-a-kind devices, implants, scaffolds for tissue engineering, diagnostic platforms, and drug delivery systems. Fueled by the recent explosion in public interest and access to affordable printers, there is renewed interest to combine stem cells with custom {3D} scaffolds for personalized regenerative medicine. Before {3D} Printing can be used routinely for the regeneration of complex tissues (e.g. bone, cartilage, muscles, vessels, nerves in the craniomaxillofacial complex), and complex organs with intricate {3D} microarchitecture (e.g. liver, lymphoid organs), several technological limitations must be addressed. In this review, the major materials and technology advances within the last five years for each of the common {3D} Printing technologies (Three Dimensional Printing, Fused Deposition Modeling, Selective Laser Sintering, Stereolithography, and {3D} Plotting/Direct-Write/Bioprinting) are described. Examples are highlighted to illustrate progress of each technology in tissue engineering, and key limitations are identified to motivate future research and advance this fascinating field of advanced manufacturing.},
}
@article{ChianetalCURRENT-BIOLOGY-11,
       author = {Chiang, Ann-Shyn and Lin, Chih-Yung and Chuang, Chao-Chun and Chang, Hsiu-Ming and Hsieh, Chang-Huain and Yeh, Chang-Wei and Shih, Chi-Tin and Wu, Jian-Jheng and Wang, Guo-Tzau and Chen, Yung-Chang and Wu, Cheng-Chi and Chen, Guan-Yu and Ching, Yu-Tai and Lee, Ping-Chang and Lin, Chih-Yang and Lin, Hui-Hao and Wu, Chia-Chou and Hsu, Hao-Wei and Huang, Yun-Ann and Chen, Jing-Yi and Chiang, Hsin-Jung and Lu, Chun-Fang and Ni, Ru-Fen and Yeh, Chao-Yuan and Hwang, Jenn-Kang},
        title = {Three-Dimensional Reconstruction of Brain-wide Wiring Networks in <em>Drosophila</em> at Single-Cell Resolution},
      journal = {Current Biology},
         year = {2011},
    publisher = {Elsevier},
       volume = {21},
        issue = {1},
        pages = {1-11},
     abstract = {Animal behavior is governed by the activity of interconnected brain circuits. Comprehensive brain wiring maps are thus needed in order to formulate hypotheses about information flow and also to guide genetic manipulations aimed at understanding how genes and circuits orchestrate complex behaviors.},
}
@article{ChiangetalBIOXRIV-19,
       author = {Chiang, Jeffrey N. and Peng, Yujia and Lu, Hongjing and Holyoak, Keith J. and Monti, Martin M.},
        title = {Distributed Code for Semantic Relations Predicts Neural Similarity},
      journal = {bioRxiv},
         year = {2019},
    publisher = {Cold Spring Harbor Laboratory},
     abstract = {The ability to generate and process semantic relations is central to many aspects of human cognition. Theorists have long debated whether such relations are coded as atomistic links in a semantic network, or as distributed patterns over some core set of abstract relations. The form and content of the conceptual and neural representations of semantic relations remains to be empirically established. The present study combined computational modeling and neuroimaging to investigate the representation and comparison of abstract semantic relations in the brain. By using sequential presentation of verbal analogies, we decoupled the neural activity associated with encoding the representation of the first-order semantic relation between words in a pair from that associated with the second-order comparison of two relations. We tested alternative computational models of relational similarity in order to distinguish between rival accounts of how semantic relations are coded and compared in the brain. Analyses of neural similarity patterns supported the hypothesis that semantic relations are coded, in the parietal cortex, as distributed representations over a pool of abstract relations specified in a theory-based taxonomy. These representations, in turn, provide the immediate inputs to the process of analogical comparison, which draws on a broad frontoparietal network. This study sheds light not only on the form of relation representations but also on their specific content.Significance Relations provide basic building blocks for language and thought. For the past half century, cognitive scientists exploring human semantic memory have sought to identify the code for relations. In a neuroimaging paradigm, we tested alternative computational models of relation processing that predict patterns of neural similarity during distinct phases of analogical reasoning. The findings allowed us to draw inferences not only about the form of relation representations, but also about their specific content. The core of these distributed representations is based on a relatively small number of abstract relation types specified in a theory-based taxonomy. This study helps to resolve a longstanding debate concerning the nature of the conceptual and neural code for semantic relations in the mind and brain.},
}
@article{ChiangetalNATURE-96,
       author = {Chiang, Chin and Litingtung, Ying and Lee, Eric and Young, Keith E. and Corden, Jeffrey L. and Westphal, Heiner and Beachy, Philip A.},
        title = {Cyclopia and defective axial patterning in mice lacking Sonic hedgehog gene function},
      journal = {Nature},
       volume = 383,
        issue = 6599,
         year = 1996,
        pages = {407-413},
}
@article{ChiangetalPHYSICAL-REVIEW-LETTERS-77,
       author = {Chiang, C. K. and Fincher, Jr., C. R. and Park, Y. W. and Heeger, A. J. and Shirakawa, H. and Louis, E. J. and Gau, S. C. and MacDiarmid, A. G.},
        title = {Electrical Conductivity in Doped Polyacetylene},
      journal = {Physical Review Letters},
       volume = 39,
         year = 1977,
        pages = {1098-1101},
     abstract = {Doped polyacetylene forms a new class of conducting polymers in which the electrical conductivity can be systematically and continuously varied over a range of eleven orders of magnitude. Transport studies and far-infrared transmission measurements imply a metal-to-insulator transition at dopant concentrations near 1\%.}
}
@article{ChicaetalBSF-18,
       author = {Chica, Ana B. and Thiebaut de Schotten, Michel and Bartolomeo, Paolo and Paz-Alonso, Pedro M.},
        title = {White matter microstructure of attentional networks predicts attention and consciousness functional interactions},
      journal = {Brain Structure and Function},
       volume = {223},
       number = {2},
         year = {2018},
        pages = {653-668},
     abstract = {Attention is considered as one of the pre-requisites of conscious perception. Phasic alerting and exogenous orienting improve conscious perception of near-threshold information through segregated brain networks. Using a multimodal neuroimaging approach, combining data from functional MRI (fMRI) and diffusion-weighted imaging (DWI), we investigated the influence of white matter properties of the ventral branch of superior longitudinal fasciculus (SLF III) in functional interactions between attentional systems and conscious perception. Results revealed that (1) reduced integrity of the left hemisphere SLF III was predictive of the neural interactions observed between exogenous orienting and conscious perception, and (2) increased integrity of the left hemisphere SLF III was predictive of the neural interactions observed between phasic alerting and conscious perception. Our results combining fMRI and DWI data demonstrate that structural properties of the white matter organization determine attentional modulations over conscious perception.}
}
@inproceedings{ChienandWeissmanIJCAI-75,
       author = {Chien, R. T. and Weissman, S.},
        title = {Planning and Execution in Incompletely Specified Environments},
    booktitle = {Proceedings {IJCAI} 4},
     location = {Tbilisi, Georgia, USSR},
 organization = {IJCAII},
         year = 1975,
        pages = {169-174},
}
@article{ChikkeruretalVR-10,
        title = {What and where: A {B}ayesian inference theory of attention},
       author = {S. Chikkerur, T. Serre, C. Tan and T. Poggio},
      journal = {Vision Research (submitted)},
       volume = 0,
        issue = 0,
        pages = 0,
         year = 2010,
}
@misc{Child_Development_TrackerPBS-17,
        title = {Child Development Tracker},
       author = {Public Broadcasting System},
         year = {2017},
 howpublished = {PBS Parents Resources},
          url = {http://www.pbs.org/parents/childdevelopmenttracker/eight/mathematics.html},
}
@inproceedings{ChinandCooperUAI-87,
       author = {Howard Chin and Gregory F. Cooper},
        title = {Stochastic Simulation of {B}ayesian Belief Networks},
    booktitle = UAI-87,
     location = {Seattle, Washington},
         year = 1987,
}
@article{ChinetalJCN-14,
       author = {Chin, An-Lun and Lin, Chih-Yung and Fu, Tsai-Feng and Dickson, Barry J. and Chiang, Ann-Shyn},
        title = {Diversity and wiring variability of visual local neurons in the Drosophila medulla M6 stratum},
      journal = {Journal Comparative Neurology},
    publisher = {BlackWell Publishing Ltd},
       volume = {522},
        issue = {17},
         year = {2014},
        pages = {3795-3816},
     abstract = {Local neurons in the vertebrate retina are instrumental in transforming visual inputs to extract contrast, motion, and color information and in shaping bipolar-to-ganglion cell transmission to the brain. In Drosophila, UV vision is represented by R7 inner photoreceptor neurons that project to the medulla M6 stratum, with relatively little known of this downstream substrate. Here, using R7 terminals as references, we generated a 3D volume model of the M6 stratum, which revealed a retinotopic map for UV representations. Using this volume model as a common 3D framework, we compiled and analyzed the spatial distributions of more than 200 single M6-specific local neurons (M6-LNs). Based on the segregation of putative dendrites and axons, these local neurons were classified into two families, directional and nondirectional. Neurotransmitter immunostaining suggested a signal routing model in which some visual information is relayed by directional M6-LNs from the anterior to the posterior M6 and all visual information is inhibited by a diverse population of nondirectional M6-LNs covering the entire M6 stratum. Our findings suggest that the Drosophila medulla M6 stratum contains diverse LNs that form repeating functional modules similar to those found in the vertebrate inner plexiform layer.},
}
@book{ChippandCorreia1968,
        title = {Henry Moore, "The Sculptor Speaks", 1937},
       author = {Chipp, H.B. and Correia, A.},
    publisher = {Ketterer and Kunst},
       series = {Period in art history: Surrealism to conceptualism},
         year = {1968}
}
@inproceedings{ChistyakovetalICLR-17,
       author = {Alexander Chistyakov and Ekaterina Lobacheva and Arseny Kuznetsov and Alexey Romanenko},
        title = {Semantic embeddings for program behaviour patterns},
    booktitle = {ICLR Workshop},
         year = {2017},
     abstract = {In this paper, we propose a new feature extraction technique for program execution logs. First, we automatically extract complex patterns from a program's behaviour graph. Then, we embed these patterns into a continuous space by training an autoencoder. We evaluate the proposed features on a real-world malicious software detection task. We also find that the embedding space captures interpretable structures in the space of pattern parts.}
}
@article{ChklovskiietalNEURON-02,
    author = {Chklovskii, D. B. and Schikorski, T. and Stevens, C. F.},
     title = {Wiring optimization in cortical circuits},
   journal = {Neuron},
    volume = 34,
     issue = 3,
      year = 2002,
     pages = {341-347},
  abstract = {Wiring a brain presents a formidable problem because neural circuits require an enormous number of fast and durable connections. We propose that evolution was likely to have optimized neural circuits to minimize conduction delays in axons, passive cable attenuation in dendrites, and the length of "wire" used to construct circuits, and to have maximized the density of synapses. Here we ask the question: "What fraction of the volume should be taken up by axons and dendrites (i.e., wire) when these variables are at their optimal values?" The biophysical properties of axons and dendrites dictate that wire should occupy 3/5 of the volume in an optimally wired gray matter. We have measured the fraction of the volume occupied by each cellular component and find that the volume of wire is close to the predicted optimal value.},
}
@article{Choetal98,
     author = {Junghoo Cho and Hector Garc{\'\i}a-Molina and Lawrence Page},
      title = {Efficient Crawling Through {URL} Ordering},
    journal = {Computer Networks and ISDN Systems},
     volume = 30,
      issue = {1-7},
      pages = {161-172},
        url = {http://citeseer.nj.nec.com/30730.html},
       year = 1998,
}
@inproceedings{Choetal98b,
     author = {J. Cho and H. Garcia-Molina and L. Page},
      title = {Efficient crawling through {URL} ordering},
  booktitle = {Proceedings of the 7th International Conference on World Wide Web},
       year = 1998,
}
@article{ChoetalCoRR-14,
       author = {K. Cho and B. Merri\"{e}nboer and C. Gulcehre and F. Bougares and H. Schwenk and Y. Bengio},
        title = {Learning phrase representations using RNN encoder-decoder for statistical machine translation},
      journal = {CoRR},
       volume = {arXiv:406.1078},
         year = 2014,
          url = {http://arxiv.org/abs/406.1078},
     abstract = {In this paper, we propose a novel neural network model called RNN Encoder-Decoder that consists of two recurrent neural networks (RNN). One RNN encodes a sequence of symbols into a fixed-length vector representation, and the other decodes the representation into another sequence of symbols. The encoder and decoder of the proposed model are jointly trained to maximize the conditional probability of a target sequence given a source sequence. The performance of a statistical machine translation system is empirically found to improve by using the conditional probabilities of phrase pairs computed by the RNN Encoder-Decoder as an additional feature in the existing log-linear model. Qualitatively, we show that the proposed model learns a semantically and syntactically meaningful representation of linguistic phrases.},
}
@article{ChoetalCoRR-15,
        title = {On the Properties of Neural Machine Translation: Encoder-Decoder Approaches},
       author = {Kyunghyun Cho and Bart van Merri\"{e}nboer and Dzmitry Bahdanau and Yoshua Bengio},
      journal = {CoRR},
       volume = {arXiv:1409.1259},
         year = 2015,
     abstract = {Neural machine translation is a relatively new approach to statistical machine translation based purely on neural networks. The neural machine translation models often consist of an encoder and a decoder. The encoder extracts a fixed-length representation from a variable-length input sentence, and the decoder generates a correct translation from this representation. In this paper, we focus on analyzing the properties of the neural machine translation using two models; RNN Encoder-Decoder and a newly proposed gated recursive convolutional neural network. We show that the neural machine translation performs relatively well on short sentences without unknown words, but its performance degrades rapidly as the length of the sentence and the number of unknown words increase. Furthermore, we find that the proposed gated recursive convolutional network learns a grammatical structure of a sentence automatically.},
}
@inproceedings{ChoetalEMNLP-14,
       author = {Cho, Kyunghyun and van Merri{\"{e}}nboer, Bart and G{\"{u}}l{\c c}ehre, {\c C}alar and Bahdanau, Dzmitry and Bougares, Fethi and Schwenk, Holger and Bengio, Yoshua},
        title = {Learning Phrase Representations using RNN Encoder--Decoder for Statistical Machine Translation},
    booktitle = {Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP)},
         year = {2014},
        pages = {1724-1734},
    publisher = {Association for Computational Linguistics},
      address = {Doha, Qatar},
}
@article{ChoietalNATURE-07,
      title = {Renal Clearance of Nanoparticles},
     author = {Hak Soo Choi and Wenhao Liu and Preeti Misra and Eiichi Tanaka and John P. Zimmer and Binil Itty Ipe and Moungi G. Bawendi and John V. Frangioni},
    journal = {Nature Biotechnology},
     volume = 25,
      issue = 10,
       year = 2007,
      pages = {1165-1170},
}
@book{Chomsky86,
   author = {Noam Chomsky},
    title = {Knowledge of Language: Its Nature, Origin and Use},
publisher = {Praeger},
  address = {New York, NY},
     year = 1986,
}
@book{Chomsky86philosophy,
       author = {Noam Chomsky},
        title = {Knowledge of Language: {I}ts Nature, Origin and Use},
    publisher = {Praeger},
      address = {Santa Barbara, CA},
         year = 1986,
}
@article{ChoromanskaetalFNC-12,
       author = {Choromanska, A. and Chang, S. F. and Yuste, R.},
        title = {{{A}utomatic reconstruction of neural morphologies with multi-scale tracking}},
      journal = {Front Neural Circuits},
         year = 2012,
       volume = 6,
        pages = 25,
     abstract = {Neurons have complex axonal and dendritic morphologies that are the structural building blocks of neural circuits. The traditional method to capture these morphological structures using manual reconstructions is time-consuming and partly subjective, so it appears important to develop automatic or semi-automatic methods to reconstruct neurons. Here we introduce a fast algorithm for tracking neural morphologies in 3D with simultaneous detection of branching processes. The method is based on existing tracking procedures, adding the machine vision technique of multi-scaling. Starting from a seed point, our algorithm tracks axonal or dendritic arbors within a sphere of a variable radius, then moves the sphere center to the point on its surface with the shortest Dijkstra path, detects branching points on the surface of the sphere, scales it until branches are well separated and then continues tracking each branch. We evaluate the performance of our algorithm on preprocessed data stacks obtained by manual reconstructions of neural cells, corrupted with different levels of artificial noise, and unprocessed data sets, achieving 90\% precision and 81\% recall in branch detection. We also discuss limitations of our method, such as reconstructing highly overlapping neural processes, and suggest possible improvements. Multi-scaling techniques, well suited to detect branching structures, appear a promising strategy for automatic neuronal reconstructions.}
}
@book{Chorost2011world,
        title = {World Wide Mind: The Coming Integration of Humanity, Machines, and the Internet},
       author = {Chorost, Michael},
         year = {2011},
    publisher = {Free Press}
}
@inproceedings{ChotimongkolandRudnickyEMNLP-08,
       author = {Ananlada Chotimongkol and Alexander I. Rudnicky},
        title = {Acquiring Domain-Specific Dialog Information from Task-Oriented Human-Human Interaction through an Unsupervised Learning},
    booktitle = {Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing},
         year = 2011,
        pages = {955-964},
     location = {Honolulu, Hawaii},
    publisher = {Association for Computational Linguistics},
     abstract = {We describe an approach for acquiring the domain-specific dialog knowledge required to configure a task-oriented dialog system that uses human-human interaction data. The key aspects of this problem are the design of a dialog information representation and a learning approach that supports capture of domain information from in-domain dialogs. To represent a dialog for a learning purpose, we based our representation, the form-based dialog structure representation, on an observable structure. We show that this representation is sufficient for modeling phenomena that occur regularly in several dissimilar task-oriented domains, including information-access and problem-solving. With the goal of ultimately reducing human annotation effort, we examine the use of unsupervised learning techniques in acquiring the components of the form-based representation (i.e. task, subtask, and concept). These techniques include statistical word clustering based on mutual information and Kullback-Liebler distance, TextTiling, HMM-based segmentation, and bisecting K-mean document clustering. With some modifications to make these algorithms more suitable for inferring the structure of a spoken dialog, the unsupervised learning algorithms show promise.}
}
@article{ChouetalNATURE-NEUROSCIENCE-10,
       author = {Chou, Ya-Hui and Spletter, Maria L. and Yaksi, Emre and Leong, Jonathan C. S. and Wilson, Rachel I. and Luo, Liqun},
        title = {Diversity and Wiring Variability of Olfactory Local Interneurons in the Drosophila Antennal Lobe},
      journal = {Nature Neuroscience},
         year = {2010},
       volume = {13},
        issue = {4},
        pages = {439-449},
     abstract = {Local interneurons play essential roles in information processing by neural circuits. Here we present a comprehensive genetic, anatomical, and electrophysiological analysis of local interneurons (LNs) in the Drosophila antennal lobe, the first olfactory processing center in the brain. We find that LNs are diverse in their neurotransmitter profiles, connectivity, and physiological properties. Analysis of >1500 individual LNs reveals major morphological classes characterized by coarsely stereotyped glomerular innervation patterns. Some of these morphological classes exhibit distinct physiological properties. However, the finer-scale connectivity of an individual LN varies considerably across brains and there is notable physiological variability within each morphological or genetic class. Finally, we show that LN innervation requires interaction with olfactory receptor neurons during development, and some individual variability also likely reflects LN-LN interactions. Our results reveal an unexpected degree of complexity and individual variation in an invertebrate neural circuit, a result that creates challenges for solving the Drosophila connectome.},
}
@article{ChowandBoydenSTM-13,
        title = {Optogenetics and Translational Medicine},
       author = {Chow, Brian Y. and Boyden, Edward S.},
      journal = {Science Translational Medicine},
       volume = 5,
        issue = 177,
        pages = {177ps5},
         year = 2013,
     abstract = {Optogenetic tools enable light-mediated control of cellular excitability and signaling in vivo. By manipulating biological processes, scientists can determine the roles played by these processes in intact biological systems, such as the brain. Such cellular-level control has greatly affected basic science. Here, we discuss how optogenetic tools might be translated into clinical impact through identification of new molecular and circuit-level targets and provide temporally precise interventions for defined biochemical or cellular events.},
}
@article{ChowandLiuIEEE-68,
       author = {C. Chow and C. Liu},
        title = {Approximating Discrete Probability Distributions with Dependence Trees},
      journal = {{IEEE} Transactions on Information Theory},
       volume = 114,
         year = 1968,
        pages = {462-467},
}
@article{ChowdhuryandMemoliCoRR-17,
        title = {Persistent Path Homology of Directed Networks},
       author = {Samir Chowdhury and Facundo M\'{e}́moli},
      journal = {CoRR},
       volume = {arXiv:1701.00565},
         year = 2017,
     abstract = {While standard persistent homology has been successful in extracting information from metric datasets, its applicability to more general data, e.g. directed networks, is hindered by its natural insensitivity to asymmetry. We study a construction of homology of digraphs due to Grigor'yan, Lin, Muranov and Yau, and extend this construction to the persistent framework. The result, which we call persistent path homology, can provide information about the digraph structure of a directed network at varying resolutions. Moreover, this method encodes a rich level of detail about the asymmetric structure of the input directed network. We test our method on both simulated and real-world directed networks and conjecture some of its characteristics.}
}
@article{ChristensenNEURON-02,
     title = {A primary culture system for functional analysis of C. elegans neurons and muscle cells},
    author = {Christensen, M. and Estevez, A. and Yin, X. and Fox, R. and Morrison, R. and McDonnell, M. and Gleason, C. and Miller 3rd, D.M. and Strange, K.},
   journal = {Neuron},
    volume = 33,
     issue = 4,
     pages = {503-14},
      year = 2002,
  abstract = {C.\ elegans has provided important insights into neuromuscular system function and development. However, the animal's small size limits access to individual neurons and muscle cells for physiological, biochemical, and molecular study. We describe here primary culture methods that allow C. elegans embryonic cells to differentiate into neurons and muscle cells in vitro. Morphological  electrophysiological, and GFP reporter studies demonstrate that the differentiation and functional properties of cultured cells are similar to those observed in vivo. Enriched populations of cells expressing specific GFP reporters can be generated by fluorescence-activated cell sorting. Addition of double-stranded RNA to the culture medium induces dramatic knockdown of targeted gene expression. Primary nematode cell culture provides a new foundation for a wide variety of experimental opportunities heretofore unavailable in the field.}
}
@book{ChristianandGriffiths2016algorithms,
        title = {Algorithms to Live By: The Computer Science of Human Decisions},
       author = {Brian Christian and Tom Griffiths},
         year = {2016},
    publisher = {HarperCollins Publishers}
}
@article{ChristiansenandChaterBBS-08,
       author = {Christiansen, M. H. and Chater, N.},
        title = {Language as shaped by the brain},
      journal = {Behavior Brain Science},
         year = {2008},
       volume = {31},
       number = {5},
        pages = {489-508},
     abstract = {It is widely assumed that human learning and the structure of human languages are intimately related. This relationship is frequently suggested to derive from a language-specific biological endowment, which encodes universal, but communicatively arbitrary, principles of language structure (a Universal Grammar or UG). How might such a UG have evolved? We argue that UG could not have arisen either by biological adaptation or non-adaptationist genetic processes, resulting in a logical problem of language evolution. Specifically, as the processes of language change are much more rapid than processes of genetic change, language constitutes a "moving target" both over time and across different human populations, and, hence, cannot provide a stable environment to which language genes could have adapted. We conclude that a biologically determined UG is not evolutionarily viable. Instead, the original motivation for UG--the mesh between learners and languages--arises because language has been shaped to fit the human brain, rather than vice versa. Following Darwin, we view language itself as a complex and interdependent "organism," which evolves under selectional pressures from human learning and processing mechanisms. That is, languages themselves are shaped by severe selectional pressure from each generation of language users and learners. This suggests that apparently arbitrary aspects of linguistic structure may result from general learning and processing biases deriving from the structure of thought processes, perceptuo-motor factors, cognitive limitations, and pragmatics.}
}
@article{ChristiansenandChaterCS-99,
        title = {Toward a connectionist model of recursion in human linguistic performance},
       author = {Morten H. Christiansen and Nick Chater},
      journal = {Cognitive Science},
       volume = {23},
         year = {1999},
        pages = {157-205}
}
@article{ChristiansenandChaterML-94,
       author = {M. H. Christiansen and Nick Chater},
        title = {Generalization and Connectionist Language Learning},
      journal = {Mind and Language},
       volume = {9},
       number = {3},
         year = {1994},
        pages = {273-87},
}
@inproceedings{ChristiansenandGoldberg90,
   author =    {Alan Christiansen and Ken Goldberg},
   title =     {Robotic Manipulation Planning with Stochastic Actions},
   booktitle = {DARPA Workshop on Innovative Approaches to Planning, Scheduling
                and Control},
   location =  {San Diego,California},
   year =      1990,
   page =      {3-8},
}
@article{ChristieetalJNE-15,
       author = {Christie B.\ P. and Tat D.\ M. and Irwin Z.\ T. and Gilja V. and Nuyujukian P. and Foster J.\ D. and Ryu S.\ I. and Shenoy K.\ V. and Thompson D.\ E. and Chestek C.\ A.},
        title = {Comparison of spike sorting and thresholding of voltage waveforms for intracortical brain-machine interface performance},
      journal = {Journal of Neural Engineering},
       volume = 12,
        issue = 1,
         year = 2015,
        pages = 016009,
     abstract = {Objective. For intracortical brain-machine interfaces (BMIs) action potential voltage waveforms are often sorted to separate out individual neurons. If these neurons contain independent tuning information this process could increase BMI performance. However the sorting of action potentials (spikes) requires high sampling rates and is computationally expensive. To explicitly define the difference between spike sorting and alternative methods we quantified BMI decoder performance when using threshold-crossing events versus sorted action potentials. Approach. We used data sets from 58 experimental sessions from two rhesus macaques implanted with Utah arrays. Data were recorded while the animals performed a center-out reaching task with seven different angles. For spike sorting neural signals were sorted into individual units by using a mixture of Gaussians to cluster the first four principal components of the waveforms. For thresholding events spikes that simply crossed a set threshold were retained. We decoded the data offline using both a Naive Bayes classifier for reaching direction and a linear regression to evaluate hand position. Main results. We found the highest performance for thresholding when placing a threshold between -3 and -4.5 x Vrms. Spike sorted data outperformed thresholded data for one animal but not the other. The mean Naive Bayes classification accuracy for sorted data was 88.5\% and changed by 5\% on average when data were thresholded. The mean correlation coefficient for sorted data was 0.92 and changed by 0.015 on average when thresholded. Significance. For prosthetics applications these results imply that when thresholding is used instead of spike sorting only a small amount of performance may be lost. The utilization of threshold-crossing events may significantly extend the lifetime of a device because these events are often still detectable once single neurons are no longer isolated.},
}
@article{ChuaIEEE-71,
       author = {Chua, L.O.},
      journal = {{IEEE} Transactions on Circuit Theory},
        title = {Memristor-The missing circuit element},
         year = {1971},
       volume = {18},
        issue = {5},
        pages = {507-519},
     abstract = {A new two-terminal circuit element-called the memristor characterized by a relationship between the charge q(t)\equiv \int_{-\infty }^{t} i(\tau ) d \tau and the flux-linkage \varphi (t)\equiv \int_{- \infty }^{t} v(\tau ) d \tau is introduced as the fourth basic circuit element. An electromagnetic field interpretation of this relationship in terms of a quasi-static expansion of Maxwell's equations is presented. Many circuit-theoretic properties of memistors are derived. It is shown that this element exhibits some peculiar behavior different from that exhibited by resistors, inductors, or capacitors. These properties lead to a number of unique applications which cannot be realized with RLC networks alone. Although a physical memristor device without internal power supply has not yet been discovered, operational laboratory models have been built with the help of active circuits. Experimental results are presented to demonstrate the properties and potential applications of memristors.},
}
@article{ChuaandKangIEEE-76,
       author = {Chua, L.O. and Sung Mo Kang},
      journal = {Proceedings of the {IEEE}},
        title = {Memristive devices and systems},
         year = {1976},
       volume = {64},
        issue = {2},
        pages = {209-223},
     abstract = {A broad generalization of memristors--a recently postulated circuit element--to an interesting class of nonlinear dynamical systems called memristive systems is introduced. These systems are unconventional in the sense that while they behave like resistive devices, they can be endowed with a rather exotic variety of dynamic characteristics. While possessing memory and exhibiting small-signal inductive or capacitive effects, they are incapable of energy discharge and they introduce no phase shift between the input and output waveforms. This zero-crossing property gives rise to a Lissajous figure which always passes through the origin. Memristive systems are hysteretic in the sense that their Lissajous figures vary with the excitation frequency. At very low frequencies, memristive systems are indistinguishable from nonlinear resistors while at extremely high frequencies, they reduce to linear resistors. These anomalous properties have misled and prevented the identification of many memristive devices and systems-including the thermistor, the Hodgkin-Huxley membrane circuit model, and the discharge tubes. Generic properties of memristive systems are derived and a canonic dynamical system model is presented along with an explicit algorithm for identifying the model parameters and functions.},
}
@article{ChuaetalIJBC-12,
       author = {Chua, Leon and Sbitnev, Valery and Kim, Hyongsuk},
        title = {The {Hodgkin-Huxley} axon is made of memristors},
      journal = {International Journal of Bifurcation and Chaos},
       volume = {22},
        issue = {3},
        pages = {1230011},
         year = {2012},
}
@article{ChubykinetalNEURON-13,
       author = {Chubykin, Alexander A. and Roach, Emma B. and Bear, Mark F. and Shuler, Marshall G.  Hussain},
        title = {A Cholinergic Mechanism for Reward Timing within Primary Visual Cortex},
      journal = {Neuron},
    publisher = {Elsevier},
       volume = {77},
        issue = {4},
         year = {2012},
        pages = {723-735},
     abstract = {Neurons in rodent primary visual cortex (V1) relate operantly conditioned stimulus-reward intervals with modulated patterns of spiking output, but little is known about the locus or mechanism of this plasticity. Here we show that cholinergic basal forebrain projections to V1 are necessary for the neural acquisition, but not the expression, of reward timing in the visual cortex of awake, behaving animals. We then mimic reward timing in vitro by pairing white matter stimulation with muscarinic receptor activation at a fixed interval and show that this protocol results in the prolongation of electrically evoked spike train durations out to the conditioned interval. Together, these data suggest that V1 possesses the circuitry and plasticity to support reward time prediction learning and the cholinergic system serves as an important reinforcement signal which, in vivo, conveys to the cortex the outcome of behavior.},
}
@inproceedings{ChuetalCVPR-03,
     author = {Chi-Wei Chu and Odest Chadwicke Jenkins and Maja J Matari\'{c}},
      title = {Markerless Kinematic Model and Motion Capture from Volume Sequences},
  booktitle = {Proceedings of {IEEE} Computer Vision and Pattern Recognition},
     volume = 2,
   location = {Madison, Wisconsin},
      pages = {475-482},
       year = 2003,
}
@article{ChungetalCSHL-15,
       author = {Chung, Won-Suk and Allen, Nicola J. and Eroglu, Cagla},
        title = {Astrocytes Control Synapse Formation, Function, and Elimination},
      journal = {Cold Spring Harbor Perspectives in Biology},
         year = {2015},
    publisher = {Cold Spring Harbor Laboratory Press},
       volume = {7},
        issue = {9},
        pages = {a020370},
     abstract = {Astrocytes, through their close associations with synapses, can monitor and alter synaptic function, thus actively controlling synaptic transmission in the adult brain. Besides their important role at adult synapses, in the last three decades a number of critical findings have highlighted the importance of astrocytes in the establishment of synaptic connectivity in the developing brain. In this article, we will review the key findings on astrocytic control of synapse formation, function, and elimination. First, we will summarize our current structural and functional understanding of astrocytes at the synapse. Then, we will discuss the cellular and molecular mechanisms through which developing and mature astrocytes instruct the formation, maturation, and refinement of synapses. Our aim is to provide an overview of astrocytes as important players in the establishment of a functional nervous system.},
}
@article{ChungetalCoRR-15,
       author = {Junyoung Chung and {\c{C}}aglar G{\"{u}}l{\c{c}}ehre and KyungHyun Cho and Yoshua Bengio},
        title = {Gated Feedback Recurrent Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1502.02367},
         year = {2015},
     abstract = {Developing neural network image classification models often requires significant architecture engineering. In this paper, we attempt to automate this engineering process by learning the model architectures directly on the dataset of interest. As this approach is expensive when the dataset is large, we propose to search for an architectural building block on a small dataset and then transfer the block to a larger dataset. Our key contribution is the design of a new search space which enables transferability. In our experiments, we search for the best convolutional layer (or "cell") on the CIFAR-10 dataset and then apply this cell to the ImageNet dataset by stacking together more copies of this cell, each with their own parameters. Although the cell is not searched for directly on ImageNet, an architecture constructed from the best cell achieves, among the published works, state-of-the-art accuracy of 82.7\% top-1 and 96.2\% top-5 on ImageNet. Our model is 1.2\% better in top-1 accuracy than the best human-invented architectures while having 9 billion fewer FLOPS -- a reduction of 28\% in computational demand from the previous state-of-the-art model. When evaluated at different levels of computational cost, accuracies of our models exceed those of the state-of-the-art human-designed models. For instance, a smaller network constructed from the best cell also achieves 74\% top-1 accuracy, which is 3.1\% better than equivalently-sized, state-of-the-art models for mobile platforms. On CIFAR-10, an architecture constructed from the best cell achieves 2.4\% error rate, which is also state-of-the-art. Finally, the image features learned from image classification can also be transferred to other computer vision problems. On the task of object detection, the learned features used with the Faster-RCNN framework surpass state-of-the-art by 4.0\% achieving 43.1\% mAP on the COCO dataset.}
}
@inproceedings{ChungetalICML-15,
       author = {Chung, Junyoung and Gulcehre, {\c{C}}aglar and Cho, Kyunghyun and Bengio, Yoshua},
        title = {Gated Feedback Recurrent Neural Networks},
    booktitle = {Proceedings of the 32Nd International Conference on International Conference on Machine Learning - Volume 37},
         year = {2015},
     location = {Lille, France},
        pages = {2067-2075},
}
@article{ChungetalIPMI-09,
       author = {Chung M. K. and Bubenik P. and Kim P.T.},
        title = {Persistence diagrams of cortical surface data},
      journal = {Information Processing Medical Imaging},
       volume = {21},
         year = {2009},
        pages = {386-397},
     abstract = {We present a novel framework for characterizing signals in images using techniques from computational algebraic topology. This technique is general enough for dealing with noisy multivariate data including geometric noise. The main tool is persistent homology which can be encoded in persistence diagrams. These diagrams visually show how the number of connected components of the sublevel sets of the signal changes. The use of local critical values of a function differs from the usual statistical parametric mapping framework, which mainly uses the mean signal in quantifying imaging data. Our proposed method uses all the local critical values in characterizing the signal and by doing so offers a completely new data reduction and analysis framework for quantifying the signal. As an illustration, we apply this method to a 1D simulated signal and 2D cortical thickness data. In case of the latter, extra homological structures are evident in an control group over the autistic group.}
}
@article{ChungetalNATURE-13,
    author = {Kwanghun Chung and Jenelle Wallace and Sung-Yon Kim and Sandhiya Kalyanasundaram and Aaron S. Andalman and
              Thomas J. Davidson and Julie J. Mirzabekov and  Kelly A. Zalocusky and Joanna Mattis, Aleksandra K. Denisin and  
              Sally Pak and Hannah Bernstein and Charu Ramakrishnan and Logan Grosenick and Viviana Gradinaru and Karl Deisseroth},
     title = {Structural and molecular interrogation of intact biological systems},
   journal = {Nature},
    volume = 497, 
      year = 2013,
     pages = {332-337},
  abstract = {Obtaining high-resolution information from a complex system, while maintaining the global perspective needed to understand system function, represents a key challenge in biology. Here we address this challenge with a method (termed CLARITY) for the transformation of intact tissue into a nanoporous hydrogel-hybridized form (crosslinked to a three-dimensional network of hydrophilic polymers) that is fully assembled but optically transparent and macromolecule-permeable. Using mouse brains, we show intact-tissue imaging of long-range projections, local circuit wiring, cellular relationships, subcellular structures, protein complexes, nucleic acids and neurotransmitters. CLARITY also enables intact-tissue in situ hybridization, immunohistochemistry with multiple rounds of staining and de-staining in non-sectioned tissue, and antibody labelling throughout the intact adult mouse brain. Finally, we show that CLARITY enables fine structural analysis of clinical samples, including non-sectioned human tissue from a neuropsychiatric-disease setting, establishing a path for the transmutation of human tissue into a stable, intact and accessible form suitable for probing structural and molecular underpinnings of physiological function and disease.}
}
@techreport{ChungetalTR-87,
       author = {Chung, Jen-Yao and Liu, Jane W.S. and Lin, Kwei-Jay},
        title = {Scheduling Periodic Jobs Using Imprecise Results},
        issue = {UIUCDCS-R-87-1307},
  institution = {University of Illinois at Urbana-Champaign Department of Computer Science},
         year = 1987,
}
@article{ChuongPLoS-18,
       author = {Chuong, Edward B.},
        title = {The placenta goes viral: Retroviruses control gene expression in pregnancy},
      journal = {PLOS Biology},
    publisher = {Public Library of Science},
       volume = {16},
         year = {2018},
        pages = {1-7},
     abstract = {The co-option of endogenous retroviruses (ERVs) is increasingly recognized as a recurrent theme in placental biology, which has far-reaching implications for our understanding of mammalian evolution and reproductive health. Most research in this area has focused on ERV-derived proteins, which have been repeatedly co-opted to promote cell-cell fusion and immune modulation in the placenta. ERVs also harbor regulatory sequences that can potentially control placental gene expression, but there has been limited evidence to support this role. In a recent study, Dunn-Fletcher and colleagues discover a striking example of an ERV-derived enhancer element that has been co-opted to regulate a gene important for human pregnancy. Using genomic and experimental approaches, they firmly establish that a primate-specific ERV functions as a placenta-specific enhancer for corticotropin-releasing hormone (CRH), a hormone linked to the control of birth timing in humans. Their findings implicate an extensive yet understudied role for retroviruses in shaping the evolution of placental gene regulatory networks.},
}
@article{ChurchSCIAM-06,
    author = {George M. Church},
     title = {Genomes for All},
   journal = {Scientific American},
    volume = 294,
      year = 2006,
     pages = {46-54},
}
@article{ChurchandPatilAJCL-82,
       author = {Church, Kenneth and Patil, Ramesh},
        title = {Coping with Syntactic Ambiguity or How to Put a Block in the Box on the Table},
      journal = {American Journal of Computational Linguistics},
       volume = 8,
        issue = {3-4},
         year = 1982,
        pages = {139-149},
}
@book{ChurchandRegis2012,
       author = {George Church and Ed Regis},
        title = {Regenesis: How Synthetic Biology Will Reinvent Nature and Ourselves},
    publisher = {Basic Books},
         year = 2012
}
@misc{ChurchandShendurePATENT-03,
        title = {Nucleic acid memory device},
       author = {Church, George and Shendure, Jay},
         year = 2003,
 howpublished = {US Patent 20030228611},
}
@article{ChurchlandandAbbottNATURE-NEUROSCIENCE-16,
       author = {Churchland, Anne K. and Abbott, L. F.},
        title = {Conceptual and technical advances define a key moment for theoretical neuroscience},
      journal = {Nature Neuroscience},
         year = {2016},
    publisher = {Nature Publishing Group},
       volume = {19},
        issue = {3},
        pages = {348-349},
}
@book{ChurchlandandSejnowski92,
       author = {Churchland, Patricia S. and Sejnowski, Terrence J.},
        title = {The Computational Brain},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1992,
}
@article{ChurchlandetalNATURE-12,
       author = {Churchland, M. M. and Cunningham, J. P. and Kaufman, M. T. and Foster, J. D. and Nuyujukian, P. and Ryu, S. I. and Shenoy, K. V.},
        title = {Neural population dynamics during reaching},
      journal = {Nature},
       volume = {487},
       number = {7405},
         year = {2012},
        pages = {51-56},
     abstract = {Most theories of motor cortex have assumed that neural activity represents movement parameters. This view derives from what is known about primary visual cortex, where neural activity represents patterns of light. Yet it is unclear how well the analogy between motor and visual cortex holds. Single-neuron responses in motor cortex are complex, and there is marked disagreement regarding which movement parameters are represented. A better analogy might be with other motor systems, where a common principle is rhythmic neural activity. Here we find that motor cortex responses during reaching contain a brief but strong oscillatory component, something quite unexpected for a non-periodic behaviour. Oscillation amplitude and phase followed naturally from the preparatory state, suggesting a mechanistic role for preparatory neural activity. These results demonstrate an unexpected yet surprisingly simple structure in the population response. This underlying structure explains many of the confusing features of individual neural responses.}
}
@book{Chvatal80,
        title = {Linear Programming},
       author = {Chvatal, Vasek},
     location = {New York},
    publisher = {W. H. Freeman and Company},
         year = 1980,
}
@inproceedings{CiaHofJoh03,
       author = {M. Ciaramita and T. Hofmann and M. Johnson},
        title = {Hierarchical Semantic Classification: Word Sense Disambiguation with World Knowledge},
    booktitle = {Proceedings of the 18th International Joint Conference on Artificial Intelligence (IJCAI-03)},
         year = 2003,
}
@article{CiaramelliFiP-13,
       author = {Ciaramelli, Elisa and Bernardi, Francesco and Moscovitch, Morris},
        title = {Individualized Theory of Mind (iToM): When Memory Modulates Empathy},
      journal = {Frontiers in Psychology},
       volume = {4},
        pages = {4},
         year = {2013},
     abstract = {Functional neuroimaging studies have noted that brain regions supporting theory of mind (ToM) overlap remarkably with those underlying episodic memory, suggesting a link between the two processes. The present study shows that memory for others' past experiences modulates significantly our appraisal of, and reaction to, what is happening to them currently. Participants read the life story of two characters; one had experienced a long series of love-related failures, the other a long series of work-related failures. In a later faux pas recognition task, participants reported more empathy for the character unlucky in love in love-related faux pas scenarios, and for the character unlucky at work in work-related faux pas scenarios. The memory-based modulation of empathy correlated with the number of details remembered from the characters' life story. These results suggest that individuals use memory for other people's past experiences to simulate how they feel in similar situations they are currently facing. The integration of ToM and memory processes allows adjusting mental state inferences to fit unique social targets, constructing an individualized ToM (iToM).}
}
@incollection{CiresanetalNIPS-12,
        title = {Deep Neural Networks Segment Neuronal Membranes in Electron Microscopy Images},
       author = {Dan Ciresan and Alessandro Giusti and Luca M. Gambardella and Schmidhuber, J{\"{u}}rgen},
    booktitle = {Advances in Neural Information Processing Systems 25},
       editor = {F. Pereira and C.J.C. Burges and L. Bottou and K.Q. Weinberger},
         year = 2012,
        pages = {2843-2851},
    publisher = {Curran Associates, Inc.},
     abstract = {We address a central problem of neuroanatomy, namely, the automatic segmentation of neuronal structures depicted in stacks of electron microscopy (EM) images. This is necessary to efficiently map 3D brain structure and connectivity. To segment biological neuron membranes, we use a special type of deep artificial neural network as a pixel classifier. The label of each pixel (membrane or nonmembrane) is predicted from raw pixel values in a square window centered on it. The input layer maps each window pixel to a neuron. It is followed by a succession of convolutional and max-pooling layers which preserve 2D information and extract features with increasing levels of abstraction. The output layer produces a calibrated probability for each class. The classifier is trained by plain gradient descent on a 512 \% 512 \% 30 stack with known ground truth, and tested on a stack of the same size (ground truth unknown to the authors) by the organizers of the ISBI 2012 EM Segmentation Challenge. Even without problem-specific postprocessing, our approach outperforms competing techniques by a large margin in all three considered metrics, i.e. rand error, warping error and pixel error. For pixel error, our approach is the only one outperforming a second human observer.} 
}
@article{CisekAPP-19,
       author = {Cisek, Paul},
        title = {{{R}esynthesizing behavior through phylogenetic refinement}},
      journal = {Attention, Perception \& Psychophysics},
       volume = {81},
       number = {7},
         year = {2019},
        pages = {2265-2287},
     abstract = {This article proposes that biologically plausible theories of behavior can be constructed by following a method of "phylogenetic refinement," whereby they are progressively elaborated from simple to complex according to phylogenetic data on the sequence of changes that occurred over the course of evolution. It is argued that sufficient data exist to make this approach possible, and that the result can more effectively delineate the true biological categories of neurophysiological mechanisms than do approaches based on definitions of putative functions inherited from psychological traditions. As an example, the approach is used to sketch a theoretical framework of how basic feedback control of interaction with the world was elaborated during vertebrate evolution, to give rise to the functional architecture of the mammalian brain. The results provide a conceptual taxonomy of mechanisms that naturally map to neurophysiological and neuroanatomical data and that offer a context for defining putative functions that, it is argued, are better grounded in biology than are some of the traditional concepts of cognitive science.}
}
@article{Clark01,
       author = {D. Clark},
        title = {Face-to-Face with Peer-to-Peer Networking},
      journal = {Computer},
       volume = {34:1},
         year = 2001,
}
@book{Clark03,
       author = {Andy Clark},
        title = {Natural-Born Cyborgs: Minds, Technologies, and the Future of Human Intelligence},
    publisher = {Oxford University Press},
      address = {Oxford},
         year = 2003,
}
@book{Clark2009first,
        title = {First Language Acquisition},
       author = {Clark, E.V.},
         year = {2009},
    publisher = {Cambridge University Press},
      comment = {Eve V. Clark is Richard W. Lyman Professor in the Humanities and Professor of Linguistics in the Department of Linguistics at Stanford University.}
}
@book{Clark2016embodied,
        title = {Surfing Uncertainty: Prediction, Action, and the Embodied Mind},
       author = {Clark, Andy},
         year = {2016},
    publisher = {Oxford University Press}
}
@article{ClarkBBS-13,
       author = {Clark, Andy},
        title = {Whatever next? {P}redictive brains, situated agents, and the future of cognitive science},
      journal = {Behavioral and Brain Science},
       volume = {36},
       number = {3},
         year = {2013},
        pages = {181-204},
     abstract = {Brains, it has recently been argued, are essentially prediction machines. They are bundles of cells that support perception and action by constantly attempting to match incoming sensory inputs with top-down expectations or predictions. This is achieved using a hierarchical generative model that aims to minimize prediction error within a bidirectional cascade of cortical processing. Such accounts offer a unifying model of perception and action, illuminate the functional role of attention, and may neatly capture the special contribution of cortical processing to adaptive success. This target article critically examines this "hierarchical prediction machine" approach, concluding that it offers the best clue yet to the shape of a unified science of mind and action. Sections 1 and 2 lay out the key elements and implications of the approach. Section 3 explores a variety of pitfalls and challenges, spanning the evidential, the methodological, and the more properly conceptual. The paper ends (sections 4 and 5) by asking how such approaches might impact our more general vision of mind, experience, and agency.}
}
@article{ClarkCLS-02,
       author = {Clark, Eve},
        title = {Grounding and attention in language acquisition},
      journal = {Papers from the 37th meeting of the Chicago Linguistic Society},
       volume = {1},
         year = {2002},
        pages = {95-116},
}

@inbook{ClarkCOMMON-GROUND-15,
       author = {Clark, Eve V.},
    publisher = {John Wiley \& Sons, Ltd},
        title = {Common Ground},
    booktitle = {The Handbook of Language Emergence},
         year = {2015},
        pages = {328-353},
     abstract = {Where does common ground come from? When do infants and young children start to make use of common ground as they communicate? How do they assess what their interlocutor knows? And how do they mark information as given (in common ground) vs. new (to be added to common ground)? These are some of the questions addressed in the present chapter. Common ground, I suggest, originates in infants' everyday settings and routines where they can rely on caregivers' knowledge. Common ground depends from the first on joint attention, physical co-presence, and conversational co-presence, so it is intrinsic to communication from the start. But children have to learn how to indicate linguistically that information is given or new, how to assess what their addressees do and don't know on each occasion, and how to design utterances based on such assessments.}
}
@article{ClarkDISCOURSE-PROCESSES-20,
       author = {Eve V. Clark},
        title = {Conversational Repair and the Acquisition of Language},
      journal = {Discourse Processes},
       volume = {57},
       number = {5-6},
         year = {2020},
        pages = {441-459},
    publisher = {Routledge},
     abstract = {In this article, I examine how repairs in adult-child conversations guide children's acquisition of language. Children make unprompted self-repairs to their utterances. They also respond to prompts for repair, whether open (Hm?, What?) or restricted (You hid what?), and to restricted offers (Child: I falled, Adult: You fell?). Children respond to clarification requests with self-repairs in the next turn, and make use of the feedback offered. The contrast between their utterance and the adult utterance identifies the locus of the error (negative feedback), while the adult's offer presents a conventional version of the child's utterance (positive feedback). I describe the use of restricted offers in conversations with children acquiring English and French, then present two case studies of how these inform children about homophonous French verb forms and early opaque Hebrew verb uses. These findings demonstrate the fundamental role of repair in the acquisition of a first language.}
}
@article{ClarkFiP-13,
       author = {Clark, Andy},
        title = {{T}he many faces of precision ({R}eplies to commentaries on '{W}hatever next? {N}eural prediction, situated agents, and the future of cognitive science')},
      journal = {Frontiers in Psychology},
         year = {2013},
       volume = {4},
        pages = {270},
     abstract = {An appreciation of the many roles of "precision-weighting" (upping the gain on select populations of prediction error units) opens the door to better accounts of planning and "offline simulation," makes suggestive contact with large bodies of work on embodied and situated cognition, and offers new perspectives on the "active brain". Combined with the complex affordances of language and culture, and operating against the essential backdrop of a variety of more biologically basic ploys and stratagems, the result is a maximally context-sensitive, restless, constantly self-reconfiguring architecture.}
}
@article{ClarkNOUS-17,
    author = {Clark, Andy},
     title = {Busting out: predictive brains, embodied minds, and the puzzle of the evidentiary veil},
   journal = {Nous},
    volume = {51},
     issue = {4},
      year = {2017},
     pages = {727-753},
     abstract = {Biological brains are increasingly cast as'prediction machines': evolved organs whose core operating principle is to learn about the world by trying to predict their own patterns of sensory stimulation. This, some argue, should lead us to  embrace a brain-bound'neurocentric' vision of the mind. The mind, such views suggest, consists entirely in the skull-bound activity of the predictive brain. In this paper I reject the inference from predictive brains to skull-bound minds. Predictive brains, I hope to show, can be apt participants in larger cognitive circuits. The path is thus cleared for a new synthesis in which predictive brains act as entry-points for'extended minds', and embodiment and action contribute constitutively to knowing contact with the world.}
}
@article{ClarkTiCS-04,
       author = {Eve V. Clark},
        title = {How language acquisition builds on cognitive development},
      journal = {Trends in Cognitive Sciences},
       volume = {8},
       number = {10},
         year = {2004},
        pages = {472-478},
     abstract = {When children acquire a first language, they build on what they know - conceptual information that discriminates and helps create categories for the objects, relations and events they experience. This provides the starting point for language from the age of 12 months on. So children first set up conceptual representations, then add linguistic representations for talking about experience. Do they then discard earlier conceptual representations in favour of linguistic ones, or do they retain them? Recent research on the coping strategies that young children (and adults) rely on when they are unable to draw on language suggest that they retain both types of representations for use as needed.}
}
@book{ClarkUSING-LANGUAGE-96,
       author = {Clark, Herbert H.},
        title = {Using Language},
    publisher = {Cambridge University Press},
      address = {Cambridge},
         year = {1996},
}     
@article{ClarkandChalmersANALYSIS-98,
       author = {Clark, Andy and Chalmers, David},
        title = {{The Extended Mind}},
      journal = {Analysis},
       volume = {58},
       number = {1},
         year = {1998},
        pages = {7-19},
}
@book{ClarkandRussell00,
       author = {David P. Clark and Lonnie D. Russell},
        title = {Molecular Biology Made Simple and Fun},
    publisher = {Cache River Press},
      address = {Vienna, Illinois},
         year = 2000,
}
@incollection{ClarkeMODEL-CHECKING-08,
       author = {Clarke, Edmund M.},
        title = {The Birth of Model Checking},
    booktitle = {25 Years of Model Checking},
       editor = {Grumberg, Orna and Veith, Helmut},
         year = {2008},
        pages = {1-26},
    publisher = {Springer-Verlag},
      address = {Berlin, Heidelberg},
}
@incollection{Clarkeetal01,
       author = {I. Clarke and O. Sandberg and B. Wiley and T.W. Hong},
        title = {Freenet: A Distributed Anonymous Information Storage and Retrieval System},
    booktitle = {Designing Privacy Enhancing Technologies: International Workshop on Design Issues in Anonymity and Unobservability},
 series       = {LNCS},
        issue = 2009,
       editor = {H. Federrath},
    publisher = {Springer},
      address = {New York},
         year = 2001,
}
@article{ClarketalJoN-13,
       author = {Clark, Luke and Averbeck, Bruno and Payer, Doris and Sescousse, Guillaume and Winstanley, Catharine A. and Xue, Gui},
        title = {Pathological Choice: The Neuroscience of Gambling and Gambling Addiction},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {33},
        issue = {45},
         year = {2013},
        pages = {17617-17623},
     abstract = {Gambling is pertinent to neuroscience research for at least two reasons. First, gambling is a naturalistic and pervasive example of risky decision making, and thus gambling games can provide a paradigm for the investigation of human choice behavior and 'irrationality'. Second, excessive gambling involvement (i.e., pathological gambling) is currently conceptualized as a behavioral addiction, and research on this condition may provide insights into addictive mechanisms in the absence of exogenous drug effects. This article is a summary of topics covered in a Society for Neuroscience minisymposium, focusing on recent advances in understanding the neural basis of gambling behavior, including translational findings in rodents and nonhuman primates, which have begun to delineate neural circuitry and neurochemistry involved.},
}
@incollection{ClaytonandDickinsonEAB-10,
       author = {Mental Time Travel: Can Animals Recall the Past and Plan for the Future?},
        title = {N. S. Clayton and A. Dickinson},
       editor = {Breed, Michael D.  and Moore, Janice},
    booktitle = {Encyclopedia of Animal Behavior},
    publisher = {Academic Press},
         year = {2010},
        pages = {438-442},
     abstract = {According to the mental time travel hypothesis, only humans can mentally dissociate themselves from the present, traveling backward in time to recollect specific past events about what happened where and when (episodic memory) and traveling forward in time to anticipate future needs (future planning). A series of studies of the mnemonic capabilities of food-caching western scrub-jays question this assumption. In terms of the retrospective component of episodic memory, these birds remember the 'what, where, and when' of specific past caching episodes; they keep track of how long ago they cached different types of perishable foods that decay at different rates, and also remember whether another individual was present at the time of caching, and if so, which bird was watching when. Recent work demonstrates that the jays also make provision for a future need, caching more food in places in which they will not be given breakfast the next morning than in places where they will be receive breakfast the next morning even though there is plenty of food available to them at the time when they cache the food. Taken together these results challenge the mental time travel hypothesis by showing that some elements of both retrospective and prospective mental time travel appear not to be uniquely human.}
}
@article{Clemente-PerezetalCELL-REPORTS-17,
       author = {Clemente-Perez, A. and Makinson, S.R. and Higashikubo, B. and Brovarney, S. and Cho, F.S. and Urry, A. and Holden, S.S. and Wimer, M. and David, C. and Fenno, L.E. and Acsady, L. and Deisseroth, K. and Paz, J. T.},
        title = {Distinct Thalamic Reticular Cell Types Differentially Modulate Normal and Pathological Cortical Rhythms},
      journal = {Cell Reports},
       volume = {19},
       number = {10},
         year = {2017},
        pages = {2130-2142},
     abstract = {Integrative brain functions depend on widely distributed, rhythmically coordinated computations. Through its long-ranging connections with cortex and most senses, the thalamus orchestrates the flow of cognitive and sensory information. Essential in this process, the nucleus reticularis thalami (nRT) gates different information streams through its extensive inhibition onto other thalamic nuclei, however, we lack an understanding of how different inhibitory neuron subpopulations in nRT function as gatekeepers. We dissociated the connectivity, physiology, and circuit functions of neurons within rodent nRT, based on parvalbumin (PV) and somatostatin (SOM) expression, and validated the existence of such populations in human nRT. We found that PV, but not SOM, cells are rhythmogenic, and that PV and SOM neurons are connected to and modulate distinct thalamocortical circuits. Notably, PV, but not SOM, neurons modulate somatosensory behavior and disrupt seizures. These results provide a conceptual framework for how nRT may gate incoming information to modulate brain-wide rhythms.}
}
@article{Clifford83,
       author = {Clifford, James and Warren, David S.},
        title = {Formal Semantics for Time in Databases},
      journal = {{ACM} Transactions on Database Systems},
       volume = 8,
        issue = 2,
 month        = {June},
         year = 1983,
        pages = {214-254},
}
@misc{Clip2,
      author = {Clip2},
howpublished = {\url{http://www.clip2.com/}},
        year = 2001,
}
@book{Clocksin87,
       author = {Clocksin, W. F. and Mellish, C. S.},
        title = {Programming in Prolog, Third Edition},
    publisher = {Springer-Verlag},
      address = {New York},
         year = 1987,
}
@inproceedings{CoatesetalIROS-09,
     author = {Adam Coates and Paul Baumstarck and Quoc Le and Andrew Y. Ng}, 
      title = {Scalable Learning for Object Detection with {GPU} Hardware}, 
  booktitle = {Proceedings of the International Conference on Intelligent Robots and Systems},
       year = 2009,
}
@article{CogganetalSCIENCE-05,
    author = {Coggan, Jay S. and Bartol, Thomas M. and Esquenazi, Eduardo and Stiles, Joel R. and Lamont, Stephan and Martone, Maryann E. and Berg, Darwin K. and Ellisman, Mark H. and Sejnowski, Terrence J.}, 
     title = {Evidence for ectopic neurotransmission at a neuronal synapse},
   journal = {Science}, 
    volume = 309, 
     issue = 5733, 
     pages = {446-451}, 
      year = 2005, 
  abstract = {Neurotransmitter release is well known to occur at specialized synaptic regions that include presynaptic active zones and postsynaptic densities. At cholinergic synapses in the chick ciliary ganglion, however, membrane formations and physiological measurements suggest that release distant from postsynaptic densities can activate the predominantly extrasynaptic $\alpha$7 nicotinic receptor subtype. We explored such ectopic neurotransmission with a novel model synapse that combines Monte Carlo simulations with high-resolution serial electron microscopic tomography. Simulated synaptic activity is consistent with experimental recordings of miniature excitatory postsynaptic currents only when ectopic transmission is included in the model, broadening the possibilities for mechanisms of neuronal communication.}, 
}
@book{Cohen95,
       author = {Cohen, Paul R.},
        title = {Empirical Methods for Artificial Intelligence},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1995,
}
@article{CohenSCIENCE-07,
    author = {Cohen, Jon}, 
     title = {Relative Differences: {T}he Myth of 1\%}, 
   journal = {Science}, 
    volume = 316, 
     issue = 5833, 
     pages = 1836, 
      year = 2007, 
}
@incollection{CohenUAI-88,
       author = {Cohen, Marvin S.},
        title = {A Framework for Non-Monotonic Reasoning About Probabilistic Assumptions},
     crossref = UAI-88,
}
@book{CohenandLefebvreHANDBOOK0-05,
        title = {Handbook of Categorization in Cognitive Science (First Edition)},
       author = {Cohen, H. and Lefebvre, C.},
         year = {2005},
    publisher = {Elsevier}
}
@incollection{CohenandLevesquePW-86,
       author = {Cohen, Philip and Levesque, Hector},
        title = {Persistence, Intention, and Commitment},
     crossref = {GeorgeffandLanskyPW-86},
        pages = {297-340},
}
@article{Cohenetal99,
   author = {Cohen, W.W. and Schapire, R.E. and Singer, Y.},
    title = {Learning to Order Things},
  journal = {Journal of Artificial Intelligence Research},
   volume = 10,
    pages = {243-270},
     year = 1999,
}
@techreport{CohenetalBBN-81,
       author = {Cohen, P. and Perrault, C. R. and Allen, J.},
        title = {{B}eyond {Q}uestion-{A}nswering},
  institution = {Bolt Beranek and Newman},
       number = 4644,
         year = 1981
}
@inproceedings{CohnHofmann01,
       author = {David Cohn and Thomas Hofmann},
        title = {The Missing Link: A Probabilistic Model of Document Content and Hypertext Connectivity},
    booktitle = {Advances in Neural Information Processing Systems (NIPS 13)},
         year = 2001,
    publisher = {MIT Press}
}
@inproceedings{CohnHofmannNIPS-02,
       author = {Cohn, D. and Hofmann, T.},
        title = {The Missing Link: A Probabilistic Model of Document Content and Hypertext Connectivity},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 14,
         year = 2002,
}
@article{CohnetalML-94,
       author = {Cohn, David and Atlas, Les and Ladner, Richard},
        title = {Improving Generalization with Active Learning},
      journal = {Machine Learning},
       volume = 15,
        issue = 2,
        pages = {201-221},
         year = 1994,
}
@incollection{CohnetalNIPS-90,
       author = {Cohn, David and Atlas, Les and Ladner, Richard and El-Sharkawi,
                M. A. and Marks, R. J. and Aggoune, M.E. and Park, D. C.},
        title = {Training Connectionist Networks with Queries and Selective Sampling},
    booktitle = {Neural Information Processing Systems},
    publisher = {Morgan Kaufmann},
         year = 1990,
}
@incollection{Cole2020chineseroom,
       author = {Cole, David},
        title = {The Chinese Room Argument},
    booktitle = {The Stanford Encyclopedia of Philosophy},
    publisher = {Metaphysics Research Lab, Stanford University},
       editor = {Edward N. Zalta},
         year = {2020},
}
@article{CollinsandFrankCOGNITION-16,
       author = {Anne Gabrielle Eva Collins and Michael Joshua Frank},
        title = {Neural signature of hierarchically structured expectations predicts clustering and transfer of rule sets in reinforcement learning},
      journal = {Cognition},
       volume = {152},
         year = {2016},
        pages = {160-169},
      comment = {http://ski.clps.brown.edu/papers/CollinsFrank_Cognition.pdf},
     abstract = {Often the world is structured such that distinct sensory contexts signify the same abstract rule set. Learning from feedback thus informs us not only about the value of stimulus-action associations but also about which rule set applies. Hierarchical clustering models suggest that learners discover structure in the environment, clustering distinct sensory events into a single latent rule set. Such structure enables a learner to transfer any newly acquired information to other contexts linked to the same rule set, and facilitates re-use of learned knowledge in novel contexts. Here, we show that humans exhibit this transfer, generalization and clustering during learning. Trial-by-trial model-based analysis of EEG signals revealed that subjects' reward expectations incorporated this hierarchical structure; these structured neural signals were predictive of behavioral transfer and clustering. These results further our understanding of how humans learn and generalize flexibly by building abstract, behaviorally relevant representations of the complex, high-dimensional sensory environment.},
}
@article{CollinsandFrankPSYCHOLOGICAL-REVIEW-13,
       author = {Collins, Anne G. E. and Frank, Michael J.},
        title = {Cognitive control over learning: creating, clustering, and generalizing task-set structure},
      journal = {Psychological Review},
       volume = {120},
        issue = {1},
         year = {2013},
        pages = {190-229},
      comment = {http://ski.clps.brown.edu/papers/CollinsFrank_psyrev.pdf},
     abstract = {Learning and executive functions such as task-switching share common neural substrates, notably prefrontal cortex and basal ganglia. Understanding how they interact requires studying how cognitive control facilitates learning but also how learning provides the (potentially hidden) structure, such as abstract rules or task-sets, needed for cognitive control. We investigate this question from 3 complementary angles. First, we develop a new context-task-set (C-TS) model, inspired by nonparametric Bayesian methods, specifying how the learner might infer hidden structure (hierarchical rules) and decide to reuse or create new structure in novel situations. Second, we develop a neurobiologically explicit network model to assess mechanisms of such structured learning in hierarchical frontal cortex and basal ganglia circuits. We systematically explore the link between these modeling levels across task demands. We find that the network provides an approximate implementation of high-level C-TS computations, with specific neural mechanisms modulating distinct C-TS parameters. Third, this synergism yields predictions about the nature of human optimal and suboptimal choices and response times during learning and task-switching. In particular, the models suggest that participants spontaneously build task-set structure into a learning problem when not cued to do so, which predicts positive and negative transfer in subsequent generalization tests. We provide experimental evidence for these predictions and show that C-TS provides a good quantitative fit to human sequences of choices. These findings implicate a strong tendency to interactively engage cognitive control and learning, resulting in structured abstract representations that afford generalization opportunities and, thus, potentially long-term rather than short-term optimality.},
}
@article{CollinsandFrankPSYCHOLOGICAL-REVIEW-14,
       author = {Collins, Anne G. E. and Frank, M. J.},
        title = {Opponent actor learning ({OpAL}): modeling interactive effects of striatal dopamine on reinforcement learning and choice incentive},
      journal = {Psychological Review},
       volume = {121},
       number = {3},
         year = {2014},
        pages = {337-366},
     abstract = {The striatal dopaminergic system has been implicated in reinforcement learning (RL), motor performance, and incentive motivation. Various computational models have been proposed to account for each of these effects individually, but a formal analysis of their interactions is lacking. Here we present a novel algorithmic model expanding the classical actor-critic architecture to include fundamental interactive properties of neural circuit models, incorporating both incentive and learning effects into a single theoretical framework. The standard actor is replaced by a dual opponent actor system representing distinct striatal populations, which come to differentially specialize in discriminating positive and negative action values. Dopamine modulates the degree to which each actor component contributes to both learning and choice discriminations. In contrast to standard frameworks, this model simultaneously captures documented effects of dopamine on both learning and choice incentive-and their interactions-across a variety of studies, including probabilistic RL, effort-based choice, and motor skill learning.}
}
@article{CollinsandFrankPSYCHOLOGY-REVIEW-14,
       author = {Collins, A. G. and Frank, M. J.},
        title = {{O}pponent actor learning ({O}p{A}{L}): modeling interactive effects of striatal dopamine on reinforcement learning and choice incentive},
      journal = {Psychology Review},
       volume = {121},
       number = {3},
         year = {2014},
        pages = {337-366},
     abstract = {The striatal dopaminergic system has been implicated in reinforcement learning (RL), motor performance, and incentive motivation. Various computational models have been proposed to account for each of these effects individually, but a formal analysis of their interactions is lacking. Here we present a novel algorithmic model expanding the classical actor-critic architecture to include fundamental interactive properties of neural circuit models, incorporating both incentive and learning effects into a single theoretical framework. The standard actor is replaced by a dual opponent actor system representing distinct striatal populations, which come to differentially specialize in discriminating positive and negative action values. Dopamine modulates the degree to which each actor component contributes to both learning and choice discriminations. In contrast to standard frameworks, this model simultaneously captures documented effects of dopamine on both learning and choice incentive-and their interactions-across a variety of studies, including probabilistic RL, effort-based choice, and motor skill learning.}
}
@article{CollinsandKoechlinPLoS-12,
       author = {Collins, Anne and Koechlin, Etienne},
      journal = {PLOS Biology},
    publisher = {Public Library of Science},
        title = {Reasoning, Learning, and Creativity: Frontal Lobe Function and Human Decision-Making},
       volume = {10},
       number = {3},
         year = {2012},
        pages = {1-16},
     abstract = {Computational modeling and behavioral experimentation suggest that human frontal lobe function is capable of monitoring three or four concurrent behavioral strategies in order to select the most suitable one during decision-making.},
}
@inproceedings{CollinsandPryorAAAI-92,
       author = {Collins, Gregg and Pryor, Louise},
        title = {Achieving the Functionality of Filter Conditions in a Partial Order Planner},
    booktitle = {Proceedings AAAI-92},
     location = {San Jose, California},
 organization = {AAAI},
        pages = {375-380},
         year = 1992,
}
@article{CollinsetalML-02,
     author = {Collins, Michael and Schapire, Robert E. and Singer, Yoram},
      title = {Logistic Regression, AdaBoost and Bregman Distances},
    journal = {Machine Learning},
     volume = 48,
      issue = {1-3},
       year = 2002,
      pages = {253-285},
  publisher = {Kluwer Academic Publishers},
    address = {Hingham, MA, USA},
}
@article{CollmanetalJN-15,
       author = {Collman, Forrest and Buchanan, JoAnn and Phend, Kristen D. and Micheva, Kristina D. and Weinberg, Richard J. and Smith, Stephen J},
        title = {Mapping Synapses by Conjugate Light-Electron Array Tomography},
      journal = {The Journal of Neuroscience},
       volume = 35,
        issue = 14,
        pages = {5792-5807},
         year = 2015,
     abstract = {Synapses of the mammalian CNS are diverse in size, structure, molecular composition, and function. Synapses in their myriad variations are fundamental to neural circuit development, homeostasis, plasticity, and memory storage. Unfortunately, quantitative analysis and mapping of the brain's heterogeneous synapse populations has been limited by the lack of adequate single-synapse measurement methods. Electron microscopy (EM) is the definitive means to recognize and measure individual synaptic contacts, but EM has only limited abilities to measure the molecular composition of synapses. This report describes conjugate array tomography (AT), a volumetric imaging method that integrates immunofluorescence and EM imaging modalities in voxel-conjugate fashion. We illustrate the use of conjugate AT to advance the proteometric measurement of EM-validated single-synapse analysis in a study of mouse cortex.},
}
@article{CollobertetalJMLR-11,
       author = {Collobert, Ronan and Weston, Jason and Bottou, L{\'e}on and Karlen, Michael and Kavukcuoglu, Koray and Kuksa, Pavel},
        title = {Natural Language Processing (Almost) from Scratch},
      journal = {Journal Machine Learning Research},
       volume = 12,
         year = 2011,
        pages = {2493-2537},
}
@article{ColonnierandOKusky81,
       author = {M. Colonnier and J. O'Kusky},
        title = {Number of neurons and synapses in the visual cortex of different species},
      journal = {Revue Canadienne de Biologie},
         year = 1981,
       volume = 40,
        issue = 1,
        pages = {91-99},
          url = {http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?CMD=Display&DB=pubmed}
}
@article{ColungaandSmithPR-05,
        title = {From the lexicon to expectations about kinds: a role for associative learning},
       author = {Eliana Colunga and Linda B. Smith},
      journal = {Psychological Review},
       volume = {112},
       number = {2},
         year = {2005},
        pages = {347-382},
     abstract = {In the novel noun generalization task, 2 1/2-year-old children display generalized expectations about how solid and nonsolid things are named, extending names for never-before-encountered solids by shape and for never-before-encountered nonsolids by material. This distinction between solids and nonsolids has been interpreted in terms of an ontological distinction between objects and substances. Nine simulations and behavioral experiments tested the hypothesis that these expectations arise from the correlations characterizing early learned noun categories. In the simulation studies, connectionist networks were trained on noun vocabularies modeled after those of children. These networks formed generalized expectations about solids and nonsolids that match children's performances in the novel noun generalization task in the very different languages of English and Japanese. The simulations also generate new predictions supported by new experiments with children. Implications are discussed in terms of children's development of distinctions between kinds of categories and in terms of the nature of this knowledge.}
}
@article{ComaniciuandMeerPAMI-02,
       author = {Dorin Comaniciu and Peter Meer},
        title = {Mean Shift: A Robust Approach Toward Feature Space Analysis},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
       volume = 24,
        issue = 5,
         year = 2002,
        pages = {603-619},
}
@article{ComaniciuetalPAMI-03,
       author = {Dorin Comaniciu and Visvanathan Ramesh and Peter Meer},
        title = {Kernel-Based Object Tracking},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
       volume = 25,
        issue = 5,
         year = 2003,
        pages = {564-575},
}
@incollection{Compton88,
       author = {K. J. Compton},
        title = {0-1 Laws in Logic and Combinatorics},
    booktitle = {{NATO} Adv. Study Inst. on Algorithms and Order},
       editor = {I. Rival},
    publisher = {D. Reidel Publishing Company},
        pages = {353-383},
         year = 1988,
}
@article{CondonNATURE-REVIEWS-GENETICS-06,
       author = {Condon, Anne},
        title = {Designed {DNA} molecules: principles and applications of molecular nanotechnology},
      journal = {Nature Reviews Genetics},
       volume = 7,
        issue = 7,
         year = 2006,
        pages = {565-575},
}
@article {CongetalBIORXIV-17,
       author = {Cong, Lin and Wang, Zeguan and Chai, Yuming and Hang, Wei and Shang, Chunfeng and Yang, Wenbin and Bai, Lu and Du, Jiulin and Wang, Kai and Wen, Quan},
        title = {Rapid Whole Brain Imaging Of Neural Activities In Freely Behaving Larval Zebrafish},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Labs Journals},
         year = {2017},
     abstract = {The internal brain dynamics that links sensation and action is arguably better studied during animal{\textquoteright}s natural behaviors. Here we report novel volume imaging and 3D tracking techniques that perform whole brain imaging of neural activities in freely swimming larval zebrafish. We demonstrated the capability of our system through functional imaging of neural activities during visually evoked behavior and prey capture behavior in larval zebrafish.},
}
@article{ConroyetalAR-09,
       author = {Conroy, Joseph and Gremillion, Gregory and Ranganathan, Badri and Humbert, J. Sean},
        title = {Implementation of wide-field integration of optic flow for autonomous quadrotor navigation},
      journal = {Autonomous Robots},
         year = {2009},
       volume = {27},
       number = {3},
        pages = {189},
     abstract = {Insects are capable of robust visual navigation in complex environments using efficient information extraction and processing approaches. This paper presents an implementation of insect inspired visual navigation that uses spatial decompositions of the instantaneous optic flow to extract local proximity information. The approach is demonstrated in a corridor environment on an autonomous quadrotor micro-air-vehicle (MAV) where all the sensing and processing, including altitude, attitude, and outer loop control is performed on-board. The resulting methodology has the advantages of computation speed and simplicity, hence are consistent with the stringent size, weight, and power requirements of MAVs.},
}
@article{ConstantinidisandQietalFiN-18,
       author = {Constantinidis, C. and Qi, X. L.},
        title = {Representation of Spatial and Feature Information in the Monkey Dorsal and Ventral Prefrontal Cortex},
      journal = {Frontiers Integrative Neuroscience},
         year = {2018},
       volume = {12},
        pages = {31},
     abstract = {The primate prefrontal cortex (PFC) is critical for executive functions including working memory, task switching and response selection. The functional organization of this area has been a matter of debate over a period of decades. Early models proposed segregation of spatial and object information represented in working memory in the dorsal and ventral PFC, respectively. Other models emphasized the integrative ability of the entire PFC depending on task demands, not necessarily tied to working memory. An anterior-posterior hierarchy of specialization has also been speculated, in which progressively more abstract information is represented more anteriorly. Here we revisit this debate, updating these arguments in light of recent evidence in non-human primate neurophysiology studies. We show that spatial selectivity is predominantly represented in the posterior aspect of the dorsal PFC, regardless of training history and task performed. Objects of different features excite both dorsal and ventral prefrontal neurons, however neurons highly specialized for feature information are located predominantly in the posterior aspect of the ventral PFC. In accordance with neuronal selectivity, spatial working memory is primarily impaired by inactivation or lesion of the dorsal PFC and object working memory by ventral inactivation or lesion. Neuronal responses are plastic depending on task training but training too has dissociable effects on ventral and dorsal PFC, with the latter appearing to be more plastic. Despite the absence of an overall topography, evidence exists for the orderly localization of stimulus information at a sub-millimeter scale, within the dimensions of a cortical column. Unresolved questions remain, regarding the existence or not of a functional map at the areal and columnar scale, and the link between behavior and neuronal activity for different prefrontal subdivisions.}
}
@article{ConstantinidisetalNATURE-NEUROSCIENCE-01,
       author = {Constantinidis, Christos and Franowicz, Matthew N. and Goldman-Rakic, Patricia S.},
        title = {The sensory nature of mnemonic representation in the primate prefrontal cortex},
      journal = {Nature Neuroscience},
         year = {2001},
       volume = {4},
        issue = {3},
        pages = {311-316},
     abstract = {A long-standing issue concerning the function of the primate dorsolateral prefrontal cortex is whether the activity of prefrontal neurons reflects the perceived sensory attributes of a remembered stimulus, or the decision to execute a motor response. To distinguish between these possibilities, we recorded neuronal activity from monkeys trained to make a saccade toward the brighter of two memoranda, under conditions of varied luminance. Our results indicated that during the delay period when sensory information was no longer available, neuronal discharge was modulated by the luminance of the stimulus appearing in the receptive field, and was directly correlated with psychophysical performance in the task. The findings suggest that although prefrontal cortex codes for a diversity of representations, including the decision for an impending response, a population of neurons maintains the dimensional attributes of remembered stimuli throughout the delay period, which allows for flexibility in the outcome of a mnemonic process.},
}
@incollection{ConwayHBN-08,
       author = {Martin Conway},
       editor = {Dere, E. and Easton, A. and Nadel, L. and Huston, J.P.},
        title = {Chapter 1.2 Exploring Episodic Memory},
    booktitle = {Handbook of Behavioral Neuroscience},
    publisher = {Elsevier Science},
         year = {2008},
        pages = {34-45}
}
@article{ConwayVISNEURO-13,
       author = {Conway, B. R.},
        title = {Color signals through dorsal and ventral visual pathways},
      journal = {Visual Neuroscience},
         year = 2013,
        pages = {1-13},
     abstract = {Explanations for color phenomena are often sought in the retina, lateral geniculate nucleus, and V1, yet it is becoming increasingly clear that a complete account will take us further along the visual-processing pathway. Working out which areas are involved is not trivial. Responses to S-cone activation are often assumed to indicate that an area or neuron is involved in color perception. However, work tracing S-cone signals into extrastriate cortex has challenged this assumption: S-cone responses have been found in brain regions, such as the middle temporal (MT) motion area, not thought to play a major role in color perception. Here, we review the processing of S-cone signals across cortex and present original data on S-cone responses measured with fMRI in alert macaque, focusing on one area in which S-cone signals seem likely to contribute to color (V4/posterior inferior temporal cortex) and on one area in which S signals are unlikely to play a role in color (MT). We advance a hypothesis that the S-cone signals in color-computing areas are required to achieve a balanced neural representation of perceptual color space, whereas those in noncolor-areas provide a cue to illumination (not luminance) and confer sensitivity to the chromatic contrast generated by natural daylight (shadows, illuminated by ambient sky, surrounded by direct sunlight). This sensitivity would facilitate the extraction of shape-from-shadow signals to beneﬁ t global scene analysis and motion perception.},
}
@inproceedings{CookFOCS-71,
       author = {Cook, S. A.},
        title = {The Complexity of Theorem-Proving Procedures},
    booktitle = {Proceedings of the Third {IEEE} Symposium on the Foundations of
                Computer Science},
        pages = {151-158},
         year = 1971,
}
@article{CoolsIRCS-16,
       author = {Cools, Roshan},
        title = {The costs and benefits of brain dopamine for cognitive control},
      journal = {Wiley Interdisciplinary Reviews Cognitive Science},
       volume = {7},
       number = {5},
         year = {2016},
        pages = {317-329},
     abstract = {Cognitive control helps us attain our goals by resisting distraction and temptations. Dopaminergic drugs are well known to enhance cognitive control. However, there is great variability in the effects of dopaminergic drugs across different contexts, with beneficial effects on some tasks but detrimental effects on other tasks. The mechanisms underlying this variability across cognitive task demands remain unclear. I aim to elucidate this across-task variability in dopaminergic drug efficacy by going beyond classic models that emphasize the importance of dopamine in the prefrontal cortex for cognitive control and working memory. To this end, I build on recent advances in cognitive neuroscience that highlight a role for dopamine in cost-benefit decision making. Specifically, I reconceptualize cognitive control as involving not just prefrontal dopamine but also modulation of cost-benefit decision making by striatal dopamine. This approach will help us understand why we sometimes fail to (choose to) exert cognitive control while also identifying mechanistic factors that predict dopaminergic drug effects on cognitive control.}
}
@article{CoolsetalJOURNAL-NEUROSCIENCE-09,
       author = {Cools, R. and Frank, M. J. and Gibbs, S. E. and Miyakawa, A. and Jagust, W. and D'Esposito, M.},
        title = {Striatal dopamine predicts outcome-specific reversal learning and its sensitivity to dopaminergic drug administration},
      journal = {Journal Neuroscience},
         year = {2009},
       volume = {29},
       number = {5},
        pages = {1538-1543},
     abstract = {Individual variability in reward-based learning has been ascribed to quantitative variation in baseline levels of striatal dopamine. However, direct evidence for this pervasive hypothesis has hitherto been unavailable. We demonstrate that individual differences in reward-based reversal learning reflect variation in baseline striatal dopamine synthesis capacity, as measured with neurochemical positron emission tomography. Subjects with high baseline dopamine synthesis in the striatum showed relatively better reversal learning from unexpected rewards than from unexpected punishments, whereas subjects with low baseline dopamine synthesis in the striatum showed the reverse pattern. In addition, baseline dopamine synthesis predicted the direction of dopaminergic drug effects. The D(2) receptor agonist bromocriptine improved reward-based relative to punishment-based reversal learning in subjects with low baseline dopamine synthesis capacity, while impairing it in subjects with high baseline dopamine synthesis capacity in the striatum. Finally, this pattern of drug effects was outcome-specific, and driven primarily by drug effects on punishment-, but not reward-based reversal learning. These data demonstrate that the effects of D(2) receptor stimulation on reversal learning in humans depend on task demands and baseline striatal dopamine synthesis capacity.}
}
@article{CooperAIJ-90,
       author = {Cooper, Gregory F.},
        title = {The computational complexity of probabilistic inference using
              {B}ayesian belief networks},
      journal = AIJ,
       volume = 42,
        pages = {393-405},
         year = 1990,
}
@phdthesis{CooperPhD,
       author = {Gregory F. Cooper},
        title = {{NESTOR}: A computer-based medical diagnostic aid that integrates
                 causal and probabilistic knowledge},
 department   = {Medical Information Sciences},
 school       = {Stanford University},
 month        = {November},
         year = 1984,
}
@techreport{CooperTR-87,
       author = {Cooper, Gregory F.},
        title = {Probabilistic Inference Using Belief Networks is \mbox{NP-Hard}},
        issue = {KSL-87-27},
  institution = {Stanford Knowledge Systems Laboratory},
         year = 1987,
}
@inproceedings{CooperandFrieze01,
       author = {C. Cooper and A. Frieze},
        title = {On a general model of {W}eb graphs},
    booktitle = {Submitted},
        pages = {},
         year = 2001,
}

@article{CooperandHerskovitsML-92,
       author = {Cooper, Greg F. and Herskovits, Edward},
        title = {A {B}ayesian Method for the Induction of Probabilistic Networks
              from Data},
      journal = {Machine Learning},
       volume = 9,
         year = 1992,
        pages = {309-347},
}
@inproceedings{CooperetalIUW-90,
      author = {Cooper, D. B. and Dean, T. L. and Wolovich, W. A.},
       title = {Image Understanding Research at Brown},
   booktitle = {Proceedings of the DARPA Image Understanding Workshop},
    location = {Pittsburgh, Pennsylvania},
organization = {DARPA},
       pages = {131-133},
        year = 1990,
}
@article{CooperetalNATURE-14,
       author = {Cooper, A. and Magesan, E. and Yum, H. N. and Cappellaro, P.},
        title = {Time-resolved magnetic sensing with electronic spins in diamond},
      journal = {Nature Communications},
    publisher = {Nature Publishing Group},
       volume = 5,
         year = 2014,
     abstract = {Quantum probes can measure time-varying fields with high sensitivity and spatial resolution, enabling the study of biological, material and physical phenomena at the nanometre scale. In particular, nitrogen-vacancy centres in diamond have recently emerged as promising sensors of magnetic and electric fields. Although coherent control techniques have measured the amplitude of constant or oscillating fields, these techniques are not suitable for measuring time-varying fields with unknown dynamics. Here we introduce a coherent acquisition method to accurately reconstruct the temporal profile of time-varying fields using Walsh sequences. These decoupling sequences act as digital filters that efficiently extract spectral coefficients while suppressing decoherence, thus providing improved sensitivity over existing strategies. We experimentally reconstruct the magnetic field radiated by a physical model of a neuron using a single electronic spin in diamond and discuss practical applications. These results will be useful to implement time-resolved magnetic sensing with quantum probes at the nanometre scale.},
}
@techreport{CooperetalTR-88,
       author = {Cooper, Gregory F. and Horvitz, Eric J. and Heckerman, David E.},
        title = {A Method for Temporal Probabilistic Reasoning},
        issue = {KSL-88-30},
  institution = {Stanford Knowledge Systems Laboratory},
         year = 1988,
}
@inproceedings{CoppersmithSTOC-87,
       author = {Don Coppersmith and Schmuel Winograd},
        title = {Matrix multiplication via arithmetic progressions},
    booktitle = {Proceedings of 19th Annual {ACM} Symposium on Theory of Computing},
        pages = {1-6},
         year = 1987,
}
@book{Corballis2017truth,
        title = {The Truth about Language: What It Is and Where It Came From},
       author = {Corballis, Michael C.},
         year = {2017},
    publisher = {University of Chicago Press}
}
@article{CormackandHorspoolCJ-87,
       author = {Cormack, G. V. and Horspool, R. N. S.},
        title = {Data Compression Using Dynamic Markov Modelling},
      journal = {Computer Journal},
    publisher = {Oxford University Press},
       volume = {30},
       number = {6},
         year = {1987},
        pages = {541-550},
     abstract = {A method of dynamically constructing Markov chain models that describe the characteristics of binary messages is developed. Such models can be used to predict future message characters and can therefore be used as a basis for data compression. To this end, the Markov modelling technique is combined with Guazzo's arithmetic coding scheme to produce a powerful method of data compression. The method has the advantage of being adaptive: messages may be encoded or decoded with just a single pass through the data. Experimental results reported here indicate that.the Markov modelling approach generally achieves much better data compression than that observed with competing methods on typical computer data.}
}
@book{CormanLeisersonRivest,
       author = {Corman, T. H. and Leiserson, C. E. and Rivest, R. L.},
        title = {Introduction to Algorithms},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1991,
}
@book{Cormenetal01,
     author = {Thomas H. Cormen and Charles E. Leiserson and Ronald L. Rivest and Clifford Stein},
      title = {Introduction to Algorithms},
  publisher = {Second edition, MIT Press},
    address = {Cambridge, MA},
       year = 2001,
}
@article {Correia-MeloetalEMBO-16,
       author = {Correia-Melo, Clara and Marques, Francisco D. M. and Anderson, Rhys and Hewitt, Graeme and Hewitt, Rachael and Cole, John and Carroll, Bernadette M. and Miwa, Satomi and Birch, Jodie and Merz, Alina and Rushton, Michael D. and Charles, Michelle and Jurk, Diana and Tait, Stephen W. G. and Czapiewski, Rafal and Greaves, Laura and Nelson, Glyn and Bohlooly-Y, Mohammad and Rodriguez-Cuenca, Sergio and Vidal-Puig, Antonio and Mann, Derek and Saretzki, Gabriele and Quarato, Giovanni and Green, Douglas R, and Adams, Peter D, and von Zglinicki, Thomas and Korolchuk, Viktor I, and Passos, Jo{\~a}o F,},
        title = {Mitochondria are required for pro-ageing features of the senescent phenotype},
      journal = {The EMBO Journal},
         year = {2016},
    publisher = {EMBO Press},
     abstract = {Cell senescence is an important tumour suppressor mechanism and driver of ageing. Both functions are dependent on the development of the senescent phenotype, which involves an overproduction of pro-inflammatory and pro-oxidant signals. However, the exact mechanisms regulating these phenotypes remain poorly understood. Here, we show the critical role of mitochondria in cellular senescence. In multiple models of senescence, absence of mitochondria reduced a spectrum of senescence effectors and phenotypes while preserving ATP production via enhanced glycolysis. Global transcriptomic analysis by RNA sequencing revealed that a vast number of senescent-associated changes are dependent on mitochondria, particularly the pro-inflammatory phenotype. Mechanistically, we show that the ATM, Akt and mTORC1 phosphorylation cascade integrates signals from the DNA damage response (DDR) towards PGC-1\hmbeta{}-dependent mitochondrial biogenesis, contributing to a ROS-mediated activation of the DDR and cell cycle arrest. Finally, we demonstrate that the reduction in mitochondrial content in~vivo, by either mTORC1 inhibition or PGC-1hat the ATM, Akt and mTORC1 phosphorylation cascade integrates signals from the DNA damage response (DDR) towards PGC-1\hmbeta{} deletion, prevents senescence in the ageing mouse liver. Our results suggest that mitochondria are a candidate target for interventions to reduce the deleterious impact of senescence in ageing tissues.},
}
@article{CortesandVapnikML-95,
       author = {Cortes, Corinna and Vapnik, Vladimir},
        title = {Support Vector Networks},
      journal = {Machine Learning},
       volume = {20},
        issue = {3},
         year = {1995},
        pages = {273-297},
     abstract = {The support-vector network is a new learning machine for two-group classification problems. The machine conceptually implements the following idea: input vectors are non-linearly mapped to a very high-dimension feature space. In this feature space a linear decision surface is constructed. Special properties of the decision surface ensures high generalization ability of the learning machine. The idea behind the support-vector network was previously implemented for the restricted case where the training data can be separated without errors. We here extend this result to non-separable training data.}
}
@article{CortesetalJoCN-12,
       author = {Cortes, J. M. and Marinazzo, D. and Series, P. and Oram, M. W. and Sejnowski, T. J. and van Rossum, M. C.},
        title = {The effect of neural adaptation on population coding accuracy},
      journal = {Journal of Computational Neuroscience},
       volume = {32},
       number = {3},
         year = {2012},
        pages = {387-402},
     abstract = {Most neurons in the primary visual cortex initially respond vigorously when a preferred stimulus is presented, but adapt as stimulation continues. The functional consequences of adaptation are unclear. Typically a reduction of firing rate would reduce single neuron accuracy as less spikes are available for decoding, but it has been suggested that on the population level, adaptation increases coding accuracy. This question requires careful analysis as adaptation not only changes the firing rates of neurons, but also the neural variability and correlations between neurons, which affect coding accuracy as well. We calculate the coding accuracy using a computational model that implements two forms of adaptation: spike frequency adaptation and synaptic adaptation in the form of short-term synaptic plasticity. We find that the net effect of adaptation is subtle and heterogeneous. Depending on adaptation mechanism and test stimulus, adaptation can either increase or decrease coding accuracy. We discuss the neurophysiological and psychophysical implications of the findings and relate it to published experimental data.}
}
@incollection{CosmidesEVOLUTION_of_INTELLIGENCE-02,
       author = {Cosmides, Leda and Tooby, John},
        title = {Unraveling the enigma of human intelligence: Evolutionary psychology and the multimodular mind.},
    booktitle = {The evolution of intelligence},
    publisher = {Lawrence Erlbaum Associates Publishers},
         year = {2002},
        pages = {145-198},
     abstract = {Evolution brought brains and minds into a world initially devoid of intelligent life. The evolutionary process designed the neural machinery that generates intelligent behavior, and important insights into how this machinery works can be gained by understanding how evolution constructs organisms. Evolutionary psychology was founded on interlocking contributions from evolutionary biology, cognitive science, psychology, anthropology, and neuroscience. It reflects an attempt to think through, from first principles, how current knowledge from these various fields can be integrated into a single, consistent, scientific framework for the study of the mind and brain.}
}
@incollection{CosmidesandTooby1992evolutionary,
       author = {Leda Cosmides and John Tooby},
        title = {Cognitive Adaptations for Social Exchange},
    booktitle = {The adapted mind: {E}volutionary psychology and the generation of culture},
       editor = {J. H. Barkow and L. Cosmides and J. Tooby and George C. Williams},
    publisher = {Oxford University Press},
      address = {New York, New York},
         year = {1992},
        pages = {163-228},
}
@article{CossartetalNATURE-03,
       author = {Cossart, R. and Aronov, D. and Yuste, R.},
        title = {Attractor dynamics of network {U}{P} states in the neocortex},
      journal = {Nature},
         year = {2003},
       volume = {423},
       number = {6937},
        pages = {283-288},
     abstract = {The cerebral cortex receives input from lower brain regions, and its function is traditionally considered to be processing that input through successive stages to reach an appropriate output. However, the cortical circuit contains many interconnections, including those feeding back from higher centres, and is continuously active even in the absence of sensory inputs. Such spontaneous firing has a structure that reflects the coordinated activity of specific groups of neurons. Moreover, the membrane potential of cortical neurons fluctuates spontaneously between a resting (DOWN) and a depolarized (UP) state, which may also be coordinated. The elevated firing rate in the UP state follows sensory stimulation and provides a substrate for persistent activity, a network state that might mediate working memory. Using two-photon calcium imaging, we reconstructed the dynamics of spontaneous activity of up to 1,400 neurons in slices of mouse visual cortex. Here we report the occurrence of synchronized UP state transitions ('cortical flashes') that occur in spatially organized ensembles involving small numbers of neurons. Because of their stereotyped spatiotemporal dynamics, we conclude that network UP states are circuit attractors--emergent features of feedback neural networks that could implement memory states or solutions to computational problems.}
}
@article{CosselletalNATURE-15,
       author = {Cossell, Lee and Iacaruso, Maria Florencia and Muir, Dylan R. and Houlton, Rachael and Sader, Elie N. and Ko, Ho and Hofer, Sonja B. and Mrsic-Flogel, Thomas D.},
        title = {Functional organization of excitatory synaptic strength in primary visual cortex},
      journal = {Nature},
         year = 2015,
    publisher = {Nature Publishing Group},
       volume = 518,
        issue = 7539,
        pages = {399-403},
     abstract = {The strength of synaptic connections fundamentally determines how neurons influence each other's firing. Excitatory connection amplitudes between pairs of cortical neurons vary over two orders of magnitude, comprising only very few strong connections among many weaker ones. Although this highly skewed distribution of connection strengths is observed in diverse cortical areas, its functional significance remains unknown: it is not clear how connection strength relates to neuronal response properties, nor how strong and weak inputs contribute to information processing in local microcircuits. Here we reveal that the strength of connections between layer 2/3 (L2/3) pyramidal neurons in mouse primary visual cortex (V1) obeys a simple rule--the few strong connections occur between neurons with most correlated responses, while only weak connections link neurons with uncorrelated responses. Moreover, we show that strong and reciprocal connections occur between cells with similar spatial receptive field structure. Although weak connections far outnumber strong connections, each neuron receives the majority of its local excitation from a small number of strong inputs provided by the few neurons with similar responses to visual features. By dominating recurrent excitation, these infrequent yet powerful inputs disproportionately contribute to feature preference and selectivity. Therefore, our results show that the apparently complex organization of excitatory connection strength reflects the similarity of neuronal responses, and suggest that rare, strong connections mediate stimulus-specific response amplification in cortical microcircuits.}
}
@article{CostaetalCB-04,
     title = {Differential Corticostriatal Plasticity During Fast and Slow Motor Skill Learning in Mice},
    author = {Rui M. Costa and Dana Cohen and and Miguel A.L. Nicolelis},
   journal = {Current Biology},
    volume = 14,
     issue = 13,
      year = 2004,
     pages = {1124-1134},
  abstract = {Results: By using multielectrode arrays, we recorded the simultaneous activity of neuronal ensembles in motor cortex and dorsal striatum of mice during the different phases of skill learning on an accelerating rotarod. Mice exhibited fast improvement in the task during the initial session and also slow improvement across days. Throughout training, a high percentage of striatal (57\%) and motor cortex (55\%) neurons were task related; i.e   changed their firing rate while mice were running on the rotarod   Improvement in performance was accompanied by substantial plastic changes in both striatum and motor cortex. We observed parallel recruitment of task-related neurons in both structures specifically during the first session. Conversely, during slow learning across sessions we observed differential refinement of the firing patterns in each structure. At the neuronal ensemble level  we observed considerable changes in activity within the first session that became less evident during subsequent sessions   Conclusions: These data indicate that cortical and striatal circuits exhibit remarkable but dissociable plasticity during fast and slow motor skill learning and suggest that distinct neural processes mediate the different phases of motor skill learning.},
}
@inproceedings{CotteretalIVLSI-14,
       author = {M. J. Cotter and Y. Fang and S. P. Levitan and D. M. Chiarulli and V. Narayanan},
    booktitle = {{IEEE} Computer Society Annual Symposium on {VLSI}},
        title = {Computational Architectures Based on Coupled Oscillators},
         year = {2014},
        pages = {130-135},
     abstract = {Recent advances in device technology have opened the door to exploration of new computing paradigms. These paradigms are geared to exploit the behavior of emerging devices such as coupled oscillator arrays. Coupled oscillators are one such device technology that offloads some of the computational complexity to the devices exploiting the physics of the device to perform computation, rather than relying purely on Boolean logic. In this work, we explore the variety of available coupled oscillator architectures. Additionally, we evaluate some basic computation using these architectures in the image processing domain, with an eye towards more complex algorithms and applications.},
}
@article{CoulonetalPFLUGERS-12,
       author = {Coulon, P. and Budde, T. and Pape, H. C. },
        title = {The sleep relay--the role of the thalamus in central and decentral sleep regulation},
      journal = {Pflugers Archive /  European Journal of Physiology},
       volume = {463},
       number = {1},
         year = {2012},
        pages = {53-71},
     abstract = {Surprisingly, the concept of sleep, its necessity and function, the mechanisms of action, and its elicitors are far from being completely understood. A key to sleep function is to determine how and when sleep is induced. The aim of this review is to merge the classical concepts of central sleep regulation by the brainstem and hypothalamus with the recent findings on decentral sleep regulation in local neuronal assemblies and sleep regulatory substances that create a scenario in which sleep is both local and use dependent. The interface between these concepts is provided by thalamic cellular and network mechanisms that support rhythmogenesis of sleep-related activity. The brainstem and the hypothalamus centrally set the pace for sleep-related activity throughout the brain. Decentral regulation of the sleep-wake cycle was shown in the cortex, and the homeostat of non-rapid-eye-movement sleep is made up by molecular networks of sleep regulatory substances, allowing individual neurons or small neuronal assemblies to enter sleep-like states. Thalamic neurons provide state-dependent gating of sensory information via their ability to produce different patterns of electrogenic activity during wakefulness and sleep. Many mechanisms of sleep homeostasis or sleep-like states of neuronal assemblies, e.g. by the action of adenosine, can also be found in thalamic neurons, and we summarize cellular and network mechanisms of the thalamus that may elicit non-REM sleep. It is argued that both central and decentral regulators ultimately target the thalamus to induce global sleep-related oscillatory activity. We propose that future studies should integrate ideas of central, decentral, and thalamic sleep generation.},
}
@article{CourtneyCABN-04,
       author = {Courtney, S. M.},
        title = {Attention and cognitive control as emergent properties of information representation in working memory},
      journal = {Cognitive, Affective and Behavioral Neuroscience},
         year = {2004},
       volume = {4},
       number = {4},
        pages = {501-516},
     abstract = {A hallmark of primate, and particularly human, behavior is cognitive control, the ability to integrate information from a multitude of sources and use that information to flexibly guide behavior in order to achieve an infinite number of goals. The neural mechanisms of cognitive control have yet to be fully elucidated, although the prefrontal cortex is known to play a critical role. Here, I review evidence suggesting that a unifying principle regarding the role of various portions of the prefrontal cortex in a wide range of cognitive tasks is the active maintenance in working memory of different types of currently relevant information from specific stimulus features, to instructional cues, to motivational goals and contexts. I argue that the key to demonstrating the existence of this domain-dependent organization lies in a better understanding of the nature of the representation of this information and the ways in which this information itself controls cognition and behavior.}
}
@article{CoutonetalCURRENT-BIOLOGY-15,
       author = {Couton, Louise and Mauss, Alex S. and Yunusov, Temur and Diegelmann, Soeren and Evers, Jan Felix and Landgraf, Matthias},
        title = {Development of Connectivity in a Motoneuronal Network in Drosophila Larvae},
      journal = {Current Biology},
         year = {2015},
    publisher = {Cell Press},
       volume = {25},
        issue = {5},
        pages = {568-576},
     abstract = {BACKGROUND: Much of our understanding of how neural networks develop is based on studies of sensory systems, revealing often highly stereotyped patterns of connections, particularly as these diverge from the presynaptic terminals of sensory neurons. We know considerably less about the wiring strategies of motor networks, where connections converge onto the dendrites of motoneurons. Here, we investigated patterns of synaptic connections between identified motoneurons with sensory neurons and interneurons in the motor network of the Drosophila larva and how these change as it develops. RESULTS: We find that as animals grow, motoneurons increase the number of synapses with existing presynaptic partners. Different motoneurons form characteristic cell-type-specific patterns of connections. At the same time, there is considerable variability in the number of synapses formed on motoneuron dendrites, which contrasts with the stereotypy reported for presynaptic terminals of sensory neurons. Where two motoneurons of the same cell type contact a common interneuron partner, each postsynaptic cell can arrive at a different connectivity outcome. Experimentally changing the positioning of motoneuron dendrites shows that the geography of dendritic arbors in relation to presynaptic partner terminals is an important determinant in shaping patterns of connectivity. CONCLUSIONS: In the Drosophila larval motor network, the sets of connections that form between identified neurons manifest an unexpected level of variability. Synapse number and the likelihood of forming connections appear to be regulated on a cell-by-cell basis, determined primarily by the postsynaptic dendrites of motoneuron terminals.}
}
@inproceedings{CovellandBalujaICASP-09,
     author = {Michele Covell and Shumeet Baluja},
      title = {{LSH} banding for large-scale retrieval with memory and recall constraints},
  booktitle = {{IEEE} International Conference on Acoustics, Speech, and Signal Processing},
       year = 2009,
      pages = {1865-1868},
}
@book{Cover91,
       author = {Thomas M. Cover and Joy A. Thomas},
        title = {Elements of Information Theory},
    publisher = {John Wiley and Sons},
         year = 1991,
}
@article{CowanPBR-08,
       author = {Cowan, Nelson},
        title = {What are the differences between long-term, short-term, and working memory?},
      journal = {Progress in Brain Research},
       volume = {169},
         year = {2008},
        pages = {323-338},
     abstract = {In the recent literature there has been considerable confusion about the three types of memory: long-term, short-term, and working memory. This chapter strives to reduce that confusion and makes up-to-date assessments of these types of memory. Long- and short-term memory could differ in two fundamental ways, with only short-term memory demonstrating (1) temporal decay and (2) chunk capacity limits. Both properties of short-term memory are still controversial but the current literature is rather encouraging regarding the existence of both decay and capacity limits. Working memory has been conceived and defined in three different, slightly discrepant ways: as short-term memory applied to cognitive tasks, as a multi-component system that holds and manipulates information in short-term memory, and as the use of attention to manage short-term memory. Regardless of the definition, there are some measures of memory in the short term that seem routine and do not correlate well with cognitive aptitudes and other measures (those usually identified with the term ``working memory'') that seem more attention demanding and do correlate well with these aptitudes.},
}
@article{CowardBICA-14,
       author = {Coward, Andrew},
        title = {Brain Computational Primitives},
      journal = {Procedia Computer Science},
       volume = {41},
         year = {2014},
        pages = {164-175},
     abstract = {The brain uses computational primitives that are analogous with but qualitatively different from the computational primitives used in electronic computer systems. The primary computational primitives of the brain are described, and their implementation in anatomy and physiology discussed. Combinations and sequences of these primitives implement cognitive tasks. Many of the primitives have also been implemented electronically. The brain is a very effective general learning system, and although an artificial general intelligence system will be required to learn a different range of behaviours from the brain, the computational primitives used by the brain are the best available guide to appropriate primitives for such an AGI system.}
}
@book{Cowell99,
       author = {R. G. Cowell and A. P. Dawid and S. L. Lauritzen and D. J. Spiegelhalter},
        title = {Probabilistic Networks and Expert Systems},
    publisher = {Springer},
         year = 1999,
}
@book{Cowelletal99,
       author = {Cowell, R. G. and Dawid, A. P. and Lauritzen, S. L. and Spiegelhalter, D. J.},
        title = {Probabilistic networks and expert systems},
    publisher = {Springer-Verlag},
      address = {New York, NY},
         year = 1999,
}
@article{CowenetalNEUROIMAGE-14,
       author = {Cowen, A. S. and Chun, M. M. and Kuhl, B. A.},
        title = {Neural portraits of perception: reconstructing face images from evoked brain activity},
      journal = {Neuroimage},
         year = 2014,
       volume = 94,
        pages = {12-22},
     abstract = {Recent neuroimaging advances have allowed visual experience to be reconstructed from patterns of brain activity. While neural reconstructions have ranged in complexity, they have relied almost exclusively on retinotopic mappings between visual input and activity in early visual cortex. However, subjective perceptual information is tied more closely to higher-level cortical regions that have not yet been used as the primary basis for neural reconstructions. Furthermore, no reconstruction studies to date have reported reconstructions of face images, which activate a highly distributed cortical network. Thus, we investigated (a) whether individual face images could be accurately reconstructed from distributed patterns of neural activity, and (b) whether this could be achieved even when excluding activity within occipital cortex. Our approach involved four steps. (1) Principal component analysis (PCA) was used to identify components that efficiently represented a set of training faces. (2) The identified components were then mapped, using a machine learning algorithm, to fMRI activity collected during viewing of the training faces. (3) Based on activity elicited by a new set of test faces, the algorithm predicted associated component scores. (4) Finally, these scores were transformed into reconstructed images. Using both objective and subjective validation measures, we show that our methods yield strikingly accurate neural reconstructions of faces even when excluding occipital cortex. This methodology not only represents a novel and promising approach for investigating face perception, but also suggests avenues for reconstructing 'offline' visual experiences-including dreams, memories, and imagination-which are chiefly represented in higher-level cortical areas.}
}
@book{Cox2018talking,
        title = {Now You're Talking: {H}uman Conversation from the Neanderthals to Artificial Intelligence},
       author = {Cox, Trevor},
         year = {2018},
    publisher = {Counterpoint Press}
}
@article{CoxCoRR-16,
       author = {David Cox},
        title = {Clique Topology Reveals Intrinsic Geometric Structure in Neural Correlations: An Overview},
      journal = {CoRR},
       volume = {arXiv:1608.03463},
         year = {2016},
     abstract = {Gusti et al. present clique topology - a novel matrix analysis technique to extract structural features from neural activity data that contains hidden nonlinearities. This represents a significant improvement upon the previous eigenvalue-based approaches that fail under similar conditions. Structural features can yield significant insight into neurological processes. Their findings are verified by showing that the geometric structure of neural correlations in the rat hippocampus is determined by hippocampal circuits rather than being a consequence of positional coding.}
}
@article{CoxandDeanCURRENT-BIOLOGY-14,
    author = {David Daniel Cox and Thomas Dean},
     title = {Neural Networks and Neuroscience-Inspired Computer Vision},
   journal = {Current Biology},
    volume = {24},
     issue = {18},
      year = {2014},
     pages = {921-929},
     abstract = {Brains are, at a fundamental level, biological computing machines. They transform a torrent of complex and ambiguous sensory information into coherent thought and action, allowing an organism to perceive and model its environment, synthesize and make decisions from disparate streams of information, and adapt to a changing environment. Against this backdrop, it is perhaps not surprising that computer science, the science of building artificial computational systems, has long looked to biology for inspiration. However, while the opportunities for cross-pollination between neuroscience and computer science are great, the road to achieving brain-like algorithms has been long and rocky. Here, we review the historical connections between neuroscience and computer science, and we look forward to a new era of potential collaboration, enabled by recent rapid advances in both biologically-inspired computer vision and in experimental neuroscience methods. In particular, we explore where neuroscience-inspired algorithms have succeeded, where they still fail, and we identify areas where deeper connections are likely to be fruitful.}
}
@article{CoxandDiCarloJN-08,
     author = {David D. Cox and James J. DiCarlo},
      title = {Does Learned Shape Selectivity in Inferior Temporal Cortex Automatically Generalize Across Retinal Position?},
    journal = {Journal of Neuroscience},
     volume = 28, 
      issue = 40,
      pages = {10045-10055},
      month = {November},
       year = 2008,
}
@article{CoxetalJoN-08,
    author = {Cox, D. D. and Papanastassiou, A. M. and Oreper, D. and Andken, B. B. and DiCarlo, J. J.},
     title = {High-resolution three-dimensional microelectrode brain mapping using stereo microfocal x-ray imaging},
   journal = {Journal of Neurophysiology},  
    volume = 100,
     issue = 5,
      year = 2008,
     pages = {2966-2976},
  abstract = {Much of our knowledge of brain function has been gleaned from studies using microelectrodes to characterize the response properties of individual neurons in vivo. However, because it is difficult to accurately determine the location of a microelectrode tip within the brain, it is impossible to systematically map the fine three-dimensional spatial organization of many brain areas, especially in deep structures. Here, we present a practical method based on digital stereo microfocal X-ray imaging that makes it possible to estimate the three-dimensional position of each and every microelectrode recording site in "real time" during experimental sessions. We determined the system"s ex vivo localization accuracy to be better than 50 microns, and we show how we have used this method to coregister hundreds of deep-brain microelectrode recordings in monkeys to a common frame of reference with median error of <150 microns. We further show how we can coregister those sites with magnetic resonance images (MRIs), allowing for comparison with anatomy, and laying the groundwork for more detailed electrophysiology/functional MRI comparison. Minimally, this method allows one to marry the single-cell specificity of microelectrode recording with the spatial mapping abilities of imaging techniques; furthermore, it has the potential of yielding fundamentally new kinds of high-resolution maps of brain function.}
}
@article{CraigheadNATURE-06,
     title = {Future lab-on-a-chip technologies for interrogating individual molecules},
    author = {Craighead, H.},
   journal = {Nature},
    volume = 442,
     issue = 7101,
     pages = {387-393},
      year = 2006,
  abstract = {Advances in technology have allowed chemical sampling with high spatial resolution and the manipulation and measurement of individual molecules. Adaptation of these approaches to lab-on-a-chip formats is providing a new class of research tools for the investigation of biochemistry and life processes.}
}
@book{Craik1943,
        title = {The Nature of Explanation},
       author = {Craik, Kenneth J. W.},
    publisher = {Cambridge University Press},
      comment = {See PDF for Chapter V only},
         year = {1943},
}
@article{CraneIRE-60,
       author = {H. D. Crane},
        title = {The Neuristor},
      journal = {{IRE} Transactions on Electronic Computing},
       volume = {EC-9},
        issue = 3,
         year = 1960,
        pages = {370-371},
}
@article{CretonBBR-09,
       author = {Robbert Creton},
        title = {Automated analysis of behavior in zebrafish larvae},
      journal = {Behavioural Brain Research},
       volume = {203},
       number = {1},
         year = {2009},
        pages = {127-136},
     abstract = {Zebrafish larvae have become a popular model system to examine genetic and environmental factors that affect behavior. However, studying complex behavior in large numbers of fish larvae can be challenging. The present study describes a novel high-resolution imaging system that is unique in its ability to automatically analyze the location and orientation of zebrafish larvae in multiwell plates. The system revealed behaviors in zebrafish larvae that would have been missed by more manual approaches, including a preference to face a threatening stimulus from a distance and a clockwise orientation in a two-fish assay. The clockwise orientation of the larvae correlates with a clockwise orientation of molecular structures during early development. Larvae with reversed embryonic asymmetries display a counter-clockwise orientation in the two-fish assay, suggesting that embryonic asymmetry and chiral behavior are regulated by the same developmental mechanisms. The developed imaging techniques may be used in large-scale screens to identify genes, pharmaceuticals, and environmental toxicants that influence complex behaviors.}
}
@article {Creutzfeldt1977,
    author = {Creutzfeldt, Otto D.},
     title = {Generality of the functional structure of the neocortex},
   journal = {Naturwissenschaften},
     pages = {507-517},
    volume = 64,
     issue = 10,
      year = 1977
}
@article{CrickSCIAM-79,
    author = {Thinking about the brain},
     title = {Crick, F. H.},
   journal = {Scientific American},
    volume = 241,
     issue = 3,
      year = 1979,
     pages = {219-232},
}
@article{CrickandKochNATURE-NEUROSCIENCE-03,
       author = {Crick, Francis and Koch, Christof},
        title = {A framework for consciousness},
      journal = {Nature Neuroscience},
         year = {2003},
       volume = {6},
        issue = {2},
        pages = {119-126},
     abstract = {Here we summarize our present approach to the problem of consciousness. After an introduction outlining our general strategy, we describe what is meant by the term "framework" and set it out under ten headings. This framework offers a coherent scheme for explaining the neural correlates of (visual) consciousness in terms of competing cellular assemblies. Most of the ideas we favor have been suggested before, but their combination is original. We also outline some general experimental approaches to the problem and, finally, acknowledge some relevant aspects of the brain that have been left out of the proposed framework.},
}
@article{CrickandKochPTRSB-05,
     title = {{What is the function of the claustrum?}},
    author = {Crick, Francis C. and Koch, Christof},
   journal = {Philosophical Transactions of the Royal Society B: Biological Sciences},
 publisher = {The Royal Society},
    volume = 360,
     issue = 1458,
      year = 2005,
     pages = {1271-1279},
  abstract = {The claustrum is a thin, irregular, sheet-like neuronal structure hidden beneath the inner surface of the neocortex in the general region of the insula. Its function is enigmatic. Its anatomy is quite remarkable in that it receives input from almost all regions of cortex and projects back to almost all regions of cortex. We here briefly summarize what is known about the claustrum, speculate on its possible relationship to the processes that give rise to integrated conscious percepts, propose mechanisms that enable information to travel widely within the claustrum and discuss experiments to address these questions.},
}
@article {CrittendenetalJoN-16,
       author = {Crittenden, Ben M. and Mitchell, Daniel J. and Duncan, John},
        title = {Task Encoding across the Multiple Demand Cortex Is Consistent with a Frontoparietal and Cingulo-Opercular Dual Networks Distinction},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {36},
       number = {23},
         year = {2016},
        pages = {6147-6155},
     abstract = {Multiple-demand (MD) regions of the human brain show coactivation during many different kinds of task performance. Previous work based on resting-state functional magnetic resonance imaging (fMRI) has shown that MD regions may be divided into two closely coupled subnetworks centered around the lateral frontoparietal (FP) and cingulo-opercular cortex. Here, we used on-task fMRI to test whether this division is apparent during the performance of an executive task. Furthermore, we investigated whether there is a difference in the encoding of task between the two subnetworks. Using connectivity methods, we found that activity across the entire MD cortex is correlated during task performance. Meanwhile, however, there was significantly stronger connectivity within each of the subnetworks than between them. Using multivoxel pattern analysis, we also found that, although we were able to decode task-relevant information from all regions of the MD cortex, classification accuracy scores were significantly higher in the FP subnetwork. These results suggest a nested picture with MD regions as a whole showing coactivation and broad rule representation, but with significant functional distinctions between component subnetworks.SIGNIFICANCE STATEMENT Multiple-demand (MD) regions of frontal and parietal cortex appear essential for the orchestration of goal-directed behavior and problem solving. Understanding the relative specialization of regions within the MD cortex is crucial to understanding how we can coordinate and execute complex action plans. By examining functional connectivity during task performance, we extend previous findings suggesting that the MD cortex can be divided into two subnetworks centered around the frontoparietal (FP) and cingulo-opercular (CO) cortex. Furthermore, using multivoxel pattern analysis, we show that, compared with the CO subnetwork, the FP subnetwork manifests more differentiated coding of specific task events.},
}
@article{CroftetalNP-15,
       author = {Wayne Croft and Katharine L. Dobson and Tomas C. Bellamy},
        title = {Plasticity of Neuron-Glial Transmission: Equipping Glia for Long-Term Integration of Network Activity},
      journal = {Neural Plasticity},
       volume = {2015},
        issue = {765792},
         year = {2015},
        pages = {1-11},
     abstract = {The capacity of synaptic networks to express activity-dependent changes in strength and connectivity is essential for learning and memory processes. In recent years, glial cells (most notably astrocytes) have been recognized as active participants in the modulation of synaptic transmission and synaptic plasticity, implicating these electrically nonexcitable cells in information processing in the brain. While the concept of bidirectional communication between neurons and glia and the mechanisms by which gliotransmission can modulate neuronal function are well established, less attention has been focused on the computational potential of neuron-glial transmission itself. In particular, whether neuron-glial transmission is itself subject to activity-dependent plasticity and what the computational properties of such plasticity might be has not been explored in detail. In this review, we summarize current examples of plasticity in neuron-glial transmission, in many brain regions and neurotransmitter pathways. We argue that induction of glial plasticity typically requires repetitive neuronal firing over long time periods (minutes-hours) rather than the short-lived, stereotyped trigger typical of canonical long-term potentiation. We speculate that this equips glia with a mechanism for monitoring average firing rates in the synaptic network, which is suited to the longer term roles proposed for astrocytes in neurophysiology.}
}
@incollection{CrollAIP-75,
       author = {Neil A. Croll},
        title = {Behavioural Analysis of Nematode Movement},
       editor = {Ben Dawes},
    booktitle = {Advances in Parasitology},
    publisher = {Academic Press},
         year = {1975},
       volume = {13},
        pages = {71-122},
     abstract = {Publisher Summary This chapter presents the behavioral physiology of the nematode level of organization. Nematodes are generally considered to be a phylum or a class of the phylum Aschelminthes. Their exact status is debatable, but the nematodes are at a more complex level of organization than the coelenterates and platyhelminths and simpler than the coelomate groups: annelids, molluscs, and arthropods. The neuromuscular organization of nematodes is basically bilaterally symmetrical, but the amphids, photoreceptors, phasmids, deirids, and hemizonids are arranged perpendicularly to the dorsoventral direction of muscular contraction. The chapter discusses the nature of locomotory activity and reviews the relative contributions of exogenous and endogenous components in the integrated activities of nematodes. The main behavioral events investigated with respect to locomotion are hatching, moulting, feeding, and mechanisms of orientation. Many species of phytoparasitic nematode are observed feeding, and the mechanism of integrated feeding are discussed. Feeding in phytoparasites usually consists of active movement prior to contact, contact of the nematode head with the plant, reduced movement, head waving and spasmodic stylet thrusting, cessation of bodily waves, deliberate stylet thrusting and adoption of body posture, rapid bulb pulsation and ingestion, and end of feeding and movement away. Nematodes, as other organisms, may disperse in three main ways: by drifting in currents of water or air, by being carried on moving objects, or by swimming or creeping and thus, dispersing by their own efforts.}
}
@misc{Crosslinking_Reagents_Handbook-15,
        title = {Crosslinking Reagents Handbook},
       author = {Thermo Scientific},
         year = 2015,
          url = {https://tools.lifetechnologies.com/content/sfs/brochures/1602163-Crosslinking-Reagents-Handbook.pdf},
 howpublished = {Thermo Scientific Technical Report},
}    
@article{CrottiandRansohoffIMMUNITY-16,
       author = {Andrea Crotti and Richard M. Ransohoff},
        title = {Microglial Physiology and Pathophysiology: Insights from Genome-wide Transcriptional Profiling},
      journal = {Immunity},
       volume = {44},
       number = {3},
        pages = {505-515},
         year = {2016},
     abstract = {Microglia originate from erythromyeloid progenitors (EMPs) in the yolk sac and develop in the forming CNS. Microglia are fundamental for the development and function of a healthy brain. By contrast, their role in immune host defense of the \{CNS\} remains speculative, given the immune privilege of this organ. Alterations in microglia functionality are involved in brain aging, as well as in neurodegenerative disease severity and progression. The combination of their ontogeny with the influence of the complex environment of the \{CNS\} makes microglia a unique cell population. Recent observations about microglia ontogeny combined with extensive gene expression profiling allow us to better capture the variety of nuances that microglia can manifest. Here, we provide a contemporary appraisal of microglial uniqueness based on their origin, functions, and expression profiles. Furthermore, we give an overview of the impact of aging and neurodegenerative diseases on microglia transcriptomes.}
}  
@article{CrottiandRansohoffIMMUNITY-18,
       author = {Crotti, Andrea and Ransohoff, Richard M.},
        title = {Microglial Physiology and Pathophysiology: Insights from Genome-wide Transcriptional Profiling},
      journal = {Immunity},
         year = {2018},
    publisher = {Elsevier},
       volume = {44},
        issue = {3},
        pages = {505-515},
     abstract = {Microglia originate from erythromyeloid progenitors (EMPs) in the yolk sac and develop in the forming CNS. Microglia are fundamental for the development and function of a healthy brain. By contrast, their role in immune host defense of the CNS remains speculative, given the immune privilege of this organ. Alterations in microglia functionality are involved in brain aging, as well as in neurodegenerative disease severity and progression. The combination of their ontogeny with the influence of the complex environment of the CNS makes microglia a unique cell population. Recent observations about microglia ontogeny combined with extensive gene expression profiling allow us to better capture the variety of nuances that microglia can manifest. Here, we provide a contemporary appraisal of microglial uniqueness based on their origin, functions, and expression profiles. Furthermore, we give an overview of the impact of aging and neurodegenerative diseases on microglia transcriptomes.},
}
@article{CrouseetalCoRR-20,
        title = {Neural Analogical Matching},
       author = {Maxwell Crouse and Constantine Nakos and Ibrahim Abdelaziz and Kenneth Forbus},
         year = {2020},
       volume = {arXiv:2004.03573},
      journal = {CoRR},
     abstract = {Analogy is core to human cognition. It allows us to solve problems based on prior experience, it governs the way we conceptualize new information, and it even influences our visual perception. The importance of analogy to humans has made it an active area of research in the broader field of artificial intelligence, resulting in data-efficient models that learn and reason in human-like ways. While cognitive perspectives of analogy and deep learning have generally been studied independently of one another, the integration of the two lines of research is a promising step towards more robust and efficient learning techniques. As part of a growing body of research on such an integration, we introduce the Analogical Matching Network: a neural architecture that learns to produce analogies between structured, symbolic representations that are largely consistent with the principles of Structure-Mapping Theory.}
}
@article{CrupietalICR-12,
       author = {Michael Crupi and Laxman Pradhan and Stuart Tozer},
        title = {Modelling Neural Plasticity with Memristors},
      journal = {IEEE Canadian Review},
       volume = 68,
         year = 2012,
        pages = {10-14},
abstract = {The goal of this project was to explore the use of memristors for modelling the behaviour of neurons. This paper outlines the approaches used to build a neuromorphic circuit using only passive electronic components, including memristive devices fabricated at Carleton University. Specifically, a model of real-time pattern recognition in the human visual system was used to demonstrate how memristors could be used to simulate memory storage in the brain.}
}
@incollection{Crutchfield91,
       author = {Crutchfield, James P.},
        title = {Reconstructing Language Hierarchies},
       editor = {Atmanspacher, H. and Scheingraber, H.},
    booktitle = {Information Dynamics},
    publisher = {Plenum Press},
     location = {New York},
         year = 1991,
        pages = {45-60},
}
@incollection{CrutchfieldandYoung90,
       author = {Crutchfield, James P. and Young, Karl},
        title = {Computation at the Onset of Chaos},
    booktitle = {Complexity, Entropy, and the Physics of Information},
    publisher = {Addison-Wesley},
       volume = 8,
         year = 1990,
        pages = {223-268},
       editor = {W. H. Zurek},
}
@article{CruzetalNEUROSCIENCE-05,
       author = {Cruz, L. and Buldyrev, S. V. and Peng, S. and Roe, D. L. and Urbanc, B. and Stanley, H. E. and Rosene, D. L.},
        title = {A statistically based density map method for identification and quantification of regional differences in microcolumnarity in the monkey brain},
      journal = {Journal Neuroscience Methods},
         year = 2005,
       volume = 141,
        issue = 2,
        pages = {321-332},
     abstract = {We present a statistical density map method derived from condensed matter physics to quantify microcolumns, the fundamental computational unit of the cerebral cortex. This method provides measures for microcolumnar strength, width, spacing, length, and periodicity. We applied this method to Nissl-stained 30 microm thick frozen sections from areas 46, TE, and TL of rhesus monkey brains, areas that differ visually in microcolumnarity and are associated with different cognitive functions. Our results indicate that microcolumns in these areas are similar in width, spacing, and periodicity, but are stronger (possess a higher neuronal density) in area TE, as compared to areas TL and 46. We modeled the effect of section orientation on microcolumnar spacing and demonstrated that this method provides an adequate estimate of spacing. We also modeled disruption of microcolumnarity by performing simulations that randomly displace neurons and demonstrated that displacements of only one neuronal diameter effectively eliminate microcolumnar organization. These results indicate that our density map method is sensitive enough to detect and quantify subtle differences in microcolumnar organization that may occur in the context of development, aging, and neuropathology, as well as between areas and species.}
}
@techreport{CsajiMSc-01,
       author = {Bal\'{a}zs Csan\'{a}d Cs\'{a}ji},
        title = {Approximation with Artificial Neural Networks},
  institution = {Faculty of Sciences; E\"{o}tv\"{o}s Lor\'{a}nd University, Hungary},
         year = 2015,
}
@article{CsiszarandTusnadySD-84,
        title = {Information geometry and alternating minimization procedures},
       author = {Csisz\'{a}r, I. and Tusn\'{a}dy, G.},
      journal = {Statistics and Decisions},
       editor = {E. F. Dedewicz},
         year = {1984},
        pages = {205-237}
}
@article{CumminsetalNEUROCOMPUTING-03,
       author = {G.I. Cummins and S.M. Crook and A.G. Dimitrov and T. Ganje and G.A. Jacobs and J.P. Miller},
        title = {Structural and biophysical mechanisms underlying dynamic sensitivity of primary sensory interneurons in the cricket cercal sensory system },
      journal = {Neurocomputing},
       volume = {52-54},
         year = {2003},
        pages = {45-52},
     abstract = {We constructed probabilistic models of afferent inputs and compartmental models of interneurons in the cricket cercal system to examine the effects of dendritic morphology, distribution of synaptic inputs, and membrane properties on interneuron directional tuning properties. The mean directional tuning of afferent inputs to an interneuron was an excellent predictor of its directional tuning. Location of the synapses on the interneurons' dendrites was not essential to determining tuning characteristics, but had a substantial effect on sensitivity. Thus, we conclude that both sampling of the afferent population, and anatomical distribution of synaptic inputs are important determinants of an interneuron's directional response.}
}
@article{CunninghamandYuNATURE-NEUROSCIENCE-14,
       author = {Cunningham, J. P. and Yu, B. M.},
        title = {Dimensionality reduction for large-scale neural recordings},
      journal = {Nature Neuroscience},
         year = {2014},
       volume = {17},
       number = {11},
        pages = {1500-1509},
     abstract = {Most sensory, cognitive and motor functions depend on the interactions of many neurons. In recent years, there has been rapid development and increasing use of technologies for recording from large numbers of neurons, either sequentially or simultaneously. A key question is what scientific insight can be gained by studying a population of recorded neurons beyond studying each neuron individually. Here, we examine three important motivations for population studies: single-trial hypotheses requiring statistical power, hypotheses of population response structure and exploratory analyses of large data sets. Many recent studies have adopted dimensionality reduction to analyze these populations and to find features that are not apparent at the level of individual neurons. We describe the dimensionality reduction methods commonly applied to population activity and offer practical advice about selecting methods and interpreting their outputs. This review is intended for experimental and computational researchers who seek to understand the role dimensionality reduction has had and can have in systems neuroscience, and who seek to apply these methods to their own data.}
}
@article{CurrieandTateAIJ-91,
       author = {Currie, K. and Tate, A.},
        title = {{O-Plan}: The Open Planning Architecture},
      journal = AIJ,
       volume = 51,
        issue = 1,
        pages = {49-86},
         year = 1991,
}
@misc{CurtoCEB-10,
       author = {Carina Curto},
        title = {What can topology tell us about the neural code?},
 affiliation  = {American Mathematical Society Joint Mathematics Meetings},
         year = {2016},
     abstract = {Neuroscience is undergoing a period of rapid experimental progress and expansion. New mathematical tools, previously unknown in the neuroscience community, are now being used to tackle fundamental questions and analyze emerging data sets. Consistent with this trend, the last decade has seen an uptick in the use of topological ideas and methods in neuroscience. In this talk I will survey recent applications of topology in neuroscience, and explain why topology is an especially natural tool for understanding neural codes.}
}
@article{CurtoetalCMB-13,
        title = {The neural ring: an algebraic tool for analyzing the intrinsic structure of neural codes},
       author = {Curto, Carina and Itskov, Vladimir and Veliz-Cuba, Alan and Youngs, Nora},
      journal = {Bulletin of Mathematical Biology},
    publisher = {Springer},
       volume = {75},
       number = {9},
         year = {2013},
        pages = {1571-1611},
     abstract = {Neurons in the brain represent external stimuli via neural codes. These codes often arise from stereotyped stimulus-response maps, associating to each neuron convex receptive field. An important problem confronted by the brain is to infer properties of a represented stimulus space without knowledge of the receptive fields, using only the intrinsic structure of the neural code. How does the brain do this? To address this question, it is important to determine what stimulus space features n--in principle--be extracted from neural codes. This motivates us to define the neural ring and a related neural ideal, algebraic objects that encode the full combinatorial data of a neural code. Our main finding is that these objects can be expressed in a "canonical form" that directly translates to a minimal description of the receptive field structure intrinsic to the code. We also find connections to Stanley-Reisner rings, and use ideas similar to those in the theory of monomial ideals to obtain an algorithm for computing the primary decomposition of pseudo-monomial ideals. This allows us to algorithmically extract the canonical form associated to any neural code, avoiding the groundwork for inferring stimulus space features from neural activity alone.}, 
}
@article{CurtoetalJoN-09,
       author = {Curto, C. and Sakata, S. and Marguet, S. and Itskov, V. and Harris, K. D.},
        title = {A simple model of cortical dynamics explains variability and state dependence of sensory responses in urethane-anesthetized auditory cortex},
      journal = {Journal Neuroscience},
         year = {2009},
       volume = {29},
       number = {34},
        pages = {10600-10612},
     abstract = {The responses of neocortical cells to sensory stimuli are variable and state dependent. It has been hypothesized that intrinsic cortical dynamics play an important role in trial-to-trial variability; the precise nature of this dependence, however, is poorly understood. We show here that in auditory cortex of urethane-anesthetized rats, population responses to click stimuli can be quantitatively predicted on a trial-by-trial basis by a simple dynamical system model estimated from spontaneous activity immediately preceding stimulus presentation. Changes in cortical state correspond consistently to changes in model dynamics, reflecting a nonlinear, self-exciting system in synchronized states and an approximately linear system in desynchronized states. We propose that the complex and state-dependent pattern of trial-to-trial variability can be explained by a simple principle: sensory responses are shaped by the same intrinsic dynamics that govern ongoing spontaneous activity.}
}
@book{Cutsuridis2011perception,
        title = {Perception-Action Cycle: Models, Architectures, and Hardware},
       author = {Cutsuridis, V. and Hussain, A. and Taylor, J.G.},
       series = {Springer Series in Cognitive and Neural Systems},
         year = {2011},
	  url = {https://link.springer.com/content/https://link.springer.com/content/pdf/10.1007/978-1-4419-1452-1.pdf},
    publisher = {Springer New York}
}
@article{CybenkoMCSS-89,
        title = {Approximation by superpositions of a sigmoidal function},
       author = {Cybenko, George},
      journal = {Mathematics of Control, Signals, and Systems (MCSS)},
       volume = {2},
       number = {4},
        pages = {303-314},
         year = {1989},
    publisher = {Springer London}
}
@article{CybulskietalFCN-15,
       author = {Thaddeus R. Cybulski and Joshua I. Glaser and Adam H. Marblestone and Bradley M. Zamft and Edward S. Boyden and George M. Church and Konrad P. Kording},
        title = {Spatial Information in Large-scale Neural Recordings},
      journal = {Frontiers in Computational Neuroscience},
       volume = {8},
        issue = {172},
         year = {2015},
        pages = {1-16},
     abstract = {To record from a given neuron, a recording technology must be able to separate the activity of that neuron from the activity of its neighbors. Here, we develop a Fisher information based framework to determine the conditions under which this is feasible for a given technology. This framework combines measurable point spread functions with measurable noise distributions to produce theoretical bounds on the precision with which a recording technology can localize neural activities. If there is sufficient information to uniquely localize neural activities, then a technology will, from an information theoretic perspective, be able to record from these neurons. We (1) describe this framework, and (2) demonstrate its application in model experiments. This method generalizes to many recording devices that resolve objects in space and should be useful in the design of next-generation scalable neural recording systems.}
}
@incollection{CzechetalMCELL-09,
       author = {Czech, Jacob and Dittrich, Markus and Stiles, Joel R.},
        title = {Rapid Creation, Monte Carlo Simulation, and Visualization of Realistic {3D} Cell Models},
    booktitle = {Systems Biology},
       editor = {Maly, Ivan V.},
       series = {Methods in Molecular Biology},
    publisher = {Humana Press},
       volume = 500,
         year = 2009,
        pages = {237-287},
     abstract = {Spatially realistic diffusion-reaction simulations supplement traditional experiments and provide testable hypotheses for complex physiological systems. To date, however, the creation of realistic 3D cell models has been difficult and time-consuming, typically involving hand-reconstruction from electron microscopic images. Here we present a complementary approach that is much simpler and faster, because the cell architecture (geometry) is created directly in silico using 3D modeling software like that used for commercial film animations. We show how a freely available open source program (Blender) can be used to create the model geometry, which then can be read by our Monte Carlo simulation and visualization software (MCell and DReAMM, respectively). This new workflow allows rapid prototyping and development of realistic computational models, and thus should dramatically accelerate their use by a wide variety of computational and experimental investigators. Using two self-contained examples based on synaptic transmission, we illustrate the creation of 3D cellular geometry with Blender, addition of molecules, reactions, and other run-time conditions using MCell's Model Description Language (MDL), and subsequent MCell simulations and DReAMM visualizations. In the first example, we simulate calcium influx through voltage-gated channels localized on a presynaptic bouton, with subsequent intracellular calcium diffusion and binding to sites on synaptic vesicles. In the second example, we simulate neurotransmitter release from synaptic vesicles as they fuse with the presynaptic membrane, subsequent transmitter diffusion into the synaptic cleft, and binding to postsynaptic receptors on a dendritic spine.}
}
@article{DEpenouxMS-63,
       author = {{D'E}penoux, F.},
        title = {Sur un probl\'{e}me de production et de stockage dans l'aleatoire},
      journal = {Management Science},
       volume = 10,
        pages = {98-108},
         year = 1963,
}
@article{DEspositoetalNATURE-95,
       author = {D'Esposito, M. and Detre, J. A. and Alsop, D. C. and Shin, R. K. and Atlas, S. and Grossman, M.},
        title = {The neural basis of the central executive system of working memory},
      journal = {Nature},
       volume = {378},
       number = {6554},
         year = {1995},
        pages = {279-281},
     abstract = {Working memory refers to a system for temporary storage and manipulation of information in the brain, a function critical for a wide range of cognitive operations. It has been proposed that working memory includes a central executive system (CES) to control attention and information flow to and from verbal and spatial short-term memory buffers. Although the prefrontal cortex is activated during both verbal and spatial passive working memory tasks, the brain regions involved in the CES component of working memory have not been identified. We have used functional magnetic resonance imaging (fMRI) to examine brain activation during the concurrent performance of two tasks, which is expected to engage the CES. Activation of the prefrontal cortex was observed when both tasks are performed together, but not when they are performed separately. These results support the view that the prefrontal cortex is involved in human working memory.}
}
@misc{DMOZ,
       title = {{O}pen {D}irectory {P}roject},
      author = {Netscape, Communication Corporation},
howpublished = {\url{http://dmoz.org/}},
        year = 2003,
}
@article{DabneyetalNATURE-20,
       author = {Dabney, Will and Kurth-Nelson, Zeb and Uchida, Naoshige and Starkweather, Clara Kwon and Hassabis, Demis and Munos, R{\`{e}}mi and Botvinick, Matthew},
        title = {A distributional code for value in dopamine-based reinforcement learning},
      journal = {Nature},
         year = {2020},
       volume = {577},
        issue = {7792},
        pages = {671-675},
     abstract = {Since its introduction, the reward prediction error theory of dopamine has explained a wealth of empirical phenomena, providing a unifying framework for understanding the representation of reward and value in the brain. According to the now canonical theory, reward predictions are represented as a single scalar quantity, which supports learning about the expectation, or mean, of stochastic outcomes. Here we propose an account of dopamine-based reinforcement learning inspired by recent artificial intelligence research on distributional reinforcement learning. We hypothesized that the brain represents possible future rewards not as a single mean, but instead as a probability distribution, effectively representing multiple future outcomes simultaneously and in parallel. This idea implies a set of empirical predictions, which we tested using single-unit recordings from mouse ventral tegmental area. Our findings provide strong evidence for a neural realization of distributional reinforcement learning.},
}
@incollection{DaganetalPASCAL-06,
        title = {The PASCAL Recognising Textual Entailment Challenge},
       author = {Ido Dagan and Oren Glickman and Bernardo Magnini},
    booktitle = {Proceedings of the 2005 Machine Learning Challenges Workshop},
       volume = 3944,
       series = {Lecture Notes in Artificial Intelligence},
       editor = {Quinonero-Candela et al.},
    publisher = {Springer-Verlag},
         year = 2006, 
          url = {http://pascallin.ecs.soton.ac.uk/Challenges/RTE3/},
        pages = {177-190},
}
@article{DagumandLubyAIJ-93,
       author = {Dagum, P. and Luby, M.},
        title = {Approximating Probabilistic Inference in {B}ayesian Belief Networks is {NP}-hard},
      journal = AIJ,
       volume = 60,
         year = 1993,
        pages = {141-153},
}
@inproceedings{DagumetalFOCS-95,
       author = {Dagum, P. and Karp, R. and Luby, M. and Ross, Sheldon M.},
        title = {An Optimal Stopping Rule for Monte Carlo Estimation},
    booktitle = {Proceedings of the 1995 Symposium on the Foundations of Computer Science},
         year = 1995,
}
@inproceedings{DalalandTriggsCVPR-05,
       author = {Dalal, N. and Triggs, B.},
        title = {Histograms of Oriented Gradients for Human Detection},
    booktitle = {{IEEE} Computer Society Conference on Computer Vision and Pattern Recognition},
         year = 2005,
       volume = 1,
        pages = {886-893},
}
@incollection{Dalhlaus89,
       author = {Dalhlaus, E.},
        title = {Reductions to {NP}-complete problems by interpretations},
    booktitle = {Logic and Machines: Decision Problems and Complexity},
       editor = {E. B\"{o}rger {\em et al.,\/}},
    publisher = {Springer-Verlag},
        pages = {357-365},
         year = 1989,
}
@book{Damasio03,
     author = {Antonio Damasio},
      title = {Looking for Spinoza: Joy, Sorrow, and the Feeling Brain},
  publisher = {Harcourt},
    address = {New York, NY},
       year = 2003,
}
@book{Damasio1999,
     author = {Antonio Damasio},
      title = {The Feeling of What Happens: Body and 
               Emotion in the Making of Consciousness},
  publisher = {Harcourt},
    address = {New York, NY},
       year = 1999,
}
@book{Damasio2010,
     author = {Antonio Damasio},
      title = {Self Comes to Mind: Constructing the Conscious Brain},
  publisher = {Albert A. Knopf},
    address = {New York, NY},
       year = 2010,
} 
@article{DanemanandMeriklePBR-96,
       author = {Daneman, Meredyth and Merikle, Philip M.},
        title = {Working memory and language comprehension: A meta-analysis},
      journal = {Psychonomic Bulletin \& Review},
       volume = {3},
        issue = {4},
        pages = {422-433},
         year = {1996},
     abstract = {This paper presents a meta-analysis of the data from 6,179 participants in 77 studies that investigated the association between working-memory capacity and language comprehension ability. A primary goal of the meta-analysis was to compare the predictive power of the measures of working memory developed by Daneman and Carpenter (1980) with the predictive power of other measures of working memory. The results of the meta-analysis support Daneman and Carpenter's (1980) claim that measures that tap the combined processing and storage capacity of working memory (e.g., reading span, listening span) are better predictors of comprehension than are measures that tap only the storage capacity (e.g., word span, digit span). The meta-analysis also showed that math process plus storage measures of working memory are good predictors of comprehension. Thus, the superior predictive power of the process plus storage measures is not limited to measures that involve the manipulation of words and sentences.},
}
@book{Daniel83,
       author = {Daniel, Lesley},
        title = {Planning and operations research},
    booktitle = {Artificial Intelligence: Tools, Techniques, and Applications},
    publisher = {Harper and Row},
      address = {New York},
         year = 1983,
}
@misc{Daniel_Sieta_DQN_Deep_Q-Network,
       author = {Daniel Sieta},
        title = {{Understanding Deep-Q-Networks}},
 howpublished = {https://danieltakeshi.github.io/2016/12/01/going-deeper-into-reinforcement-learning-understanding-dqn/},
         year = {2016}
}
@article{DanihelkaetalCoRR-16,
       author = {Ivo Danihelka and Greg Wayne and Benigno Uria and Nal Kalchbrenner and Alex Graves},
        title = {Associative Long Short-Term Memory},
      journal = {CoRR},
       volume = {arXiv:1602.03032},
         year = {2016},
          url = {http://arxiv.org/abs/1602.03032},
     abstract = {We investigate a new method to augment recurrent neural networks with extra memory without increasing the number of network parameters. The system has an associative memory based on complex-valued vectors and is closely related to Holographic Reduced Representations and Long Short-Term Memory networks. Holographic Reduced Representations have limited capacity: as they store more information, each retrieval becomes noisier due to interference. Our system in contrast creates redundant copies of stored information, which enables retrieval with reduced noise. Experiments demonstrate faster learning on multiple memorization tasks.}
}
@book{Dantzig1930number,
        title = {Number: The Language of Science},
       author = {Dantzig, Tobias},
         year = {1930},
    publisher = {Macmillan Company},
      address = {New York, USA},
}
@book{Dantzig63,
       author = {Dantzig, George},
        title = {Linear Programming and Extensions},
    publisher = {Princeton University Press},
     location = {Princeton, New Jersey},
         year = 1963,
}
@article{DantzigandWolfeOR-60,
       author = {Dantzig, George and Wolfe, Philip},
        title = {Decomposition Principle for Dynamic Programs},
      journal = {Operations Research},
       volume = 8,
        issue = 1,
        pages = {101-111},
         year = 1960,
}
@inproceedings{DarwicheandGoldszmidtUAI-94,
       author = {Darwiche, Adnan and Goldszmidt, Moises},
        title = {Action Networks: A Framework for Reasoning About Actions and Change Under Uncertainty},
    booktitle = {Tenth Conference on Uncertainty in Artificial Intelligence},
     location = {Seattle, Washington},
        pages = {136-144},
         year = 1994,
}
@inproceedings{DarwicheandPearlSSS-94,
       author = {Darwiche, Adnan and Pearl, Judea},
        title = {Symbolic Causal Networks for Reasoning About Actions and Plans},
    booktitle = {Proceedings of the {AAAI} Spring Symposium on Decision Theoretic Planning},
     location = {Stanford, California},
         year = 1994,
}
@book{Darwin1859,
       author = {Charles Darwin},
        title = {The Origin of Species by Means of Natural Selection, or the Preservation of Favoured Races in the Struggle for Life},
      comment = {http://www.zoo.uib.no/classics/origin.html},
    publisher = {John Murray},
      address = {London},
         year = 1859,
}
@book{Darwin1872,
       author = {Charles Darwin},
        title = {The expression of the emotions in man and animals},
    publisher = {John Murray},
      address = {London},
         year = 1872,
} 
@misc{DasPhD-12,
        title = {Semi-Supervised and Latent-Variable Models of Natural Language Semantics},
       author = {Dipanjan Das},
         year = 2012,
 howpublished = {Technical Report CMU-LTI-12-007, May 2012},
}
@article{DasenetalNATURE-03,
        title = {Motor neuron columnar fate imposed by sequential phases of {Hox-c} activity},
       author = {Dasen, J.S. and Liu, J. and Jessell, T.M.},
      journal = {Nature},
       volume = 425,
        issue = 6961,
        pages = {926-933},
         year = 2003,
     abstract = {The organization of neurons into columns is a prominent feature of central nervous system structure and function. In many regions of the central nervous system the grouping of neurons into columns links cell-body position to axonal trajectory, thus contributing to the establishment of topographic neural maps. This link is prominent in the developing spinal cord, where columnar sets of motor neurons innervate distinct targets in the periphery. We show here that sequential phases of Hox-c protein expression and activity control the columnar differentiation of spinal motor neurons. Hox expression in neural progenitors is established by graded fibroblast growth factor signalling and translated into a distinct motor neuron Hox pattern. Motor neuron columnar fate then emerges through cell autonomous repressor and activator functions of Hox proteins. Hox proteins also direct the expression of genes that establish motor topographic projections, thus implicating Hox proteins as critical determinants of spinal motor neuron identity and organization.}
}
@article{DasetalCL-13,
       author = {Das, Dipanjan and Chen, Desai and Martins, Andr\'{e} F. T. and Schneider, Nathan and Smith, Noah A.},
        title = {Frame-Semantic Parsing},
      journal = {Computational Linguistics},
       volume = 40,
        issue = 1,
         year = 2013,
        pages = {9-56},
}
@misc{DasetalCMU-10,
        title = {{SEMAFOR} 1.0: A Probabilistic Frame-Semantic Parser},
       author = {Dipanjan Das, Nathan Schneider, Desai Chen, and Noah A. Smith},
         year = 2010,
 howpublished = {Carnegie Mellon University Technical Report CMU-LTI-10-001},
}
@inproceedings{DasetalCVPR-17,
        title = {{V}isual {D}ialog},
       author = {Abhishek Das and Satwik Kottur and Khushi Gupta and Avi Singh and Deshraj Yadav and Jos\'e M.F. Moura and Devi Parikh and Dhruv Batra},
    booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
         year = {2017},
     abstract = {We introduce the task of Visual Dialog, which requires an AI agent to hold a meaningful dialog with humans in natural, conversational language about visual content. Specifically, given an image, a dialog history, and a question about the image, the agent has to ground the question in image, infer context from history, and answer the question accurately. Visual Dialog is disentangled enough from a specific downstream task so as to serve as a general test of machine intelligence, while being grounded in vision enough to allow objective evaluation of individual responses and benchmark progress. We develop a novel two-person chat data-collection protocol to curate a large-scale Visual Dialog dataset (VisDial). VisDial v0.9 has been released and contains 1 dialog with 10 question-answer pairs on \approx{}120k images from COCO, with a total of \approx{}1.2M dialog question-answer pairs. We introduce a family of neural encoder-decoder models for Visual Dialog with 3 encoders \emdash{} Late Fusion, Hierarchical Recurrent Encoder and Memory Network \emdash{} and 2 decoders (generative and discriminative), which outperform a number of sophisticated baselines. We propose a retrieval-based evaluation protocol for Visual Dialog where the AI agent is asked to sort a set of candidate answers and evaluated on metrics such as mean-reciprocal-rank of human response. We quantify gap between machine and human performance on the Visual Dialog task via human studies. Putting it all together, we demonstrate the first 'visual chatbot'! Our dataset, code, trained models and visual chatbot are available on this {\urlh{https://visualdialog.org/}{URL}}.},
}
@article{DasetalCoRR-17,
       author = {Abhishek Das and Satwik Kottur and Jos{\'{e}} M. F. Moura and Stefan Lee and Dhruv Batra},
        title = {Learning Cooperative Visual Dialog Agents with Deep Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1703.06585},
         year = {2017},
     abstract = {We introduce the first goal-driven training for visual question answering and dialog agents. Specifically, we pose a cooperative 'image guessing' game between two agents -- Qbot and Abot -- who communicate in natural language dialog so that Qbot can select an unseen image from a lineup of images. We use deep reinforcement learning (RL) to learn the policies of these agents end-to-end -- from pixels to multi-agent multi-round dialog to game reward. We demonstrate two experimental results.

First, as a 'sanity check' demonstration of pure RL (from scratch), we show results on a synthetic world, where the agents communicate in ungrounded vocabulary, i.e., symbols with no pre-specified meanings (X, Y, Z). We find that two bots invent their own communication protocol and start using certain symbols to ask/answer about certain visual attributes (shape/color/style). Thus, we demonstrate the emergence of grounded language and communication among 'visual' dialog agents with no human supervision.

Second, we conduct large-scale real-image experiments on the VisDial dataset, where we pretrain with supervised dialog data and show that the RL 'fine-tuned' agents significantly outperform SL agents. Interestingly, the RL Qbot learns to ask questions that Abot is good at, ultimately resulting in more informative dialog and a better team.}
}
@article{DasguptaetalBIORXIV-17,
       author = {Sanjoy Dasgupta and Charles F. Stevens and Saket Navlakha},
        title = {A neural algorithm for a fundamental computing problem},
      journal = {bioRxiv},
         year = {2016},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {Similarity search, such as identifying similar images in a database or similar documents on the Web, is a fundamental computing problem faced by many large-scale information retrieval systems. We discovered that the fly's olfactory circuit solves this problem using a novel variant of a traditional computer science algorithm (called locality-sensitive hashing). The fly's circuit assigns similar neural activity patterns to similar input stimuli (odors), so that behaviors learned from one odor can be applied when a similar odor is experienced. The fly's algorithm, however, uses three new computational ingredients that depart from traditional approaches. We show that these ingredients can be translated to improve the performance of similarity search compared to traditional algorithms when evaluated on several benchmark datasets. Overall, this perspective helps illuminate the logic supporting an important sensory function (olfaction), and it provides a conceptually new algorithm for solving a fundamental computational problem.}
}
@article{DasguptaetalSCIENCE-17,
       author = {Sanjoy Dasgupta and Charles F. Stevens and Saket Navlakha},
        title = {A neural algorithm for a fundamental computing problem},
      journal = {Science},
       volume = {358},
         year = {2017},
        pages = {793-796},
     abstract = {Similarity search, such as identifying similar images in a database or similar documents on the Web, is a fundamental computing problem faced by many large-scale information retrieval systems. We discovered that the fly's olfactory circuit solves this problem using a novel variant of a traditional computer science algorithm (called locality-sensitive hashing). The fly's circuit assigns similar neural activity patterns to similar input stimuli (odors), so that behaviors learned from one odor can be applied when a similar odor is experienced. The fly's algorithm, however, uses three new computational ingredients that depart from traditional approaches. We show that these ingredients can be translated to improve the performance of similarity search compared to traditional algorithms when evaluated on several benchmark datasets. Overall, this perspective helps illuminate the logic supporting an important sensory function (olfaction), and it provides a conceptually new algorithm for solving a fundamental computational problem.}
}
@inproceedings{DattaetalADAC-14,
       author = {Datta, Suman and Shukla, Nikhil and Cotter, Matthew and Parihar, Abhinav and Raychowdhury, Arijit},
        title = {Neuro Inspired Computing with Coupled Relaxation Oscillators},
    booktitle = {Proceedings of 51st Annual Design Automation Conference on Design Automation Conference},
    publisher = {ACM},
         year = 2014,
        pages = {74:1-74:6},
     abstract = {Computing with networks of synchronous oscillators has attracted wide-spread attention as novel materials and device topologies have enabled realization of compact, scalable and low-power coupled oscillatory systems. Of particular interest are compact and low-power relaxation oscillators that have been recently demonstrated using MIT (metal- insulator-transition) devices using properties of correlated oxides. This paper presents an analysis of the dynamics and synchronization of a system of two such identical coupled relaxation oscillators implemented with MIT devices. We focus on two implementations of the oscillator: (a) a D-D configuration where complementary MIT devices (D) are connected in series to provide oscillations and (b) a D-R configuration where it is composed of a resistor (R) in series with a voltage-triggered state changing MIT device (D). The MIT device acts like a hysteresis resistor with different resistances in the two different states. The synchronization dynamics of such a system has been analyzed with purely charge based coupling using a resistive (Rc) and a capacitive (Cc) element in parallel. It is shown that in a D-D configuration symmetric, identical and capacitively coupled relaxation oscillator system synchronizes to an anti-phase locking state, whereas when coupled resistively the system locks in phase. Further, we demonstrate that for certain range of values of Rc and Cc, a bistable system is possible which can have potential applications in associative computing. In D-R configuration, we demonstrate the existence of rich dynamics including non-monotonic flows and complex phase relationship governed by the ratios of the coupling impedance. Finally, the developed theoretical formulations have been shown to explain experimentally measured waveforms of such pairwise coupled relaxation oscillators.}
}
@incollection{DaugmanBMBT-93,
       author = {Daugman, John G.},
        title = {Brain Metaphor and Brain Theory},
    booktitle = {Computational Neuroscience},
       editor = {Schwartz, Eric L.},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
         year = {1993},
        pages = {9-18},
}
@article{DauphinetalCoRR-17,
       author = {Yann N. Dauphin and Angela Fan and Michael Auli and David Grangier},
        title = {Language Modeling with Gated Convolutional Networks},
      journal = {CoRR},
       volume = {arXiv:1612.08083},
         year = {2017},
     abstract = {The pre-dominant approach to language modeling to date is based on recurrent neural networks. Their success on this task is often linked to their ability to capture unbounded context. In this paper we develop a finite context approach through stacked convolutions, which can be more efficient since they allow parallelization over sequential tokens. We propose a novel simplified gating mechanism that outperforms Oord et al (2016) and investigate the impact of key architectural decisions. The proposed approach achieves state-of-the-art on the WikiText-103 benchmark, even though it features long-term dependencies, as well as competitive results on the Google Billion Words benchmark. Our model reduces the latency to score a sentence by an order of magnitude compared to a recurrent baseline. To our knowledge, this is the first time a non-recurrent approach is competitive with strong recurrent models on these large scale language tasks.},
} 
@article{DavalosetalNATURE-NEUROSCIENCE-05,
       author = {Davalos, D. and Grutzendler, J. and Yang, G. and Kim, J. V. and Zuo, Y. and Jung, S. and Littman, D. R. and Dustin, M. L. and Gan, W. B.},
        title = {{ATP} mediates rapid microglial response to local brain injury in vivo},
      journal = {Nature Neuroscience},
         year = {2005},
       volume = {8},
       number = {6},
        pages = {752-758},
     abstract = {Parenchymal microglia are the principal immune cells of the brain. Time-lapse two-photon imaging of GFP-labeled microglia demonstrates that the fine termini of microglial processes are highly dynamic in the intact mouse cortex. Upon traumatic brain injury, microglial processes rapidly and autonomously converge on the site of injury without cell body movement, establishing a potential barrier between the healthy and injured tissue. This rapid chemotactic response can be mimicked by local injection of ATP and can be inhibited by the ATP-hydrolyzing enzyme apyrase or by blockers of G protein-coupled purinergic receptors and connexin channels, which are highly expressed in astrocytes. The baseline motility of microglial processes is also reduced significantly in the presence of apyrase and connexin channel inhibitors. Thus, extracellular ATP regulates microglial branch dynamics in the intact brain, and its release from the damaged tissue and surrounding astrocytes mediates a rapid microglial response towards injury.}
}
@misc{David_Silver_Value_Function_Lecture,
       author = {David Silver},
        title = {{Introduction to Reinforcement Learning}},
 howpublished = {http://www0.cs.ucl.ac.uk/staff/D.Silver/web/Teaching.html},
         year = {2015}
}
@inproceedings{DavidsonandFehlingSSS-94,
       author = {Davidson, Ron and Fehling, Michael},
        title = {A Structured, Probabilistic Model of Action},
    booktitle = {Proceedings of the {AAAI} Spring Symposium on Decision Theoretic Planning},
     location = {Stanford, California},
         year = 1994,
}
@book{Davies2010,
       author = {Kevin Davies},
        title = {The \$1,000 Genome: The Revolution in DNA Sequencing and the New Era of Personalized Medicine},
    publisher = {Free Press},
      address = {New York, NY},
         year = 2010,
}
@book{Davis00,
       author = {Martin Davis},
        title = {The Universal Computer: The Road from Leibniz to Turing},
    publisher = {W. W. Norton},
      address = {New York, NY},
         year = 2000,
}
@book{Davis01,
       author = {Martin Davis},
        title = {Engines of Logic: Mathematicians and the Origin of the Computer},
    publisher = {W. W. Norton},
      address = {New York, NY},
         year = 2001,
}
@book{Davis2004,
       author = {Fred J. Davis},
        title = {Polymer Chemistry},
    publisher = {Oxford University Press},
         year = 2004,
      address = {Oxford, UK},
}
@incollection{Davis82,
       author = {Davis, Randall},
        title = {Teiresias: Applications of Meta-Level Knowledge},
       editor = {Davis, Randall and Lenat, Douglas B.},
    booktitle = {Knowledge-Based Systems in Artificial Intelligence},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1982,
        pages = {227-490},
}
@techreport{Davis85,
       author = {Davis, Ernest},
        title = {Constraint Propagation of Real-Valued Quantities},
        issue = 189,
  institution = {New York University Department of Computer Science},
         year = 1985,
}
@book{Davis90,
       author = {Davis, Ernest},
        title = {Representations of Commonsense Knowledge},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1990,
}
@inproceedings{DavisAAAI-88,
       author = {Davis, Ernest},
        title = {Inferring Ignorance from the Locality of Visual Perception},
    booktitle = {Proceedings AAAI-88},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
         year = 1988,
        pages = {786-790},
}
@article{DavisAIJ-87,
       author = {Davis, Ernest},
        title = {Constraint Propagation with Interval Labels},
      journal = AIJ,
       volume = 32,
         year = 1987,
        pages = {281-331},
}
@book{DavisPhD,
       author = {Davis, Ernest},
        title = {Representing and Acquiring Geographic Knowledge},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1986,
}
@book{DavisandMeindl2012,
        title = {Interconnect Technology and Design for Gigascale Integration},
       author = {Davis, J.A. and Meindl, J.D.},
         year = {2012},
    publisher = {Springer US}
}
@article{DavisandPutnamJACM-60,
       author = {Davis, M. and Putnam, H. A.},
        title = {A Computing Procedure for Quantification Theory},
      journal = {Journal of the {ACM}},
       volume = 7,
        pages = {201-215},
         year = 1960,
}
@article{DavisetalAIAA-91,
       author = {Davis, Thomas Jr. and Erzberger, H. and Green, S. M. and Nedell, W.},
        title = {Design and Evaluation of an Air Traffic Control Final Approach Spacing Tool},
      journal = {AIAA Journal of Guidance, Control, and Dynamics},
       volume = 14,
        pages = {848-854},
         year = 1991,
}
@article{DawandFrankCOGNITION-09,
       author = {Daw, N.D. and Frank, M.J.},
        title = {Reinforcement learning and higher level cognition: introduction to special issue},
      journal = {Cognition},
       volume = {113},
       number = {5},
         year = {2009},
        pages = {259-261},
}
@article{DawetalNATURE-NEUROSCIENCE-05,
       author = {Daw, Nathaniel D. and Niv, Yael and Dayan, Peter},
        title = {Uncertainty-based competition between prefrontal and dorsolateral striatal systems for behavioral control},
      journal = {Nature Neuroscience},
       volume = {8},
        issue = {12},
         year = {2005},
        pages = {1704-1711},
     abstract = {A broad range of neural and behavioral data suggests that the brain contains multiple systems for behavioral choice, including one associated with prefrontal cortex and another with dorsolateral striatum. However, such a surfeit of control raises an additional choice problem: how to arbitrate between the systems when they disagree. Here, we consider dual-action choice systems from a normative perspective, using the computational theory of reinforcement learning. We identify a key trade-off pitting computational simplicity against the flexible and statistically efficient use of experience. The trade-off is realized in a competition between the dorsolateral striatal and prefrontal systems. We suggest a Bayesian principle of arbitration between them according to uncertainty, so each controller is deployed when it should be most accurate. This provides a unifying account of a wealth of experimental evidence about the factors favoring dominance by either system.},
}
@book{Dawkins89,
       author = {Richard Dawkins},
        title = {The Selfish Gene},
    publisher = {Oxford University Press},
      address = {New York},
         year = 1989,
}
@article{DayanNC-93b,
        title = {Improving generalization for temporal difference learning: The successor representation},
       author = {Dayan, Peter},
      journal = {Neural Computation},
       volume = {5},
        issue = {4},
        pages = {613-624},
         year = {1993},
     abstract = {Estimation of returns over time, the focus of temporal difference (TD) algorithms, imposes particular constraints on good function approximators or representations. Appropriate generalisation between states is determined by how similar their successors are, and representations should follow suit. This paper shows how TD machinery can be used to learn such representations, and illustrates, using a navigation task, the appropriately distributed nature of the result.}
}
@book{DayanandAbbott01,
       author = {Dayan, Peter. and Abbott, Larry F.},
        title = {Theoretical Neuroscience},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2001,
}
@incollection{DayanandHintonNIPS-93,
       author = {Dayan, Peter and Hinton, Geoffrey E.},
        title = {Feudal Reinforcement Learning},
       editor = {Hanson, S. J. and Cowan, J. D. and Giles, C. L.},
    booktitle = {Advances in Neural Information Processing 5},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1993,
        pages = {271-278},
} 
@article{DayanandHintonNN-96,
        title = {Varieties of Helmholtz Machine},
       author = {Peter Dayan and Geoffrey E. Hinton},
      journal = {Neural Networks},
       volume = {9},
       number = {8},
        pages = {1385-1403},
         year = {1996},
     abstract = {The Helmholtz machine is a new unsupervised learning architecture that uses top-down connections to build probability density models of input and bottom-up connections to build inverses to those models. The wake-sleep learning algorithm for the machine involves just the purely local delta rule. This paper suggests a number of different varieties of Helmholtz machines, each with its own strengths and weaknesses, and relates them to cortical information processing.},
}
@article{DayanetalNC-95,
   author = {Dayan, P. and Hinton, G.E. and Neal, R.M. and Zemel, R.S.},
    title = {The {H}elmholtz Machine},
  journal = {Neural Computation},
   volume = 7,
    issue = 8,
      url = {http://www.cs.utoronto.ca/~hinton/absps/helmholtz.pdf},
    pages = {889-904},
     year = 1995,
     abstract = {Discovering the structure inherent in a set of patterns is a fundamental aim of statistical inference or learning. One fruitful approach is to build a parameterised stochastic generative model, independent draws from which are likely to produce the patterns. For all but the simplest generative models, each pattern can be generated in exponentially many ways. It is thus intractable to adjust the parameters to maximize the probability of the observed patterns, We describe a way of finessing this combinatorial explosion by maximising an easily computed lower bound on the probability of the observations. Our method can be viewed as a form of hierarchical self-supervised learning that may relate to the function of bottom-up and top-down cortical processing pathways.}
}
@article{DeAngelis93,
       author = {G.C. De{A}ngelis and I. Ohzawa and R.D. Freeman},
        title = {The spatiotemporal organization of simple cell receptive fields in the cat's striate cortex. {II}. {L}inearity of temporal and spatial summation},
      journal = {Journal of Neurophysiology},
         year = 1993,
       volume = 69,
        pages = {1118-1135},
}
@article{DeBondtandThalerTJF-84,
       author = {Werner F. M. De Bondt and Richard Thaler},
        title = {Does the Stock Market Overreact?},
      journal = {The Journal of Finance},
       volume = {40},
        issue = {3},
         year = {1984},
        pages = {793-805},
}
@book{DeBotton1997,
        title = {How Proust Can Change Your Life},
       author = {De Botton, Alain},
         year = {1997},
    publisher = {Picador}
}
@article{DeBrigardetalNEUROIMAGE-15,
       author = {De Brigard, Felipe and Nathan Spreng, R. and Mitchell, Jason P. and Schacter, Daniel L.},
        title = {Neural activity associated with self, other, and object-based counterfactual thinking},
      journal = {Neuroimage},
       volume = {109},
         year = {2015},
        pages = {12-26},
     abstract = {Previous research has shown that autobiographical episodic counterfactual thinking-i.e., mental simulations about alternative ways in which one's life experiences could have occurred-engages the brain's default network (DN). However, it remains unknown whether or not the DN is also engaged during impersonal counterfactual thoughts, specifically those involving other people or objects. The current study compares brain activity during counterfactual simulations involving the self, others and objects. In addition, counterfactual thoughts involving others were manipulated in terms of similarity and familiarity with the simulated characters. The results indicate greater involvement of DN during person-based (i.e., self and other) as opposed to object-based counterfactual simulations. However, the involvement of different regions of the DN during other-based counterfactual simulations was modulated by how close and/or similar the simulated character was perceived to be by the participant. Simulations involving unfamiliar characters preferentially recruited dorsomedial prefrontal cortex. Simulations involving unfamiliar similar characters, characters with whom participants identified personality traits, recruited lateral temporal gyrus. Finally, our results also revealed differential coupling of right hippocampus with lateral prefrontal and temporal cortex during counterfactual simulations involving familiar similar others, but with left transverse temporal gyrus and medial frontal and inferior temporal gyri during counterfactual simulations involving either oneself or unfamiliar dissimilar others. These results suggest that different brain mechanisms are involved in the simulation of personal and impersonal counterfactual thoughts, and that the extent to which regions associated with autobiographical memory are recruited during the simulation of counterfactuals involving others depends on the perceived similarity and familiarity with the simulated individuals.},
} 
@article{DeBruyckereetalFiMN-18,
       author = {De Bruyckere, Elodie and Simon, Ruth and Nestel, Sigrun and Heimrich, Bernd and K\"{a}tzel, Dennis and Egorov, Alexei V. and Liu, Pentao and Jenkins, Nancy A. and Copeland, Neal G. and Schwegler, Herbert and Draguhn, Andreas and Britsch, Stefan},
        title = {{Stability and Function of Hippocampal Mossy Fiber Synapses Depend on Bcl11b/Ctip2}},
      journal = {Frontiers in Molecular Neuroscience},
       volume = {11},
         year = {2018},
        pages = {103},
     abstract = {Structural and functional plasticity of synapses are critical neuronal mechanisms underlying learning and memory. While activity-dependent regulation of synaptic strength has been extensively studied, much less is known about the transcriptional control of synapse maintenance and plasticity. Hippocampal mossy fiber synapses connect dentate granule cells to CA3 pyramidal neurons and are important for spatial memory formation and -consolidation. The transcription factor Bcl11b/Ctip2 is expressed in dentate granule cells and required for postnatal hippocampal development. Ablation of Bcl11b/Ctip2 in the adult hippocampus results in impaired adult neurogenesis and spatial memory. The molecular mechanisms underlying the behavioral impairment remained unclear. Here we show that selective deletion of Bcl11b/Ctip2 in the adult mouse hippocampus leads to a rapid loss of excitatory synapses in CA3 as well as reduced ultrastructural complexity of remaining mossy fiber boutons. Moreover, a dramatic decline of LTP of the dentate gyrus-CA3 projection is caused by adult loss of Bcl11b/Ctip2. Differential transcriptomics revealed the deregulation of genes associated with synaptic transmission in mutants. Together, our data suggest Bcl11b/Ctip2 to regulate maintenance and function of mossy fiber synapses in the adult hippocampus.}
}
@article{DeFelipeetalJNC-02,
       author = {DeFelipe, Javier and Elston, Guy N. and Fujita, Ichiro and Fuster, Joaqu{\'i}n and Harrison, Kimberly H. and Hof, Patrick R. and Kawaguchi, Yasuo and Martin, Kevan A.C. and Rockland, Kathleen S. and Thomson, Alex M. and Wang, Samuel S.-H. and White, Edward L. and Yuste, Rafael},
        title = {Neocortical circuits: Evolutionary aspects and specificity versus non-specificity of synaptic connections. Remarks, main conclusions and general comments and discussion},
      journal = {Journal of Neurocytology},
       volume = {31},
       number = {3},
         year = {2002},
        pages = {387-416},
     abstract = {In this article the main conclusions, and some of what the guest editor (Javier DeFelipe) considers the most interesting remarks, have been extracted from each of the individual articles. These commentaries are not necessarily directly derived from the original work of the authors, and may be the result of the collective work of several different laboratories. As indicated in the Preface to this Special Issue, the aim of this summary is to provide the reader with a synopsis of current thoughts about cortical circuitry and function, by transcribing the authors' ideas in their own words. The article presents the remarks and the main conclusions of each of the individual articles, followed by a section dedicated to the general comments and discussion of the issues raised therein. The authors who have participated in this article are listed in alphabetical order.},
}
@book{DeGroot70,
       author = {DeGroot, M.},
        title = {Optimal Statistical Decisions},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1970,
}
@book{DeGroot86,
     author = {Morris H. DeGroot},
      title = {Probability and Statistics},
  publisher = {Second edition, Addison-Wesley},
    address = {Reading, MA},
       year = 1986,
}
@article{DeLaPenaetalJIM-09,
     title = {Artificial exosomes as tools for basic and clinical immunology},
    author = {De La Pe\~{n}a, H. and Madrigal, J.A. and Rusakiewicz, S. and Bencsik, M. and Cave, G.W.V. and Selman, A. and Rees, R.C. and Travers, P.J. and Dodi, I.A.},
   journal = {Journal of Immunological Methods},
    volume = 344,
     issue = 2,
     pages = {121-32},
      year = 2009,
  abstract = {Dendritic cell derived exosomes are able to mediate and modulate immune responses in vivo by semi-direct T cell activation. T cells can eradicate primary, metastatic, relapsed tumours and ameliorate otherwise fatal viral infections. Not surprisingly activation and expansion of T cells has become one of the main focuses for immunotherapy. Using nanotechnology, we have developed targeted and traceable in vivo artificial exosomes by coating liposomes (FDA approved) with an optimized number of MHC Class I/peptide complexes and a selected specific range of ligands for adhesion, early activation, late activation and survival T cell receptors. These targeted artificial exosomes are traceable both in vitro and in vivo via fluorescent and Magnetic Resonance Imaging and facilitate imaging of specific areas by applying localised nuclear magnetic interactions of hydrogens via super paramagnetic labels. Here we show that artificial exosomes activate and expand functional antigen specific T cells at sufficient levels. This novel system has potential basic and clinical applications in immunology where the study of membrane interactions is desired.}
}
@article{DeMarseetalAR-01,
       author = {Thomas B. DeMarse and Daniel A. Wagenaar and Axel W. Blau and Steve M. Potter},
        title = {The Neurally Controlled Animat: Biological Brains Acting with Simulated Bodies},
      journal = {Autonomous Robots},
       volume = 11,
        issue = 3,
        pages = {1573-7527},
         year = 2001,
      comment = {The brain is perhaps the most advanced and robust computation system known. We are creating a method to study how information is processed and encoded in living cultured neuronal networks by interfacing them to a computer-generated animal, the Neurally-Controlled Animat, within a virtual world. Cortical neurons from rats are dissociated and cultured on a surface containing a grid of electrodes (multi-electrode arrays, or MEAs) capable of both recording and stimulating neural activity. Distributed patterns of neural activity are used to control the behavior of the Animat in a simulated environment. The computer acts as its sensory system providing electrical feedback to the network about the Animat's movement within its environment. Changes in the Animat's behavior due to interaction with its surroundings are studied in concert with the biological processes (e.g., neural plasticity) that produced those changes, to understand how information is processed and encoded within a living neural network. Thus, we have created a hybrid real-time processing engine and control system that consists of living, electronic, and simulated components. Eventually this approach may be applied to controlling robotic devices, or lead to better real-time silicon-based information processing and control algorithms that are fault tolerant and can repair themselves.}
}
@inproceedings{DeMenthon02,
        title = {Spatio-temporal Segmentation of Video by Hierarchical Mean-shift Analysis},
       author = {Daniel DeMenthon},
    booktitle = {Proceedings of Statistical Methods in Video Processing Workshop},
     location = {Copenhagen, Denmark},
         year = 2002,
}
@article{DeSchutteretalTiN-05,
       author = {Erik De Schutter and \"{O}rjan Ekeberg and Jeanette Hellgren Kotaleski and Pablo Achard and Anders Lansner},
        title = {Biophysically detailed modelling of microcircuits and beyond},
      journal = {Trends in Neurosciences},
       volume = {28},
       number = {10},
        pages = {562-569},
         year = {2005},
     abstract = {Realistic bottom-up modelling has been seminal to understanding which properties of microcircuits control their dynamic behaviour, such as the locomotor rhythms generated by central pattern generators. In this article of the \{TINS\} Microcircuits Special Feature, we review recent modelling work on the leech-heartbeat and lamprey-swimming pattern generators as examples. Top-down mathematical modelling also has an important role in analyzing microcircuit properties but it has not always been easy to reconcile results from the two modelling approaches. Most realistic microcircuit models are relatively simple and need to be made more detailed to represent complex processes more accurately. We review methods to add neuromechanical feedback, biochemical pathways or full dendritic morphologies to microcircuit models. Finally, we consider the advantages and challenges of full-scale simulation of networks of microcircuits.}
}
@article{DeSousaWebberCoRR-15,
       author = {Francisco De Sousa Webber},
        title = {Semantic Folding Theory And its Application in Semantic Fingerprinting},
      journal = {CoRR},
       volume = {arXiv:1511.08855},
         year = {2015},
     abstract = {Human language is recognized as a very complex domain since decades. No computer system has been able to reach human levels of performance so far. The only known computational system capable of proper language processing is the human brain. While we gather more and more data about the brain, its fundamental computational processes still remain obscure. The lack of a sound computational brain theory also prevents the fundamental understanding of Natural Language Processing. As always when science lacks a theoretical foundation, statistical modeling is applied to accommodate as many sampled real-world data as possible. An unsolved fundamental issue is the actual representation of language (data) within the brain, denoted as the Representational Problem. Starting with Jeff Hawkins' Hierarchical Temporal Memory (HTM) theory, a consistent computational theory of the human cortex, we have developed a corresponding theory of language data representation: The Semantic Folding Theory. The process of encoding words, by using a topographic semantic space as distributional reference frame into a sparse binary representational vector is called Semantic Folding and is the central topic of this document. Semantic Folding describes a method of converting language from its symbolic representation (text) into an explicit, semantically grounded representation that can be generically processed by Hawkins' HTM networks. As it turned out, this change in representation, by itself, can solve many complex NLP problems by applying Boolean operators and a generic similarity function like the Euclidian Distance. Many practical problems of statistical NLP systems, like the high cost of computation, the fundamental incongruity of precision and recall , the complex tuning procedures etc., can be elegantly overcome by applying Semantic Folding.}
}
@article {DeWindBIORXIV-19,
       author = {DeWind, Nicholas K.},
        title = {The number sense is an emergent property of a deep convolutional neural network trained for object recognition},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Laboratory},
         year = {2019},
     abstract = {Humans and many non-human animals have the 'number sense,' an ability to estimate the number of items in a set without counting. This innate sense of number is hypothesized to provide a foundation for more complex numerical and mathematical concepts. Here I investigated whether we also share the number sense with a deep convolutional neural network (DCNN) trained for object recognition. These in silico networks have revolutionized machine learning over the last seven years, allowing computers to reach human-level performance on object recognition tasks for the first time. Their architecture is based on the structure of mammalian visual cortex, and after they are trained, they provide a highly predictive model of responses in primate visual cortex, suggesting deep homologies. I found that the DCNN demonstrates three key hallmarks of the number sense: numerosity-selective units (analogous to biological neurons), the behavioral ratio effect, and ordinality over representational space. Because the DCNN was not trained to enumerate, I conclude that the number sense is an emergent property of the network, the result of some combination of the network architecture and the constraint to develop the complex representational structure necessary for object recognition. By analogy I conclude that the number sense in animals was not necessarily the result of direct selective pressure to enumerate but might have 'come for free' with the evolution of a complex visual system that evolved to identify objects and scenes in the real world.},
} 
@article{DeWindetalCOGNITION-15,
        title = {Modeling the approximate number system to quantify the contribution of visual stimulus features},
       author = {Nicholas K. DeWind and Geoffrey K. Adams and Michael L. Platt and Elizabeth M. Brannon},
      journal = {Cognition},
       volume = {142},
         year = {2015},
        pages = {247-265},
     abstract = {The approximate number system (ANS) subserves estimation of the number of items in a set. Typically, ANS function is assessed by requiring participants to compare the number of dots in two arrays. Accuracy is determined by the numerical ratio of the sets being compared, and each participant's Weber fraction (w) provides a quantitative index of ANS acuity. When making numerical comparisons, however, performance is also influenced by non-numerical features of the stimuli, such as the size and spacing of dots. Current models of numerosity comparison do not account for these effects and consequently lead to different estimates of w depending on the methods used to control for non-numerical features. Here we proffer a new model that teases apart the effects of ANS acuity from the effects of non-numerical stimulus features. The result is an estimate of w that is a more theoretically valid representation of numerical acuity and novel terms that denote the degree to which a participant's perception of number is affected by non-numerical features. We tested this model in a sample of 20 adults and found that, by correctly attributing errors due to non-numerical stimulus features, the w obtained was more reliable across different stimulus conditions. We found that although non-numerical features biased numerosity discriminations in all participants, number was the primary feature driving discriminations in most of them. Our findings support the idea that, while numerosity is a distinct visual quantity, the internal representation of number is tightly bound to the representation of other magnitudes. This tool for identifying the different effects of the numerical and non-numerical features of a stimulus has important implications not only for the behavioral investigation of the ANS, but also for the collection and analyses of neural data sets associated with ANS function.}
}
@article{DeWindetalCORTEX-19,
       author = {Nicholas K. DeWind and Joonkoo Park and Marty G. Woldorff and Elizabeth M. Brannon},
        title = {Numerical encoding in early visual cortex},
      journal = {Cortex},
       volume = {114},
         year = {2019},
        pages = {76-89},
     abstract = {The ability to estimate numerosity in a visual array arose early in evolution, develops early in human development, and is correlated with mathematical ability. Previous work with visually presented arrays indicates that the intraparietal sulcus (IPS) represents number. However, it is not clear if the number signal originates in IPS or is propagated from earlier visual areas. Previous work from our group has demonstrated a rapidly instantiated representation of number in low-level regions of visual cortex using the high temporal resolution of event-related electro-encephalography (EEG). Here, we use a rapid event-related functional magnetic resonance imaging (fMRI) paradigm and find convergent evidence for a number signal in low-level visual cortex (areas V1, V2, and V3). Employing a stringent set of stimulus controls, we demonstrate that this signal cannot be explained by the total extent of the array, the density of the items in the array, the aggregate visual area of the items, the size of individual items, the proportion of the array covered by items, nor the overall scale of the array and items. Our findings thus provide strong support for the hypothesis that number is rapidly and directly encoded early in the visual processing stream.}
}
@article{DeWitetalJoN-10,
       author = {de-Wit, Lee and Machilsen, Bart and Putzeys, Tom},
        title = {Predictive coding and the neural response to predictable stimuli},
      journal = {The Journal of neuroscience: the official journal of the Society for Neuroscience},
    publisher = {Society for Neuroscience},
         year = {2010},
       volume = {30},
        issue = {26},
        pages = {8702-8703},
     abstract = {An influential theoretical and computational framework, closely related to efficient coding, is predictive coding. This framework provides a novel means of understanding the interactions between distinct brain areas. Predictive coding posits that the brain actively predicts upcoming sensory input rather than passively registering it. Predictive coding is efficient in the sense that the brain does not need to maintain multiple versions of the same information at different levels of the processing hierarchy. Indeed, activation in early sensory areas no longer represents sensory information per se, but only that part of the input that has not been successfully predicted by higher level areas. The activity in lower level areas can therefore be considered an "error signal" that updates the predictions at higher areas and guides learning.}
}
@book{Deacon1998symbolic,
        title = {The Symbolic Species: The Co-evolution of Language and the Brain},
       author = {Deacon, Terrence W.},
         year = {1998},
    publisher = {W. W. Norton}
}
@book{Deacon2012incomplete,
        title = {Incomplete Nature: How Mind Emerged from Matter},
       author = {Deacon, Terrence W.},
         year = {2012},
    publisher = {W. W. Norton}
}
@misc{Deacon@exonerated,
        title = {Investigation Exonerates Terry Deacon},
       author = {Robert M. Price},
 howpublished = {https://terrydeacon.berkeley.edu/plagiarism-investigation-exonerates-terrence-w-deacon},
         year = {2007},
}
@article{DeahaeneetalTiCS-17,
       author = {Dehaene, Stanislas and Changeux, Jean-Pierre and Naccache, Lionel and Sackur, J\`{e}r\^{o}me and Sergent, Claire},
        title = {Conscious, preconscious, and subliminal processing: a testable taxonomy},
      journal = {Trends in Cognitive Sciences},
         year = {2017},
    publisher = {Elsevier},
       volume = {10},
        issue = {5},
        pages = {204-211},
     abstract = {Of the many brain events evoked by a visual stimulus, which are specifically associated with conscious perception, and which merely reflect non-conscious processing? Several recent neuroimaging studies have contrasted conscious and non-conscious visual processing, but their results appear inconsistent. Some support a correlation of conscious perception with early occipital events, others with late parieto-frontal activity. Here we attempt to make sense of these dissenting results. On the basis of the global neuronal workspace hypothesis, we propose a taxonomy that distinguishes between vigilance and access to conscious report, as well as between subliminal, preconscious and conscious processing. We suggest that these distinctions map onto different neural mechanisms, and that conscious perception is systematically associated with surges of parieto-frontal activity causing top-down amplification.},
}
@incollection{DeakerBCN-14,
       author = {De\'{a}k, M. O.},
       editor = {P. Brooks \& V. Kampe},
        title = {Interrelations of language and cognitive development},
    booktitle = {The International Encyclopedia of the Social \& Behavioral Sciences},
    publisher = {SAGE},
         year = {2014},
        pages = {284-291},
}
@book{Dean04,
       author = {Dean, Thomas},
        title = {Talking With Computers},
    publisher = {Cambridge University Press},
      address = {New York},
         year = 2005,
} 
@book{Dean2017inverted,
        title = {The Inverted Matrix: A Vision of Hope and High Aspiration},
       author = {Thomas Dean},
         year = {2017},
          url = {https://web.stanford.edu/class/cs379c/resources/inverted/paper.pdf},
    publisher = {Under Review}
}
@book{Dean99,
       editor = {Dean, Thomas},
        title = {Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence, {IJCAI}-99, {S}tockholm, {S}weden, {J}uly 31 -- {A}ugust 6, 1999, 2 {V}olumes},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
         year = 1999,
}
@inproceedings{DeanAAAI-05,
      author = {Dean, Thomas},
       title = {A Computational Model of the Cerebral Cortex},
   booktitle = {Proceedings of AAAI-05},
   publisher = {MIT Press},
     address = {Cambridge, Massachusetts},
    location = {Pittsburgh, Pennsylvania},
       pages = {938-943},
        year = 2005,
}
@inproceedings{DeanAIMATH-06,
       author = {Dean, Thomas},
        title = {Scalable Inference in Hierarchical Generative Models},
    booktitle = {Proceedings of the Ninth International Symposium on Artificial Intelligence and Mathematics},
         year = 2006,
}
@article{DeanAMAI-06,
       author = {Dean, Thomas},
        title = {Learning Invariant Features Using Inertial Priors},
      journal = {Annals of Mathematics and Artificial Intelligence},
       volume = 47,
        issue = {3-4},
      comment = {http://springerlink.metapress.com/content/q42w12xr0323/?p=93c0422fadce4af783e5e9f7dbebc415&pi=0},
        pages = {223-250},
         year = 2006,
}
@article{DeanAOR-88,
       author = {Dean, Thomas},
        title = {An Approach to Reasoning About the Effects of Actions for Automated Planning Systems},
      journal = {Annals of Operations Research},
       volume = 12,
        issue = {1-4},
        pages = {147-167},
         year = 1988,
}
@techreport{DeanCS-05-13,
       author = {Thomas Dean},
        title = {Hierarchical Expectation Refinement for Learning Generative Perception Models},
        issue = {CS-05-13},
  institution = {Brown University Department of Computer Science},
         year = 2005,
}
@inproceedings{DeanCSCSI-84,
       author = {Dean, Thomas},
        title = {Managing Time Maps},
    booktitle = {Proceedings of the Canadian Society for Computational Studies of Intelligence},
 organization = {CSCSI},
         year = 1984,
}
@article{DeanCSUR-96,
       author = {Dean, Thomas},
        title = {Integrating Theory and Practice in Planning},
      journal = {{ACM} Computing Surveys},
       volume = 28,
        issue = 4,
        pages = 2,
         year = 1996,
}
@article{DeanCoRR-17,
        title = {Inferring Mesoscale Models of Neural Computation},
       author = {Thomas Dean},
      journal = {CoRR},
       volume = {arXiv:1710.05183},
         year = {2017},
          url = {http://arxiv.org/abs/1710.05183},
      comment = {password: caf93},
     abstract = {Recent years have seen dramatic progress in the development of techniques for measuring the activity and connectivity of large populations of neurons in the brain. However, as these techniques grow ever more powerful---allowing us to even contemplate measuring every neuron in entire brain---a new problem arises: how do we make sense of the mountains of data that these techniques produce? Here, we argue that the time is ripe for building an intermediate or "mesoscale" computational theory that can bridge between single-cell (microscale) accounts of neural function and behavioral (macroscale) accounts of animal cognition and environmental complexity. Just as digital accounts of computation in conventional computers abstract away the non-essential dynamics of the analog circuits that implementing gates and registers, so too a computational account of animal cognition can afford to abstract from the non-essential dynamics of neurons. We argue that the geometry of neural circuits is essential in explaining the computational limitations and technological innovations inherent in biological information processing. We propose a blueprint for how to employ tools from modern machine learning to automatically infer a satisfying mesoscale account of neural computation that combines functional and structural data, with an emphasis on learning and exploiting regularity and repeating motifs in neuronal circuits. Rather than suggest a specific theory, we present a new class of scientific instruments that can enable neuroscientists to design, propose, implement and test mesoscale theories of neural computation.}
}
@misc{DeanCoRR-18,
       author = {Thomas Dean},
        title = {{The Inverted Matrix: A Vision of Hope and High Aspiration}},
       volume = {arXiv:1912.00421},
         year = {2018},
}
@inproceedings{DeanDARPA-87a,
       author = {Dean, Thomas},
        title = {Planning Paradigms},
    booktitle = {Proceedings of the {DARPA} Knowledge-Based Planning Workshop (also appeared in the Proceedings of the 3rd Annual Workshop of Israel Association for Artificial Intelligence, and in the {AI} Magazine, Fall 1988)},
     location = {Austin, Texas},
 organization = {DARPA},
         year = 1988,
}
@inproceedings{DeanDARPA-87b,
       author = {Dean, Thomas},
        title = {Planning, Execution, and Control},
    booktitle = {Proceedings of the {DARPA} Knowledge-Based Planning Workshop},
     location = {Austin, Texas},
 organization = {DARPA},
        pages = {29:1-10},
         year = 1987,
}
@misc{DeanDIACRITICAL-14,
        title = {Interaction and Negotiation in Learning and Understanding Dialog},
       author = {Thomas Dean},
         year = {2014},
 howpublished = {{\tt{https://web.stanford.edu/class/cs379c/resources/dialogical/zanax_DOC.dir/index.html}}},
     abstract = {Interaction and negotiation are an essential component of natural language understanding in conversation. We argue this is particularly the case in building artificial agents that rely primarily on language to interact with humans. Rather than thinking about misunderstanding-thinking the user said one thing when he said another-and non-understanding-not having a clue what the user was talking about-as a problem to be overcome, it makes more sense to think of such events as opportunities to learn something and a natural part of understanding that becomes essential when the agent trying to understand has a limited language understanding capability. Moreover, many of the same strategies that are effective in situations in which the agent's limited language facility fails also apply to the agent actively engaging the user in an unobtrusive manner to collect data and ground truth in order to extend its repertoire of services that it can render and to improve its existing language understanding capabilities. In the case of developing an agent to engage users in conversations about music, actively soliciting information from users about their music interests and resolving misunderstandings on both sides about what services the application can offer and what service in particular the user wants now is already a natural part of the conversation. Data collected from thousands or millions of users would provide an invaluable resource for training NLP components that could be used to build more sophisticated conversational agents.}
}
@article{DeanDSS-86,
       author = {Dean, Thomas},
        title = {Handling Shared Resources in a Temporal Database Management System},
      journal = {Decision Support Systems},
       volume = {2},
        issue = {2},
        pages = {135-143},
         year = 1986,
}
@inproceedings{DeanIEEE-84,
       author = {Dean, Thomas},
        title = {Planning and Temporal Reasoning under Uncertainty},
    booktitle = {Proceedings of the {IEEE} Workshop on Principles of Knowledge-Based Systems},
 organization = {IEEE},
         year = 1984,
        pages = {131-138},
}
@inproceedings{DeanIJCAI-85,
       author = {Dean, Thomas},
        title = {Temporal Reasoning Involving Counterfactuals and Disjunctions},
    booktitle = {Proceedings {IJCAI}-85},
     location = {Los Angeles, California},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
 organization = {IJCAI},
         year = 1985,
        pages = {860-866},
}
@inproceedings{DeanIJCAI-87,
       author = {Dean, Thomas},
        title = {Large-Scale Temporal Databases for Planning in Complex Domains},
    booktitle = {Proceedings {IJCAI}-87},
     location = {Milan, Italy},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
 organization = {IJCAI},
        pages = {860-866},
         year = 1987,
}
@article{DeanIJIS-91,
       author = {Dean, Thomas},
        title = {Decision-Theoretic Control of Inference for Time-Critical Applications},
      journal = {International Journal Intelligent Systems},
       volume = 6,
        issue = 4,
        pages = {417-441},
         year = 1991,
}
@article{DeanJACM-89,
       author = {Dean, Thomas},
        title = {Using Temporal Hierarchies to Efficiently Maintain Large Temporal Databases},
      journal = {Journal of the {ACM}},
       volume = 36,
        issue = 4,
        pages = {687-718},
         year = 1989,
}     
@misc{DeanMISCELLANEOUS-18,
       author = {Thomas Dean},
        title = {{The Inverted Matrix: A Vision of Hope and High Aspiration}},
 howpublished = {{https://web.stanford.edu/class/cs379c/resources/inverted/paper.pdf}},
         year = {2018},
     abstract = {How might humans, cyborgs and artificial intelligence work together to create a shared vision for the future and what could we do now to make this possible. Some experts believe it is inevitable that humans will one day share this planet with biologically enhanced humans, cyborgs and artificially intelligent robots. We've been told the age of AI might lead to massive unemployment, advanced weapon systems, pervasive surveillance and law enforcement, and even the rise of artificially intelligent overlords. These bleak visions of the future may be possible but they need not be inevitable. The technology of artificial intelligence could be applied to realize a future in which intelligent systems of many sorts create a global framework for planet-wide collective decision making, share in the governance and stewardship of our planet, amicably accommodate their differences and resolve disputes, and ultimately forge a path in which technology and shared purpose replace natural selection as the engines of human evolution.

We discuss some of most relevant advances in cognitive and systems neuroscience, AI and machine learning research leading to advances in technology that could transform society and significantly improve the lot of human beings by unshackling us from instincts evolved millions of years ago to enable us to survive a kill-or-be-killed jungle but are now counterproductive to our happiness and wellbeing. Soon we will have technology that will enable physically unaltered humans to substantially extend their cognitive capabilities, and eventually physically augment their bodies to achieve even more efficient interfaces with machines and expand their capabilities still further. Looking beyond the next decade, unaltered humans, augmented (cybernetically enhanced) humans and nonbiological AI systems will populate earth and it is (theoretically) within our capacity to ensure that such heterogeneous populations co-habit this planet amicably, learn from one another and govern themselves so as to allow all to prosper and realize their potential. This optimistic future is not certain, but will have a greater chance of success if companies begin planning and producing products and services that create an environment in which relevant innovation will thrive, or, better yet, create consortia to independently pursue and freely disseminate the relevant enabling technology.

The truth is that without Human AI as partners, proxies and prosthetic thinking appliances, we - members of Homo Sapiens 1.0 cannot scale our innate social-bonding and decision-making capacity to solve the problems facing us today. Perhaps that seems overly pessimistic, but human history, psychology and neurobiology provide little evidence to discourage such a dour outlook. Already, it seems that we are ratcheting up our xenophobic and (parochially speaking) misanthropic distrust of strangers to vilify robots and AI systems. Shackled AI systems are like stockpiled nuclear weapons just waiting to be detonated, stolen, or used as threats to further the ends of dictators, despots and terrorists, and once we can deploy other AI systems to pick the locks and hack the boot ROMs we might as well just hand over the keys. We are warmongering the citizenry, stirring up the troops and rattling the sabers even as the enemy is still waddling around in diapers. It is time that we are honest with ourselves, admit our weaknesses, and, as we have done so many times before, invent technology to overcome our shortcomings and control our destiny.}
}
@article{DeanNETWORK-NEUROSCIENCE-17,
       author = {Dean, Thomas},
        title = {Inferring Mesoscale Models of Neural Computation},
      journal = {Network Neuroscience},
         year = {2017},
} 
@inproceedings{DeanOIS-86,
       author = {Dean, Thomas},
        title = {Decision Support for Coordinated Multi-Agent Planning},
    booktitle = {Proceedings of the 3rd International {ACM} Conference on Office Information Systems},
         year = 1986,
}
@article{DeanOMNI-92,
       author = {Dean, Thomas},
        title = {Anticipating Tomorrow's Technology Needs},
      journal = {Omni Magazine},
        month = {November},
         year = {1992},
}
@inproceedings{DeanPLANNING-WORKSHOP-87,
       author = {Dean, Thomas},
        title = {Experiments in Planning and Control},
    booktitle = {Proceedings of the AAAI-87 Workshop on Planning for Autonomous Mobile Robots},
     location = {Seattle, Washington},
         year = 1987,
}
@incollection{DeanPW-86,
       author = {Dean, Thomas},
        title = {Intractability and Time-Dependent Planning},
       editor = {Georgeff, Michael P. and Lansky, Amy L.},
    booktitle = {Reasoning About Actions and Plans},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
        pages = {245-266},
         year = 1986,
}
@book{DeanPhD,
       author = {Dean, Thomas},
        title = {Temporal Imagery: {A}n Approach to Reasoning about Time for Planning and Problem Solving ({P}h.{D}.\ {T}hesis)},
    publisher = {Bell and Howell Information and Learning},
      comment = {Also published as Yale Computer Science technical report: Yale/CSD/RR #433, October, 1985},
      address = {Skokie, Illinois},
         year = 1985,
}
@article{DeanSIGART-96,
       author = {Dean, Thomas},
        title = {Context-Dependent Computational Components},
      journal = {{SIGART} Bulletin},
       volume = 7,
        issue = 2,
         year = 1996,
        pages = {3-6},
}

@string{AIJ   = "Artificial Intelligence Journal"}
@string{BROWN = "Brown University Department of Computer Science"}
@string{CIJ   = "Computational Intelligence Journal"}
@string{CVGIP = "Computer Vision Graphics and Image Processing"}
@string{CVIU  = "Computer Vision and Image Understanding"}
@string{CVPR  = "Computer Vision and Pattern Recognition"} 
@string{ECCV  = "European Conference on Computer Vision"}
@string{FOCS  = "Symposium on Foundations of Computer Science"}
@string{ICANN = "International Conference on Artificial Neural Networks"}
@string{ICCV  = "International Conference on Computer Vision"}  
@string{IJCAI = "International Joint Conference on Artificial Intelligence"}
@string{IJCV  = "International Journal on Computer Vision"}
@string{IJCNN = "International Joint Conference on Neural Networks"}
@string{IWANN = "International Workshop on Artificial and Neural Networks"}
@string{JAIR  = "Journal of Artificial Intelligence Research"}
@string{JRSS  = "Journal of the Royal Statistical Society"}
@string{NIPS  = "Neural Information Processing Society"}
@string{PAMI  = "IEEE Transactions on Pattern Analysis and Machine Intelligence"}
@string{SIGART = "SIGART Newsletter"}
@string{STANFORD = "Stanford University Department of Computer Science"}
@string{STOC  = "ACM Symposium on Theoretical Computing"}
@string{UAI-87 = "Proceedings of the 1987 Workshop on Uncertainty in Artificial Intelligence"}
@string{UAI-88 = "Proceedings of the 1988 Workshop on Uncertainty in Artificial Intelligence"}
@string{UAI-89 = "Proceedings of the 1989 Workshop on Uncertainty in Artificial Intelligence"}
@string{YALE  = "Yale University Department of Computer Science"}

@inproceedings{DeanSIMULATION-WORKSHOP-87,
       author = {Dean, Thomas},
        title = {Benchmarks for Research in Planning},
    booktitle = {Proceedings of the 2nd Workshop on {AI} and Simulation},
     location = {Seattle, Washington},
         year = 1987,
}
@inproceedings{DeanSPIE-87,
       author = {Dean, Thomas},
        title = {High-Level Planning and Low-Level Control},
    booktitle = {Proceedings of the {SPIE} Conference on Advances in Intelligent Robotic Systems},
 organization = {SPIE},
     location = {Cambridge, Massachusetts},
         year = 1987,
        pages = {496-501},
}
@inproceedings{DeanSPIE-88,
       author = {Dean, Thomas},
        title = {On the Complexity of Integrating Spatial Measurements},
    booktitle = {Proceedings of the {SPIE} Conference on Advances in Intelligent Robotic Systems},
 organization = {SPIE},
     location = {Cambridge, Massachusetts},
         year = 1988,
}
@inproceedings{DeanTIME-94,
       author = {Dean, Thomas},
        title = {Meditations on Time and Space: {E}xpediting Temporal Reasoning by Exploiting Time and Space ({A}bstract)},
       editor = {S.D. Goodwin and H.J. Hamilton},
    booktitle = {Proceedings of the First International Workshop on Temporal Representation and Reasoning ({TIME}-94)},
     location = {Pensacala, Florida},
        pages = {3-12},
         year = 1994,
}
@inproceedings{DeanandBoddyAAAI-87,
       author = {Dean, Thomas and Mark Boddy},
        title = {Incremental Causal Reasoning},
    booktitle = {Proceedings {AAAI}-87},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
     location = {Seattle, Washington},
 organization = {AAAI},
        pages = {196-201},
         year = 1987,
}
@inproceedings{DeanandBoddyAAAI-88,
       author = {Dean, Thomas and Boddy, Mark},
        title = {An Analysis of Time-Dependent Planning},
    booktitle = {Proceedings {AAAI}-88},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
        pages = {49-54},
         year = 1988,
}
@article{DeanandBoddyAIJ-88,
       author = {Dean, Thomas and Boddy, Mark},
        title = {Reasoning About Partially Ordered Events ({A}lso appears in "{R}eadings in {Q}ualitative {R}easoning {A}bout {P}hysical {S}ystems" ({M}organ {K}aufmann), edited by {D}an {W}eld and {J}ohan {D}e {K}leer)},
      journal = AIJ,
       volume = 36,
        issue = 3,
        pages = {375-399},
         year = 1988,
}
@article{DeanandBonassoAIM-93,
       author = {Dean, Thomas and Bonasso, R. Peter},
        title = {1992 {AAAI} Robot Exhibition and Competition},
      journal = {{AI} Magazine},
       volume = 14,
        issue = 1,
        pages = {35-48},
         year = 1993,
}
@article{DeanandCoxNETWORK-NEUROSCIENCE-17,
       author = {Dean, Thomas and Cox, David},
        title = {Inferring Mesoscale Models of Neural Computation},
      journal = {Network Neuroscience},
       volume = {SUBMITTED},
          url = {https://web.stanford.edu/class/cs379c/resources/networks/mesoscale.pdf},
         year = {2017},
     abstract = {Recent years have seen dramatic progress in the development of techniques for measuring the activity and connectivity of large populations of neurons in the brain.  However, as these techniques grow ever more powerful---allowing us to even contemplate measuring every neuron in entire brains---a new problem arises: how do we make sense of the mountains of data that these techniques produce? Here, we argue that the time is ripe for building an intermediate or mesoscale computational theory that can bridge between single-cell (microscale) accounts of neural function and behavioral (macroscale) accounts of animal cognition and environmental complexity. Just as digital accounts of computation in conventional computers abstract away the non-essential dynamics of the analog circuits that implementing gates and registers, so too a computational account of animal cognition can afford to abstract from the non-essential dynamics of neurons. We argue that the geometry of neural circuits is essential in explaining the computational limitations and technological innovations inherent in biological information processing. We propose a blueprint for how to employ tools from modern machine learning to automatically infer a satisfying mesoscale account of neural computation that combines functional and structural data, with an emphasis on learning and exploiting regularity and repeating motifs in neuronal circuits. Rather than suggest a specific theory, we present a new class of scientific instruments that can enable neuroscientists to propose, implement and test mesoscale theories of neural computation.},
}
@inproceedings{DeanandGivanAAAI-97,
       author = {Dean, Thomas and Givan, Robert},
        title = {Model Minimization in {M}arkov Decision Processes},
    booktitle = {Proceedings {AAAI}-97},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
     location = {Providence, Rhode Island},
 organization = {AAAI},
        pages = {106-111},
         year = 1997,
}
@inproceedings{DeanandGreenwaldAIPS-94,
       author = {Greenwald, Lloyd and Dean, Thomas},
        title = {Solving Time-Critical Decision-Making Problems with Predictable Computational Demands},
    booktitle = {Proceedings of the 2nd International Conference on {AI} Planning Systems ({ICAPS}-94)},
    publisher = {{AAAI} Press},
     location = {Chicago, Illinois},
       editor = {Kristian Hammond},
        pages = {25-30},
         year = 1994,
}
@inproceedings{DeanandGreenwaldARPI-94,
       author = {Greenwald, Lloyd and Dean, Thomas},
        title = {{M}onte {C}arlo Simulation and Bottleneck-Centered Heuristics for Time-critical Scheduling in Stochastic Domains},
    booktitle = {{ARPI} Planning Initiative Workshop},
         year = 1994,
}
@inproceedings{DeanandGreenwaldWAFR-94,
       author = {Greenwald, Lloyd and Dean, Thomas},
        title = {Anticipating Computational Demands when Solving Time-Critical Decision-Making Problems},
    booktitle = {The Algorithmic Foundations of Robotics},
    publisher = {A. K. Peters, Boston, Massachusetts},
       editor = {K. Goldberg and D. Halperin and J.C. Latombe and R. Wilson},
         year = 1995,
}
@inproceedings{DeanandKanazawaAAAI-88,
       author = {Dean, Thomas and Kanazawa, Keiji},
        title = {Probabilistic Temporal Reasoning},
    booktitle = {Proceedings {AAAI}-88},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
        pages = {524-528},
         year = 1988,
}
@article{DeanandKanazawaCIJ-89,
       author = {Dean, Thomas and Kanazawa, Keiji},
        title = {A Model for Reasoning About Persistence and Causation},
      journal = {Computational Intelligence Journal},
       volume = 5,
        issue = 3,
        pages = {142-150},
         year = 1989,
}
@inproceedings{DeanandKanazawaCSCSI-88,
       author = {Dean, Thomas and Kanazawa, Keiji},
        title = {Probabilistic Causal Reasoning},
    booktitle = {Proceedings of the Canadian Society for Computational Studies of Intelligence},
 organization = {CSCSI},
        pages = {125-132},
         year = 1988,
}
@article{DeanandKanazawaIEEE-89,
       author = {Dean, Thomas and Kanazawa, Keiji},
        title = {Persistence and Probabilistic Inference},
      journal = {{IEEE} Transactions on Systems, Man, and Cybernetics},
       volume = 19,
        issue = 3,
        pages = {574-585},
         year = 1989,
} 
@inproceedings{DeanandKirmanIUW-92,
       author = {Dean, Thomas and Kirman, Jak},
        title = {Representation Issues in {B}ayesian Decision Theory for Planning and Active Perception},
    booktitle = {Proceedings of the DARPA Image Understanding Workshop},
     location = {San Diego, California},
 organization = {DARPA},
        pages = {763-768},
         year = 1992,
}
@inproceedings{DeanandLinIJCAI-95,
       author = {Dean, Thomas and Lin, Shieu-Hong},
        title = {Decomposition Techniques for Planning in Stochastic Domains},
    booktitle = {Proceedings {IJCAI}-95},
     location = {Montreal, Canada},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
 organization = {IJCAI},
        pages = {1121-1127},
         year = 1995,
     abstract = {This paper is concerned with modeling planning problems involving uncertainty as discrete-time, finite-state stochastic automata. Solving planning problems is reduced to computing policies for Markov decision processes. Classical methods for solving Markov decision processes cannot cope with the size of the state spaces for typical problems encountered in practice. As an alternative, we investigate methods that decompose global planning problems into a number of local problems, solve the local problems separately and then combine the local solutions to generate a global solution. We present algorithms that decompose planning problems into smaller problems given an arbitrary partition of the state space. The local problems are interpreted as Markov decision processes and solutions to the local problems are interpreted as policies restricted to the subsets of the state space defined by the partition. One algorithm relies on constructing and solving an abstract version of the original decision problem. A second algorithm iteratively approximates parameters of the local problems to converge to an optimal solution. We show how properties of a specified partition affect the time and storage required for these algorithms.}
}
@incollection{DeanandMarion-97,
       author = {Dean, Thomas and Marion, Jean-Luc},
        title = {Planning and Navigation in Stochastic Environments},
    booktitle = {Visual Navigation: From Biological Systems to Unmanned Ground Vehicles},
       editor = {Yiannis Aloimonos},
      address = {Mahwah, New Jersey},
    publisher = {Lawrence Erlbaum Associates, Publishers},
         year = 1997,
}
@article{DeanandMcDermottAIJ-87,
       author = {Dean, Thomas and McDermott, Drew V.},
        title = {Temporal Data Base Management ({A}lso appears in "{R}eadings in {P}lanning" ({M}organ {K}aufmann), edited by {J}ames {A}llen, {J}ames {H}endler, and {A}ustin {T}ate)},
      journal = AIJ,
       volume = 32,
        issue = 1,
        pages = {1-55},
         year = 1987,
}
@book{DeanandMcKeown91,
       editor = {Dean, Thomas and McKeown, Kathy},
        title = {Proceedings of the Ninth National Conference on Artificial Intelligence, {AAAI}-91, {A}naheim, {C}alifornia, {J}uly 14-19, 1991, 2 {V}olumes},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1991,
}
@inproceedings{DeanandSiegleAAAI-90,
       author = {Dean, Thomas and Siegle, Greg},
        title = {An Approach to Reasoning About Continuous Change for Applications in Planning},
    booktitle = {Proceedings {AAAI}-90},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
     location = {Boston, Massachusetts},
 organization = {AAAI},
        pages = {132-137},
         year = 1990,
}
@book{DeanandWellman91,
       author = {Dean, Thomas and Wellman, Michael},
        title = {Planning and Control},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
         year = 1991,
}
@inproceedings{DeanandWellmanRPW-89,
       author = {Dean, Thomas and Wellman, Michael},
        title = {On the Value of Goals},
       editor = {Tenenberg, Josh and Weber, Jay and Allen, James},
    booktitle = {Proceedings from the Rochester Planning Workshop: From Formal Systems to Practical Systems},
         year = 1989,
        pages = {129-140},
}
@incollection{Deanetal92a,
       author = {Dean, Thomas and Basye, Ken and Shewchuk, John},
        title = {Reinforcement Learning for Planning and Control},
    booktitle = {Machine Learning Methods for Planning and Scheduling},
       editor = {Minton, Steve},
         year = 1992,
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
}
@incollection{Deanetal92b,
       author = {Dean, Thomas and Basye, Ken and Kaelbling, Leslie},
        title = {Uncertainty in Graph-based Map Learning},
    booktitle = {Robot Learning},
       editor = {Connell, Jonathon and Mahadevan, Sridhar},
    publisher = {Kluwer},
         year = 1992,
}
@book{Deanetal95,
       author = {Dean, Thomas and Allen, James and Aloimonos, Yiannis},
        title = {{Artificial Intelligence: Theory and Practice}},
    publisher = {Addison-Wesley Publishing Company},
      address = {Reading, Massachusetts},
         year = 1995,
}
@inproceedings{DeanetalAAAI-07,
       author = {Dean, Thomas and Glenn Carroll and Rich Washington},
        title = {On the Prospects for Building a Working Model of the Visual Cortex},
    booktitle = {Proceedings of AAAI-07},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
     location = {Vancouver, British Columbia},
        pages = {1597-1600},
         year = 2007,
}
@inproceedings{DeanetalAAAI-12,
       author = {Thomas Dean and Greg S.\ Corrado and Jonathon Shlens},
        title = {Three Controversial Hypotheses Concerning Computation in the Primate Cortex},
    booktitle = {Proceedings of the Twenty-Sixth {AAAI} Conference on Artificial Intelligence},
         year = 2012,
          url = {http://research.google.com/pubs/pub41100.html}
}
@inproceedings{DeanetalAAAI-90,
       author = {Dean, Thomas and Basye, Kenneth and Chekaluk, Robert and Hyun, Seungseok and Lejter, Moises and Randazza, Margaret},
        title = {Coping with Uncertainty in a Control System for Navigation and Exploration},
    booktitle = {Proceedings AAAI-90},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
     location = {Boston, Massachusetts},
 organization = {AAAI},
        pages = {1010-1015},
         year = 1990,
}
@inproceedings{DeanetalAAAI-92,
       author = {Dean, Thomas and Angluin, Dana and Basye, Kenneth and Engelson, Sean and Kaelbling, Leslie and Kokkevis, Evangelos and Maron, Oded},
        title = {Inferring Finite Automata with Stochastic Output Functions and an Application to Map Learning},
    booktitle = {Proceedings {AAAI}-92},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
     location = {San Jose, California},
 organization = {AAAI},
        pages = {208-214},
         year = 1992,
}
@inproceedings{DeanetalAAAI-93,
       author = {Dean, Thomas and Kaelbling, Leslie and Kirman, Jak and Nicholson, Ann},
        title = {Planning With Deadlines in Stochastic Domains},
    booktitle = {Proceedings {AAAI}-93},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
     location = {Washington, DC},
 organization = {AAAI},
        pages = {574-579},
         year = 1993,
}
@article{DeanetalAIJ-95,
       author = {Dean, Thomas and Kaelbling, Leslie and Kirman, Jak and Nicholson, Ann},
        title = {Planning Under Time Constraints in Stochastic Domains},
      journal = AIJ,
       volume = 76,
        issue = {1-2},
        pages = {35-74},
         year = 1995,
}
@incollection{DeanetalAIPS-92,
       author = {T. Dean and J. Kirman and K. Kanazawa},
        title = {Probabilistic Network Representations of Continuous-Time Stochastic Processes for Applications in Planning and Control},
    booktitle = {Proceedings of the First International Conference on Artificial Intelligence Planning Systems ({ICAPS}-92)},
       editor = {James Hendler},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
        pages = {273-274},
         year = 1992,
}
@inproceedings{DeanetalAIPS-98,
       author = {Dean, Thomas and Givan, Robert and Kim, Kee-Eung},
        title = {Solving Planning Problems with Large State and Action Spaces},
    booktitle = {Proceedings of the 4th International Conference on Artificial Intelligence Planning Systems ({ICAPS}-98)},
     location = {Pittsburgh, Pennsylvania},
        pages = {102-110},
         year = 1998,
     abstract = {Planning methods for deterministic planning problems traditionally exploit factored representations to encode the dynamics of problems in terms of a set of parameters, e.g., the location of a robot or the status of a piece of equipment. Factored representations achieve economy of representation by taking advantage of structure in the form of dependency relationships among these parameters. In recent work, we have addressed the problem of achieving the same economy of representation and exploiting the resulting encoding of structure for stochastic planning problems represented as Markov decision processes. In this paper, we extend our earlier work on reasoning about such factored representations to handle problems with large action spaces that are also represented in factored form, where the parameters in this case might correspond to the control parameters for different effectors on a robot or the allocations for a set of resources. The techniques described in this paper employ factored representations for Markov decision processes to identify and exploit regularities in the dynamics to expedite inference. These regularities are in the form of sets of states (described for example by boolean formulas) that behave the same with respect to sets of actions where these sets are thought of as aggregate states and aggregate actions respectively. We present theoretical foundations, describe algorithms, provide examples in which our techniques provide leverage and examples in which they fail to do so, and summarize the results of experiments with a preliminary implementation}
}
@article{DeanetalARXIV-13,
        title = {On the Technology Prospects and Investment Opportunities for Scalable Neuroscience},
       author = {Thomas Dean and Biafra Ahanonu and Mainak Chowdhury and Anjali Datta and Andre Esteva and Daniel Eth and Nobie Redmon and Oleg Rumyantsev and Ysis Tarter},
      journal = {CoRR},
       volume = {arXiv:1307.7302},
         year = 2013,
          url = {http://arxiv.org/abs/1307.7302},
}
@article{DeanetalCIJ-88,
       author = {Dean, Thomas and R. James Firby and David P. Miller},
        title = {Hierarchical Planning Involving Deadlines, Travel Time and Resources ({A}lso appears in "{R}eadings in {P}lanning" ({M}organ {K}aufmann), edited by {J}ames {A}llen, {J}ames {H}endler, and {A}ustin {T}ate, and in "{A}utonomous {M}obile {R}obots: {C}ontrol, {P}lanning, and {A}rchitecture" ({IEEE} {C}omputer {S}ociety {P}ress), edited by {S}. {S}. {I}yengar and {A}lberto {E}lfes)},
      journal = CIJ,
       volume = 4,
        issue = 4,
        pages = {381-398},
         year = 1988,
}
@inproceedings{DeanetalCVPR-13,
       author = {Thomas Dean and Mark A. Ruzon and Mark Segal and Jonathon Shlens and Sudheendra Vijayanarasimhan and Jay Yagnik},
        title = {Fast, Accurate Detection of 100,000 Object Classes on a Single Machine},
    booktitle = {2013 IEEE Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 2013,
        pages = {1814-1821},
     abstract = {Many object detection systems are constrained by the time required to convolve a target image with a bank of filters that code for different aspects of an object's appearance, such as the presence of component parts. We exploit locality-sensitive hashing to replace the dot-product kernel operator in the convolution with a fixed number of hash-table probes that effectively sample all of the filter responses in time independent of the size of the filter bank. To show the effectiveness of the technique, we apply it to evaluate 100,000 deformable-part models requiring over a million (part) filters on multiple scales of a target image in less than 20 seconds using a single multi-core processor with 20GB of RAM. This represents a speed-up of approximately 20,000 times---four orders of magnitude---when compared with performing the convolutions explicitly on the same hardware. While mean average precision over the full set of 100,000 object classes is around 0.16 due in large part to the challenges in gathering training data and collecting ground truth for so many classes, we achieve a mAP of at least 0.20 on a third of the classes and 0.30 or better on about 20\% of the classes.},
}
@article{DeanetalCoRR-13,
        title = {On the Technology Prospects and Investment Opportunities for Scalable Neuroscience},
       author = {Thomas Dean and Biafra Ahanonu and Mainak Chowdhury and Anjali Datta and Andre Esteva and Daniel Eth and Nobie Redmon and Oleg Rumyantsev and Ysis Tarter},
      journal = {CoRR},
       volume = {arXiv:1307.7302},
         year = 2013,
          url = {http://arxiv.org/abs/1307.7302},
}
@inproceedings{DeanetalCoRR-13b,
       author = {Dean, Thomas and Givan, Robert and Leach, Sonia},
        title = {Model Reduction Techniques for Computing Approximately Optimal Solutions for {M}arkov Decision Processes},
    booktitle = {Proceedings of the 13th Conference on Uncertainty in Artificial Intelligence},
       editor = {Geiger, Dan and Shenoy, Prakesh Pundalik},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
 organization = {AUAI},
     location = {Providence, Rhode Island},
          url = {https://arxiv.org/abs/1302.1533},
         year = 1997,
        pages = {124-131},
     abstract = {We present a method for solving implicit (factored) Markov decision processes (MDPs) with very large state spaces. We introduce a property of state space partitions which we call epsilon-homogeneity. Intuitively, an epsilon-homogeneous partition groups together states that behave approximately the same under all or some subset of policies. Borrowing from recent work on model minimization in computer-aided software verification, we present an algorithm that takes a factored representation of an MDP and an 0<=epsilon<=1 and computes a factored epsilon-homogeneous partition of the state space. This partition defines a family of related MDPs - those MDPs with state space equal to the blocks of the partition, and transition probabilities "approximately" like those of any (original MDP) state in the source block. To formally study such families of MDPs, we introduce the new notion of a "bounded parameter MDP" (BMDP), which is a family of (traditional) MDPs defined by specifying upper and lower bounds on the transition probabilities and rewards. We describe algorithms that operate on BMDPs to find policies that are approximately optimal with respect to the original MDP. In combination, our method for reducing a large implicit MDP to a possibly much smaller BMDP using an epsilon-homogeneous partition, and our methods for selecting actions in BMDPs constitute a new approach for analyzing large implicit MDPs. Among its advantages, this new approach provides insight into existing algorithms to solving implicit MDPs, provides useful connections to work in automata theory and model minimization, and suggests methods, which involve varying epsilon, to trade time and space (specifically in terms of the size of the corresponding state space) for solution quality.},
}
@article{DeanetalCoRR-18,
        title = {{Amanuensis: The Programmer's Apprentice}},
       author = {Thomas Dean and Maurice Chiang and Marcus Gomez and Nate Gruver and Yousef Hindy and Michelle Lam and Peter Lu and Sophia Sanchez and Rohun Saxena and Michael Smith and Lucy Wang and Catherine Wong},
      journal = {CoRR},
       volume = {arXiv:1807.00082},
         year = 2018,
          url = {http://arxiv.org/abs/1807.00082},
}
@article{DeanetalCoRR-19,
        title = {Biological Blueprints for Next Generation {AI} Systems},
       author = {Thomas Dean and Chaofei Fan and Francis E. Lewis and Megumi Sano},
      journal = {CoRR},
       volume = {arXiv:1912.00421},
         year = 2019,
          url = {http://arxiv.org/abs/1912.00421},
	  abstract = {Diverse subfields of neuroscience have enriched artificial intelligence for many decades. With recent advances in machine learning and artificial neural networks, many neuroscientists are partnering with AI researchers and machine learning experts to analyze data and construct models. This paper attempts to demonstrate the value of such collaborations by providing examples of how insights derived from neuroscience research are helping to develop new ma- chine learning algorithms and artificial neural network architectures. We survey the relevant neuroscience necessary to appreciate these insights and then describe how we can translate our current understanding of the relevant neurobiology into algorithmic techniques and architectural designs. Finally, we characterize some of the major challenges facing current AI technology and suggest avenues for overcoming these challenges that draw upon research in developmental and comparative cognitive neuroscience.}
}
@inproceedings{DeanetalDARPA-90,
       author = {Dean, Thomas and Basye, Kenneth and Lejter, Moises},
        title = {Planning and Active Perception},
    booktitle = {Proceedings of the DARPA Workshop on Innovative Approaches to Planning, Scheduling, and Control},
 organization = {DARPA},
        pages = {271-276},
         year = 1990,
}
@inproceedings{DeanetalIEEE-09,
        title = {Recursive Sparse Spatiotemporal Coding},
       author = {Thomas Dean and Greg Corrado and Richard Washington},
    booktitle = {Proceedings of the Fifth IEEE International Workshop on Multimedia Information Processing and Retrieval},
         year = {2009},
        pages = {657-661},
}
@inproceedings{DeanetalIEEE-90,
       author = {Dean, Thomas and Kanazawa, Keiji and Shewchuk, John},
        title = {Prediction, Observation, and Estimation in Planning and Control},
    booktitle = {Proceedings of the 5th {IEEE} International Symposium on Intelligent Control},
     location = {Philadelphia, Pennsylvania},
        pages = {645-650},
         year = 1990,
}
@inproceedings{DeanetalIUW-90,
       author = {Dean, Thomas and Camus, Theodore and Kirman, Jak},
        title = {Sequential Decision Making for Active Perception},
    booktitle = {Proceedings of the DARPA Image Understanding Workshop},
     location = {Pittsburgh, Pennsylvania},
 organization = {DARPA},
        pages = {889-894},
         year = 1990,
}
@inproceedings{DeanetalMIPR-09,
       author = {Thomas Dean and Greg Corrado and Rich Washington},
        title = {Recursive Sparse, Spatiotemporal Coding},
    booktitle = {Proceedings of the Fifth IEEE International Workshop on Multimedia Information Processing and Retrieval},
     location = {San Diego, California},
        month = {December},
         year = 2009,
}
@article{DeanetalML-95,
       author = {Dean, Thomas and Angluin, Dana and Basye, Kenneth and Engelson, Sean and Kaelbling, Leslie and Kokkevis, Evangelos and Maron, Oded},
        title = {Inferring Finite Automata with Stochastic Output Functions and an Application to Map Learning},
      journal = {Machine Learning},
       volume = 18,
        issue = 1,
        pages = {81-108},
         year = 1995,
}
@incollection{DeanetalNIPS-12,
        title = {Large Scale Distributed Deep Networks},
       author = {Jeffrey Dean and Greg S. Corrado and Rajat Monga and Kai Chen and Matthieu Devin and Quoc V. Le and Mark Z. Mao and Marc'Aurelio Ranzato and Andrew Senior and Paul Tucker and Ke Yang and and Andrew Y. Ng},
    booktitle = {Advances in Neural Information Processing Systems 25},
       editor = {F. Pereira and C.J.C. Burges and L. Bottou and K.Q. Weinberger},
    publisher = {Curran Associates, Inc.},
         year = 2012,
        pages = {1223-1231},
     abstract = {Recent work in unsupervised feature learning and deep learning has shown that being able to train large models can dramatically improve performance. In this paper, we consider the problem of training a deep network with billions of parameters using tens of thousands of CPU cores. We have developed a software framework called DistBelief that can utilize computing clusters with thousands of machines to train large models. Within this framework, we have developed two algorithms for large-scale distributed training: (i) Downpour SGD, an asynchronous stochastic gradient descent procedure supporting a large number of model replicas, and (ii) Sandblaster, a framework that supports a variety of distributed batch optimization procedures, including a distributed implementation of L-BFGS. Downpour SGD and Sandblaster L-BFGS both increase the scale and speed of deep network training. We have successfully used our system to train a deep network 30x larger than previously reported in the literature, and achieves state-of-the-art performance on ImageNet, a visual object recognition task with 16 million images and 21k categories. We show that these same techniques dramatically accelerate the training of a more modestly-sized deep network for a commercial speech recognition service. Although we focus on and report performance of these methods as applied to training large neural networks, the underlying algorithms are applicable to any gradient-based machine learning algorithm.},
}
@inproceedings{DeanetalSOSDI-04,
       author = {Dean, J. and Ghemawat, S.},
        title = {{M}ap{R}educe: Simplified Data Processing on Large Clusters},
    booktitle = {Proceedings of the 6th Symposium on Operating Systems Design and Implementation},
     location = {San Francisco, CA},
        pages = {137-150},
         year = 2004,
}
@techreport{DeanetalTECHREPORT-13,
        title = {On the Technology Prospects and Investment Opportunities for Scalable Neuroscience},
       author = {Thomas Dean and Biafra Ahanonu and Mainak Chowdhury and Anjali Datta and Andre Esteva and Daniel Eth and Nobie Redmon and Oleg Rumyantsev and Ysis Tarter},
  institution = {Google},
        issue = 41324,
         year = 2013,
          url = {http://research.google.com/pubs/pub41324.html},
}
@techreport{DeanetalTR-09,
       author = {Thomas Dean and Greg Corrado and Rich Washington},
        title = {Recursive Sparse, Spatiotemporal Coding},
        issue = {Technical Report CS-09-08},
  institution = {Brown University Department of Computer Science},
         year = 2009,
}
@inproceedings{DeanetalUAI-93,
       author = {Dean, Thomas and Kaelbling, Leslie and Kirman, Jak and Nicholson, Ann},
        title = {Deliberation Scheduling for Time-Critical Sequential Decision Making},
    booktitle = {Proceedings of the 9th Conference on Uncertainty in Artificial Intelligence},
 organization = {AUAI},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
     location = {Washington, DC},
        pages = {309-316},
         year = 1993,
}
@inproceedings{DeanetalUAI-97,
       author = {Dean, Thomas and Givan, Robert and Leach, Sonia},
        title = {Model Reduction Techniques for Computing Approximately Optimal Solutions for {M}arkov Decision Processes},
    booktitle = {Proceedings of the 13th Conference on Uncertainty in Artificial Intelligence},
       editor = {Geiger, Dan and Shenoy, Prakesh Pundalik},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
 organization = {AUAI},
     location = {Providence, Rhode Island},
         year = 1997,
        pages = {124-131},
     abstract = {We present a method for solving implicit (factored) Markov decision processes (MDPs) with very large state spaces. We introduce a property of state space partitions which we call epsilon-homogeneity. Intuitively, an epsilon-homogeneous partition groups together states that behave approximately the same under all or some subset of policies. Borrowing from recent work on model minimization in computer-aided software verification, we present an algorithm that takes a factored representation of an MDP and an 0<=epsilon<=1 and computes a factored epsilon-homogeneous partition of the state space. This partition defines a family of related MDPs - those MDPs with state space equal to the blocks of the partition, and transition probabilities "approximately" like those of any (original MDP) state in the source block. To formally study such families of MDPs, we introduce the new notion of a "bounded parameter MDP" (BMDP), which is a family of (traditional) MDPs defined by specifying upper and lower bounds on the transition probabilities and rewards. We describe algorithms that operate on BMDPs to find policies that are approximately optimal with respect to the original MDP. In combination, our method for reducing a large implicit MDP to a possibly much smaller BMDP using an epsilon-homogeneous partition, and our methods for selecting actions in BMDPs constitute a new approach for analyzing large implicit MDPs. Among its advantages, this new approach provides insight into existing algorithms to solving implicit MDPs, provides useful connections to work in automata theory and model minimization, and suggests methods, which involve varying epsilon, to trade time and space (specifically in terms of the size of the corresponding state space) for solution quality.},
}
@incollection{Debreu54,
       author = {Debreu, Gerard},
        title = {Representation of a preference ordering by a numerical function},
    booktitle = {Decision Processes},
       editor = {Thrall, R. M. and Coombs, C. H. and Davis, R. L.},
         year = 1954,
    publisher = {John Wiley and Sons},
      address = {New York},
}
@article{DechterandPearlAIJ-87,
       author = {Dechter, R. and Pearl, J.},
        title = {Network-based Heuristics for Constraint-Satisfaction Problems},
      journal = AIJ,
       volume = 34,
        issue = 1,
         year = 1987,
        pages = {1-38},
}
@article{DeenetalNATURE-COMMUNICATIONS-17,
       author = {Deen, B. and Richardson, H. and Dilks, D. D. and Takahashi, A. and Keil, B. and Wald, L. L. and Kanwisher, N. and Saxe, R.},
        title = {Organization of high-level visual cortex in human infants},
      journal = {Nature Communications},
         year = {2017},
       volume = {8},
        pages = {13995},
     abstract = {How much of the structure of the human mind and brain is already specified at birth, and how much arises from experience? In this article, we consider the test case of extrastriate visual cortex, where a highly systematic functional organization is present in virtually every normal adult, including regions preferring behaviourally significant stimulus categories, such as faces, bodies, and scenes. Novel methods were developed to scan awake infants with fMRI, while they viewed multiple categories of visual stimuli. Here we report that the visual cortex of 4-6-month-old infants contains regions that respond preferentially to abstract categories (faces and scenes), with a spatial organization similar to adults. However, precise response profiles and patterns of activity across multiple visual categories differ between infants and adults. These results demonstrate that the large-scale organization of category preferences in visual cortex is adult-like within a few months after birth, but is subsequently refined through development.}
}
@article{DefferardetalNIPS-16,
       author = {Micha{\"{e}}l Defferrard and Xavier Bresson and Pierre Vandergheynst},
        title = {Convolutional Neural Networks on Graphs with Fast Localized Spectral Filtering},
      journal = {CoRR},
       volume = {arXiv:1606.09375},
         year = {2016},
}
@book{DegrootandSchervish02,
       author = {Morris H. DeGroot and Mark J. Schervish},
        title = {Probability and Statistics},
    publisher = {Third edition, Addison-Wesley},
      address = {Reading, MA},
         year = 2002,
}
@article{Dehaene-LambertzandSpelkeNEURON-15,
       author = {G. Dehaene-Lambertz and E.S. Spelke},
        title = {The Infancy of the Human Brain},
      journal = {Neuron},
       volume = {88},
       number = {1},
        pages = {93-109},
         year = {2015},
     abstract = {The human infant brain is the only known machine able to master a natural language and develop explicit, symbolic, and communicable systems of knowledge that deliver rich representations of the external world. With the emergence of noninvasive brain imaging, we now have access to the unique neural machinery underlying these early accomplishments. After describing early cognitive capacities in the domains of language and number, we review recent findings that underline the strong continuity between human infants' and adults' neural architecture, with notably early hemispheric asymmetries and involvement of frontal areas. Studies of the strengths and limitations of early learning, and of brain dynamics in relation to regional maturational stages, promise to yield a better understanding of the sources of human cognitive achievements.},
}
@book{Dehaene1999,
        title = {The Number Sense: How the Mind Creates Mathematics},
       author = {Stanislas Dehaene},
    publisher = {Oxford University Press},
         year = 1999,
}
@book{Dehaene2009,
        title = {Reading in the Brain: The Science and Evolution of a Human Invention},
       author = {Stanislas Dehaene},
    publisher = {Viking Press},
         year = 2009,
}
@book{Dehaene2014,
        title = {Consciousness and the Brain: Deciphering How the Brain Codes Our Thoughts},
       author = {Stanislas Dehaene},
    publisher = {Viking Press},
         year = 2014,
}
@article{DehaeneChangeuxNEURON-11,
       author = {Dehaene, Stanislas and Changeux, Jean-Pierre},
        title = {Experimental and Theoretical Approaches to Conscious Processing},
      journal = {Neuron},
         year = {2017},
    publisher = {Elsevier},
       volume = {70},
        issue = {2},
        pages = {200-227},
     abstract = {Recent experimental studies and theoretical models have begun to address the challenge of establishing a causal link between subjective conscious experience and measurable neuronal activity. The present review focuses on the well-delimited issue of how an external or internal piece of information goes beyond nonconscious processing and gains access to conscious processing, a transition characterized by the existence of a reportable subjective experience. Converging neuroimaging and neurophysiological data, acquired during minimal experimental contrasts between conscious and nonconscious processing, point to objective neural measures of conscious access: late amplification of relevant sensory activity, long-distance cortico-cortical synchronization at beta and gamma frequencies, and "ignition" of a large-scale prefronto-parietal network. We compare these findings to current theoretical models of conscious processing, including the Global Neuronal Workspace (GNW) model according to which conscious access occurs when incoming information is made globally available to multiple brain systems through a network of neurons with long-range axons densely distributed in prefrontal, parieto-temporal, and cingulate cortices. The clinical implications of these results for general anesthesia, coma, vegetative state, and schizophrenia are discussed.},
}
@book{DehaeneandBrannon2011mathminds,
        title = {Space, Time and Number in the Brain: Searching for the Foundations of Mathematical Thought},
       author = {Dehaene, Stanislas and Brannon, Elizabeth},
         year = {2011},
    publisher = {Elsevier Science}
}
@article{DehaeneandChangeuxPBR-00,
       author = {Dehaene, S. and Changeux, J. P.},
        title = {Reward-dependent learning in neuronal networks for planning and decision making},
      journal = {Progress in Brain Research},
         year = {2000},
       volume = {126},
        pages = {217-229},
     abstract = {Neuronal network models have been proposed for the organization of evaluation and decision processes in prefrontal circuitry and their putative neuronal and molecular bases. The models all include an implementation and simulation of an elementary reward mechanism. Their central hypothesis is that tentative rules of behavior, which are coded by clusters of active neurons in prefrontal cortex, are selected or rejected based on an evaluation by this reward signal, which may be conveyed, for instance, by the mesencephalic dopaminergic neurons with which the prefrontal cortex is densely interconnected. At the molecular level, the reward signal is postulated to be a neurotransmitter such as dopamine, which exerts a global modulatory action on prefrontal synaptic efficacies, either via volume transmission or via targeted synaptic triads. Negative reinforcement has the effect of destabilizing the currently active rule-coding clusters; subsequently, spontaneous activity varies again from one cluster to another, giving the organism the chance to discover and learn a new rule. Thus, reward signals function as effective selection signals that either maintain or suppress currently active prefrontal representations as a function of their current adequacy. Simulations of this variation-selection have successfully accounted for the main features of several major tasks that depend on prefrontal cortex integrity, such as the delayed-response test, the Wisconsin card sorting test, the Tower of London test and the Stroop test. For the more complex tasks, we have found it necessary to supplement the external reward input with a second mechanism that supplies an internal reward; it consists of an auto-evaluation loop which short-circuits the reward input from the exterior. This allows for an internal evaluation of covert motor intentions without actualizing them as behaviors, by simply testing them covertly by comparison with memorized former experiences. This element of architecture gives access to enhanced rates of learning via an elementary process of internal or covert mental simulation. We have recently applied these ideas to a new model, developed with M. Kerszberg, which hypothesizes that prefrontal cortex and its reward-related connections contribute crucially to conscious effortful tasks. This model distinguishes two main computational spaces within the human brain: a unique global workspace composed of distributed and heavily interconnected neurons with long-range axons, and a set of specialized and modular perceptual, motor, memory, evaluative and attentional processors. We postulate that workspace neurons are mobilized in effortful tasks for which the specialized processors do not suffice; they selectively mobilize or suppress, through descending connections, the contribution of specific processor neurons. In the course of task performance, workspace neurons become spontaneously co-activated, forming discrete though variable spatio-temporal patterns subject to modulation by vigilance signals and to selection by reward signals. A computer simulation of the Stroop task shows workspace activation to increase during acquisition of a novel task, effortful execution, and after errors. This model makes predictions concerning the spatio-temporal activation patterns during brain imaging of cognitive tasks, particularly concerning the conditions of activation of dorsolateral prefrontal cortex and anterior cingulate, their relation to reward mechanisms, and their specific reaction during error processing.},
}
@article{DehaeneandChangeuxPNAS-97,
        title = {A hierarchical neuronal network for planning behavior},
       author = {Dehaene, Stanislas and Changeux, Jean-Pierre},
      journal = {Proceedings of the National Academy of Sciences},
       volume = 94,
        issue = 24,
         year = 1997,
        pages = {13293-13298},
     abstract = {Planning a goal-directed sequence of behavior is a higher function of the human brain that relies on the integrity of prefrontal cortical areas. In the Tower of London test, a puzzle in which beads sliding on pegs must be moved to match a designated goal configuration, patients with lesioned prefrontal cortex show deficits in planning a goal-directed sequence of moves. We propose a neuronal network model of sequence planning that passes this test and, when lesioned, fails in a way that mimics prefrontal patients' behavior. Our model comprises a descending planning system with hierarchically organized plan, operation, and gesture levels, and an ascending evaluative system that analyzes the problem and computes internal reward signals that index the correct/erroneous status of the plan. Multiple parallel pathways connecting the evaluative and planning systems amend the plan and adapt it to the current problem. The model illustrates how specialized hierarchically organized neuronal assemblies may collectively emulate central executive or supervisory functions of the human brain.},
}
@article{DehaeneandCohenNEURON-07,
       author = {Stanislas Dehaene and Laurent Cohen},
        title = {Cultural Recycling of Cortical Maps},
      journal = {Neuron},
       volume = {56},
       number = {2},
        pages = {384-398},
         year = {2007},
     abstract = {Part of human cortex is specialized for cultural domains such as reading and arithmetic, whose invention is too recent to have influenced the evolution of our species. Representations of letter strings and of numbers occupy reproducible locations within large-scale macromaps, respectively in the left occipito-temporal and bilateral intraparietal cortex. Furthermore, recent fMRI studies reveal a systematic architecture within these areas. To explain this paradoxical cerebral invariance of cultural maps, we propose a neuronal recycling hypothesis, according to which cultural inventions invade evolutionarily older brain circuits and inherit many of their structural constraints.}
}
@article{DehaeneandNaccacheCOGNITION-01,
        title = {Towards a cognitive neuroscience of consciousness: basic evidence and a workspace framework},
       author = {Stanislas Dehaene and Lionel Naccache},
      journal = {Cognition}, 
       volume = {79},
       number = {1},
        pages = {1-37},
         year = {2001},
     abstract = {This introductory chapter attempts to clarify the philosophical, empirical, and theoretical bases on which a cognitive neuroscience approach to consciousness can be founded. We isolate three major empirical observations that any theory of consciousness should incorporate, namely (1) a considerable amount of processing is possible without consciousness, (2) attention is a prerequisite of consciousness, and (3) consciousness is required for some specific cognitive tasks, including those that require durable information maintenance, novel combinations of operations, or the spontaneous generation of intentional behavior. We then propose a theoretical framework that synthesizes those facts: the hypothesis of a global neuronal workspace. This framework postulates that, at any given time, many modular cerebral networks are active in parallel and process information in an unconscious manner. An information becomes conscious, however, if the neural population that represents it is mobilized by top-down attentional amplification into a brain-scale state of coherent activity that involves many neurons distributed throughout the brain. The long-distance connectivity of these 'workspace neurons' can, when they are active for a minimal duration, make the information available to a variety of processes including perceptual categorization, long-term memorization, evaluation, and intentional action. We postulate that this global availability of information through the workspace is what we subjectively experience as a conscious state. A complete theory of consciousness should explain why some cognitive and cerebral representations can be permanently or temporarily inaccessible to consciousness, what is the range of possible conscious contents, how they map onto specific cerebral circuits, and whether a generic neuronal mechanism underlies all of them. We confront the workspace model with those issues and identify novel experimental predictions. Neurophysiological, anatomical, and brain-imaging data strongly argue for a major role of prefrontal cortex, anterior cingulate, and the areas that connect to them, in creating the postulated brain-scale workspace.}
}
@incollection{Dehaeneetal2011,
       author = {Dehaene, Stanislas and Changeux, Jean-Pierre and Naccache, Lionel},
       editor = {Dehaene, Stanislas and Christen, Yves},
        title = {The Global Neuronal Workspace Model of Conscious Access: From Neuronal Architectures to Clinical Applications},
    booktitle = {Characterizing Consciousness: From Cognition to the Clinic?},
         year = {2011},
    publisher = {Springer Berlin Heidelberg},
      address = {Berlin, Heidelberg},
        pages = {55-84},
     abstract = {While a considerable body of experimental data has been accumulated on the differences between conscious and non-conscious processing, a theory is needed to bridge the neuro-psychological gap and establish a causal relationship between objective neurophysiological data and subjective reports. In the present review, we first briefly outline the detailed postulates and predictions of our working hypothesis, referred to as the global neuronal workspace (GNW) model. We then compare these predictions to experimental studies that have attempted to delineate the physiological signatures of conscious sensory perception by contrasting it with subliminal processing, using a variety of methods: behavioral, PET and fMRI imaging, time-resolved imaging with ERP and MEG, and finally single-cell electrophysiology. In a final section, we examine the relevance of these findings for pathologies of consciousness in coma and vegetative states.},
}
@article{DehaeneetalCOGNITIVE-NEUROPSYCHOLOGY-03,
       author = {Dehaene, S. and Piazza, M. and Pinel, P. and Cohen, L.},
        title = {{T}hree parietal circuits for number processing},
      journal = {Cognitive Neuropsychology},
         year = {2003},
       volume = {20},
       number = {3},
        pages = {487-506},
     abstract = {Did evolution endow the human brain with a predisposition to represent and acquire knowledge about numbers? Although the parietal lobe has been suggested as a potential substrate for a domain-specific representation of quantities, it is also engaged in verbal, spatial, and attentional functions that may contribute to calculation. To clarify the organisation of number-related processes in the parietal lobe, we examine the three-dimensional intersection of fMRI activations during various numerical tasks, and also review the corresponding neuropsychological evidence. On this basis, we propose a tentative tripartite organisation. The horizontal segment of the intraparietal sulcus (HIPS) appears as a plausible candidate for domain specificity: It is systematically activated whenever numbers are manipulated, independently of number notation, and with increasing activation as the task puts greater emphasis on quantity processing. Depending on task demands, we speculate that this core quantity system, analogous to an internal "number line," can be supplemented by two other circuits. A left angular gyrus area, in connection with other left-hemispheric perisylvian areas, supports the manipulation of numbers in verbal form. Finally, a bilateral posterior superior parietal system supports attentional orientation on the mental number line, just like on any other spatial dimension.}
}
@article{DehaeneetalPNAS-98,
        title = {A neuronal model of a global workspace in effortful cognitive tasks},
       author = {Dehaene, Stanislas and Kerszberg, Michel and Changeux, Jean-Pierre},
      journal = {Proceedings of the National Academy of Sciences},
       volume = 95,
        issue = 24,
         year = 1998,
        pages = {14529-14534},
     abstract = {A minimal hypothesis is proposed concerning the brain processes underlying effortful tasks. It distinguishes two main computational spaces: a unique global workspace composed of distributed and heavily interconnected neurons with long-range axons, and a set of specialized and modular perceptual, motor, memory, evaluative, and attentional processors. Workspace neurons are mobilized in effortful tasks for which the specialized processors do not suffice. They selectively mobilize or suppress, through descending connections, the contribution of specific processor neurons. In the course of task performance, workspace neurons become spontaneously coactivated, forming discrete though variable spatio-temporal patterns subject to modulation by vigilance signals and to selection by reward signals. A computer simulation of the Stroop task shows workspace activation to increase during acquisition of a novel task, effortful execution, and after errors. We outline predictions for spatio-temporal activation patterns during brain imaging, particularly about the contribution of dorsolateral prefrontal cortex and anterior cingulate to the workspace.},
}
@article{DehaeneetalSCIENCE-17,
       author = {Dehaene, Stanislas and Lau, Hakwan and Kouider, Sid},
        title = {What is consciousness, and could machines have it?},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {358},
       number = {6362},
         year = {2017},
        pages = {486-492},
     abstract = {The controversial question of whether machines may ever be conscious must be based on a careful consideration of how consciousness arises in the only physical system that undoubtedly possesses it: the human brain. We suggest that the word 'consciousness' conflates two different types of information-processing computations in the brain: the selection of information for global broadcasting, thus making it flexibly available for computation and report (C1, consciousness in the first sense), and the self-monitoring of those computations, leading to a subjective sense of certainty or error (C2, consciousness in the second sense). We argue that despite their recent successes, current machines are still mostly implementing computations that reflect unconscious processing (C0) in the human brain. We review the psychological and neural science of unconscious (C0) and conscious computations (C1 and C2) and outline how they may inspire novel machine architectures.},
}
@article{DeisenrothetalPAMI-15,
       author = {Deisenroth, M.P. and Fox, D. and Rasmussen, C.E.},
      journal = {Pattern Analysis and Machine Intelligence, IEEE Transactions on},
        title = {Gaussian Processes for Data-Efficient Learning in Robotics and Control},
         year = {2015},
       volume = {37},
       number = {2},
        pages = {408-423},
     abstract = {Autonomous learning has been a promising direction in control and robotics for more than a decade since data-driven learning allows to reduce the amount of engineering knowledge, which is otherwise required. However, autonomous reinforcement learning (RL) approaches typically require many interactions with the system to learn controllers, which is a practical limitation in real systems, such as robots, where many interactions can be impractical and time consuming. To address this problem, current learning approaches typically require task-specific knowledge in form of expert demonstrations, realistic simulators, pre-shaped policies, or specific knowledge about the underlying dynamics. In this paper, we follow a different approach and speed up learning by extracting more information from data. In particular, we learn a probabilistic, non-parametric Gaussian process transition model of the system. By explicitly incorporating model uncertainty into long-term planning and controller learning our approach reduces the effects of model errors, a key problem in model-based learning. Compared to state-of-the art RL our model-based policy search method achieves an unprecedented speed of learning. We demonstrate its applicability to autonomous learning in real robot and control tasks.},
}
@incollection{DekeletalNIPS-04,
       author = {Ofer Dekel and Shai Shalev-Shwartz and Yoram Singer},
        title = {The Power of Selective Memory: Self-Bounded Learning of Prediction Suffix Trees},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 17,
    publisher = {MIT Press},
       editor = {Lawrence K. Saul and Yair Weiss and {L\'{e}on} Bottou},
      address = {Cambridge, MA},
         year = 2004,
}
@incollection{DekeletalNIPS-18,
        title = {The Forgetron: A Kernel-Based Perceptron on a Fixed Budget},
       author = {Ofer Dekel and Shai Shalev-Shwartz and Yoram Singer},
    booktitle = {Advances in Neural Information Processing Systems 18},
       editor = {Y. Weiss and B. Sch\"{o}lkopf and J. Platt},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2006
}
@article{DelCuletalBRAIN-09,
       author = {Del Cul, A.  and Dehaene, S.  and Reyes, P.  and Bravo, E.  and Slachevsky, A.},
        title = {Causal role of prefrontal cortex in the threshold for access to consciousness},
      journal = {Brain},
         year = {2009},
       volume = {132},
       number = {9},
        pages = {2531-2540},
     abstract = {What neural mechanisms support our conscious perception of briefly presented stimuli? Some theories of conscious access postulate a key role of top-down amplification loops involving prefrontal cortex (PFC). To test this issue, we measured the visual backward masking threshold in patients with focal prefrontal lesions, using both objective and subjective measures while controlling for putative attention deficits. In all conditions of temporal or spatial attention cueing, the threshold for access to consciousness was systematically shifted in patients, particular after a lesion of the left anterior PFC. The deficit affected subjective reports more than objective performance, and objective performance conditioned on subjective visibility was essentially normal. We conclude that PFC makes a causal contribution to conscious visual perception of masked stimuli, and outline a dual-route signal detection theory of objective and subjective decision making.}
}
@inproceedings{DelfinadoandEdelsbrunnerACM-93,
       author = {Delfinado, Cecil Jose A. and Edelsbrunner, Herbert},
        title = {An Incremental Algorithm for Betti Numbers of Simplicial Complexes},
    booktitle = {Proceedings of the Ninth Annual Symposium on Computational Geometry},
    publisher = {ACM},
      address = {New York, NY, USA},
         year = {1993},
        pages = {232-239},
     abstract = {A general and direct method for computing the betti numbers of the homology groups of a finite simplicial complex is given. For subcomplexes of a triangulation of S3 this method has implementations that run in time O(ndgr;(n)) and O(n), where n is the number of simplices in the triangulation. If applied to the family of dgr;-shapes of a finite point set in R3 it takes time O(ndgr;(n)) to compute the betti numbers of all dgr;-shapes.}
}
@article{DeminandNekhaevFiN-18,
       author = {Demin, Vyacheslav and Nekhaev, Dmitry},
        title = {Recurrent Spiking Neural Network Learning Based on a Competitive Maximization of Neuronal Activity},
      journal = {Frontiers in Neuroinformatics},
       volume = {12},
        pages = {79},
         year = {2018},
     abstract = {Spiking neural networks (SNNs) are believed to be highly computationally and energy efficient for specific neurochip hardware real-time solutions. However, there is a lack of learning algorithms for complex SNNs with recurrent connections, comparable in efficiency with back-propagation techniques and capable of unsupervised training. Here we suppose that each neuron in a biological neural network tends to maximize its activity in competition with other neurons, and put this principle at the basis of a new SNN learning algorithm. In such a way, a spiking network with the learned feed-forward, reciprocal and intralayer inhibitory connections, is introduced to the MNIST database digit recognition. It has been demonstrated that this SNN can be trained without a teacher, after a short supervised initialization of weights by the same algorithm. Also, it has been shown that neurons are grouped into families of hierarchical structures, corresponding to different digit classes and their associations. This property is expected to be useful to reduce the number of layers in deep neural networks and modeling the formation of various functional structures in a biological nervous system. Comparison of the learning properties of the suggested algorithm, with those of the Sparse Distributed Representation approach shows similarity in coding but also some advantages of the former. The basic principle of the proposed algorithm is believed to be practically applicable to the construction of much more complicated and diverse task solving SNNs. We refer to this new approach as "Family-Engaged Execution and Learning of Induced Neuron Groups", or FEELING.}
}
@article{DemirhanJoBO-11,
       author = {Kobat, Demirhan and Horton, Nicholas G. and Xu, Chris},
        title = {In vivo two-photon microscopy to 1.6-mm depth in mouse cortex},
      journal = {Journal of Biomedical Optics},
       volume = {16},
       number = {10},
         year = {2011},
        pages = {106014},
     abstract = {Deep tissue in vivo two-photon fluorescence imaging of cortical vasculature in a mouse brain using 1280-nm excitation is presented. A record imaging depth of 1.6 mm in mouse cortex is achieved in vivo, approximately reaching the fundamental depth limit in scattering tissue.},
}
@article{DempsteretalJRSS-77,
       author = {Dempster, A. and Laird, N. and Rubin, D.},
        title = {Maximum likelihood from incomplete data via the {EM} algorithm},
      journal = {Journal of the Royal Statistical Society},
       volume = {39 (Series B)},
        pages = {1-38},
         year = 1977,
}
@article{DenOudenetalFiP-12,
       author = {Den Ouden, Hanneke and Kok, Peter and De Lange, Floris},
        title = {How Prediction Errors Shape Perception, Attention, and Motivation},
      journal = {Frontiers in Psychology},
       volume = {3},
        pages = {548},
         year = {2012},
     abstract = {Prediction errors (PE) are a central notion in theoretical models of reinforcement learning, perceptual inference, decision-making and cognition, and prediction error signals have been reported across a wide range of brain regions and experimental paradigms. Here, we will make an attempt to see the forest for the trees and consider the commonalities and differences of reported PE signals in light of recent suggestions that the computation of PE forms a fundamental mode of brain function. We discuss where different types of PE are encoded, how they are generated, and the different functional roles they fulfill. We suggest that while encoding of PE is a common computation across brain regions, the content and function of these error signals can be very different and are determined by the afferent and efferent connections within the neural circuitry in which they arise.}
} 
@article{DeneveandMachensNATURE-NEUROSCIENCE-16,
       author = {Deneve, Sophie and Machens, Christian K.},
        title = {Efficient codes and balanced networks},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {19},
        issue = {3},
         year = {2016},
        pages = {375-382},
     abstract = {Recent years have seen a growing interest in inhibitory interneurons and their circuits. A striking property of cortical inhibition is how tightly it balances excitation. Inhibitory currents not only match excitatory currents on average, but track them on a millisecond time scale, whether they are caused by external stimuli or spontaneous fluctuations. We review, together with experimental evidence, recent theoretical approaches that investigate the advantages of such tight balance for coding and computation. These studies suggest a possible revision of the dominant view that neurons represent information with firing rates corrupted by Poisson noise. Instead, tight excitatory/inhibitory balance may be a signature of a highly cooperative code, orders of magnitude more precise than a Poisson rate code. Moreover, tight balance may provide a template that allows cortical neurons to construct high-dimensional population codes and learn complex functions of their inputs.},
}
@article{DengetalBSF-14,
       author = {Deng, F.  and Jiang, X.  and Zhu, D.  and Zhang, T.  and Li, K.  and Guo, L.  and Liu, T. },
        title = {{{A} functional model of cortical gyri and sulci}},
      journal = {Brain Structure and Function},
       volume = {219},
       number = {4},
         year = {2014},
        pages = {1473-1491},
     abstract = {Diffusion tensor imaging (DTI) and high angular resolution diffusion imaging (HARDI) have been broadly used in the neuroimaging field to investigate the macro-scale fiber connection patterns in the cerebral cortex. Our recent analyses of DTI and HARDI data demonstrated that gyri are connected by denser, streamlined fibers than sulci are. Inspired by this finding and motivated by the fact that DTI-derived fibers provide the structural substrates for functional connectivity, we hypothesize that gyri are global functional connection centers and sulci are local functional units. To test this functional model of gyri and sulci, we examined the structural and functional connectivity among the landmarks on the selected gyral/sulcal areas in the frontal/parietal lobe and in the whole cerebral cortex via multimodal DTI and resting state fMRI (R-fMRI) datasets. Our results demonstrate that functional connectivity is strong among gyri, weak among sulci, and moderate between gyri and sulci. These results suggest that gyri are functional connection centers that exchange information among remote structurally connected gyri and neighboring sulci, while sulci communicate directly with their neighboring gyri and indirectly with other cortical regions through gyri. This functional model of gyri and sulci has been supported by a series of experiments, and provides novel perspectives on the functional architecture of the cerebral cortex.}
}
@misc{DenkPOSITIONAL-ENCODING-19,
        title = {Linear Relationships in the Transformer's Positional Encoding},
       author = {Timo Denk},
 howpublished = {\url{https://timodenk.com/blog/linear-relationships-in-the-transformers-positional-encoding/}},
         year = {2019},
}
@article{DenketalSCIENCE-90,
        title = {Two-photon laser scanning fluorescence microscopy},
       author = {Denk, Winfried and Strickler, James H. and Webb, Watt W. and others},
      journal = {Science},
       volume = {248},
        issue = {4951},
        pages = {73-76},
         year = {1990}
}
@book{Dennett03,
       author = {Daniel Dennett},
        title = {Freedom Evolves},
    publisher = {Viking},
      address = {New York, NY},
         year = 2003,
}
@book{Dennett15,
       author = {Daniel Dennett},
        title = {Elbow Room: The Varieties of Free Will Worth Wanting (Second Edition)},
       series = {Bradford Books},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2015,
}
@book{Dennett17,
       author = {Daniel Dennett},
        title = {From Bacteria to Bach and Back: The Evolution of Minds},
    publisher = {W.W. Norton},
      address = {New York, NY},
         year = 2017,
}
@book{Dennett1989,
       author = {Dennett, Daniel C.},
        title = {The Intentional Stance},
       series = {Bradford Books},
    publisher = {MIT Press},
         year = {1989},
}
@book{Dennett84,
       author = {Daniel Dennett},
        title = {Elbow Room: The Varieties of Free Will Worth Wanting (First Edition)},
       series = {Bradford Books},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1984,
}
@book{Dennett91,
       author = {Daniel Dennett},
        title = {Consciousness Explained},
    publisher = {Penguin},
      address = {London},
         year = 1991,
}
@book{Dennett95,
       author = {Daniel Dennett},
        title = {Darwin's Dangerous Idea: Evolution and the Meanings of Life},
    publisher = {Simon \& Schuster},
      address = {New York, NY},
         year = 1995,
}
@book{Dennett98,
       author = {Daniel Dennett},
        title = {Brainchildren: Essays on Designing Minds},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1998,
}
@book{DennettFREE_WILL_WORTH_WANTING-15,
       author = {Daniel Dennett},
        title = {Elbow Room: The Varieties of Free Will Worth Wanting (Second Edition)},
       series = {Bradford Books},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2015,
}
@article{DennettPTRSB-18,
       author = {Daniel C. Dennett},
        title = {Facing up to the hard question of consciousness},
      journal = {Philosophical Transactions of the Royal Society B: Biological Sciences},
       volume = {373},
       number = {1755},
        pages = {20170342},
         year = {2018},
     abstract = {The so-called hard problem of consciousness is a chimera, a distraction from the hard question of consciousness, which is once some content reaches consciousness, 'then what happens?'. This question is seldom properly asked, for reasons good and bad, but when asked it opens up avenues of research that promise to dissolve the hard problem and secure a scientifically sound theory of how the human brain produces the (sometimes illusory) convictions that mislead us.}
}
@article{Dennett_Aching_Voids_and_Making_Voids,
       author = {Daniel C. Dennett},
        title = {Aching Voids and Making Voids: A review of Incomplete Nature: How Mind Emerged from Matter by Terrence W. Deacon Review},
      journal = {Quarterly Review of Biology},
    publisher = {The University of Chicago Press},
       volume = {88},
        issue = {4},
         year = {2013},
        pages = {321-324},
}
@misc{Dennett_on_Chalmers_and_Illusionism_Draft_04-15-19,
        title = {Welcome to Strong Illusionism},
       author = {Daniel, Dennett},
         year = 2015,
}
@article{DenningUBIQUITY-10,
       author = {Denning, Peter J.},
        title = {Ubiquity Symposium {"What is Computation?"}: Opening Statement},
      journal = {Ubiquity},
    publisher = {ACM},
      address = {New York, NY, USA},
       volume = {2010},
       number = {November},
         year = {2010},
}
@article{DensmoreandHassounIEEE-12,
       author = {Douglas Densmore and Soha Hassoun},
        title = {Design Automation for Synthetic Biological Systems},
      journal = {{IEEE} Design and Testing of Computers},
       volume = 29,
        issue = 3,
         year = 2003,
        pages = {7-20}
}
@inproceedings{DeorasetalEMNLP-11,
       author = {Anoop Deoras and Tom\`{a}s Mikolov and Kenneth Church},
        title = {A Fast Re-scoring Strategy to Capture Long-Distance Dependencies},
        pages = {1116-1127},
    booktitle = {Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing},
         year = 2011,
     location = {Edinburgh, GB},
    publisher = {Association for Computational Linguistics},
}
@book{Dereetal2008,
        title = {Handbook of Episodic Memory},
       editor = {Ekrem Dere and Alexander Easton and Lynn Nadel and Joseph P. Huston},
       series = {Handbook of Behavioral Neuroscience},
    publisher = {Elsevier},
       volume = {18},
         year = {2008}
}
@book{DereetalHANDBOOK-EPISODIC-MEMORY-08,
        title = {Handbook of Episodic Memory},
       editor = {Ekrem Dere and Alexander Easton and Lynn Nadel and Joseph P. Huston},
       series = {Handbook of Behavioral Neuroscience},
    publisher = {Elsevier},
       volume = {18},
         year = {2008}
}
@book{Derman70,
       author = {Derman, Cyrus},
        title = {Finite State {M}arkovian Decision Processes},
    publisher = {Cambridge University Press},
      address = {New York},
         year = 1970,
}
@inproceedings{DesaietalICSE-16,
       author = {Desai, Aditya and Gulwani, Sumit and Hingorani, Vineet and Jain, Nidhi and Karkare, Amey and Marron, Mark and R, Sailesh and Roy, Subhajit},
        title = {Program Synthesis Using Natural Language},
    booktitle = {Proceedings of the 38th International Conference on Software Engineering},
    publisher = {Association for Computing Machinery},
         year = {2016},
        pages = {345–356},
	abstract = {Interacting with computers is a ubiquitous activity for millions of people. Repetitive or specialized tasks often require creation of small, often one-off, programs. End-users struggle with learning and using the myriad of domain-specific languages (DSLs) to effectively accomplish these tasks. We present a general framework for constructing program synthesizers that take natural language (NL) inputs and produce expressions in a target DSL. The framework takes as input a DSL definition and training data consisting of NL/DSL pairs. From these it constructs a synthesizer by learning optimal weights and classifiers (using NLP features) that rank the outputs of a keywordprogramming based translation. We applied our framework to three domains: repetitive text editing, an intelligent tutoring system, and flight information queries. On 1200+ English descriptions, the respective synthesizers rank the desired program as the top-1 and top3 for 80\% and 90\% descriptions respectively}
}
@article{DeschenesetalBRR-98,
       author = {Martin Desch\'{e}nes and Pierre Veinante and Zhong-Wei Zhang},
        title = {The organization of corticothalamic projections: reciprocity versus parity},
      journal = {Brain Research Reviews},
       volume = 28,
        issue = 3,
        pages = {286-308},
         year = 1998,
}
@article{DeshpandeandMontanariFoCM-13,
        title = {Finding Hidden Cliques of Size $\backslash$ sqrt $\{$N/e$\}$ in Nearly Linear Time},
       author = {Deshpande, Yash and Montanari, Andrea},
      journal = {Foundations of Computational Mathematics},
    publisher = {Springer US},
         year = 2013,
        pages = {1-60},
     abstract = {Consider an Erd\"{o}s-Renyi random graph in which each edge is present independently with probability\(1/2\), except for a subset\(\ mathsf {C}_N\) of the vertices that form a clique (a completely connected subgraph). We consider the problem of identifying the clique, given a realization of such a random graph. The algorithm of Dekel et al. (ANALCO. SIAM, pp 67-75, 2011) provably identifies the clique\(\ mathsf {C}_N\) in linear time, provided\(|\ mathsf {C}_N|\ \geq 1.261 \sqrt {N}\). Spectral methods can be shown to fail on cliques smaller than \sqrt N. In this paper we describe a nearly linear time algorithm that succeeds with high probability for |CN | \geq (1 + \epsilon)p N/e for any \epsilon > 0. This is the first algorithm that provably improves over spectral methods. We further generalize the hidden clique problem to other background graphs (the standard case corresponding to the complete graph on N vertices). For large girth regular graphs of degree (\Delta + 1) we prove that 'local' algorithms succeed if | CN | \geq (1 + \epsilon)N/\sqrt e\Delta and fail if |CN | \leq (1 − \epsilon)N/\sqrt e\Delta.}
}
@book{Deutscher2006evolution,
        title = {The Unfolding of Language: {A}n Evolutionary Tour of Mankind's Greatest Invention},
       author = {Deutscher, Guy},
      address = {New York City, NY},
    publisher = {Henry Holt and Company},
         year = {2006},
}
@article{DevlinetalCoRR-17,
       author = {Jacob Devlin and Jonathan Uesato and Surya Bhupatiraju and Rishabh Singh and Abdel{-}rahman Mohamed and Pushmeet Kohli},
        title = {RobustFill: Neural Program Learning under Noisy {I/O}},
      journal = {CoRR},
       volume = {arXiv:1703.07469},
         year = {2017},
     abstract = {The problem of automatically generating a computer program from some specification has been studied since the early days of AI. Recently, two competing approaches for automatic program learning have received significant attention: (1) neural program synthesis, where a neural network is conditioned on input/output (I/O) examples and learns to generate a program, and (2) neural program induction, where a neural network generates new outputs directly using a latent program representation. Here, for the first time, we directly compare both approaches on a large-scale, real-world learning task. We additionally contrast to rule-based program synthesis, which uses hand-crafted semantics to guide the program generation. Our neural models use a modified attention RNN to allow encoding of variable-sized sets of I/O pairs. Our best synthesis model achieves 92\% accuracy on a real-world test set, compared to the 34\% accuracy of the previous best neural synthesis approach. The synthesis model also outperforms a comparable induction model on this task, but we more importantly demonstrate that the strength of each approach is highly dependent on the evaluation metric and end-user application. Finally, we show that we can train our neural models to remain very robust to the type of noise expected in real-world data (e.g., typos), while a highly-engineered rule-based system fails entirely.}
}
@article{DevlinetalCoRR-18,
       author = {Jacob Devlin and Ming{-}Wei Chang and Kenton Lee and Kristina Toutanova},
        title = {{BERT:} Pre-training of Deep Bidirectional Transformers for Language Understanding},
      journal = {CoRR},
       volume = {arXiv:1810.04805},
         year = {2018},
     abstract = {We introduce a new language representation model called BERT, which stands for Bidirectional Encoder Representations from Transformers. Unlike recent language representation models, BERT is designed to pre-train deep bidirectional representations from unlabeled text by jointly conditioning on both left and right context in all layers. As a result, the pre-trained BERT model can be fine-tuned with just one additional output layer to create state-of-the-art models for a wide range of tasks, such as question answering and language inference, without substantial task-specific architecture modifications.}
}
@inproceedings{DevlinetalICLR-18,
        title = {Semantic Code Repair using Neuro-Symbolic Transformation Networks},
       author = {Jacob Devlin and Jonathan Uesato and Rishabh Singh and Pushmeet Kohli},
    booktitle = {International Conference on Learning Representations},
         year = {2018},
     abstract = {We study the problem of semantic code repair, which can be broadly defined as automatically fixing non-syntactic bugs in source code. The majority of past work in semantic code repair assumed access to unit tests against which candidate repairs could be validated. In contrast, the goal here is to develop a strong statistical model to accurately predict both bug locations and exact fixes without access to information about the intended correct behavior of the program. Achieving such a goal requires a robust contextual repair model, which we train on a large corpus of real-world source code that has been augmented with synthetically injected bugs. Our framework adopts a two-stage approach where first a large set of repair candidates are generated by rule-based processors, and then these candidates are scored by a statistical model using a novel neural network architecture which we refer to as Share, Specialize, and Compete. Specifically, the architecture (1) generates a  shared encoding of the source code using an RNN over the abstract syntax tree, (2) scores each candidate repair using specialized network modules, and (3) then normalizes these scores together so they can compete against one another in comparable probability space. We evaluate our model on a real-world test set gathered from GitHub containing four common categories of bugs. Our model is able to predict the exact correct repair 41\% of the time with a single guess, compared to 13\% accuracy for an attentional sequence-to-sequence model.}
}
@inproceedings{DevlinetalICML-17,
       author = {Devlin, Jacob and Uesato, Jonathan and Bhupatiraju, Surya and Singh, Rishabh and Mohamed, Abdel-rahman and Kohli, Pushmeet},
        title = {RobustFill: Neural Program Learning Under Noisy I/O},
    booktitle = {Proceedings of the 34th International Conference on Machine Learning - Volume 70},
    publisher = {JMLR.org},
     location = {Sydney, NSW, Australia},
         year = {2017},
        pages = {990-998},
}
@incollection{DevlinetalNIPS-17,
        title = {Neural Program Meta-Induction},
       author = {Devlin, Jacob and Bunel, Rudy R and Singh, Rishabh and Hausknecht, Matthew and Kohli, Pushmeet},
    booktitle = {Advances in Neural Information Processing Systems 30},
       editor = {I. Guyon and U. V. Luxburg and S. Bengio and H. Wallach and R. Fergus and S. Vishwanathan and R. Garnett},
    publisher = {Curran Associates, Inc.},
         year = {2017},
        pages = {2077-2085},
     abstract = {Most recently proposed methods for Neural Program induction work under the assumption of having a large set of input/output (I/O) examples for learning any given input-output mapping. This paper aims to address the problem of data and computation efficiency of program induction by leveraging information from related tasks. Specifically, we propose two novel approaches for cross-task knowledge transfer to improve program induction in limited-data scenarios. In our first proposal, portfolio adaptation, a set of induction models is pretrained on a set of related tasks, and the best model is adapted towards the new task using transfer learning. In our second approach, meta program induction, a k-shot learning approach is used to make a model generalize to new tasks without additional training. To test the efficacy of our methods, we constructed a new benchmark of programs written in the Karel programming language. Using an extensive experimental evaluation on the Karel benchmark, we demonstrate that our proposals dramatically outperform the baseline induction method that does not use knowledge transfer. We also analyze the relative performance of the two approaches and study conditions in which they perform best. In particular, meta induction outperforms all existing approaches under extreme data sparsity (when a very small number of examples are available), i.e., fewer than ten. As the number of available I/O examples increase (i.e. a thousand or more), portfolio adapted program induction becomes the best approach. For intermediate data sizes, we demonstrate that the combined method of adapted meta program induction has the strongest performance.}
}
@article{DeweyetalNATURE-14,
       author = {Dewey, Daniel C. and Strulson, Christopher A. and Cacace, David N. and Bevilacqua, Philip C. and Keating, Christine D.},
        title = {Bioreactor droplets from liposome-stabilized all-aqueous emulsions},
      journal = {Nature Communications},
    publisher = {Nature Publishing Group},
       volume = 5,
         year = 2014,
     abstract = {Artificial bioreactors are desirable for in vitro biochemical studies and as protocells. A key challenge is maintaining a favourable internal environment while allowing substrate entry and product departure. We show that semipermeable, size-controlled bioreactors with aqueous, macromolecularly crowded interiors can be assembled by liposome stabilization of an all-aqueous emulsion. Dextran-rich aqueous droplets are dispersed in a continuous polyethylene glycol (PEG)-rich aqueous phase, with coalescence inhibited by adsorbed {\textasciitilde}130-nm diameter liposomes. Fluorescence recovery after photobleaching and dynamic light scattering data indicate that the liposomes, which are PEGylated and negatively charged, remain intact at the interface for extended time. Inter-droplet repulsion provides electrostatic stabilization of the emulsion, with droplet coalescence prevented even for submonolayer interfacial coatings. RNA and DNA can enter and exit aqueous droplets by diffusion, with final concentrations dictated by partitioning. The capacity to serve as microscale bioreactors is established by demonstrating a ribozyme cleavage reaction within the liposome-coated droplets.},
}
@article{DeyandHaqueIJDAR-09,
       author = {Dey, Lipika and Haque, S K Mirajul},
        title = {Opinion mining from noisy text data},
      journal = {Proceedings of the second workshop on Analytics for noisy unstructured text data},
       volume = 12,
        issue = 3,
         year = 2009,
        pages = {205-226},
     abstract = {The proliferation of Internet has not only led to the generation of huge volumes of unstructured information in the form of web documents, but a large amount of text is also generated in the form of emails, blogs, and feedbacks, etc. The data generated from online communication acts as potential gold mines for discovering knowledge, particularly for market researchers. Text analytics has matured and is being successfully employed to mine important information from unstructured text documents. The chief bottleneck for designing text mining systems for handling blogs arise from the fact that online communication text data are often noisy. These texts are informally written. They suffer from spelling mistakes, grammatical errors, improper punctuation and irrational capitalization. This paper focuses on opinion extraction from noisy text data. It is aimed at extracting and consolidating opinions of customers from blogs and feedbacks, at multiple levels of granularity. We have proposed a framework in which these texts are first cleaned using domain knowledge and then subjected to mining. Ours is a semi-automated approach, in which the system aids in the process of knowledge assimilation for knowledge-base building and also performs the analytics. Domain experts ratify the knowledge base and also provide training samples for the system to automatically gather more instances for ratification. The system identifies opinion expressions as phrases containing opinion words, opinionated features and also opinion modifiers. These expressions are categorized as positive or negative with membership values varying from zero to one. Opinion expressions are identified and categorized using localized linguistic techniques. Opinions can be aggregated at any desired level of specificity i.e. feature level or product level, user level or site level, etc. We have developed a system based on this approach, which provides the user with a platform to analyze opinion expressions crawled from a set of pre-defined blogs.}
}
@article{DheenetalCMC-07,
       author = {Dheen, S. T. and Kaur, C. and Ling, E. A.},
        title = {Microglial activation and its implications in the brain diseases},
      journal = {Current Medicinal Chemistry},
       volume = {14},
       number = {11},
         year = {2007},
        pages = {1189-1197},
     abstract = {An inflammatory process in the central nervous system (CNS) is believed to play an important role in the pathway leading to neuronal cell death in a number of neurodegenerative diseases including Parkinson's disease, Alzheimer's disease, prion diseases, multiple sclerosis and HIV-dementia. The inflammatory response is mediated by the activated microglia, the resident immune cells of the CNS, which normally respond to neuronal damage and remove the damaged cells by phagocytosis. Activation of microglia is a hallmark of brain pathology. However, it remains controversial whether microglial cells have beneficial or detrimental functions in various neuropathological conditions. The chronic activation of microglia may in turn cause neuronal damage through the release of potentially cytotoxic molecules such as proinflammatory cytokines, reactive oxygen intermediates, proteinases and complement proteins. Therefore, suppression of microglia-mediated inflammation has been considered as an important strategy in neurodegenerative disease therapy. Several anti-inflammatory drugs of various chemical ingredients have been shown to repress the microglial activation and to exert neuroprotective effects in the CNS following different types of injuries. However, the molecular mechanisms by which these effects occur remain unclear. In recent years, several research groups including ours have attempted to explain the potential mechanisms and signaling pathways for the repressive effect of various drugs, on activation of microglial cells in CNS injury. We provide here a comprehensive review of recent findings of mechanisms and signaling pathways by which microglial cells are activated in CNS inflammatory diseases. This review article further summarizes the role of microglial cells in neurodegenerative diseases and various forms of potential therapeutic options to inhibit the microglial activation which amplifies the inflammation-related neuronal injury in neurodegenerative diseases.}
}
@article{DiCarloandCoxTCS-07,
       author = {James J. DiCarlo and David D. Cox},
        title = {Untangling invariant object recognition},
      journal = {Trends in Cognitive Sciences},
       volume = 11,
        issue = 8,
        pages = {333-341},
         year = 2007,
}
@article {DiCarloetalNEURON-12,
        title = {How does the brain solve visual object recognition?},
       author = {James J {DiCarlo} and Davide Zoccolan and Nicole C Rust},
      journal = {Neuron},
       volume = 73,
         year = 2012,
        pages = {415-34},
     abstract = {Mounting evidence suggests that core object recognition, the ability to rapidly recognize objects despite substantial appearance variation, is solved in the brain via a cascade of reflexive, largely feedforward computations that culminate in a powerful neuronal representation in the inferior temporal cortex. However, the algorithm that produces this solution remains poorly understood. Here we review evidence ranging from individual neurons and neuronal populations to behavior and computational models. We propose that understanding this algorithm will require using neuronal and psychophysical data to sift through many computational models, each based on building blocks of small, canonical subnetworks with a common functional goal.},
} 
@book{Diamond2013,
        title = {Collapse: How Societies Choose to Fail or Survive},
       author = {Diamond, Jared},
         year = {2013},
    publisher = {Penguin Books Limited}
}
@book{Dick1968,
       author = {Philip K. Dick},
        title = {Do Androids Dream of Electric Sheep?},
         year = 1968,
    publisher = {Knopf Doubleday},
      address = {New York, NY},
}
@incollection{DiegoandHamprechtNIPS-13,
        title = {Learning Multi-level Sparse Representations},
       author = {Diego Andilla, Ferran and Hamprecht, Fred A},
    booktitle = {Advances in Neural Information Processing Systems 26},
       editor = {C. J. C. Burges and L. Bottou and M. Welling and Z. Ghahramani and K. Q. Weinberger},
    publisher = {Curran Associates, Inc.},
        pages = {818-826},
         year = {2013},
     abstract = {Bilinear approximation of a matrix is a powerful paradigm of unsupervised learning. In some applications, however, there is a natural hierarchy of concepts that ought to be reflected in the unsupervised analysis. For example, in the neurosciences image sequence considered here, there are the semantic concepts of pixel → neuron → assembly that should find their counterpart in the unsupervised analysis. Driven by this concrete problem, we propose a decomposition of the matrix of observations into a product of more than two sparse matrices, with the rank decreasing from lower to higher levels. In contrast to prior work, we allow for both hierarchical and heterarchical relations of lower-level to higher-level concepts. In addition, we learn the nature of these relations rather than imposing them. Finally, we describe an optimization scheme that allows to optimize the decomposition over all levels jointly, rather than in a greedy level-by-level fashion. The proposed bilevel SHMF (sparse heterarchical matrix factorization) is the first formalism that allows to simultaneously interpret a calcium imaging sequence in terms of the constituent neurons, their membership in assemblies, and the time courses of both neurons and assemblies. Experiments show that the proposed model fully recovers the structure from difficult synthetic data designed to imitate the experimental data. More importantly, bilevel SHMF yields plausible interpretations of real-world Calcium imaging data.}
}
@incollection{DiegoandHamprechtNIPS-14,
        title = {Sparse Space-Time Deconvolution for Calcium Image Analysis},
       author = {Diego Andilla, Ferran and Hamprecht, Fred A.},
    booktitle = {Advances in Neural Information Processing Systems 27},
       editor = {Z. Ghahramani and M. Welling and C. Cortes and N. D. Lawrence and K. Q. Weinberger},
    publisher = {Curran Associates, Inc.},
         year = {2014},
        pages = {64-72},
     abstract = {We describe a unified formulation and algorithm to find an extremely sparse rep- resentation for Calcium image sequences in terms of cell locations, cell shapes, spike timings and impulse responses. Solution of a single optimization problem yields cell segmentations and activity estimates that are on par with the state of the art, without the need for heuristic pre- or postprocessing. Experiments on real and synthetic data demonstrate the viability of the proposed method.},
}
@inproceedings{DiegoetalISBI-13,
       author = {F. Diego and S. Reichinnek and M. Both and F. A. Hamprecht},
    booktitle = {Biomedical Imaging (ISBI), 2013 IEEE 10th International Symposium on},
        title = {Automated identification of neuronal activity from calcium imaging by sparse dictionary learning},
         year = {2013},
        pages = {1058-1061},
     abstract = {We present ADINA, an automated pipeline for analyzing and identifying neuronal activity from calcium imaging data to investigate neuronal activity patterns. This entails the detection and classification of cell centroids and of calcium transients (events) that reappeared during different activity periods as memory consolidation. Specifically, the pipeline implements a sparse dictionary learning to infer the most relevant Ca2+ patterns, an image segmentation procedure using a wavelet- transform and watershed to identify single cells, and an estimation of the transient signals by means of sparse coding exploiting spatial and temporal sparsity. We validate our automated approach on artificial and two different calcium imaging sequences from mice hippocampal slice cultures acquired with fluorescence and confocal microscopes. Our approach achieves ca. 94\% sensitivity on average for correctly detecting events, thus improving significantly the estimation of cell signals relative to published procedures.},
}
@article{DietterichJAIR-00,
       author = {Thomas G. Dietterich},
        title = {Hierarchical reinforcement learning with the {MAXQ} value function decomposition},
      journal = {Journal of Artificial Intelligence Research},
       volume = {13},
         year = {2000},
        pages = {227-303},
     abstract = {This paper presents a new approach to hierarchical reinforcement learning based on decomposing the target Markov decision process (MDP) into a hierarchy of smaller MDPs and decomposing the value function of the target MDP into an additive combination of the value functions of the smaller MDPs. The decomposition, known as the MAXQ decomposition, has both a procedural semantics---as a subroutine hierarchy---and a declarative semantics---as a representation of the value function of a hierarchical policy. MAXQ unifies and extends previous work on hierarchical reinforcement learning by Singh, Kaelbling, and Dayan and Hinton. It is based on the assumption that the programmer can identify useful subgoals and define subtasks that achieve these subgoals. By defining such subgoals, the programmer constrains the set of policies that need to be considered during reinforcement learning. The MAXQ value function decomposition can represent the value function of any policy that is consistent with the given hierarchy. The decomposition also creates opportunities to exploit state abstractions, so that individual MDPs within the hierarchy can ignore large parts of the state space. This is important for the practical application of the method. This paper defines the MAXQ hierarchy, proves formal results on its representational power, and establishes five conditions for the safe use of state abstractions. The paper presents an online model-free learning algorithm, MAXQ-Q, and proves that it converges with probability 1 to a kind of locally-optimal policy known as a recursively optimal policy, even in the presence of the five kinds of state abstraction. The paper evaluates the MAXQ representation and MAXQ-Q through a series of experiments in three domains and shows experimentally that MAXQ-Q (with state abstractions) converges to a recursively optimal policy much faster than flat Q learning. The fact that MAXQ learns a representation of the value function has an important benefit: it makes it possible to compute and execute an improved, non-hierarchical policy via a procedure similar to the policy improvement step of policy iteration. The paper demonstrates the effectiveness of this non-hierarchical execution experimentally. Finally, the paper concludes with a comparison to related work and a discussion of the design tradeoffs in hierarchical reinforcement learning.}
}
@incollection{DietterichREVIEW,
       author = {Dietterich, Thomas G.},
        title = {Machine Learning},
    booktitle = {Annual Review of Computer Science, Volume 4},
    publisher = {Annual Review Incorporated},
         year = 1990,
        pages = {255-306},
}
@inproceedings{DietterichRL-MAXQ-98,
       author = {Thomas G. Dietterich},
        title = {The MAXQ Method for Hierarchical Reinforcement Learning},
    booktitle = {In Proceedings of the Fifteenth International Conference on Machine Learning},
    publisher = {Morgan Kaufmann},
         year = {1998},
        pages = {118-126},
     abstract = {This paper presents a new approach to hierarchical reinforcement learning based on the MAXQ decomposition of the value function. The MAXQ decomposition has both a procedural semantics---as a subroutine hierarchy---and a declarative semantics---as a representation of the value function of a hierarchical policy. MAXQ unifies and extends previous work on hierarchical reinforcement learning by Singh, Kaelbling, and Dayan and Hinton. Conditions under which the MAXQ decomposition can represent the optimal value function are derived. The paper defines a hierarchical Q learning algorithm, proves its convergence, and shows experimentally that it can learn much faster than ordinary "flat" Q learning. Finally, the paper discusses some interesting issues that arise in hierarchical reinforcement learning including the hierarchical credit assignment problem and non-hierarchical execution of the MAXQ hierarchy.}
}
@inproceedings{DietterichandFlannML-95,
       author = {Dietterich, Thomas G. and Flann, Nicholas S.},
        title = {Explanation-Based Learning and Reinforcement Learning: A Unified View},
    booktitle = {Proceedings Twelfth International Conference on Machine Learning},
     location = {Lake Tahoe, California},
        pages = {176-184},
         year = 1995,
}
@article{DietterichetalAIJ-97,
       author = {Thomas G. Dietterich and Richard H. Lathrop and Tom\'{a}s Lozano-P\'{e}rez},
        title = {Solving the Multiple-Instance Problem with Axis-Parallel Rectangles},
      journal = {Artificial Intelligence},
       volume = 89,
        pages = {31-71},
         year = 1997,
}
@article{DietzscholdetalCTMI-05,
       author = {Dietzschold, B. and Schnell, M. and Koprowski, H.},
        title = {Pathogenesis of rabies},
      journal = {Current Topics in Microbiology and Immunolology},
       volume = 292,
         year = 2005,
        pages = {45-56},
}
@article{DijkmansetalEJE-04,
       author = {Dijkmans, P.A and Juffermans, L.J.M and Musters, R.J.P and van Wamel, A and ten Cate, F.J and van Gilst, W and Visser, C.A and de Jong, N and Kamp, O},
        title = {Microbubbles and ultrasound: from diagnosis to therapy},
      journal = {European Journal of Echocardiography},
       volume = 5,
        issue = 4,
         year = 2004,
        pages = {245-246},
     abstract = {The development of ultrasound contrast agents, containing encapsulated microbubbles, has increased the possibilities for diagnostic imaging. Ultrasound contrast agents are currently used to enhance left ventricular opacification, increase Doppler signal intensity, and in myocardial perfusion imaging. Diagnostic imaging with contrast agents is performed with low acoustic pressure using non-linear reflection of ultrasound waves by microbubbles. Ultrasound causes bubble destruction, which lowers the threshold for cavitation, resulting in microstreaming and increased permeability of cell membranes. Interestingly, this mechanism can be used for delivery of drugs or genes into tissue. Microbubbles have been shown to be capable of carrying drugs and genes, and destruction of the bubbles will result in local release of their contents. Recent studies demonstrated the potential of microbubbles and ultrasound in thrombolysis. In this article, we will review the recent advances of microbubbles as a vehicle for delivery of drugs and genes, and discuss possible therapeutic applications in thrombolysis.},
}
@article{DijksterhuisetalSCIENCE-05,
       author = {Dijksterhuis, Ap and Bos, Maarten W. and Nordgren, Loran F. and van Baaren, Rick B.},
        title = {On Making the Right Choice: The Deliberation-Without-Attention Effect},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {311},
       number = {5763},
         year = {2006},
        pages = {1005-1007},
     abstract = {Contrary to conventional wisdom, it is not always advantageous to engage in thorough conscious deliberation before choosing. On the basis of recent insights into the characteristics of conscious and unconscious thought, we tested the hypothesis that simple choices (such as between different towels or different sets of oven mitts) indeed produce better results after conscious thought, but that choices in complex matters (such as between different houses or different cars) should be left to unconscious thought. Named the 'deliberation-without-attention' hypothesis, it was confirmed in four studies on consumer choice, both in the laboratory as well as among actual shoppers, that purchases of complex products were viewed more favorably when decisions had been made in the absence of attentive deliberation.},
}
@inproceedings{Diligenti00,
       author = {M. Diligenti and F. M. Coetzee and S. Lawrence and C. L. Giles and M. Gori},
        title = {Focused Crawling using Context Graphs},
    booktitle = {26th International Conference on Very Large Databases},
         year = 2000,
}
@article{DindoetalBC-15,
       author = {Dindo, Haris and Donnarumma, Francesco and Chersi, Fabian and Pezzulo, Giovanni},
        title = {The intentional stance as structure learning: a computational perspective on mindreading},
      journal = {Biological Cybernetics},
         year = {2015},
       volume = {109},
       number = {4},
        pages = {453-467},
     abstract = {Recent theories of mindreading explain the recognition of action, intention, and belief of other agents in terms of generative architectures that model the causal relations between observables (e.g., observed movements) and their hidden causes (e.g., action goals and beliefs). Two kinds of probabilistic generative schemes have been proposed in cognitive science and robotics that link to a "theory theory" and "simulation theory" of mindreading, respectively. The former compares perceived actions to optimal plans derived from rationality principles and conceptual theories of others' minds. The latter reuses one's own internal (inverse and forward) models for action execution to perform a look-ahead mental simulation of perceived actions. Both theories, however, leave one question unanswered: how are the generative models -- including task structure and parameters -- learned in the first place? We start from Dennett's "intentional stance" proposal and characterize it within generative theories of action and intention recognition. We propose that humans use an intentional stance as a learning bias that sidesteps the (hard) structure learning problem and bootstraps the acquisition of generative models for others' actions. The intentional stance corresponds to a candidate structure in the generative scheme, which encodes a simplified belief-desire folk psychology and a hierarchical intention-to-action organization of behavior. This simple structure can be used as a proxy for the "true" generative structure of others' actions and intentions and is continuously grown and refined -- via state and parameter learning -- during interactions. In turn -- as our computational simulations show -- this can help solve mindreading problems and bootstrap the acquisition of useful causal models of both one's own and others' goal-directed actions.},
}
@article{DiracPRS_A-30,
       author = {Dirac, Paul},
        title = {A Theory of Electrons and Protons},
      journal = {Proceedings of the Royal Society A: Mathematical, Physical, and Engineering Sciences},
       volume = {126},
       number = {801},
         year = {1930},
        pages = {360-365},
}
@inproceedings{DiuketalCRMHOB-13,
        title = {Divide and conquer: hierarchical reinforcement learning and task decomposition in humans},
       author = {Diuk, C. and Schapiro, A. and C\'{o}rdova, N. and Ribas-Fernandes, J. and Niv, Y. and Botvinick, M.},
    booktitle = {Computational and robotic models of the hierarchical organization of behavior},
    publisher = {Springer},
      address = {Berlin, Heidelberg},
         year = {2013},
        pages = {271-291},
}
@inproceedings{DivvalaetalCVPR-09,
        title = {An Empirical Study of Context in Object Detection},
       author = {Santosh K. Divvala and Derek Hoiem and James H. Hays and Alexei A. Efros and Martial Hebert},
    booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {{IEEE} Computer Society},
         year = 2009,
}
@article{DjalaliLiLT-13,
       author = {Alex J. Djalali},
        title = {Synthetic Logic},
      journal = {Linguistic Issues in Language Technology},
       volume = 9,
        issue = 1,
         year = 2013,
}
@article{DlotkoetalCoRR-16,
       author = {Pawel Dlotko and Kathryn Hess and Ran Levi and Max Nolte and Michael Reimann and Martina Scolamiero and Katharine Turner and Eilif Muller and Henry Markram},
        title = {Topological Analysis of the Connectome of Digital Reconstructions of Neural Microcircuits},
      journal = {CoRR},
       volume = {arXiv:1601.01580},
         year = {2016},
     abstract = {A recent publication provides the network graph for a neocortical microcircuit comprising 8 million connections between 31,000 neurons (H. Markram, et al., Reconstruction and simulation of neocortical microcircuitry, Cell, 163 (2015) no. 2, 456-492). Since traditional graph-theoretical methods may not be sufficient to understand the immense complexity of such a biological network, we explored whether methods from algebraic topology could provide a new perspective on its structural and functional organization. Structural topological analysis revealed that directed graphs representing connectivity among neurons in the microcircuit deviated significantly from different varieties of randomized graph. In particular, the directed graphs contained in the order of 107 simplices [in] groups of neurons with all-to-all directed connectivity. Some of these simplices contained up to 8 neurons, making them the most extreme neuronal clustering motif ever reported. Functional topological analysis of simulated neuronal activity in the microcircuit revealed novel spatio-temporal metrics that provide an effective classification of functional responses to qualitatively different stimuli. This study represents the first algebraic topological analysis of structural connectomics and connectomics-based spatio-temporal activity in a biologically realistic neural microcircuit. The methods used in the study show promise for more general applications in network science.}
}
@article{DoddsetalSCIENCE-03,
    author = {Peter Dodds and Roby Muhamad and Duncan Watts},
     title = {An Experimental Study of Search in Global Social Networks},
   journal = {Science},
    volume = 301,
     issue = 5634,
     pages = {827-829},
      year = 2003,
}
@article{DoerschCoRR-17,
       author = {Carl Doersch},
        title = {Tutorial on Variational Autoencoders},
      journal = {CoRR},
       volume = {arXiv:1606.05908},
         year = {2017},
     abstract = {In just three years, Variational Autoencoders (VAEs) have emerged as one of the most popular approaches to unsupervised learning of complicated distributions. VAEs are appealing because they are built on top of standard function approximators (neural networks), and can be trained with stochastic gradient descent. VAEs have already shown promise in generating many kinds of complicated data, including handwritten digits, faces, house numbers, CIFAR images, physical models of scenes, segmentation, and predicting the future from static images. This tutorial introduces the intuitions behind VAEs, explains the mathematics behind them, and describes some empirical behavior. No prior knowledge of variational Bayesian methods is assumed},
}
@article{DoesburgetalCORTEX-08,
       author = {Doesburg, Sam M. and Roggeveen, Alexa B. and Kitajo, Keiichi and Ward, Lawrence M.},
        title = {Large-Scale Gamma-Band Phase Synchronization and Selective Attention},
      journal = {Cerebral Cortex},
       volume = 18,
        issue = 2,
        pages = {386-396},
         year = 2008,
     abstract = {Explaining the emergence of a coherent conscious percept and an intentional agent from the activity of distributed neurons is key to understanding how the brain produces higher cognitive processes. Gamma-band synchronization has been proposed to be a mechanism for the functional integration of neural populations that together form a transitory, large-scale, task- and/or percept-specific network. The operation of this mechanism in the context of attention orienting entails that cortical regions representing attended locations should show more gamma-band synchronization with other cortical areas than would those representing unattended locations. This increased synchronization should be apparent in the same time frame as that of the deployment of attention to a particular location. In order to observe this effect, we made electroencephalogram recordings while subjects attended to one side or the other of the visual field (which we confirmed by event-related potential analysis) and calculated phase-locking statistics between the signals recorded at relevant electrode pairs. We observed increased gamma-band phase synchronization between visual cortex contralateral to the attended location and other, widespread, cortical areas approximately 240-380 ms after the directional cue was presented, confirming the prediction of a large-scale gamma synchronous network oriented to the cued location.},
}
@article{DoironetalNATURE-NEUROSCIENCE-16,
       author = {Doiron, Brent and Litwin-Kumar, Ashok and Rosenbaum, Robert and Ocker, Gabriel K. and Josic, Kresimir},
        title = {The mechanics of state-dependent neural correlations},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {19},
        issue = {3},
         year = {2016},
        pages = {383-393},
     abstract = {Simultaneous recordings from large neural populations are becoming increasingly common. An important feature of population activity is the trial-to-trial correlated fluctuation of spike train outputs from recorded neuron pairs. Similar to the firing rate of single neurons, correlated activity can be modulated by a number of factors, from changes in arousal and attentional state to learning and task engagement. However, the physiological mechanisms that underlie these changes are not fully understood. We review recent theoretical results that identify three separate mechanisms that modulate spike train correlations: changes in input correlations, internal fluctuations and the transfer function of single neurons. We first examine these mechanisms in feedforward pathways and then show how the same approach can explain the modulation of correlations in recurrent networks. Such mechanistic constraints on the modulation of population activity will be important in statistical analyses of high-dimensional neural data.},
}
@inproceedings{DolanetalACL-04,
       author = {Bill Dolan and Chris Quirk and Chris Brockett},
        title = {Unsupervised construction of large paraphrase corpora: Exploiting massively parallel news sources},
    booktitle = {Proceedings of the 20th International Conference on Computational Linguistics},
    publisher = {Association for Computational Linguistics},
         year = 2004,
}
@inproceedings{DollaretalECCV-08,
       author = {Piotr Doll{\'a}r and Boris Babenko and Serge J. Belongie and Pietro Perona and Zhuowen Tu},
        title = {Multiple Component Learning for Object Detection},
       editor = {David A. Forsyth and Philip H. S. Torr and Andrew Zisserman},
    booktitle = {10th European Conference on Computer Vision},
     location = {Marseille, France},
       volume = 2,
        pages = {211-224},
         year = 2008,
}
@inproceedings{DollaretalVS-PETS-05,
       author = {Piotr Doll\'{a}r and Vincent Rabaud and Garrison Cottrell and Serge Belongie},
        title = {Behavior Recognition via Sparse Spatio-Temporal Features},
    booktitle = {Second Joint {IEEE} International Workshop on Visual Surveillance and Performance Evaluation of Tracking and Surveillance},
        pages = 65,
     location = {Beijing},
         year = 2005,
}
@article{Dombaletal74,
       author = {{de Dombal}, F. and Leaper, D. and Horrocks, J. and Staniland, J. and McCann, A.},
        title = {Human and Computer-Aided Diagnosis of Abdominal Pain: Further Report with Emphasis on Performance},
      journal = {British Medical Journal},
       volume = 1,
        pages = {376-380},
         year = 1974,
}
@incollection{DombeckandTankCSH-11,
       author = {Daniel Dombeck and David Tank},
       editor = {Helmchen and Konnerth},
        title = {Imaging in Neuroscience},
    booktitle = {Two-Photon Imaging of Neural Activity in Awake Mobile Mice},
    publisher = {Cold Spring Harbor Press},
         year = {2011},
        pages = {355-369}
}
@article{DomenicoandRyanFiHN-17,
       author = {Di Domenico, S. I. and Ryan, R. M.},
        title = {The Emerging Neuroscience of Intrinsic Motivation: {A} New Frontier in Self-Determination Research},
      journal = {Frontiers in Human Neuroscience},
         year = {2017},
       volume = {11},
        pages = {145},
     abstract = {Intrinsic motivation refers to people's spontaneous tendencies to be curious and interested, to seek out challenges and to exercise and develop their skills and knowledge, even in the absence of operationally separable rewards. Over the past four decades, experimental and field research guided by self-determination theory (SDT; Ryan and Deci, 2017) has found intrinsic motivation to predict enhanced learning, performance, creativity, optimal development and psychological wellness. Only recently, however, have studies begun to examine the neurobiological substrates of intrinsic motivation. In the present article, we trace the history of intrinsic motivation research, compare and contrast intrinsic motivation to closely related topics (flow, curiosity, trait plasticity), link intrinsic motivation to key findings in the comparative affective neurosciences, and review burgeoning neuroscience research on intrinsic motivation. We review converging evidence suggesting that intrinsically motivated exploratory and mastery behaviors are phylogenetically ancient tendencies that are subserved by dopaminergic systems. Studies also suggest that intrinsic motivation is associated with patterns of activity across large-scale neural networks, namely, those that support salience detection, attentional control and self-referential cognition. We suggest novel research directions and offer recommendations for the application of neuroscience methods in the study of intrinsic motivation.}
}     
@book{Domingos2015,
       author = {The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World},
        title = {Pedro Domingos},
    publisher = {Basic Books},
      address = {New York, NY},
         year = 2015,
}
@misc{Don_Knuth_Yoda_of_Silicon_Valley,
       author = {Siobhan Roberts},
        title = {The Yoda of Silicon Valley},
         year = {2018},
     abstract = {Donald Knuth, master of algorithms, reflects on 50 years of his opus-in-progress, "The Art of Computer Programming."},
 howpublished = {The New York Times, December 19, 2018}
}
@article{DonahoeBEHAVIOR_AND_PHILOSOPHY-10,
       author = {John Donahoe},
        title = {{Man as Machine: A Review of "Memory and the Computational Brain" by C. R. Gallistel and A. P. King}},
      journal = {Behavior and Philosophy},
       volume = {38},
         year = {2010},
        pages = {83-101},
}     
@article{DonahueetalCoRR-14,
       author = {Jeff Donahue and Lisa Anne Hendricks and Sergio Guadarrama and Marcus Rohrbach and Subhashini Venugopalan and Kate Saenko and Trevor Darrell},
        title = {Long-term Recurrent Convolutional Networks for Visual Recognition and Description},
      journal = {CoRR},
       volume = {arXiv:1411.4389},
         year = 2014,
     abstract = {Models based on deep convolutional networks have dominated recent image interpretation tasks; we investigate whether models which are also recurrent, or "temporally deep", are effective for tasks involving sequences, visual and otherwise. We develop a novel recurrent convolutional architecture suitable for large-scale visual learning which is end-to-end trainable, and demonstrate the value of these models on benchmark video recognition tasks, image description and retrieval problems, and video narration challenges. In contrast to current models which assume a fixed spatio-temporal receptive field or simple temporal averag- ing for sequential processing, recurrent convolutional models are "doubly deep" in that they can be compositional in spatial and temporal "layers". Such models may have advantages when target concepts are complex and/or training data are limited. Learning long-term dependencies is possible when nonlinearities are incorporated into the network state updates. Long-term RNN models are appealing in that they directly can map variable-length inputs (e.g., video frames) to variable length outputs (e.g., natural lan- guage text) and can model complex temporal dynamics; yet they can be optimized with backpropagation. Our recurrent long-term models are directly connected to modern visual convnet models and can be jointly trained to simultaneously learn temporal dynamics and convolutional perceptual rep- resentations. Our results show such models have distinct advantages over state-of-the-art models for recognition or generation which are separately defined and/or optimized.}
}
@techreport{DonaldandJennings91,
       author = {Donald, Bruce and Jennings, Jim},
        title = {Sensor Interpretation and Task-Directed Planning Using Perceptual Equivalence Classes},
        issue = {CU-CS-TR},
  institution = {Cornell Unniversity Computer Science Department},
         year = 1991,
}     
@inproceedings{DongandLapataACL-16,
       author = {Li Dong and Mirella Lapata},
        title = {Language to Logical Form with Neural Attention},
    booktitle = {Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics},
        pages = {33-43},
         year = {2016},
    publisher = {Association for Computational Linguistics},
      address = {Stroudsburg, PA, USA},
     abstract = {Semantic parsing aims at mapping natural language to machine interpretable meaning representations. Traditional approaches rely on high-quality lexicons, manually-built templates, and linguistic features which are either domain- or representation-specific. In this paper we present a general method based on an attention-enhanced encoder-decoder model. We encode input utterances into vector representations, and generate their logical forms by conditioning the output sequences or trees on the encoding vectors. Experimental results on four datasets show that our approach performs competitively without using hand-engineered features and is easy to adapt across domains and meaning representations.}
}
@article {DongetalAM-15,
       author = {Dong, Ruijiao and Zhou, Yongfeng and Huang, Xiaohua and Zhu, Xinyuan and Lu, Yunfeng and Shen, Jian},
        title = {Functional Supramolecular Polymers for Biomedical Applications},
      journal = {Advanced Materials},
       volume = 27,
        issue = 3,
         year = 2015,
        pages = {498-526},
     abstract = {As a novel class of dynamic and non-covalent polymers, supramolecular polymers not only display specific structural and physicochemical properties, but also have the ability to undergo reversible changes of structure, shape, and function in response to diverse external stimuli, making them promising candidates for widespread applications ranging from academic research to industrial fields. By an elegant combination of dynamic/reversible structures with exceptional functions, functional supramolecular polymers are attracting increasing attention in various fields. In particular, functional supramolecular polymers offer several unique advantages, including inherent degradable polymer backbones, smart responsiveness to various biological stimuli, and the ease for the incorporation of multiple biofunctionalities (e.g. , targeting and bioactivity), thereby showing great potential for a wide range of applications in the biomedical field. In this Review, the trends and representative achievements in the design and synthesis of supramolecular polymers with specific functions are summarized, as well as their wide-ranging biomedical applications such as drug delivery, gene transfection, protein delivery, bio imaging and diagnosis, tissue engineering, and biomimetic chemistry. These achievements further inspire persistent efforts in an emerging interdisciplin ary research area of supramolecular chemistry, polymer science, material science, biomedical engineering, and nanotechnology.}
}
@article{DonnarummaetalAB-16,
       author = {Francesco Donnarumma and Roberto Prevete and Andrea de Giorgio and Guglielmo Montone and Giovanni Pezzulo},
        title = {Learning programs is better than learning dynamics: A programmable neural network hierarchical architecture in a multi-task scenario},
      journal = {Adaptive Behavior},
       volume = {24},
       number = {1},
        pages = {27-51},
         year = {2016},
     abstract = {Distributed and hierarchical models of control are nowadays popular in computational modeling and robotics. In the artificial neural network literature, complex behaviors can be produced by composing elementary building blocks or motor primitives, possibly organized in a layered structure. However, it is still unknown how the brain learns and encodes multiple motor primitives, and how it rapidly reassembles, sequences and switches them by exerting cognitive control. In this paper we advance a novel proposal, a hierarchical programmable neural network architecture, based on the notion of programmability and an interpreter-programmer computational scheme. In this approach, complex (and novel) behaviors can be acquired by embedding multiple modules (motor primitives) in a single, multi-purpose neural network. This is supported by recent theories of brain functioning in which skilled behaviors can be generated by combining functional different primitives embedded in "reusable" areas of "recycled" neurons. Such neuronal substrate supports flexible cognitive control, too. Modules are seen as interpreters of behaviors having controlling input parameters, or programs that encode structures of networks to be interpreted. Flexible cognitive control can be exerted by a programmer module feeding the interpreters with appropriate input parameters, without modifying connectivity. Our results in a multiple T-maze robotic scenario show how this computational framework provides a robust, scalable and flexible scheme that can be iterated at different hierarchical layers permitting to learn, encode and control multiple qualitatively different behaviors.}
}
@article {DonnarummaetalBIOXRIV-20,
       author = {Donnarumma, Francesco and Prevete, Roberto and Maisto, Domenico and Fuscone, Andrea and van der Meer, Matthijs A. A. and Kemere, Caleb and Pezzulo, Giovanni},
        title = {A framework to identify structured behavioral patterns within rodent spatial trajectories},
         year = {2020},
    publisher = {Cold Spring Harbor Laboratory},
      journal = {bioRxiv},
     abstract = {Animal behavior is highly structured. Yet, structured behavioral patterns {\textendash} or {\textquotedblleft}statistical ethograms{\textquotedblright} {\textendash} are not immediately apparent from the full spatiotemporal data that behavioral scientists usually collect. Here, we introduce a framework to characterize quantitatively rodent behavior during spatial (e.g., maze) navigation, in terms of movement building blocks or motor primitives. The hypothesis underlying this approach is that rodent behavior is characterized by a small number of motor primitives, which are combined over time to produce open-ended movements. We introduce a machine learning methodology {\textendash} dictionary learning {\textendash} which permits extracting motor primitives from rodent position and velocity data collected during spatial navigation and use them to both reconstruct past trajectories and predict novel ones. Three main results validate our approach. First, our method reconstructs rodent behavioral trajectories robustly from incomplete data, outper-forming approaches based on standard dimensionality reduction methods, such as principal component analysis. Second, the motor primitives extracted during one experimental session generalize and afford the accurate reconstruction of rodent behavior across successive experimental sessions in the same or in modified mazes. Third, the number of motor primitives that our method associates to each maze correlates with independent measures of maze complexity, hence showing that the motor primitives formalism is sensitive to essential aspects of task structure. The framework introduced here can be used by behavioral scientists and neuroscientists as an aid for behavioral and neural data analysis. Indeed, the extracted motor primitives enable the quantitative characterization of the complexity and similarity between different mazes and behavioral patterns across multiple trials (i.e., habit formation). We exemplify some uses of the method to control for confounding effects (e.g., of maze complexity on behavior and reward collection), analyze habitual or stereotyped behavior, classify or predict behavioral choices as well as place and grid cell displacement in new mazes. Keywords: maze navigation; spatial trajectories; motor primitives; ethograms},
}
@article{DonnarummaetalCS-12,
       author = {Francesco Donnarumma and Roberto Prevete and Giuseppe Trautteur},
        title = {Programming in the brain: a neural network theoretical framework},
      journal = {Connection Science},
    publisher = {Taylor \& Francis},
       volume = {24},
       number = {2-3},
         year = {2012},
        pages = {71-90},
     abstract = {Recent research shows that some brain areas perform more than one task and the switching times between them are incompatible with learning and that parts of the brain are controlled by other parts of the brain, or are "recycled", or are used and reused for various purposes by other neural circuits in different task categories and cognitive domains. All this is conducive to the notion of "programming in the brain". In this paper, we describe a programmable neural architecture, biologically plausible on the neural level, and we implement, test, and validate it in order to support the programming interpretation of the above-mentioned phenomenology. A programmable neural network is a fixed-weight network that is endowed with auxiliary or programming inputs and behaves as any of a specified class of neural networks when its programming inputs are fed with a code of the weight matrix of a network of the class. The construction is based on the "pulling out" of the multiplication between synaptic weights and neuron outputs and having it performed in "software" by specialised multiplicative-response fixed subnetworks. Such construction has been tested for robustness with respect to various sources of noise. Theoretical underpinnings, analysis of related research, detailed construction schemes, and extensive testing results are given. }
}
@article{DonnarummaetalIJNS-15,
       author = {Donnarumma, F. and Prevete, R. and Chersi, F. and Pezzulo, G.},
        title = {A Programmer-Interpreter Neural Network Architecture for Prefrontal Cognitive Control},
      journal = {International Journal Neural Systems},
         year = {2015},
       volume = {25},
       number = {6},
        pages = {1550017},
     abstract = {There is wide consensus that the prefrontal cortex (PFC) is able to exert cognitive control on behavior by biasing processing toward task-relevant information and by modulating response selection. This idea is typically framed in terms of top-down influences within a cortical control hierarchy, where prefrontal-basal ganglia loops gate multiple input-output channels, which in turn can activate or sequence motor primitives expressed in (pre-)motor cortices. Here we advance a new hypothesis, based on the notion of programmability and an interpreter-programmer computational scheme, on how the PFC can flexibly bias the selection of sensorimotor patterns depending on internal goal and task contexts. In this approach, multiple elementary behaviors representing motor primitives are expressed by a single multi-purpose neural network, which is seen as a reusable area of "recycled" neurons (interpreter). The PFC thus acts as a "programmer" that, without modifying the network connectivity, feeds the interpreter networks with specific input parameters encoding the programs (corresponding to network structures) to be interpreted by the (pre-)motor areas. Our architecture is validated in a standard test for executive function: the 1-2-AX task. Our results show that this computational framework provides a robust, scalable and flexible scheme that can be iterated at different hierarchical layers, supporting the realization of multiple goals. We discuss the plausibility of the "programmer-interpreter" scheme to explain the functioning of prefrontal-(pre)motor cortical hierarchies.}
}
@techreport{DonnerJUGGLE,
       author = {Donner, Marc D. and David H. Jameson},
        title = {A Real-Time Juggling Robot},
         type = {{IBM} Research Report},
        issue = {RC 12111 (54549)},
  institution = {IBM},
         year = 1986,
}
@book{DonnerPhD,
       author = {Donner, Marc D.},
        title = {Real-Time Control of Walking},
    publisher = {Birkhauser},
      address = {Boston, Massachusetts},
         year = 1987,
}
@article {DonosoetalSCIENCE-14,
       author = {Donoso, Ma{\"e}l and Collins, Anne G. E. and Koechlin, Etienne},
        title = {Foundations of human reasoning in the prefrontal cortex},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {344},
       number = {6191},
        pages = {1481-1486},
         year = {2014},
     abstract = {The brain{\textquoteright}s prefrontal cortex helps us to make decisions in an uncertain and constantly changing environment. Donoso et al. present a model of human reasoning as an algorithm implemented in the prefrontal cortex (see the Perspective by Hare). Brain-imaging experiments supported this model. Depending on the prevailing circumstances, human reasoning can either adapt ongoing behavioral strategies or switch to previously learned strategies. Only when neither approach is appropriate will the brain create new strategies.Science, this issue p. 1481, see also p. 1446The prefrontal cortex (PFC) subserves reasoning in the service of adaptive behavior. Little is known, however, about the architecture of reasoning processes in the PFC. Using computational modeling and neuroimaging, we show here that the human PFC has two concurrent inferential tracks: (i) one from ventromedial to dorsomedial PFC regions that makes probabilistic inferences about the reliability of the ongoing behavioral strategy and arbitrates between adjusting this strategy versus exploring new ones from long-term memory, and (ii) another from polar to lateral PFC regions that makes probabilistic inferences about the reliability of two or three alternative strategies and arbitrates between exploring new strategies versus exploiting these alternative ones. The two tracks interact and, along with the striatum, realize hypothesis testing for accepting versus rejecting newly created strategies.},
}
@book{Dorf1989,
       author = {Dorf, Richard C.},
        title = {Modern Control Systems},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1989,
}
@inproceedings{DoucetetalUAI-00,
       author = {Arnaud Doucet and Nando de Freitas and Kevin P. Murphy and Stuart J. Russell},
        title = {{R}ao-{B}lackwellised Particle Filtering for Dynamic {B}ayesian Networks},
    booktitle = {Proceedings of the 16th Conference on Uncertainty in Artificial Intelligence},
    publisher = {Morgan Kaufmann},
        pages = {176-183},
         year = 2000,
}
@article{DouglasandMartin2004,
        title = {Neuronal Circuits of the Neocortex},
       author = {Douglas, Rodney J. and Martin, Kevan A.C.},
      journal = {Annual Review of Neuroscience},
       volume = {27},
       number = {1},
         year = {2004},
        pages = {419-451},
     abstract = {We explore the extent to which neocortical circuits generalize, i.e., to what extent can neocortical neurons and the circuits they form be considered as canonical? We find that, as has long been suspected by cortical neuroanatomists, the same basic laminar and tangential organization of the excitatory neurons of the neocortex is evident wherever it has been sought. Similarly, the inhibitory neurons show characteristic morphology and patterns of connections throughout the neocortex. We offer a simple model of cortical processing that is consistent with the major features of cortical circuits: The superficial layer neurons within local patches of cortex, and within areas, cooperate to explore all possible interpretations of different cortical input and cooperatively select an interpretation consistent with their various cortical and subcortical inputs.}
}
@article{DouglasandMartin2007,
        title = {Recurrent neuronal circuits in the neocortex},
       author = {Douglas, Rodney J. and Martin, Kevan A.C.},
      journal = {Current Biology},
       volume = {18},
        issue = {13},
         year = {2007},
        pages = {496-500}
}
@article{DouglasandMartinCURRENT-BIOLOGY-12,
       author = {Rodney J. Douglas and Kevan A.C. Martin},
        title = {Behavioral architecture of the cortical sheet},
      journal = {Current Biology},
       volume = {22},
       number = {24},
         year = {2012},
        pages = {R1033-R1038},
     abstract = {The effortless ability of vertebrates to explore and exploit their environment is strongly correlated with the evolution of the most anterior part of their nervous system, the forebrain, where data from autonomic (visceral), limbic (emotive), and internal and peripheral sensors of the external world are combined to develop, decide, and deploy advantageous behaviors. The correlation of behavioral performance with forebrain expansion suggests that evolution has discovered the developmental means of building vertebrate brains to produce a scalable, special-purpose architecture for efficiently processing and expressing behavior. In mammals, the exuberant expansion of this forebrain is dominated by the growth of their cortex - the two-dimensional sheet that is the major source of their intelligent behavior, especially for primates.}
}     
@article{DouglasetalSCIENCE-95,
        title = {Recurrent Excitation in Neocortical Circuits},
       author = {Douglas, Rodney J. and Koch, Christof and Mahowald, Misha and Martin, Kevan A. C. and Suarez, Humbert H.},
      journal = {Science},
       volume = 269,
        issue = 18,
         year = 1995,
        pages = {981-985},
}
@article{DoyaCOiN-10,
       author = {Doya, Kenji},
        title = {Complementary roles of basal ganglia and cerebellum in learning and motor control},
      journal = {Current Opinion in Neurobiology},
         year = {2000},
       volume = {10},
       number = {6},
        pages = {732-739},
     abstract = {The classical notion that the basal ganglia and the cerebellum are dedicated to motor control has been challenged by the accumulation of evidence revealing their involvement in non-motor, cognitive functions. From a computational viewpoint, it has been suggested that the cerebellum, the basal ganglia, and the cerebral cortex are specialized for different types of learning: namely, supervised learning, reinforcement learning and unsupervised learning, respectively. This idea of learning-oriented specialization is helpful in understanding the complementary roles of the basal ganglia and the cerebellum in motor control and cognitive functions.}
}
@incollection{DoyaandSejnowskiNCN-99,
    author = {Kenji Doya and Terrence J. Sejnowski},
     title = {A Computational Model of Avian Song Learning},
 booktitle = {The New Cognitive Neurosciences},
    editor = {Michael S. Gazzaniga and Emilio Bizzi},
      year = 1999,
     pages = {469-481},
}
@techreport{Doyle88,
       author = {Doyle, Jon},
        title = {Artificial Intelligence and Rational Self-Government},
        issue = {CMU-CS-88-124},
  institution = {Carnegie-Mellon University Department of Computer Science},
         year = 1988,
}
@inproceedings{DoyleAAAI-83,
       author = {Doyle, Jon},
        title = {The ins and outs of reason maintenance},
    booktitle = {Proceedings of the Eighth International Joint Conference on Artificial Intelligence},
         year = 1983,
        pages = {349-351},
}
@inproceedings{DoyleAAAI-86,
       author = {Doyle, Richard and Atkinson, David and Doshi, Rajkumar},
        title = {Generating Perception Requests and Expectations to Verify the Execution of Plans},
    booktitle = {Proceedings AAAI-86},
     location = {Philadelphia, Pennsylvania},
 organization = {AAAI},
         year = 1986,
        pages = {65-69},
}
@inproceedings{DoyleIJCAI-85,
       author = {Doyle, Jon},
        title = {Reasoned Assumptions and Pareto Optimality},
    booktitle = {Proceedings {IJCAI} 9},
     location = {Los Angeles, California},
 organization = {IJCAII},
         year = 1985,
        pages = {87-90},
}
@inproceedings{DoyleMIS-89,
       author = {Doyle, Jon},
        title = {Reasoning, Representation, and Rational Self-Government},
    booktitle = {Methodologies for Intelligent Systems, 4},
         year = 1989,
       editor = {Zbigniew W. Ras},
        pages = {367-380},
    publisher = {North-Holland},
      address = {New York},
}
@techreport{DoylePhD,
       author = {Doyle, Jon},
        title = {A Model for Deliberation, Action, and Introspection},
        issue = {AI-TR-581},
  institution = {{MIT} {AI} Laboratory},
         year = 1980,
}
@article{DoyleTMS,
       author = {Doyle, Jon},
        title = {A truth maintenance system},
      journal = AIJ,
       volume = 12,
         year = 1979,
        pages = {231-272},
}
@article{DoyleandDeanAIM-97,
        title = {Strategic Directions in Artificial Intelligence ({R}eprint)},
       author = {Doyle, Jon and Dean, Thomas},
      journal = {{AI} Magazine},
       volume = 18,
        issue = 1,
        pages = {87-101},
         year = 1997,
}
@article{DoyleandDeanCSUR-96,
        title = {Strategic Directions in Artificial Intelligence ({R}eprinted in {{\em AI Magazine}} 18 (1997))},
       author = {Doyle, Jon and Dean, Thomas},
      journal = {{ACM} Computing Surveys},
       volume = 28,
        issue = 4,
        pages = {653-670},
         year = 1996,
}
@article{DoyleandWellmanAIJ-91,
       author = {Doyle, Jon and Wellman, Michael P.},
        title = {Impediments to Universal Preference-based Default Theories},
      journal = {Artificial Intelligence},
         year = 1991,
}
@incollection{DoyleandWellmanKR-89,
       author = {Doyle, Jon and Wellman, Michael P.},
        title = {Impediments to Universal Preference-Based Default Theories},
     crossref = {KR-89},
        pages = {94-102},
}
@article{DrakewetalBIORxiv-17,
       author = {Drakew, Alexander and Maier, Urban and Tippmann, Anja and Frotscher, Michael},
        title = {Single synapses control mossy cell firing},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Laboratory},
         year = {2017},
     abstract = {The function of mossy cells (MCs) in the dentate gyrus has remained elusive. Here we determined the functional impact of single mossy fibre (MF) synapses on MC firing in mouse entorhino-hippocampal slice cultures. We stimulated single MF boutons and recorded Ca2+ transients in the postsynaptic spine and unitary excitatory postsynaptic potentials (EPSPs) at the MC soma. Synaptic responses to single presynaptic stimuli varied strongly between different MF synapses, even if they were located on the same MC dendrite. Synaptic strengths ranged from subthreshold EPSPs to direct postsynaptic action potential (AP) generation. Induction of synaptic plasticity at these individual MF synapses resulted in potentiation or depression depending on the initially encountered synaptic state, indicating that synaptic transmission at MF synapses on MCs is determined by their previous functional history. With these unique functional properties MF-MC synapses control MC firing individually thereby enabling modulation of the dentate network by single granule cells.},
}
@article{DranovskyetalNEURON-11,
       author = {Alex Dranovsky and Alyssa M. Picchini and Tiffany Moadel and Alexander C. Sisti and Atsushi Yamada and Shioko Kimura and E. David Leonardo and Rene Hen},
        title = {Experience Dictates Stem Cell Fate in the Adult Hippocampus},
      journal = {Neuron},
       volume = {70},
       number = {5},
         year = {2011},
        pages = {908-923},
     abstract = {Adult hippocampal neurogenesis has been implicated in cognitive and emotional processes, as well as in response to antidepressant treatment. However, little is known about how the adult stem cell lineage contributes to hippocampal structure and function and how this process is modulated by the animal's experience. Here we perform an indelible lineage analysis and report that neural stem cells can produce expanding and persisting populations of not only neurons, but also stem cells in the adult hippocampus. Furthermore, the ratio of stem cells to neurons depends on experiences of the animal or the location of the stem cell. Surprisingly, social isolation facilitated accumulation of stem cells, but not neurons. These results show that neural stem cells accumulate in the adult hippocampus and that the stem cell-lineage relationship is under control of anatomic and experiential niches. Our findings suggest that, in the hippocampus, fate specification may act as a form of cellular plasticity for adapting to environmental changes.}
}
@inproceedings{DraperetalAIPS-94,
       author = {Draper, Denise and Hanks, Steve and Weld, Daniel},
        title = {Probabilistic Planning with Information Gathering and Contingent Execution},
    booktitle = {Second International Conference on {AI} Planning Systems},
         year = 1994,
}
@inproceedings{DraperetalSSS-94,
       author = {Draper, Denise and Hanks, Steve and Weld, Daniel},
        title = {Probabilistic Planning with Information Gathering and Contingent Execution},
    booktitle = {Proceedings of the {AAAI} Spring Symposium on Decision Theoretic Planning},
     location = {Stanford, California},
         year = 1994,
}
@article{DrewetalCSHLP-12,
        title = {Adult neurogenesis in the mammalian hippocampus: Why the dentate gyrus?},
       author = {Liam J. Drew and, Stefano Fusi and Ren\'{e} Hen},
      journal = {Learning and Memory},
         year = {2013},
       volume = {20},
        pages = {710-729},
     abstract = {In the adult mammalian brain, newly generated neurons are continuously incorporated into two networks: interneurons born in the subventricular zone migrate to the olfactory bulb, whereas the dentate gyrus (DG) of the hippocampus integrates locally born principal neurons. That the rest of the mammalian brain loses significant neurogenic capacity after the perinatal period suggests that unique aspects of the structure and function of DG and olfactory bulb circuits allow them to benefit from the adult generation of neurons. In this review, we consider the distinctive features of the DG that may account for it being able to profit from this singular form of neural plasticity. Approaches to the problem of neurogenesis are grouped as "bottom-up," where the phenotype of adult-born granule cells is contrasted to that of mature developmentally born granule cells, and "top-down," where the impact of altering the amount of neurogenesis on behavior is examined. We end by considering the primary implications of these two approaches and future directions.}
}
@article{DrewetalLEARNING-MEMORY-13,
       author = {Drew, L. J. and Fusi, S. and Hen, R.},
        title = {Adult neurogenesis in the mammalian hippocampus: {W}hy the dentate gyrus?},
      journal = {Learning and Memory},
       volume = {20},
       number = {12},
         year = {2013},
        pages = {710-729},
     abstract = {In the adult mammalian brain, newly generated neurons are continuously incorporated into two networks: interneurons born in the subventricular zone migrate to the olfactory bulb, whereas the dentate gyrus (DG) of the hippocampus integrates locally born principal neurons. That the rest of the mammalian brain loses significant neurogenic capacity after the perinatal period suggests that unique aspects of the structure and function of DG and olfactory bulb circuits allow them to benefit from the adult generation of neurons. In this review, we consider the distinctive features of the DG that may account for it being able to profit from this singular form of neural plasticity. Approaches to the problem of neurogenesis are grouped as "bottom-up," where the phenotype of adult-born granule cells is contrasted to that of mature developmentally born granule cells, and "top-down," where the impact of altering the amount of neurogenesis on behavior is examined. We end by considering the primary implications of these two approaches and future directions.}
}
@book{Drexler1992,
        title = {Nanosystems: Molecular Machinery, Manufacturing, and Computation},
       author = {K. Eric Drexler},
    publisher = {Wiley},
      address = {New York},
         year = 1992,
}
@incollection{DriessensENCYCLOPEDIA-MACHINE-LEARNING-10,
        title = {Relational Reinforcement Learning},
       author = {Driessens, Kurt},
       editor = {Sammut, Claude and Webb, Geoffrey I.},
    booktitle = {Encyclopedia of Machine Learning},
    publisher = {Springer US},
      address = {Boston, MA},
         year = {2010},
        pages = {857-862},
}
@techreport{DrineaandMitzenmacher00,
       author = {E. Drinea and M. Mitzenmacher},
        title = {Variations on random graph models of the {W}eb},
  institution = {Harvard University},
        issue = {Computer Science Technical Report TR-06-00},
         year = 2000,
}
@article{DrineasandMahoneyJMLR-05,
     author = {Drineas, Petros and Mahoney, Michael W.},
      title = {On the Nystr\"{o}m Method for Approximating a Gram Matrix for Improved Kernel-Based Learning},
    journal = {Journal of Machine Learning Research},
     volume = 6,
      pages = {2153-2175},
       year = 2005,
}
@article{DrissietalCoRR-18,
       author = {Program Language Translation Using a Grammar-Driven Tree-to-Tree Model},
        title = {Mehdi Drissi and Olivia Watkins and Aditya Khant and Vivaswat Ojha and Pedro Sandoval and Rakia Segev and Eric Weiner and Robert Keller},
      journal = {CoRR},
       volume = {arXiv:1807.01784},
         year = {2018},
     abstract = {The task of translating between programming languages differs from the challenge of translating natural languages in that programming languages are designed with a far more rigid set of structural and grammatical rules. Previous work has used a tree-to-tree encoder/decoder model to take advantage of the inherent tree structure of programs during translation. Neural decoders, however, by default do not exploit known grammar rules of the target language. In this paper, we describe a tree decoder that leverages knowledge of a language's grammar rules to exclusively generate syntactically correct programs. We find that this grammar-based tree-to-tree model outperforms the state of the art tree-to-tree model in translating between two programming languages on a previously used synthetic task.}
}
@article{DruckmannetalPLoS-11,
     title = {Effective Stimuli for Constructing Reliable Neuron Models},
    author = {Druckmann, Shaul and Berger, Thomas K. and Sch\"{u}rmann, Felix and Hill, Sean and Markram, Henry and Segev, Idan},
   journal = {{PLoS} Computational Biology},
 publisher = {Public Library of Science},
    volume = 7,
     issue = 8,
      year = 2011,
     pages = {e1002133},
  abstract = {Neurons perform complicated non-linear transformations on
  their input before producing their output - a train of action
  potentials. This input-output transformation is shaped by the
  specific composition of ion channels, out of the many possible
  types, that are embedded in the neuron's membrane. Experimentally,
  characterizing this transformation relies on injecting different
  stimuli to the neuron while recording its output; but which of the
  many possible stimuli should one apply? This combined experimental
  and theoretical study provides a general theoretical framework for
  answering this question, examining how different stimuli constrain
  the space of faithful conductance-based models of the studied
  neuron. We show that combinations of intracellular step and ramp
  currents enable the construction of models that both replicate the
  cell's response and generalize very well to novel stimuli e.g., to
  "noisy" stimuli mimicking synaptic activity. We experimentally
  verified our theoretical predictions on several cortical neuron
  types. This work presents a novel method for reliably linking the
  microscopic membrane ion channels to the macroscopic electrical
  behavior of neurons. It provides a much-needed rationale for
  selecting a particular stimulus set for studying the input-output
  properties of neurons and paves the way for standardization of
  experimental protocols along with construction of reliable neuron
  models.},
}
@incollection{DrummondKR-89,
       author = {Drummond, Mark},
        title = {Situated Control Rules},
     crossref = {KR-89},
}
@inproceedings{DrummondandBresinaAAAI-90,
       author = {Drummond, Mark and Bresina, John},
        title = {Anytime Synthetic Projection: Maximizing the Probability of Goal Satisfaction},
    booktitle = {Proceedings AAAI-90},
     location = {Boston, Massachusetts},
 organization = {AAAI},
        pages = {138-144},
         year = 1990,
}
@incollection{DrummondetalKBS-94,
       author = {Drummond, Mark and Swanson, Keith and Bresina, John},
        title = {Robust Scheduling and Execution for Automatic Telescopes},
    booktitle = {Knowledge-Based Scheduling},
       editor = {Zweben, M. and Fox, M.},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1994,
}
@article{DuanetalCoRR-17,
       author = {Yan Duan and Marcin Andrychowicz and Bradly C. Stadie and Jonathan Ho and Jonas Schneider and Ilya Sutskever and Pieter Abbeel and Wojciech Zaremba},
        title = {One-Shot Imitation Learning},
      journal = {CoRR},
       volume = {arXiv:1703.07326},
         year = {2017},
     abstract = {Imitation learning has been commonly applied to solve different tasks in isolation. This usually requires either careful feature engineering, or a significant number of samples. This is far from what we desire: ideally, robots should be able to learn from very few demonstrations of any given task, and instantly generalize to new situations of the same task, without requiring task-specific engineering. In this paper, we propose a meta-learning framework for achieving such capability, which we call one-shot imitation learning.

Specifically, we consider the setting where there is a very large set of tasks, and each task has many instantiations. For example, a task could be to stack all blocks on a table into a single tower, another task could be to place all blocks on a table into two-block towers, etc. In each case, different instances of the task would consist of different sets of blocks with different initial states. At training time, our algorithm is presented with pairs of demonstrations for a subset of all tasks.

A neural net is trained that takes as input one demonstration and the current state (which initially is the initial state of the other demonstration of the pair), and outputs an action with the goal that the resulting sequence of states and actions matches as closely as possible with the second demonstration. At test time, a demonstration of a single instance of a new task is presented, and the neural net is expected to perform well on new instances of this new task. The use of soft attention allows the model to generalize to conditions and tasks unseen in the training data. We anticipate that by training this model on a much greater variety of tasks and settings, we will obtain a general system that can turn any demonstrations into robust policies that can accomplish an overwhelming variety of tasks.}
}
@article{DubinandPatapoutianJCL-10,
       author = {Dubin, Adrienne E. and Patapoutian, Ardem},
        title = {Nociceptors: the sensors of the pain pathway},
      journal = {Journal of Clinical Investigation},
    publisher = {American Society for Clinical Investigation},
       volume = {120},
        issue = {11},
         year = {2010},
        pages = {3760-3772},
     abstract = {Specialized peripheral sensory neurons known as nociceptors alert us to potentially damaging stimuli at the skin by detecting extremes in temperature and pressure and injury-related chemicals, and transducing these stimuli into long-ranging electrical signals that are relayed to higher brain centers. The activation of functionally distinct cutaneous nociceptor populations and the processing of information they convey provide a rich diversity of pain qualities. Current work in this field is providing researchers with a more thorough understanding of nociceptor cell biology at molecular and systems levels and insight that will allow the targeted design of novel pain therapeutics.},
}
@article{DuchietalJMLR-11,
       author = {John Duchi and Elad Hazan and Yoram Singer},
        title = {Adaptive Subgradient Methods for Online Learning and Stochastic Optimization},
      journal = {Journal of Machine Learning Research},
       volume = 12,
         year = 2011,
        pages = {2121-2159},
     abstract = {We present a new family of subgradient methods that dynamically incorporate knowledge of the geometry of the data observed in earlier iterations to perform more informative gradient-based learning. Metaphorically, the adaptation allows us to find needles in haystacks in the form of very predictive but rarely seen features. Our paradigm stems from recent advances in stochastic optimization and online learning which employ proximal functions to control the gradient steps of the algorithm. We describe and analyze an apparatus for adaptively modifying the proximal function, which significantly simplifies setting a learning rate and results in regret guarantees that are provably as good as the best proximal function that can be chosen in hindsight. We give several efficient algorithms for empirical risk minimization problems with common and important regularization functions and domain constraints. We experimentally study our theoretical analysis and show that adaptive subgradient methods outperform state-of-the-art, yet non-adaptive, subgradient algorithms.}
}
@book{DudaHartandStork01,
       author = {Duda, R. O. and Hart, P. E. and David G. Stork},
        title = {Pattern Classification},
    publisher = {Second edition, John Wiley and Sons},
      address = {New York, NY},
         year = 2001,
}
@book{DudaandHart73,
       author = {Duda, R. O. and Hart, P. E.},
        title = {Pattern Classification and Scene Analysis},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1973,
}
@incollection{Dudaetal81,
       author = {Duda, R. O. and Hart, P. E. and Nilsson, N. J.},
        title = {Subjective {B}ayesian methods for rule-based inference systems},
       editor = {B.W. Webber and N.J. Nilsson},
    booktitle = {Readings in Artificial Intelligence},
    publisher = {Tioga},
      address = {Palo Alto, California},
         year = 1981,
}
@techreport{DudekTR-88,
       author = {Dudek, Gregory and Jenkins, Michael and Milios, Evangelos and Wilkes, David},
        title = {Robotic Exploration as Graph Construction},
        issue = {RBCV-TR-88-23},
  institution = {University of Toronto},
         year = 1988,
}
@article{DudewiczandDalalSANKHYA-75,
       author = {Dudewicz, E. and Dalal, S.},
        title = {Allocation of Observations in Ranking and Selection with Unequal Variances},
      journal = {Sankhya: The Indian Journal of Statistics},
       volume = 37,
        issue = {B},
        pages = {28-78},
         year = 1975,
}
@article{DueetalPLoS-11,
     title = {Multiplexed, high density electrophysiology with nanofabricated neural probes},
    author = {Du, J. and Blanche, T.J. and Harrison, R.R. and Lester, H.A. and Masmanidis, S.C.},
   journal = {{PLoS} One},
    volume = 6,
     issue = 10,
     pages = {e26204},
      year = 2011,
  abstract = {Extracellular electrode arrays can reveal the neuronal network correlates of behavior with single-cell, single-spike, and sub-millisecond resolution. However, implantable electrodes are inherently invasive, and efforts to scale up the number and density of recording sites must compromise on device size in order to connect the electrodes. Here, we report on silicon-based neural probes employing nanofabricated, high-density electrical leads. Furthermore, we address the challenge of reading out multichannel data with an application-specific integrated circuit (ASIC) performing signal amplification, band-pass filtering, and multiplexing functions. We demonstrate high spatial resolution extracellular measurements with a fully integrated, low noise 64-channel system weighing just 330 mg. The on-chip multiplexers make possible recordings with substantially fewer external wires than the number of input channels. By combining nanofabricated probes with ASICs we have implemented a system for performing large-scale, high-density electrophysiology in small, freely behaving animals that is both minimally invasive and highly scalable.}
}
@article{DuetalNATURE-17,
       author = {Du, Chao and Cai, Fuxi and Zidan, Mohammed A. and Ma, Wen and Lee, Seung Hwan and Lu, Wei D.},
        title = {Reservoir computing using dynamic memristors for temporal information processing},
      journal = {Nature Communications},
         year = {2017},
       volume = {8},
        issue = {1},
        pages = {2204},
     abstract = {Reservoir computing systems utilize dynamic reservoirs having short-term memory to project features from the temporal inputs into a high-dimensional feature space. A readout function layer can then effectively analyze the projected features for tasks, such as classification and time-series analysis. The system can efficiently compute complex and temporal data with low-training cost, since only the readout function needs to be trained. Here we experimentally implement a reservoir computing system using a dynamic memristor array. We show that the internal ionic dynamic processes of memristors allow the memristor-based reservoir to directly process information in the temporal domain, and demonstrate that even a small hardware system with only 88 memristors can already be used for tasks, such as handwritten digit recognition. The system is also used to experimentally solve a second-order nonlinear task, and can successfully predict the expected output without knowing the form of the original dynamic transfer function.},
}
@incollection{DuetalNIPS-09,
      title = {A Bayesian Model for Simultaneous Image Clustering, Annotation and Object Segmentation},
     author = {Lan Du and Lu Ren and David Dunson and Lawrence Carin},
  booktitle = {Advances in Neural Information Processing Systems 22},
     editor = {Y. Bengio and D. Schuurmans and J. Lafferty and C. K. I. Williams and A. Culotta},
      pages = {486-494},
       year = 2009
}
@article{DuistermarsetalFiBN-12,
       author = {Duistermars, Brian J. and Care, Rachel A. and Frye, Mark A.},
        title = {Binocular interactions underlying the classic optomotor responses of flying flies},
      journal = {Frontiers in Behavioral Neuroscience},
       volume = {6},
         year = {2012},
       number = {6},
     abstract = {In response to imposed course deviations, the fast optomotor reactions of animals reduce motion blur and facilitate the maintenance of stable body posture. In flies, the monocular front-to-back (progressive) and back-to-front (regressive) visual motion components generated by horizontal rotation are selectively encoded, respectively, by homo and heterolateral motion sensitive circuits in the third optic ganglion, the lobula plate. To investigate the strength of such inter-ocular interactions and their role in compensatory sensory-motor transformations, we utilize a virtual reality flight simulator to record optomotor reactions by tethered flying flies in response to imposed binocular and monocular visual rotation. With stimulus parameters generating large contrast insensitive optomotor responses to binocular rotation, we find that responses to monocular progressive motion are larger than those to panoramic rotation but contrast sensitive. Conversely, responses to monocular regressive motion are slower than those to rotation and peak at the lowest tested contrast. Together our results suggest that contrast insensitive optomotor responses to binocular rotation result from the dynamic interplay of contralateral inhibitory as well as excitatory circuit interactions and serve to maintain a stable optomotor equilibrium across a range of visual contrasts.}
}
@inproceedings{DumaisetalSIGIR-03,
      title = {{Stuff I've Seen}: a system for personal information retrieval and re-use},
     author = {Dumais, Susan and Cutrell, Edward and Cadiz, JJ and Jancke, Gavin and Sarin, Raman and Robbins, Daniel C.},
  booktitle = {Proceedings of the 26th annual international ACM SIGIR conference on Research and development in informaion retrieval},
   location = {Toronto, Canada},
  publisher = {ACM},
      pages = {72-79},
       year = 2003,
}
@article{DunbarandShultzRS-07,
    author = {Dunbar, R. I. M. and Shultz, Susanne},
     title = {{Understanding primate brain evolution}},
   journal = {Philosophical Transactions of the Royal Society B: Biological Sciences},
     issue = 1480,
     pages = {649-658},
    volume = 362,
      year = 2007,
  abstract = {We present a detailed reanalysis of the comparative brain data for primates, and develop a model using path analysis that seeks to present the coevolution of primate brain (neocortex  and sociality within a broader ecological and life-history framework. We show that body size, basal metabolic rate and life history act as constraints on brain evolution and through this influence the coevolution of neocortex size and group size. However, they do not determine either of these variables  which appear to be locked in a tight coevolutionary system. We show that, within primates, this relationship is specific to the neocortex. Nonetheless, there are important constraints on brain evolution; we use path analysis to show that, in order to evolve a large neocortex, a species must first evolve a large brain to support that neocortex and this in turn requires adjustments in diet (to provide the energy needed) and life history (to allow sufficient time both for brain growth for 'software' programming). We review a wider literature demonstrating a tight coevolutionary relationship between brain size and sociality in a range of mammalian taxa, but emphasize that the social brain hypothesis is not about the relationship between brain/neocortex size and group size; rather, it is about social complexity and we adduce evidence to support this. Finally, we consider the wider issue of how mammalian (and primate) brains evolve in order to localize the social effects.},
}
@article{DuncanTR-05,
        title = {Implanting Hope},
       author = {David Ewing Duncan},
      journal = {{MIT} Technology Review},
         year = 2005,
          url = {http://www2.technologyreview.com/articles/05/03/issue/feature_implant.asp},
}
@article{DuncanTiCS-10,
        title = {The multiple-demand (MD) system of the primate brain: mental programs for intelligent behaviour},
       author = {John Duncan},
      journal = {Trends in Cognitive Sciences},
    publisher = {Elsevier Current Trends},
       volume = {14},
       number = {4},
         year = {2010},
        pages = {172-179},
     abstract = {A common or multiple-demand (MD) pattern of frontal and parietal activity is associated with diverse cognitive demands, and with standard tests of fluid intelligence. In intelligent behaviour, goals are achieved by assembling a series of sub-tasks, creating structured mental programs. Single cell and functional magnetic resonance imaging (fMRI) data indicate a key role for MD cortex in defining and controlling the parts of such programs, with focus on the specific content of a current cognitive operation, rapid reorganization as mental focus is changed, and robust separation of successive task steps. Resembling the structured problem-solving of symbolic artificial intelligence, the mental programs of MD cortex appear central to intelligent thought and action.}
}
@book{DunnandClark74,
       author = {Dunn, Olive Jean and Clark, Virginia A.},
        title = {Applied Statistics: Analysis of Variance and Regression},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1974,
} 
@inproceedings{DunneandCampbell97,
    author = {R. A. Dunne and N. A. Campbell},
     title = {On The Pairing Of The Softmax Activation And Cross-Entropy Penalty 
             Functions And The Derivation Of The Softmax Activation Function},
 booktitle = {Proceedings of the 8th Australian Conference on Neural Networks},
  location = {Melbourne, Australia},
     pages = {181-185},
      year = 1997
}
@article {DunnetalELIFE-16,
       author = {Dunn, Timothy W. and Mu, Yu and Narayan, Sujatha and Randlett, Owen and Naumann, Eva A. and Yang, Chao-Tsung and Schier, Alexander F. and Freeman, Jeremy and Engert, Florian and Ahrens, Misha B.},
       editor = {Calabrese, Ronald L},
        title = {Brain-wide mapping of neural activity controlling zebrafish exploratory locomotion},
      journal = {eLife},
       volume = {5},
         year = {2016},
        pages = {e12741},
    publisher = {eLife Sciences Publications, Ltd},
     abstract = {In the absence of salient sensory cues to guide behavior, animals must still execute sequences of motor actions in order to forage and explore. How such successive motor actions are coordinated to form global locomotion trajectories is unknown. We mapped the structure of larval zebrafish swim trajectories in homogeneous environments and found that trajectories were characterized by alternating sequences of repeated turns to the left and to the right. Using whole-brain light-sheet imaging, we identified activity relating to the behavior in specific neural populations that we termed the anterior rhombencephalic turning region (ARTR). ARTR perturbations biased swim direction and reduced the dependence of turn direction on turn history, indicating that the ARTR is part of a network generating the temporal correlations in turn direction. We also find suggestive evidence for ARTR mutual inhibition and ARTR projections to premotor neurons. Finally, simulations suggest the observed turn sequences may underlie efficient exploration of local environments.},
}
@article{DunsworthetalPNAS-12,
       author = {Dunsworth, Holly M. and Warrener, Anna G. and Deacon, Terrence and Ellison, Peter T. and Pontzer, Herman},
        title = {Metabolic hypothesis for human altriciality},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {109},
       number = {38},
         year = {2012},
        pages = {15212-15216},
     abstract = {The classic anthropological hypothesis known as the 'obstetrical dilemma' is a well-known explanation for human altriciality, a condition that has significant implications for human social and behavioral evolution. The hypothesis holds that antagonistic selection for a large neonatal brain and a narrow, bipedal-adapted birth canal poses a problem for childbirth; the hominin 'solution' is to truncate gestation, resulting in an altricial neonate. This explanation for human altriciality based on pelvic constraints persists despite data linking human life history to that of other species. Here, we present evidence that challenges the importance of pelvic morphology and mechanics in the evolution of human gestation and altriciality. Instead, our analyses suggest that limits to maternal metabolism are the primary constraints on human gestation length and fetal growth. Although pelvic remodeling and encephalization during hominin evolution contributed to the present parturitional difficulty, there is little evidence that pelvic constraints have altered the timing of birth.},
}
@article{DunwellandPfeiferEPIGENOMICS-14,
       author = {Dunwell, T. L. and Pfeifer, G. P.},
        title = {Drosophila genomic methylation: new evidence and new questions},
      journal = {Epigenomics},
       volume = {6},
       number = {5},
         year = {2014},
        pages = {459-461},
     abstract = {The presence or absence of 5-methylcytosine in the genome of Drosophila has been a matter of disagreement in the literature for decades. We discuss some recent data that have contributed further to the complexity of this issue.}
}
@article{DupreandYusteCURRENT-BIOLOGY-17,
       author = {Dupre, C. and Yuste, R.},
        title = {{{N}on-overlapping {N}eural {N}etworks in {H}ydra vulgaris}},
      journal = {Current Biology},
         year = {2017},
       volume = {27},
       number = {8},
        pages = {1085-1097},
     abstract = {To understand the emergent properties of neural circuits, it would be ideal to record the activity of every neuron in a behaving animal and decode how it relates to behavior. We have achieved this with the cnidarian Hydra vulgaris, using calcium imaging of genetically engineered animals to measure the activity of essentially all of its neurons. Although the nervous system of Hydra is traditionally described as a simple nerve net, we surprisingly find instead a series of functional networks that are anatomically non-overlapping and are associated with specific behaviors. Three major functional networks extend through the entire animal and are activated selectively during longitudinal contractions, elongations in response to light, and radial contractions, whereas an additional network is located near the hypostome and is active during nodding. These results demonstrate the functional sophistication of apparently simple nerve nets, and the potential of Hydra and other basal metazoans as a model system for neural circuit studies.}
} 
@book{Durant1965,
        title = {The Story of Philosophy},
       author = {Durant, W.},
         year = {1965},
    publisher = {Simon and Schuster}
}
@inproceedings{DurfeeAAAI-90,
       author = {Durfee, Edmund H. and Montgomery, Thomas A.},
        title = {A Hierarchical Protocol for Coordinating Multiagent Behaviors},
    booktitle = {Proceedings AAAI-90},
     location = {Boston, Massachusetts},
 organization = {AAAI},
         year = 1990,
}    
@techreport{DurfeePhD,
       author = {Durfee, Edmund H.},
        title = {A Unified Approach to Dynamic Coordination: Planning Actions and Interactions in a Distributed Problem Solving Network},
        issue = {87-84},
  institution = {University of Massachusetts at Amherst Department of Computer and Information Science},
         year = 1987,
}
@inproceedings{DurfeeandLesserAAAI-86,
       author = {Durfee, Edmund H. and Lesser, Victor},
        title = {Incremental Planning to Control a Blackboard-Based Problem Solver},
    booktitle = {Proceedings AAAI-86},
     location = {Philadelphia, Pennsylvania},
 organization = {AAAI},
         year = 1986,
        pages = {58-64},
}
@book{Durrant-Whyte,
       author = {Durrant-Whyte, Hugh F.},
        title = {Integration, Coordination and Control of Multi-Sensor Robot Systems},
    publisher = {Kluwer},
      address = {Boston, Massachusetts},
         year = 1988,
} 
@inproceedings{DusparicandCahillATC-09,
       author = {Dusparic, I. and Cahill, V.},
        title = {Using Reinforcement Learning for Multi-policy Optimization in Decentralized Autonomic Systems: {A}n Experimental Evaluation},
       editor = {Gonz\'{a}lez Nieto J. and Reif W. and Wang G. and Indulska J.},
    booktitle = {Autonomic and Trusted Computing},
       series = {Lecture Notes in Computer Science Volume 5586},
    publisher = {Springer, Berlin, Heidelberg},
         year = {2009},
        pages = {105-119},
     abstract = {Large-scale autonomic systems are required to self-optimize with respect to high-level policies, that can differ in terms of their priority, as well as their spatial and temporal scope. Decentralized multi-agent systems represent one approach to implementing the required self-optimization capabilities. However, the presence of multiple heterogeneous policies leads to heterogeneity of the agents that implement them. In this paper we evaluate the use of Reinforcement Learning techniques to support the self-optimization of heterogeneous agents towards multiple policies in decentralized systems. We evaluate these techniques in an Urban Traffic Control simulation and compare two approaches to supporting multiple policies. Our results suggest that approaches based on W-learning, which learn separately for each policy and then select between nominated actions based on current action importance, perform better than combining policies into a single learning process over a single state space. The results also indicate that explicitly supporting multiple policies simultaneously can improve waiting times over policies dedicated to optimizing for a single vehicle type.}
}
@misc{DuvenaudEXPOSING-NEURAL-DIFFERENTIAL-EQUATIONS-19,
        title = {Noted ML Researcher Exposes His Own Paper At NeurIPS, Sets Healthy Precedent},
       author = {David Duvenaud},
 howpublished = {https://analyticsindiamag.com/neurips-2019-canada-neural-ode-david-duvenaud-researcher-flaws-ai},
         year = {2019},
}
@incollection{DuvenaudetalNIPS-15,
        title = {Convolutional Networks on Graphs for Learning Molecular Fingerprints},
       author = {Duvenaud, David K and Maclaurin, Dougal and Iparraguirre, Jorge and Bombarell, Rafael and Hirzel, Timothy and Aspuru-Guzik, Alan and Adams, Ryan P},
    booktitle = {Advances in Neural Information Processing Systems 28},
       editor = {C. Cortes and N. D. Lawrence and D. D. Lee and M. Sugiyama and R. Garnett},
    publisher = {Curran Associates, Inc.},
         year = {2015},
        pages = {2224-2232},
     abstract = {We introduce a convolutional neural network that operates directly on graphs. These networks allow end-to-end learning of prediction pipelines whose inputs are graphs of arbitrary size and shape. The architecture we present generalizes standard molecular feature extraction methods based on circular fingerprints. We show that these data-driven features are more interpretable, and have better predictive performance on a variety of tasks.}
}
@misc{DynamicMemoryNetworks-16,
        title = {Dynamic Memory Networks},
       author = {Anonymous},
         year = {2016},
 howpublished = {{LDGN} Notes and Thoughts on Machine Learning and Artificial Intelligence},
          url = {https://ldgn.wordpress.com/2016/01/04/attention-and-memory-in-deep-learning/},
     abstract = {Dynamic Memory Networks (DMN) are a general framework for question answering over inputs. Conceptually, a difference is made between inputs and questions. The DMN takes a sequence of inputs and a question and then employs an iterative attention process to compute the answer. The sequence of inputs can be seen as the history, which complements the general world knowledge (see semantic memory module). The DNM framework consists of five components: (i) input module: processes raw input and maps it to a useful representation, (ii) semantic memory module: stores general knowledge about concepts and facts. It can be instantiated by word embeddings or knowledge bases, (iii) question module: maps a question into a representation, (iv) episodic memory module: an iterative component that in each iteration focuses on different parts of the input, updates its internal state and finally outputs an answer representation (vi) answer module: generates the answer to return.}
}          
@book{Dyson1981disturbing,
        title = {Disturbing the Universe},
       author = {Dyson, Freeman},
         year = {1981},
    publisher = {Harper \& Row}
}
@book{Dyson1998darwin,
        title = {Darwin Among the Machines: The Evolution of Global Intelligence},
       author = {Dyson, George},
         year = {1998},
    publisher = {Perseus Books}
}
@book{Dyson1999origins,
        title = {Origins of Life},
       author = {Dyson, Freeman},
         year = {1999},
    publisher = {Cambridge University Press}
}
@book{Dyson2004infinite,
        title = {Infinite in All Directions},
       author = {Dyson, Freeman},
         year = {2004},
    publisher = {HarperCollins Publishers}
}
@article{Dyson2004strings,
        title = {The World on a String},
       author = {Dyson, Freeman},
      journal = {The New York Review of Books},
    publisher = {The New York Review of Books},
       volume = {51},
        issue = {8},
         year = {2004},
      comment = {Appears as a chapter in "The Scientist as Rebel"},
      comment = {/u/tld/Drive/write/articles/Dyson2004strings.txt},
}
@book{Dyson2006scientist,
        title = {The Scientist as Rebel},
       author = {Dyson, Freeman},
         year = {2006},
    publisher = {New York Review Books}
}
@book{Dyson2007many,
        title = {A Many-colored Glass: Reflections on the Place of Life in the Universe},
       author = {Dyson, Freeman},
         year = {2007},
    publisher = {University of Virginia Press}
}
@book{Dyson2012darwin,
        title = {Darwin Among the Machines: The Evolution of Global Intelligence},
       author = {Dyson, George},
         year = {2012},
    publisher = {Penguin Books Limited}
}
@book{Dyson2012turing,
        title = {Turing's Cathedral: The Origins of the Digital Universe},
       author = {Dyson, George},
         year = {2012},
    publisher = {Pantheon Books}
}           
@article{DzeroskietalML-01,
       author = {D{\v{z}}eroski, Sa{\v{s}}o and De Raedt, Luc and Driessens, Kurt},
        title = {Relational Reinforcement Learning},
      journal = {Machine Learning},
       volume = {43},
         year = {2001},
        pages = {7-52},
     abstract = {Relational reinforcement learning is presented, a learning technique that combines reinforcement learning with relational learning or inductive logic programming. Due to the use of a more expressive representation language to represent states, actions and Q-functions, relational reinforcement learning can be potentially applied to a new range of learning tasks. One such task that we investigate is planning in the blocks world, where it is assumed that the effects of the actions are unknown to the agent and the agent has to learn a policy. Within this simple domain we show that relational reinforcement learning solves some existing problems with reinforcement learning. In particular, relational reinforcement learning allows us to employ structural representations, to abstract from specific goals pursued and to exploit the results of previous learning phases when addressing new (more complex) situations.}
}
@inproceedings{DzierzanowskietalIAAI-92,
        title = {The Credit Assistant: The Second Leg in the Knowledge Highway for American Express},
       author = {Dzierzanowski, James and Hestenes, Eric and Lawson, Susan},
    booktitle = {Proceedings of the Conference on Innovative Applications of Artificial Intelligence},
        pages = {127-134},
         year = 1992,
}
@book{EAI-87,
       editor = {Shapiro, Stuart},
        title = {Encyclopedia of Artificial Intelligence},
    booktitle = {Encyclopedia of Artificial Intelligence},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1987,
}
@article{EarleyCACM-70,
       author = {Earley, J.},
        title = {An Efficient Context-Free Parsing Algorithm},
      journal = {Communications of the {ACM}},
       volume = 13,
        issue = 2,
         year = 1970,
        pages = {94-102},
}
@book{EasleyandKleinberg2011,
    author = {David Easley and Jon Kleinberg},
     title = {Networks, Crowds, and Markets: Reasoning about a Highly Connected World},
 publisher = {Cambridge University Press},
      year = 2011, 
}
@incollection{Eastonetal2008episodic,
        title = {Chapter 2.3 A new working definition of episodic memory: replacing "when" with "which"},
       author = {Alexander Easton and Madeline J. Eacott},
    booktitle = {Handbook of Episodic Memory},
       editor = {Ekrem Dere and Alexander Easton and Lynn Nadel and Joseph P. Huston},
       series = {Handbook of Behavioral Neuroscience},
    publisher = {Elsevier},
       volume = {18},
        pages = {185-196},
         year = {2008},
     abstract = {We often consider episodic memory in operational terms as what happened, where, and when. Here we consider that the memory of when an event occurred might not be especially important in defining a memory, or even be effectively processed within the episodic memory system. Therefore, we reconsider the relevance of a memory for time to episodic memory and consider an alternative that of "which occasion." This alternative can encompass elements of temporal coding when necessary or important, but does not restrict the memory to having a temporal element. Redefining episodic memory in this way allows us to develop alternative tests of episodic memory in animals, which provide an opportunity to explore the neural basis of episodic memory.}
}
@article{EbeidaetalACM-11,
       author = {Mohamed S. Ebeida and Anjul Patney and Scott A. Mitchell and Andrew Davidson and Patrick M. Knupp and John D. Owens},
        title = {Efficient Maximal {P}oisson-Disk Sampling},
      journal = {{ACM} Transactions on Graphics},
         year = 2011,
       volume = 30,
        issue = 4,
        pages = {49:1-49:12},
     abstract = {We solve the problem of generating a uniform Poisson-disk sampling that is both maximal and unbiased over bounded non-convex domains. To our knowledge this is the first provably correct algorithm with time and space dependent only on the number of points produced. Our method has two phases, both based on classical dartthrowing. The first phase uses a background grid of square cells to rapidly create an unbiased, near-maximal covering of the domain. The second phase completes the maximal covering by calculating the connected components of the remaining uncovered voids, and by using their geometry to efficiently place unbiased samples that cover them. The second phase converges quickly, overcoming a common difficulty in dart-throwing methods. The deterministic memory is O(n) and the expected running time is O(n log n), where n is the output size, the number of points in the final sample. Our serial implementation verifies that the log n dependence is minor, and nearly O(n) performance for both time and memory is achieved in practice. We also present a parallel implementation on GPUs to demonstrate the parallel-friendly nature of our method, which achieves 2.4 the performance of our serial version.}
}
@article{EberleetalJM-15,
       author = {Eberle, A.L. and Mikula, S. and Schalek, R. and Lichtman, J.W. and Tate, M.L. and Zeidler, D.},
        title = {High-resolution, high-throughput imaging with a multibeam scanning electron microscope},
      journal = {Journal of Microscopy},
         year = {2015},
     abstract = {Electron-electron interactions and detector bandwidth limit the maximal imaging speed of single-beam scanning electron microscopes. We use multiple electron beams in a single column and detect secondary electrons in parallel to increase the imaging speed by close to two orders of magnitude and demonstrate imaging for a variety of samples ranging from biological brain tissue to semiconductor wafers.},
}
@article{EbertetalCoRR-18,
       author = {Frederik Ebert and Chelsea Finn and Sudeep Dasari and Annie Xie and Alex X. Lee and Sergey Levine},
        title = {Visual Foresight: Model-Based Deep Reinforcement Learning for Vision-Based Robotic Control},
      journal = {CoRR},
       volume = {arXiv:1812.00568},
         year = {2018},
     abstract = {Deep reinforcement learning (RL) algorithms can learn complex robotic skills from raw sensory inputs, but have yet to achieve the kind of broad generalization and applicability demonstrated by deep learning methods in supervised domains. We present a deep RL method that is practical for real-world robotics tasks, such as robotic manipulation, and generalizes effectively to never-before-seen tasks and objects. In these settings, ground truth reward signals are typically unavailable, and we therefore propose a self-supervised model-based approach, where a predictive model learns to directly predict the future from raw sensory readings, such as camera images. At test time, we explore three distinct goal specification methods: designated pixels, where a user specifies desired object manipulation tasks by selecting particular pixels in an image and corresponding goal positions, goal images, where the desired goal state is specified with an image, and image classifiers, which define spaces of goal states. Our deep predictive models are trained using data collected autonomously and continuously by a robot interacting with hundreds of objects, without human supervision. We demonstrate that visual MPC can generalize to never-before-seen objects---both rigid and deformable---and solve a range of user-defined object manipulation tasks using the same model.}
}
@article{EbongandMazumderIEEE-12,
       author = {Ebong, I.E. and Mazumder, P.},
      journal = {Proceedings of the IEEE},
        title = {{CMOS} and Memristor-Based Neural Network Design for Position Detection},
         year = 2012,
       volume = 100,
        issue = 6,
        pages = {2050-2060},
     abstract = {Most hardware neural networks have a basic competitive learning rule on top of a more involved processing algorithm. This work highlights two basic learning rules/behavior: winner-take-all (WTA) and spike-timing-dependent plasticity (STDP). It also gives a design example implementing WTA combined with STDP in a position detector. A complementary metal-oxide-semiconductor (CMOS) and a memristor-MOS technology (MMOST) design simulation results are compared on the bases of power, area, and noise handling capabilities. Design and layout were done in 130-nm IBM process for CMOS, and the HSPICE model files for the process were used to simulate the CMOS part of the MMOST design. CMOS consumes area, 55-W max power, and requires a 3-dB SNR. On the other hand, the MMOST design consumes , 15-W max power, and requires a 4.8-dB SNR. There is a potential to improve upon analog computing with the adoption of MMOST designs.},
}
@book{Ecclesetal1967,
        title = {The Cerebellum as a Neuronal Machine},
       author = {Eccles, J.C. and Ito, M. and Szent{\'{a}}gothai, J.},
         year = {1967},
    publisher = {Springer-Verlag}
}
@book{Ecclesetal1967neuronal,
        title = {The Cerebellum as a Neuronal Machine},
       author = {Eccles, J.C. and Ito, M. and Szent\'{a}gothai, J.},
         year = {1967},
    publisher = {Springer-Verlag, Berlin}
}
@book{Eck95,
        title = {The Most Complex Machine: A Survey of Computers and Computing},
       author = {David Eck},
    publisher = {AK Peters Ltd},
      address = {Wellesley, MA},
         year = 1995,
}
@article{EckmannandRuelle85,
       author = {Eckmann, J.-P. and Ruelle, D.},
        title = {Ergodic Theory of Chaos and Strange Attractors},
      journal = {Rev. Mod. Phys.},
       volume = 57,
        pages = {617-656},
         year = 1985,
}
@article{EdelmanandGallyFiIN-13,
       author = {Edelman, Gerald M. and Gally, Joseph A.},
        title = {Reentry: a key mechanism for integration of brain function},
      journal = {Frontiers in Integrative Neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = {7},
        pages = {63},
         year = {2013},
     abstract = {Reentry in nervous systems is the ongoing bidirectional exchange of signals along reciprocal axonal fibers linking two or more brain areas. The hypothesis that reentrant signaling serves as a general mechanism to couple the functioning of multiple areas of the cerebral cortex and thalamus was first proposed in 1977 and 1978 (Edelman, 1978). A review of the amount and diversity of supporting experimental evidence accumulated since then suggests that reentry is among the most important integrative mechanisms in vertebrate brains (Edelman, 1993). Moreover, these data prompt testable hypotheses regarding mechanisms that favor the development and evolution of reentrant neural architectures.},
}
@book{EdelmanandTononi2000universe,
        title = {A Universe of Consciousness: How Matter Becomes Imagination},
       author = {Edelman, Gerald M. and Tononi, Giulio},
         year = {2000},
    publisher = {Basic Books}
}
@inproceedings{EdelmanetalNIPS-01,
       author = {Shimon Edelman and Benjamin P. Hiles and Hwajin Yang and Nathan Intrator},
        title = {Probabilistic principles in unsupervised learning of visual structure: human data and a model},
    booktitle = {Advances in Neural Information Processing Systems 14},
       editor = {Thomas G. Dietterich and Suzanna Becker and Zoubin Ghahramani},
         year = 2001,
        pages = {19-26},
}
@book{EdelsbrunnerandHarer2010,
        title = {Computational Topology - an Introduction},
       author = {Edelsbrunner, Herbert and Harer, John},
    publisher = {American Mathematical Society},
         year = 2010
}
@incollection{EdelsbrunnerandHarerAMS-08,
    booktitle = {Surveys on Discrete and Computational Geometry. Twenty Years Later},
        title = {Persistent Homology - a Survey},
       author = {Herbert Edelsbrunner and John Harer},
       editor = {J. E. Goodman, J. Pach and R. Pollack},
       series = {Contemporary Mathematics},
    publisher = {American Mathematical Society},
         year = {2008},
        pages = {257-282},
     abstract = {Persistent homology is an algebraic tool for measuring topological features of shapes and functions. It casts the multi-scale organization we frequently observe in nature into a mathematical formalism. Here we give a record of the short history of persistent homology and present its basic concepts. Besides the mathematics we focus on algorithms and mention the various connections to applications, including to biomolecules, biological networks, data analysis, and geometric modeling.}
}
@book{Edelson99,
      author = {Edward Edelson},
       title = {Gregor Mendel and the Roots of Genetics},
   publisher = {Oxford University Press},
     address = {Oxford},
        year = 1999,
}
@book{EdmundsandEidinow,
       author = {David Edmunds and John Eidinow},
        title = {Wittgenstein's Poker},
    publisher = {HarperCollins Publishers},
      address = {New York},
         year = 2001,
}
@article{EdwardsHF-62,
       author = {Edwards, Ward},
        title = {Dynamic Decision Theory and Probabilistic Information Processing},
      journal = {Human Factors},
       volume = 4,
         year = 1962,
        pages = {59-73},
}
@article{EdwardsetalNEUROINFORMATICS-14,
       author = {Edwards, J. and Daniel, E. and Kinney, J. and Bartol, T. and Sejnowski, T. J. and Harris, K. and Johnston, D. and Bajaj, C.},
        title = {{VolRoverN}: Automated reconstruction of cellular morphology for multiscale dynamical simulation of neural activity},
      journal = {Neuroinformatics},
       volume = 2,
        issue = 2,
         year = 2014,
        pages = {277-289},
     abstract = {Establishing meaningful relationships between cellular structure and function requires accurate morphological reconstructions. In particular, there is an unmet need for high quality surface reconstructions to model subcellular and synaptic interactions among neurons at nanometer resolution. We address this need with VolRoverN, a software package that produces accurate, efficient, and automated {3D} surface reconstructions from stacked 2D contour tracings. While many techniques and tools have been developed in the past for {3D} visualization of cellular structure, the reconstructions from VolRoverN meet specific quality criteria that are important for dynamical simulations. These criteria include manifoldness, water-tightness, lack of self- and object-object-intersections, and geometric accuracy. These enhanced surface reconstructions are readily extensible to any cell type (including glia) and are used here on complex spiny dendrites and axons from mature rat hippocampal area CA1. Both spatially realistic surface reconstructions and reduced skeletonizations are produced and formatted by VolRoverN for easy input into analysis software packages for neurophysiological simulations at multiple spatial and temporal scales ranging from ion electro-diffusion to electrical cable models.},
}
@inproceedings{EfrosandLeungICCV-99,
       author = {Alexei Efros and Thomas Leung},
        title = {Texture Synthesis by Non-parametric Sampling},
    booktitle = {Proceedings of the Seventh {IEEE} International Conference on Computer Vision},
    publisher = {IEEE Computer Society},
        pages = {1033-1038},
         year = 1999,
}
@inproceedings{EfrosetalICCV-03,
     author = {Alexei A. Efros and Er C. Berg and Greg Mori and Jitendra Malik},
      title = {Recognizing action at a distance},
  booktitle = {Proceedings of the International Conference on Computer Vision (ICCV 2003)}, 
       year = 2003,
      pages = {726-733}
}
@article{EglenetalGIGASCIENCE-14,
       author = {Eglen, Stephen and Weeks, Michael and Jessop, Mark and Simonotto, Jennifer and Jackson, Tom and Sernagor, Evelyne},
        title = {A data repository and analysis framework for spontaneous neural activity recordings in developing retina},
      journal = {GigaScience},
       volume = 3,
         year = 2014,
        issue = 1,
        pages = 3,
          url = {http://www.gigasciencejournal.com/content/3/1/3},
     abstract = {BACKGROUND: During early development, neural circuits fire spontaneously, generating activity episodes with complex spatiotemporal patterns. Recordings of spontaneous activity have been made in many parts of the nervous system over the last 25 years, reporting developmental changes in activity patterns and the effects of various genetic perturbations. RESULTS: We present a curated repository of multielectrode array recordings of spontaneous activity in developing mouse and ferret retina. The data have been annotated with minimal metadata and converted into HDF5. This paper describes the structure of the data, along with examples of reproducible research using these data files. We also demonstrate how these data can be analysed in the CARMEN workflow system. This article is written as a literate programming document; all programs and data described here are freely available. CONCLUSIONS: 1. We hope this repository will lead to novel analysis of spontaneous activity recorded in different laboratories. 2. We encourage published data to be added to the repository. 3. This repository serves as an example of how multielectrode array recordings can be stored for long-term reuse.},
}
@article{EhrlichandSchuffnyFiN-13,
       author = {Ehrlich, Matthias and Sch\"{u}ffny, Ren\'{e}},
        title = {Neural schematics as a unified formal graphical representation of large-scale Neural Network Structures},
      journal = {Frontiers in Neuroinformatics},
         year = {2013},
    publisher = {Frontiers Media},
       volume = {7},
        pages = {22},
     abstract = {One of the major outcomes of neuroscientific research are models of Neural Network Structures (NNSs). Descriptions of these models usually consist of a non-standardized mixture of text, figures, and other means of visual information communication in print media. However, as neuroscience is an interdisciplinary domain by nature, a standardized way of consistently representing models of NNSs is required. While generic descriptions of such models in textual form have recently been developed, a formalized way of schematically expressing them does not exist to date. Hence, in this paper we present Neural Schematics as a concept inspired by similar approaches from other disciplines for a generic two dimensional representation of said structures. After introducing NNSs in general, a set of current visualizations of models of NNSs is reviewed and analyzed for what information they convey and how their elements are rendered. This analysis then allows for the definition of general items and symbols to consistently represent these models as Neural Schematics on a two dimensional plane. We will illustrate the possibilities an agreed upon standard can yield on sampled diagrams transformed into Neural Schematics and an example application for the design and modeling of large-scale NNSs.},
}
@article{EichenbaumNATURE-REVIEWS-NEUROSCIENCE-14,
       author = {Eichenbaum, Howard},
        title = {Time cells in the hippocampus: {A} new dimension for mapping memories},
      journal = {Nature Reviews Neuroscience},
    publisher = {Nature Publishing Group},
         year = {2014},
        pages = {732},
       volume = {15},
     abstract = {Recent studies have revealed the existence of hippocampal neurons that fire at successive moments in temporally structured experiences. Several studies have shown that such temporal coding is not attributable to external events, specific behaviours or spatial dimensions of an experience. Instead, these cells represent the flow of time in specific memories and have therefore been dubbed 'time cells'. The firing properties of time cells parallel those of hippocampal place cells; time cells thus provide an additional dimension that is integrated with spatial mapping. The robust representation of both time and space in the hippocampus suggests a fundamental mechanism for organizing the elements of experience into coherent memories.}
}
@article{EichenbaumNEURON-04,
        title = {Hippocampus: Cognitive Processes and Neural Representations that Underlie Declarative Memory},
       author = {Howard Eichenbaum},
      journal = {Neuron},
       volume = {44},
       number = {1},
        pages = {109-120},
         year = {2004},
     abstract = {The hippocampus serves a critical role in declarative memory-our capacity to recall everyday facts and events. Recent studies using functional brain imaging in humans and neuropsychological analyses of humans and animals with hippocampal damage have revealed some of the elemental cognitive processes mediated by the hippocampus. In addition, recent characterizations of neuronal firing patterns in behaving animals and humans have suggested how neural representations in the hippocampus underlie those elemental cognitive processes in the service of declarative memory.}
}
@article{EichleretalNATURE-17,
       author = {Eichler, Katharina and Li, Feng and Litwin-Kumar, Ashok and Park, Youngser and Andrade, Ingrid and Schneider-Mizell, Casey M. and Saumweber, Timo and Huser, Annina and Eschbach, Claire and Gerber, Bertram and Fetter, Richard D. and Truman, James W. and Priebe, Carey E. and Abbott, L. F. and Thum, Andreas S. and Zlatic, Marta and Cardona, Albert},
        title = {The complete connectome of a learning and memory centre in an insect brain},
      journal = {Nature},
         year = {2017},
    publisher = {Macmillan Publishers Limited},
       volume = {548},
        issue = {7666},
        pages = {175-182},
     abstract = {Associating stimuli with positive or negative reinforcement is essential for survival, but a complete wiring diagram of a higher-order circuit supporting associative memory has not been previously available. Here we reconstruct one such circuit at synaptic resolution, the Drosophila larval mushroom body. We find that most Kenyon cells integrate random combinations of inputs but that a subset receives stereotyped inputs from single projection neurons. This organization maximizes performance of a model output neuron on a stimulus discrimination task. We also report a novel canonical circuit in each mushroom body compartment with previously unidentified connections: reciprocal Kenyon cell to modulatory neuron connections, modulatory neuron to output neuron connections, and a surprisingly high number of recurrent connections between Kenyon cells. Stereotyped connections found between output neurons could enhance the selection of learned behaviours. The complete circuit map of the mushroom body should guide future functional studies of this learning and memory centre.},
}
@article{EigenetalCoRR-13,
        title = {Learning Factored Representations in a Deep Mixture of Experts},
       author = {David Eigen and Marc'Aurelio Ranzato and Ilya Sutskever},
      journal = {CoRR},
       volume = {arXiv:1312.4314},
         year = {2013},
     abstract = {Mixtures of Experts combine the outputs of several "expert" networks, each of which specializes in a different part of the input space. This is achieved by training a "gating" network that maps each input to a distribution over the experts. Such models show promise for building larger networks that are still cheap to compute at test time, and more parallelizable at training time. In this this work, we extend the Mixture of Experts to a stacked model, the Deep Mixture of Experts, with multiple sets of gating and experts. This exponentially increases the number of effective experts by associating each input with a combination of experts at each layer, yet maintains a modest model size. On a randomly translated version of the MNIST dataset, we find that the Deep Mixture of Experts automatically learns to develop location-dependent ("where") experts at the first layer, and class-specific ("what") experts at the second layer. In addition, we see that the different combinations are in use when the model is applied to a dataset of speech monophones. These demonstrate effective use of all expert combinations.},
}
@article {EisenbergetalSCIENCE-03,
       author = {Eisenberg, Mark and Kobilo, Tali and Berman, Diego E. and Dudai, Yadin},
        title = {Stability of Retrieved Memory: Inverse Correlation with Trace Dominance},
      journal = {Science},
       volume = {301},
       number = {5636},
        pages = {1102-1104},
         year = {2003},
    publisher = {American Association for the Advancement of Science},
     abstract = {In memory consolidation, the memory trace stabilizes and becomes resistant to certain amnesic agents. The textbook account is that for any memorized item, consolidation starts and ends just once. However, evidence has accumulated that upon activation in retrieval, the trace may reconsolidate. Whereas some authors reported transient renewed susceptibility of retrieved memories to consolidation blockers, others could not detect it. Here, we report that in both conditioned taste aversion in the rat and fear conditioning in the medaka fish, the stability of retrieved memory is inversely correlated with the control of behavior by that memory. This result may explain some conflicting findings on reconsolidation of activated memories.},
}
@inproceedings {El-KahkyetalICASSP-14,
       author = {Ali El-Kahky and Derek Liu and Ruhi Sarikaya and Gokhan Tur and Dilek Hakkani-Tur and Larry Heck},
    booktitle = {{IEEE} International Conference on Acoustics, Speech, and Signal Processing},
        title = {Extending Domain Coverage of Language Understanding Systems via Intent Transfer Between Domains Using Knowledge Graphs and Search Query Click Logs},
          url = {http://research.microsoft.com/apps/pubs/default.aspx?id=210484},
         year = 2014,
     abstract = {This paper proposes a new technique to enable Natural Language Understanding (NLU) systems to handle user queries beyond their original semantic schemas defined by their intents and slots. Knowledge graph and search query logs are used to extend NLU system's coverage by transferring intents from other domains to a given domain. The transferred intents as well as existing intents are then applied to a set of new slots that they are not trained with. The knowledge graph and search click logs are used to determine whether the new slots (i.e. entities) or their attributes in the graph can be used together with the new intents without re-training the underlying NLU models with the expanded (i.e. with new intents and slots) schema. Experimental results show that the proposed technique can in fact be used in extending NLU system's domain coverage in fulfilling the user's request.},
}
@article{Elber-DorozkoandLoewensteinELIFE-18,
       author = {Elber-Dorozko, Lotem and Loewenstein, Yonatan},
        title = {Striatal action-value neurons reconsidered},
      journal = {eLife},
         year = {2018},
    publisher = {eLife Sciences Publications, Ltd},
       volume = {7},
        pages = {e34248},
     abstract = {It is generally believed that during economic decisions, striatal neurons represent the values associated with different actions. This hypothesis is based on studies, in which the activity of striatal neurons was measured while the subject was learning to prefer the more rewarding action. Here we show that these publications are subject to at least one of two critical confounds. First, we show that even weak temporal correlations in the neuronal data may result in an erroneous identification of action-value representations. Second, we show that experiments and analyses designed to dissociate action-value representation from the representation of other decision variables cannot do so. We suggest solutions to identifying action-value representation that are not subject to these confounds. Applying one solution to previously identified action-value neurons in the basal ganglia we fail to detect action-value representations. We conclude that the claim that striatal neurons encode action-values must await new experiments and analyses.},
}
@book{Eliasmith2013,
        title = {How to Build a Brain: A Neural Architecture for Biological Cognition},
       author = {Eliasmith, Chris},
       series = {Oxford Series on Cognitive Modeling},
         year = {2013},
    publisher = {Oxford University Press {USA}},
     abstract = {One goal of researchers in neuroscience, psychology, and artificial intelligence is to build theoretical models that are able to explain the flexibility and adaptiveness of biological systems. How to build a brain provides a detailed guided exploration of a new cognitive architecture that takes biological detail seriously, while addressing cognitive phenomena. The Semantic Pointer Architecture (SPA) introduced in this book provides a set of tools for constructing a wide range of biologically constrained perceptual, cognitive, and motor models.}
}
@inproceedings{ElidanetalUAI-06,
        title = {Residual Belief Propagation: {I}nformed Scheduling for Asynchronous Message Passing},
       author = {G. Elidan and I. McGraw and D. Koller},
    booktitle = {Proceedings of the Twenty-second Conference on Uncertainty in AI (UAI)},
      address = {Boston, Massachussetts},
         year = 2006,
}   
@article{EllisQJEP-A-96,
       author = {Nick C. Ellis},
        title = {Working Memory in the Acquisition of Vocabulary and Syntax: Putting Language in Good Order},
      journal = {The Quarterly Journal of Experimental Psychology Section A},
    publisher = {Routledge},
       volume = {49},
       number = {1},
         year = {1996},
        pages = {234-250},
     abstract = {This paper argues that working memory is heavily involved in language acquisition as (a) a major part of language learning is the learning of sequences, (b) working memory allows short-term maintenance of sequence information, and (c) short-term rehearsal of sequences promotes the consolidation of long-term memories of language sequences. It first reviews evidence supporting this position. Next it presents an experiment that demonstrates that subjects encouraged to rehearse foreign language (FL) utterances are better than both silent controls and subjects who are prevented from rehearsal by articulatory suppression at (a) learning to comprehend and translate FL words and phrases, (b) explicit metalinguistic knowledge of the detailed content of grammatical regularities, (c) acquisition of the FL forms of words and phrases, (d) accuracy in FL pronunciation, and (e) some aspects of productive (but not receptive) grammatical fluency and accuracy. Finally, it describes possible mechanisms underlying these effects. }
}
@inproceedings{EllisandGulwaniIJCAI-17,
       author = {Ellis, Kevin and Gulwani, Sumit},
        title = {Learning to Learn Programs from Examples: Going Beyond Program Structure},
    booktitle = {Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence, {IJCAI-17}},
         year = {2017},
     abstract = {Programming-by-example technologies let end users construct and run new programs by providing examples of the intended program behavior. But, the few provided examples seldom uniquely determine the intended program. Previous approaches to picking a program used a bias toward shorter or more naturally structured programs. Our work here gives a machine learning approach for learning to learn programs that departs from previous work by relying upon features that are independent of the program structure, instead relying upon a learned bias over program behaviors, and more generally over program execution traces. Our approach leverages abundant unlabeled data for semi-supervised learning, and incorporates simple kinds of world knowledge for common-sense reasoning during program induction. These techniques are evaluated in two programming-by-example domains, improving the accuracy of program learners.},
}
@article{EllisetalCoRR-20,
        title = {DreamCoder: Growing generalizable, interpretable knowledge with wake-sleep Bayesian program learning},
       author = {Kevin Ellis and Catherine Wong and Maxwell Nye and Mathias Sable-Meyer and Luc Cary and Lucas Morales and Luke Hewitt and Armando Solar-Lezama and Joshua B. Tenenbaum},
      journal = {CoRR},
       volume = {arXiv:2006.08381},
         year = {2020},
     abstract = {Expert problem-solving is driven by powerful languages for thinking about problems and their solutions. Acquiring expertise means learning these languages -- systems of concepts, alongside the skills to use them. We present DreamCoder, a system that learns to solve problems by writing programs. It builds expertise by creating programming languages for expressing domain concepts, together with neural networks to guide the search for programs within these languages. A ``wake-sleep'' learning algorithm alternately extends the language with new symbolic abstractions and trains the neural network on imagined and replayed problems. DreamCoder solves both classic inductive programming tasks and creative tasks such as drawing pictures and building scenes. It rediscovers the basics of modern functional programming, vector algebra and classical physics, including Newton's and Coulomb's laws. Concepts are built compositionally from those learned earlier, yielding multi-layered symbolic representations that are interpretable and transferrable to new tasks, while still growing scalably and flexibly with experience.}
}
@article{ElmanCOGNITION-93,
       author = {Jeffrey L. Elman},
        title = {Learning and development in neural networks: the importance of starting small},
      journal = {Cognition},
       volume = {48},
         year = {1993},
        pages = {71-99},
     abstract = {It is a striking fact that in humans the greatest learnmg occurs precisely at that point in time -- childhood -- when the most dramatic maturational changes also occur. This report describes possible synergistic interactions between maturational change and the ability to learn a complex domain (language), as investigated in connectionist It is a striking fact that in humans the greatest learnmg occurs precisely at that point in time -- childhood -- when the most dramatic maturational changes also occur. This report describes possible synergistic interactions between maturational change and the ability to learn a complex domain (language), as investigated in connectionist networks. The networks are trained to process complex sentences involving relative clauses, number agreement, and several types of verb argument structure. Training fails in the case of networks which are fully formed and 'adultlike' in their capacity. Training succeeds only when networks begin with limited working memory and gradually 'mature' to the adult state. This result suggests that rather than being a limitation, developmental restrictions on resources may constitute a necessary prerequisite for mastering certain complex domains. Specifically, successful learning may depend on starting small. networks. The networks are trained to process complex sentences involving relative clauses, number agreement, and several types of verb argument structure. Training fails in the case of networks which are fully formed and 'adultlike' in their capacity. Training succeeds only when networks begin with limited working memory and gradually 'mature' to the adult state. This result suggests that rather than being a limitation, developmental restrictions on resources may constitute a necessary prerequisite for mastering certain complex domains.}
}
@book{Elmanetal1998rethinking,
        title = {Rethinking Innateness: A Connectionist Perspective on Development},
       author = {Elman, J.L. and Bates, E.A. and Johnson, M.H.},
       series = {A Bradford book},
         year = {1998},
    publisher = {A Bradford Book}
}
@article{EmeretalSN-06,
       author = {Ermer, E.  and Guerin, S. A.  and Cosmides, L.  and Tooby, J.  and Miller, M. B. },
        title = {{{T}heory of mind broad and narrow: reasoning about social exchange engages {T}o{M} areas, precautionary reasoning does not}},
      journal = {Society of Neuroscience},
       volume = {1},
       number = {3-4},
         year = {2006},
        pages = {196-219},
     abstract = {Baron-Cohen (1995) proposed that the theory of mind (ToM) inference system evolved to promote strategic social interaction. Social exchange--a form of co-operation for mutual benefit--involves strategic social interaction and requires ToM inferences about the contents of other individuals' mental states, especially their desires, goals, and intentions. There are behavioral and neuropsychological dissociations between reasoning about social exchange and reasoning about equivalent problems tapping other, more general content domains. It has therefore been proposed that social exchange behavior is regulated by social contract algorithms: a domain-specific inference system that is functionally specialized for reasoning about social exchange. We report an fMRI study using the Wason selection task that provides further support for this hypothesis. Precautionary rules share so many properties with social exchange rules--they are conditional, deontic, and involve subjective utilities--that most reasoning theories claim they are processed by the same neurocomputational machinery. Nevertheless, neuroimaging shows that reasoning about social exchange activates brain areas not activated by reasoning about precautionary rules, and vice versa. As predicted, neural correlates of ToM (anterior and posterior temporal cortex) were activated when subjects interpreted social exchange rules, but not precautionary rules (where ToM inferences are unnecessary). We argue that the interaction between ToM and social contract algorithms can be reciprocal: social contract algorithms requires ToM inferences, but their functional logic also allows ToM inferences to be made. By considering interactions between ToM in the narrower sense (belief-desire reasoning) and all the social inference systems that create the logic of human social interaction--ones that enable as well as use inferences about the content of mental states--a broader conception of ToM may emerge: a computational model embodying a Theory of Human Nature (ToHN).}
}
@article{EnardetalNATURE-02,
        title = {{Molecular evolution of FOXP2, a gene involved in speech and language}},
       author = {Enard, Wolfgang and Przeworski, Molly and Fisher, Simon E. and Lai, Cecilia S. and Wiebe, Victor and Kitano, Takashi and Monaco, Anthony P. and Paabo, Svante},
      journal = {Nature},
       volume = 418,
        issue = 6900,
    publisher = {Nature Publishing Group},
        pages = {869-872},
         year = 2002,
     abstract = {Language is a uniquely human trait likely to have been a prerequisite for the development of human culture. The ability to develop articulate speech relies on capabilities, such as fine control of the larynx and mouth1, that are absent in chimpanzees and other great apes. FOXP2 is the first gene relevant to the human ability to develop language. A point mutation in FOXP2 co-segregates with a disorder in a family in which half of the members have severe articulation difficulties accompanied by linguistic and grammatical impairment. This gene is disrupted by translocation in an unrelated individual who has a similar disorder. Thus, two functional copies of FOXP2 seem to be required for acquisition of normal spoken language. We sequenced the complementary DNAs that encode the FOXP2 protein in the chimpanzee, gorilla, orang-utan, rhesus macaque and mouse, and compared them with the human cDNA. We also investigated intraspecific variation of the human FOXP2 gene. Here we show that human FOXP2 contains changes in amino-acid coding and a pattern of nucleotide polymorphism, which strongly suggest that this gene has been the target of selection during recent human evolution.},
}
@inproceedings{EngelSPIE-88,
       author = {Engel, Antonie J.},
        title = {Image Features as Virtual Beacons for Local Navigation},
    booktitle = {Proceedings of the {SPIE} Conference on Advances in Intelligent Robotic Systems},
 organization = {SPIE},
     location = {Cambridge, Massachusetts},
         year = 1988,
}
@techreport{Engelbart62,
       author = {Douglas C. Engelbart},
        title = {{\em{Augmenting the {H}uman {I}ntellect: {A} {C}onceptual {F}ramework}}},
        issue = {No.\ 3578},
  institution = {Stanford Research Institute},
      address = {Menlo Park, CA},
         year = 1962,
}
@techreport{EngelsonandMcDermott,
       author = {Engelson, Sean and McDermott, Drew},
        title = {Passive Robot Map Building with Exploration Scripts},
        issue = 898,
  institution = {Yale University Computer Science Department},
         year = 1992,
}
@article{ErcliketalNATURE-17,
       author = {Erclik, T. and Li, X. and Courgeon, M. and Bertet, C. and Chen, Z. and Baumert, R. and Ng, J. and Koo, C. and Arain, U. and Behnia, R. and del Valle Rodriguez, A. and Senderowicz, L. and Negre, N. and White, K. P. and Desplan, C.},
        title = {{{I}ntegration of temporal and spatial patterning generates neural diversity}},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = {541},
       number = {7637},
        pages = {365-370},
         year = {2017},
     abstract = {In the Drosophila optic lobes, 800 retinotopically organized columns in the medulla act as functional units for processing visual information. The medulla contains over 80 types of neuron, which belong to two classes: uni-columnar neurons have a stoichiometry of one per column, while multi-columnar neurons contact multiple columns. Here we show that combinatorial inputs from temporal and spatial axes generate this neuronal diversity: all neuroblasts switch fates over time to produce different neurons; the neuroepithelium that generates neuroblasts is also subdivided into six compartments by the expression of specific factors. Uni-columnar neurons are produced in all spatial compartments independently of spatial input; they innervate the neuropil where they are generated. Multi-columnar neurons are generated in smaller numbers in restricted compartments and require spatial input; the majority of their cell bodies subsequently move to cover the entire medulla. The selective integration of spatial inputs by a fixed temporal neuroblast cascade thus acts as a powerful mechanism for generating neural diversity, regulating stoichiometry and the formation of retinotopy.}
} 
@phdthesis{ErdmannPhd,
       author = {Michael A. Erdmann},
        title = {On Probabilistic Strategies for Robot Tasks},
       school = {MIT},
         year = 1989,
}
@article{ErdosandRenyi60,
       author = {Erd\"{o}s, P. and R\'{e}nyi, A.},
        title = {On the Evolution of Random Graphs},
      journal = {Publications of the Mathematical Institute of the Hungarian Academy of Sciences},
       volume = 5,
        pages = {17-61},
         year = 1960,
}
@article{ErdosandRenyi61,
       author = {Erd\"{o}s, P. and R\'{e}nyi, A.},
        title = {On Random Graphs {I}},
      journal = {Publ. Math. Debrecen},
       volume = 6,
        pages = {290-297},
         year = 1961,
}
@inproceedings{ErhanetalAISTATS-10,
        title = {Why Does Unsupervised Pre-training Help Deep Learning?},
       author = {Dumitru Erhan and Aaron Courville and Yoshua Bengio and Pascal Vincent},
    booktitle = {Proceedings of the Thirteenth International Conference on Artificial Intelligence and Statistics},
        pages = {201-208},
       editor = {Yee Whye Teh and Mike Titterington},
       volume = {9},
         year = {2010},
     abstract = {Much recent research has been devoted to learning algorithms for deep architectures such as Deep Belief Networks and stacks of auto-encoder   variants with impressive results being obtained in several areas, mostly   on vision and language datasets.  The best results obtained on supervised   learning tasks often involve an unsupervised learning component, usually   in an unsupervised pre-training phase. The main question investigated   here is the following: why does unsupervised pre-training work so well?   Through extensive experimentation, we explore several possible   explanations discussed in the literature including its action as a   regularizer (Erhan et al. 2009) and as an aid to optimization   (Bengio et al. 2007).  Our results build on the work of   Erhan et al. 2009, showing that unsupervised pre-training appears to   play predominantly a regularization role in subsequent supervised   training. However our results in an online setting, with a virtually unlimited   data stream, point to a somewhat more nuanced interpretation of the roles   of optimization and regularization in the unsupervised pre-training   effect.}
}
@article{EricssonandKintschPSYCHLOGICAL-REVIEW-95,
       author = {Ericsson, K. Anders and Kintsch, Walter},
      journal = {Psychological Review},
        title = {Long-term working memory},
       volume = {102},
       number = {2},
         year = {1995},
        pages = {211-245},
     abstract = {To account for the large demands on working memory during text comprehension and expert performance, the traditional models of working memory involving temporary storage must be extended to include working memory based on storage in long-term memory. In the proposed theoretical framework cognitive processes are viewed as a sequence of stable states representing end products of processing. In skilled activities, acquired memory skills allow these end products to be stored in long-term memory and kept directly accessible by means of retrieval cues in short-term memory, as proposed by skilled memory theory. These theoretical claims are supported by a review of evidence on memory in text comprehension and expert performance in such domains as mental calculation, medical diagnosis, and chess. (PsycINFO Database Record (c) 2016 APA, all rights reserved)},
}
@article{EricssonandKintschPSYCHOLOGICAL-REVIEW-95,
       author = {Ericsson, K. Anders and Kintsch, Walter},
        title = {Long-term working memory.},
      journal = {Psychological Review},
       number = {2},
       volume = {102},
         year = {1995},
        pages = {211-245},
     abstract = {To account for the large demands on working memory during text comprehension and expert performance, the traditional models of working memory involving temporary storage must be extended to include working memory based on storage in long-term memory. In the proposed theoretical framework cognitive processes are viewed as a sequence of stable states representing end products of processing. In skilled activities, acquired memory skills allow these end products to be stored in long-term memory and kept directly accessible by means of retrieval cues in short-term memory, as proposed by skilled memory theory. These theoretical claims are supported by a review of evidence on memory in text comprehension and expert performance in such domains as mental calculation, medical diagnosis, and chess.},
}
@inproceedings{ErikvandenBoogaardetal2017,
       author = {Erik van den Boogaard and Jan Treur and Maxim Turpijn},
        title = {A Neurologically Inspired Network Model for Graziano's Attention Schema Theory for Consciousness},
    booktitle = {Natural and Artificial Computation for Biomedicine and Neuroscience - International Work-Conference on the Interplay Between Natural and Artificial Computation},
        pages = {10-21},
         year = {2017},
     abstract = {This paper describes a network-oriented model based on the neuroscientist Graziano's Attention Schema Theory for consciousness. This theory describes an attention schema as an internal model of the attention process supporting the control of attention, similar to how our mind uses a body schema as an internal model of the body to control its movements. The Attention Schema Theory comes with a number of testable predictions. After designing a neurologically inspired temporal-causal network model for the Attention schema Theory, a few simulations were conducted to verify some of these predictions. One prediction is that a noticeable attention control deficit occurs when using attention without awareness. Another is that a noticeable attention control deficit occurs when using only bottom-up influence (from the sensory representations) without any top-down influence (for example, from goal or control states). The presented model is illustrated by a scenario where a hunter imagines (using internal simulation) a prey which he wants to attend to and catch, but shortly after he or she imagines a predator which he then wants to attend to and avoid. The outcomes of the simulations support the predictions that were made.}
}
@inproceedings{ErnstSNAPL-17,
       author = {Michael D. Ernst},
        title = {Natural language is a programming language: Applying natural language processing to software development},
    booktitle = {{SNAPL 2017: the 2nd Summit oN Advances in Programming Languages}},
        pages = {4:1-4:14},
      address = {Asilomar, CA, USA},
         year = {2017},
     abstract = {A powerful, but limited, way to view software is as source code alone. Treating a program as a sequence of instructions enables it to be formalized and makes it amenable to mathematical techniques such as abstract interpretation and model checking. A program consists of much more than a sequence of instructions. Developers make use of test cases, documentation, variable names, program structure, the version control repository, and more. I argue that it is time to take the blinders off of software analysis tools: tools should use all these artifacts to deduce more powerful and useful information about the program. Researchers are beginning to make progress towards this vision. This paper gives, as examples, four results that find bugs and generate code by applying natural language processing techniques to software artifacts. The four techniques use as input error messages, variable names, procedure documentation, and user questions. They use four different NLP techniques: document similarity, word semantics, parse trees, and neural networks. The initial results suggest that this is a promising avenue for future work.},
}
@book{Ernstetal69,
       author = {Ernst, G. and Allen Newell and Simon, Herbert},
        title = {{GPS}: A Case Study in Generality and Problem Solving},
    publisher = {Academic Press},
      address = {New York},
         year = 1969,
}
@article{ErokhinandFontanaCoRR-08,
       author = {Victor Erokhin and M.P. Fontana},
        title = {Electrochemically controlled polymeric device},
      journal = {CoRR},
       volume = {arXiv:0807.0333},
         year = 2008,
     abstract = {We report the fabrication and properties of a polymeric memristor, i.e. an electronic element with memory of its previous history. We show how this element can be viewed as a functional analog of a synaptic junction and how it can be used as a critical node in adaptive networks capable of bioinspired intelligent signal processing.}
}
@article{ErsuandTolleAIP-88,
       author = {Ers\"{u}, E. and Tolle, H.},
        title = {Hierarchical Learning Control - {A}n Approach With Neuron-Like Associative Memories},
      journal = {Proceedings of American Institute of Physics},
         year = {1988},
        pages = {249-261},
     abstract = {Advances in brain theory need two complementary approaches: Analytical investigations by in situ measurements and as well synthetic modelling supported by computer simulations to generate suggestive hypothesis on purposeful structures in the neural tissue. In this paper research of the second line is described: Starting from a neurophysiologically inspired model of stimulus-response (S-R) and/or associative memorization and a psychological- ly motivated ministructure for basic control tasks, pre-conditions and conditions are studied for cooperation of such units in a hierarchical organisation, as can be assumed to be the general layout of macrostructures in the brain.}
}
@article{ErsuandTolleIFAC-84,
       author = {E. Ers\"{u} and H. Tolle},
        title = {A New Concept for Learning Control Inspired by Brain Theory},
      journal = {9th IFAC World Congress: {A} Bridge Between Control Science and Technology},
       volume = {17},
       number = {2},
         year = {1984},
        pages = {1039-1044},
     abstract = {The paper explains an unconventional learning control method based on assumptions in the liturature about human problem solving and information storage in neuronal networks. The on-line learning comprises two stages: The dynamic input-output behaviour of the process to be controlled is stored stepwise in a neuron-like manner into an associative memory as a predictive process model, the control strategy planned via this model by optimization of a goal oriented performance index is then trained in the same way into a second associative memory. As a general mapping the learned behaviour is in both cases in general nonlinear, and by this such a control design is especially suited for strongly nonlinear processes. Simulation demonstrate the applicability of the new control concept.}
}
@article{EscribaetalJCMM-08,
       author = {Escriba, P. V. and Gonzalez-Ros, J. M. and Goni, F. M. and Kinnunen, P. K. and Vigh, L. and Sanchez-Magraner, L. and Fernandez, A. M. and Busquets, X. and Horvath, I. and Barcelo-Coblijn, G.},
        title = {Membranes: a meeting point for lipids, proteins and therapies},
      journal = {Journal of Cellular and Molecular Medicine},
         year = 2008,
       volume = 12,
        issue = 3,
        pages = {829-875},
     abstract = {Membranes constitute a meeting point for lipids and proteins. Not only do they define the entity of cells and cytosolic organelles but they also display a wide variety of important functions previously ascribed to the activity of proteins alone. Indeed, lipids have commonly been considered a mere support for the transient or permanent association of membrane proteins, while acting as a selective cell/organelle barrier. However, mounting evidence demonstrates that lipids themselves regulate the location and activity of many membrane proteins, as well as defining membrane microdomains that serve as spatio-temporal platforms for interacting signalling proteins. Membrane lipids are crucial in the fission and fusion of lipid bilayers and they also act as sensors to control environmental or physiological conditions. Lipids and lipid structures participate directly as messengers or regulators of signal transduction. Moreover, their alteration has been associated with the development of numerous diseases. Proteins can interact with membranes through lipid co-/post-translational modifications, and electrostatic and hydrophobic interactions, van der Waals forces and hydrogen bonding are all involved in the associations among membrane proteins and lipids. The present study reviews these interactions from the molecular and biomedical point of view, and the effects of their modulation on the physiological activity of cells, the aetiology of human diseases and the design of clinical drugs. In fact, the influence of lipids on protein function is reflected in the possibility to use these molecular species as targets for therapies against cancer, obesity, neurodegenerative disorders, cardiovascular pathologies and other diseases, using a new approach called membrane-lipid therapy.}
}
@inproceedings{EskinetalISMB-00,
       author = {Eleazar Eskin and William Noble Grundy and Yoram Singer},
        title = {Protein Family Classification Using Sparse Markov Transducers},
    booktitle = {Proceedings of the Eighth International Conference on Intelligent Systems for Molecular Biology},
        pages = {134-145},
         year = 2000,
}
@book{Esselstyn2007,
        title = {Prevent and Reverse Heart Disease: The Revolutionary, Scientifically Proven, Nutrition-Based Cure},
       author = {Esselstyn, C.B.},
         year = {2007},
    publisher = {Avery}
}
@article{EsseretalCoRR-16,
       author = {Steven K. Esser and Paul A. Merolla and John V. Arthur and Andrew S. Cassidy and Rathinakumar Appuswamy and Alexander Andreopoulos and David J. Berg and Jeffrey L. McKinstry and Timothy Melano and Davis R. Barch and Carmelo di Nolfo and Pallab Datta and Arnon Amir and Brian Taba and Myron D. Flickner and Dharmendra S. Modha},
        title = {Convolutional Networks for Fast, Energy-Efficient Neuromorphic Computing},
      journal = {CoRR},
       volume = {arXiv:1603.08270},
         year = {2016},
     abstract = {Deep networks are now able to achieve human-level performance on a broad spectrum of recognition tasks. Independently, neuromorphic computing has now demonstrated unprecedented energy-efficiency through a new chip architecture based on spiking neurons, low precision synapses, and a scalable communication network. Here, we demonstrate that neuromorphic computing, despite its novel architectural primitives, can implement deep convolution networks that (i) approach state-of-the-art classification accuracy across 8 standard datasets, encompassing vision and speech, (ii) perform inference while preserving the hardware's underlying energy-efficiency and high throughput, running on the aforementioned datasets at between 1100 and 2300 frames per second and using between 25 and 325 mW (effectively > 5000 frames / sec / W) and (iii) can be specified and trained using backpropagation with the same ease-of-use as contemporary deep learning. For the first time, the algorithmic power of deep learning can be merged with the efficiency of neuromorphic processors, bringing the promise of embedded, intelligent, brain-inspired computing one step closer.}
}
@article{EtzioniAIJ-91,
       author = {Etzioni, Oren},
        title = {Embedding Decision-Analytic Control in a Learning Architecture},
         year = 1991,
       volume = 49,
        issue = {1-3},
        pages = {129-160},
      journal = AIJ,
}
@article{EtzioniAIJ-93,
       author = {Etzioni, Oren},
        title = {Acquiring search-control knowledge via static analysis},
      journal = AIJ,
       volume = 62,
         year = 1993,
        pages = {255-302},
}
@incollection{EtzioniKR-89,
       author = {Etzioni, Oren},
        title = {Tractable Decision-Analytic Control},
     crossref = {KR-89},
        pages = {114-125},
}
@inproceedings{EtzionietalKR-92,
       author = {Etzioni, Oren and Hanks, Steve and Weld, Daniel and Draper, Denise and Lesh, Neal and Williamson, Mike},
        title = {An Approach to Planning with Incomplete Information},
    booktitle = {Proceedings of the 1992 International Conference on Principles of Knowledge Representation and Reasoning},
         year = 1992,
}
@incollection{Even2003abstraction,
        title = {Chapter 1: Digital Abstraction},
       author = {Guy Even},
    booktitle = {Computer Structure \& Introduction to Digital Computers},
       editor = {Guy Even},
    publisher = {Department of Electrical Engineering \& Systems},
      address = {Tel-Aviv University},
         year = {2003},
        pages = {1-10},
}
@book{Even2003introduction,
        title = {Computer Structure \& Introduction to Digital Computers},
       author = {Guy Even},
         year = {2003},
    publisher = {Department of Electrical Engineering \& Systems},
      address = {Tel-Aviv University}
}
@book{Even79,
       author = {Shimon Even},
        title = {Graph Algorithms},
    publisher = {W. H. Freeman},
      comment = {Originally Computer Science Press},
      address = {New York, NY},
         year = 1979,
}
@book{Everett2017language,
        title = {How Language Began: The Story of Humanity's Greatest Invention},
       author = {Everett, Daniel Leonard},
         year = {2017},
    publisher = {Boni \& Liveright},
      address = {New York, NY}
}
@misc{EveringhametalPASCAL-VOC-07,
       title = {The {PASCAL} {V}isual {O}bject {C}lasses {C}hallenge 2007 {(VOC2007)} {R}esults},
      author = {Everingham, M. and Van Gool, L. and Williams, C. K. I. and Winn, J. and Zisserman, A.},
        year = 2007,
howpublished = {http://www.pascal-network.org/challenges/VOC/voc2007/workshop/index.html},
}
@article{FabbroetalBRAIN-11,
        title = {Interfacing neurons with carbon nanotubes: (re)engineering neuronal signaling},
       author = {Fabbro, A. and Cellot, G. and Prato, M. and Ballerini, L.},
      journal = {Progress in Brain Ressearch},
       volume = {194C},
         year = 2011,
        pages = {241-252},
     abstract = {Carbon nanotubes (CNTs) are cylindrically shaped nanostructures made by sheets of graphene rolled up to form hollow tubes. Owing to their unique range of thermal, electronic, and structural properties, CNTs have been rapidly developing as a technology platform for biological and medical applications, including those designed to develop novel neuro-implantable devices. Depending on their structure, CNTs combine an incredible strength with an extreme flexibility. Further, these materials exhibit physical and chemical properties which allow them to efficiently conduit electrical current in electrochemical interfaces. CNTs can be organized in scaffolds made up of small fibers or tubes with diameters similar to those of neural processes such as axons and dendrites. Recently, CNT scaffolds have been found to promote growth, differentiation, and survival of neurons and to modify their electrophysiological properties. These features make CNTs an attractive material for the design of nano-bio hybrid systems able to govern cell-specific behaviors in cultured neuronal networks. The leading scope of this short review is to highlight how nanotube scaffolds can impact on neuronal signaling ability. In particular, we will focus on the direct and specific interactions between this synthetic nanomaterial and biological cell membranes, and on the ability of CNTs to improve interfaces developed to record or to stimulate neuronal activity. CNTs hold the potential for the development of innovative nanomaterial-based neurological implants. Therefore, it is particularly relevant to improve our knowledge on the impact on neuronal performance of interfacing nerve cells with CNTs.}
}
@article{FabbroetalCHEMICAL-NEUROSCIENCE-12,
       author = {Fabbro, Alessandra and Bosi, Susanna and Ballerini, Laura and Prato, Maurizio},
        title = {Carbon Nanotubes: Artificial Nanomaterials to Engineer Single Neurons and Neuronal Networks},
      journal = {ACS Chemical Neuroscience},
         year = 2012,
    publisher = {American Chemical Society},
       volume = 3,
        issue = 8,
        pages = {611-618},
     abstract = {In the past decade, nanotechnology applications to the nervous system have often involved the study and the use of novel nanomaterials to improve the diagnosis and therapy of neurological diseases. In the field of nanomedicine, carbon nanotubes are evaluated as promising materials for diverse therapeutic and diagnostic applications. Besides, carbon nanotubes are increasingly employed in basic neuroscience approaches, and they have been used in the design of neuronal interfaces or in that of scaffolds promoting neuronal growth in vitro. Ultimately, carbon nanotubes are thought to hold the potential for the development of innovative neurological implants. In this framework, it is particularly relevant to document the impact of interfacing such materials with nerve cells. Carbon nanotubes were shown, when modified with biologically active compounds or functionalized in order to alter their charge, to affect neurite outgrowth and branching. Notably, purified carbon nanotubes used as scaffolds can promote the formation of nanotube neuron hybrid networks, able per se to affect neuron integrative abilities, network connectivity, and synaptic plasticity. We focus this review on our work over several years directed to investigate the ability of carbon nanotube platforms in providing a new tool for nongenetic manipulations of neuronal performance and network signaling.},
}
@article{FaginJSL-76,
       author = {Fagin, R.},
        title = {Probabilities on Finite Models},
      journal = {Journal of Symbolic Logic},
       volume = 41,
        pages = {50-58},
         year = 1976,
}
@book{Faginetal94,
       author = {Fagin, Ronald and Halpern Joseph Y. and Moses, Yoram and Vardi, Moshe Y.},
        title = {Reasoning about Knowledge},
    publisher = {{MIT} Press},
         year = 1994,
}
@book{Fahlman79,
       author = {Fahlman, Scott},
        title = {NETL: A System for Representing and Using Real-World Knowledge},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1979,
}
@book{FairbankandGoldman2006,
       author = {John King Fairbank and Merle Goldman},
        title = {China: A New History (Second Edition)},
    publisher = {Harvard University Press},
      address = {Cambridge Massachusetts},
         year = 2006,
}
@article{FairetalPLoS-09,
        title = {Functional Brain Networks Develop from a Local to Distributed Organization},
       author = {Fair, D.A. and Cohen, A.L. and Power, J.D. and Dosenbach, N.U.F. and Church, J.A.},
      journal = {{PLoS} Computational Biology},
    publisher = {Public Library of Science},
       volume = 5,
        issue = 5,
         year = 2009,
}
@article{FajardoetalNEURSCIENCE-08,
        title = {Von Economo neurons are present in the dorsolateral (dysgranular) prefrontal cortex of humans},
       author = {Fajardo, C. and Escobar, M.I. and Buritic\'{a}, E. and Arteaga, G. and Umbarila, J. and Casanova, M.F. and Pimienta, H.},
      journal = {Neuroscience Letters},
       volume = 435,
        issue = 3,
         year = 2008,
        pages = {215-218},
     abstract = {Von Economo neurons (VENs), also known as spindle cells, have been described in layer V of the anterior cingulate (BA 24) and fronto-insular cortex (FI) of humans and other great apes. In the present study we used immunohistochemistry against two specific neuronal markers (NeuN and MAP2) in order to establish the presence of these cell types in Brodmann area 9 (BA 9) of the human prefrontal cortex. We evaluated tissue samples of eight human postmortem brains (age range 26-50) from BAs 9, 24, 4, 46, 45, 10 and 17. We identified a group of cells with similar morphology to that previously described for VENs in all specimens of BA 9 examined, albeit less frequently than in BA 24. This is the first description of this cell type in a human brain area with well developed granular layers (BA 9).}
}
@incollection{Falk2007,
       author = {Dean Falk},
    booktitle = {Handbook of Palaeoanthropology, {V}olume 2},
       editor = {W. Henke and H. Rothe and I. Tattersall},
    publisher = {Springer Verlag},
        title = {Primate Evolution and Human Origins},
         year = 2007,
        pages = {1133-1162},
}
@article{FalkenhaineretalAIJ-89,
       author = {Falkenhainer, Brian and Forbus, Kenneth D. and Gentner, Dedre},
        title = {{The Structure-Mapping Engine: Algorithm and Examples}},
      journal = {Artificial Intelligence},
    publisher = {Elsevier Science Publishers Ltd.},
       volume = {41},
       number = {1},
         year = {1989},
        pages = {1-63},
     abstract = {This paper describes the Structure-Mapping Engine (SME), a program for studying analogical processing. SME has been built to explore Gentner's Structure-mapping theory of analogy, and provides a "tool kit" for constructing matching algorithms consistent with this theory. Its flexibility enhances cognitive simulation studies by simplifying experimentation. Furthermore, SME is very efficient, making it a useful component in machine learning systems as well. We review the Structure-mapping theory and describe the design of the engine. We analyze the complexity of the algorithm, and demonstrate that most of the steps are polynomial, typically bounded by 0 (N2 ). Next we demonstrate some examples of its operation taken from our cognitive simulation studies and work in machine learning. Finally, we compare SME to other analogy programs and discuss several areas for future work.} 
}
@article{FalkowskiandGodfreyPTRS_B-08,
       author = {Falkowski, Paul G and Godfrey, Linda V},
        title = {Electrons, life and the evolution of Earth{\textquoteright}s oxygen cycle},
      journal = {{Philosophical Transactions of the Royal Society of London B: Biological Sciences}},
    publisher = {The Royal Society},
       volume = {363},
       number = {1504},
        pages = {2705-2716},
         year = {2008},
     abstract = {The biogeochemical cycles of H, C, N, O and S are coupled via biologically catalysed electron transfer (redox) reactions. The metabolic processes responsible for maintaining these cycles evolved over the first ca 2.3-Ga of Earth's history in prokaryotes and, through a sequence of events, led to the production of oxygen via the photobiologically catalysed oxidation of water. However, geochemical evidence suggests that there was a delay of several hundred million years before oxygen accumulated in Earth{\textquoteright}s atmosphere related to changes in the burial efficiency of organic matter and fundamental alterations in the nitrogen cycle. In the latter case, the presence of free molecular oxygen allowed ammonium to be oxidized to nitrate and subsequently denitrified. The interaction between the oxygen and nitrogen cycles in particular led to a negative feedback, in which increased production of oxygen led to decreased fixed inorganic nitrogen in the oceans. This feedback, which is supported by isotopic analyses of fixed nitrogen in sedimentary rocks from the Late Archaean, continues to the present. However, once sufficient oxygen accumulated in Earth{\textquoteright}s atmosphere to allow nitrification to out-compete denitrification, a new stable electron {\textquoteleft}market{\textquoteright} emerged in which oxygenic photosynthesis and aerobic respiration ultimately spread via endosymbiotic events and massive lateral gene transfer to eukaryotic host cells, allowing the evolution of complex (i.e. animal) life forms. The resulting network of electron transfers led a gas composition of Earth{\textquoteright}s atmosphere that is far from thermodynamic equilibrium (i.e. it is an emergent property), yet is relatively stable on geological time scales. The early coevolution of the C, N and O cycles, and the resulting non-equilibrium gaseous by-products can be used as a guide to search for the presence of life on terrestrial planets outside of our Solar System.},
}
@inproceedings{Faloutsosetal99,
    author = {M. Faloutsos and P. Faloutsos and C. Faloutsos},
     title = {On Power-Law relationships of the {I}nternet topology},
 booktitle = {Proceedings of the ACM SIGCOMM 1999 Conference},
     pages = {251-261},
      year = 1999,
}
@article{FanetalJCN-17,
       author = {Fan, D. and Wang, Q. and Su, J. and Xi, H.},
        title = {Stimulus-induced transitions between spike-wave discharges and spindles with the modulation of thalamic reticular nucleus},
      journal = {Journal of Computational Neuroscience},
         year = {2017},
       volume = {43},
       number = {3},
        pages = {203-225},
     abstract = {It is believed that thalamic reticular nucleus (TRN) controls spindles and spike-wave discharges (SWD) in seizure or sleeping processes. The dynamical mechanisms of spatiotemporal evolutions between these two types of activity, however, are not well understood. In light of this, we first use a single-compartment thalamocortical neural field model to investigate the effects of TRN on occurrence of SWD and its transition. Results show that the increasing inhibition from TRN to specific relay nuclei (SRN) can lead to the transition of system from SWD to slow-wave oscillation. Specially, it is shown that stimulations applied in the cortical neuronal populations can also initiate the SWD and slow-wave oscillation from the resting states under the typical inhibitory intensity from TRN to SRN. Then, we expand into a 3-compartment coupled thalamocortical model network in linear and circular structures, respectively, to explore the spatiotemporal evolutions of wave states in different compartments. The main results are: (i) for the open-ended model network, SWD induced by stimulus in the first compartment can be transformed into sleep-like slow UP-DOWN and spindle states as it propagates into the downstream compartments; (ii) for the close-ended model network, weak stimulations performed in the first compartment can result in the consistent experimentally observed spindle oscillations in all three compartments; in contrast, stronger periodic single-pulse stimulations applied in the first compartment can induce periodic transitions between SWD and spindle oscillations. Detailed investigations reveal that multi-attractor coexistence mechanism composed of SWD, spindles and background state underlies these state evolutions. What's more, in order to demonstrate the state evolution stability with respect to the topological structures of neural network, we further expand the 3-compartment coupled network into 10-compartment coupled one, with linear and circular structures, and nearest-neighbor (NN) coupled network as well as its realization of small-world (SW) topology via random rewiring, respectively. Interestingly, for the cases of linear and circular connetivities, qualitatively similar results were obtained in addition to the more irregularity of firings. However, SWD can be eventually transformed into the consistent low-amplitude oscillations for both NN and SW networks. In particular, SWD evolves into the slow spindling oscillations and background tonic oscillations within the NN and SW network, respectively. Our modeling and simulation studies highlight the effect of network topology in the evolutions of SWD and spindling oscillations, which provides new insights into the mechanisms of cortical seizures development.}
}
@inproceedings{FangetalIVLSI-15,
       author = {Yan Fang and  Victor V. Yashin and Donald M. Chiarulli and Steven P. Levitan},
        title = {A Simplified Phase Model for Oscillator Based Computing},
    booktitle = {{IEEE} Computer Society Annual Symposium on {VLSI}},
         year = {2015},
        pages = {231-236},
     abstract = {Building oscillator based computing systems with emerging nano-device technologies has become a promising solution for unconventional computing tasks like computer vision and pattern recognition. However, simulation and analysis of these systems is both time and compute intensive due to the nonlinearity of new devices and the complex behavior of coupled oscillators. In order to speed up the simulation of coupled oscillator systems, we propose a simplified phase model to perform phase and frequency synchronization prediction based on a synthesis of earlier models. Our model can predict the frequency locking behavior with several orders of magnitude speedup compared to direct evaluation, enabling the effective and efficient simulation of the large numbers of oscillators required for practical computing systems.}
}
@article{FangetalJMC-15,
       author = {Fang, Yan and Li, Xinming and Fang, Ying},
        title = {Organic bioelectronics for neural interfaces},
      journal = {Journal of Materials Chemistry C},
       volume = 3,
        issue = 25,
         year = 2015,
        pages = {6424-6430},
    publisher = {The Royal Society of Chemistry},
     abstract = {Organic bioelectronics offers important opportunities to study complex biological systems, such as neural networks, and develop new biomedical tools for the diagnosis and treatment of brain diseases. This Highlight is focused on recent progress in neural recording and stimulation enabled by using conducting polymers as active elements in bioelectronics, with an emphasis on the underlying mechanisms for the improved signal transduction capabilities at organic/neural interfaces. These studies are classified into two categories, electrochemical electrodes and electrochemical transistors, according to their operating principles. Future challenges and directions towards organic/neural interfaces are discussed as a conclusion.}
}
@article{FanselowandDongNEURON-10,
       author = {Fanselow, Michael S. and Dong, Hong-Wei},
        title = {Are the dorsal and ventral hippocampus functionally distinct structures?},
      journal = {Neuron},
         year = {2010},
       volume = {65},
        issue = {1},
        pages = {7-19},
     abstract = {One literature treats the hippocampus as a purely cognitive structure involved in memory; another treats it as a regulator of emotion whose dysfunction leads to psychopathology. We review behavioral, anatomical, and gene expression studies that together support a functional segmentation into three hippocampal compartments: dorsal, intermediate, and ventral. The dorsal hippocampus, which corresponds to the posterior hippocampus in primates, performs primarily cognitive functions. The ventral (anterior in primates) relates to stress, emotion, and affect. Strikingly, gene expression in the dorsal hippocampus correlates with cortical regions involved in information processing, while genes expressed in the ventral hippocampus correlate with regions involved in emotion and stress (amygdala and hypothalamus).},
}
@article{FarleyandClarkTIC-54,
    author = {Farley, B. and Clark, W.A.},
     title = {Simulation of self-Organizing Systems by Digital Computer},
   journal = {IRE Transactions on Information Theory},
    volume = 4,
      year = 1954,
     pages = {76-84},
}
@book{Farmelo2009strangest,
        title = {The Strangest Man: The Hidden Life of Paul Dirac, Quantum Genius},
       author = {Farmelo, Graham},
         year = {2009},
    publisher = {Faber \& Faber}
}
@book{Farmeretal84,
       author = {Farmer, D. and Toffoli, T. and Wolfram, S.},
        title = {Cellular Automata},
    publisher = {North Holland},
         year = 1984,
}
@article{FavaLIBRI-AND-LIBIRI-16,
        title = {Fairy Tales in Italy during the 20th Century and the Translations of Tales of Long Ago},
       author = {Sabrina Fava},
      journal = {Libri \& Liberi},
         year = 2016,
       volume = 5,
       number = 2,
        pages = {373-394},
     abstract = {Why was the book 'Tales of Long Ago' translated into Italian only in 1957, whereas it was published by Ivana Brlicc-Mazzuranicc in 1916? The paper explains that this delay depends on the development of fairy tales in Italy during the 20th century between the folktale tradition and fantastic children's literature. At the beginning of the 20th century there were two lines of evolution: on the one hand, attention was given to fairy tales coming from Northern Europe and, on the other hand, attention was given to traditional and local folktales. During the Fascist era, fantastic writing was an "unfettered realm" which prepared the development of the post-war period. In the second half of the 20th century in Italy, cultural and publishing trends were directed to reinforce national identity and build a European identity at the same time. Tales of Long Ago was translated into Italian and then this collection became part of the Enciclopedia della leggenda [Encyclopaedia of the Legend] (1957), a book edited by Giunti-Marzocco which extended the worldview of Italian children. The documentation in the Giunti Historical Archive allows us to explain the complex history of the Italian translation of Tales of Long Ago.}
}
@book{Fayer2001,
        title = {Elements of Quantum Mechanics},
       author = {Michael D. Fayer},
    publisher = {Oxford University Press},
      address = {Oxford, UK},
         year = 2001,
}        
@book{Fayer2010,
        title = {Absolutely Small: How Quantum Theory Explains our Everyday World},
       author = {Michael D. Fayer},
    publisher = {AMACOM},
      address = {New York, NY},
         year = 2010,
}
@inproceedings{FederetalSTOC-99,
       author = {Feder, Tomas and Hell, Pavol and Klein, Sulamita and Motwani, Rajeev},
        title = {Complexity of Graph Partition Problems},
    booktitle = {Proceedings of the Thirty-first Annual ACM Symposium on Theory of Computing},
    publisher = {ACM},
      address = {New York, NY, USA},
         year = {1999},
        pages = {464-472},
     abstract = {We introduce a parametrized family of graph problems that includes several well known graph partition problems as special cases. We develop tools which allow us to classify the complexity of many problems in this family and in particular lead us to a complete classification for small values of the parameters. Along the way  we obtain a variety of specific results including the following: a generalization of a communication bound on the number of clique versus independent set separators, polynomial-time algorithms to recognize generalized split graphs, and a quasi-polynomial algorithm for the Skew Cutset Problem that essentially resolves an open problem posed by Chvatal. The last two problems have interesting connections to the Strong Perfect Graph Conjecture of Berge. We also observe that the dichotomy (NP-complete versus polynomial-time solvable) conjectured for certain graph homomorphism problems would if true imply a slightly weaker dichotomy (NP-complete versus quasi-polynomial) for our graph partition problems }
} 
@article{FedorenkoandVarleyANYAS-16,
       author = {Fedorenko, Evelina and Varley, Rosemary},
        title = {Language and thought are not the same thing: evidence from neuroimaging and neurological patients},
      journal = {Annals New York Academy of Science},
       volume = {1369},
        issue = {1},
         year = {2016},
        pages = {132-153},
     abstract = {Is thought possible without language? Individuals with global aphasia, who have almost no ability to understand or produce language, provide a powerful opportunity to find out. Surprisingly, despite their near-total loss of language, these individuals are nonetheless able to add and subtract, solve logic problems, think about another person's thoughts, appreciate music, and successfully navigate their environments. Further, neuroimaging studies show that healthy adults strongly engage the brain's language areas when they understand a sentence, but not when they perform other nonlinguistic tasks such as arithmetic, storing information in working memory, inhibiting prepotent responses, or listening to music. Together, these two complementary lines of evidence provide a clear answer: many aspects of thought engage distinct brain regions from, and do not depend on, language.},
}
@book{Fedorov72,
       author = {Fedorov, V.},
        title = {Theory of Optimal Experimental Design},
    publisher = {Academic Press},
      address = {New York},
         year = 1972,
}
@article{FeeandGoldbergNEUROSCIENCE-11,
       author = {Fee, M. S. and Goldberg, J. H.},
        title = {A hypothesis for basal ganglia-dependent reinforcement learning in the songbird},
      journal = {Neuroscience},
         year = {2011},
       volume = {198},
        pages = {152-170},
     abstract = {Most of our motor skills are not innately programmed, but are learned by a combination of motor exploration and performance evaluation, suggesting that they proceed through a reinforcement learning (RL) mechanism. Songbirds have emerged as a model system to study how a complex behavioral sequence can be learned through an RL-like strategy. Interestingly, like motor sequence learning in mammals, song learning in birds requires a basal ganglia (BG)-thalamocortical loop, suggesting common neural mechanisms. Here, we outline a specific working hypothesis for how BG-forebrain circuits could utilize an internally computed reinforcement signal to direct song learning. Our model includes a number of general concepts borrowed from the mammalian BG literature, including a dopaminergic reward prediction error and dopamine-mediated plasticity at corticostriatal synapses. We also invoke a number of conceptual advances arising from recent observations in the songbird. Specifically, there is evidence for a specialized cortical circuit that adds trial-to-trial variability to stereotyped cortical motor programs, and a role for the BG in ``biasing'' this variability to improve behavioral performance. This BG-dependent ``premotor bias'' may in turn guide plasticity in downstream cortical synapses to consolidate recently learned song changes. Given the similarity between mammalian and songbird BG-thalamocortical circuits, our model for the role of the BG in this process may have broader relevance to mammalian BG function.},
}
@article{FeffermanetalCoRR-13,
        title = {Testing the Manifold Hypothesis},
       author = {Charles Fefferman and Sanjoy K. Mitter and Hariharan Narayanan},
      journal = {CoRR},
         year = {2013},
        pages = {983-1049},
     abstract = {he hypothesis that high dimensional data tend to lie in the vicinity of a low dimensional manifold is the basis of manifold learning. The goal of this paper is to develop an algorithm (with accompanying complexity guarantees) for fitting a manifold to an unknown probability distribution supported in a separable Hilbert space, only using i.i.d samples from that distribution. More precisely, our setting is the following. Suppose that data are drawn independently at random from a probability distribution P supported on the unit ball of a separable Hilbert space H. Let G(d,V,\tau{}) be the set of submanifolds of the unit ball of H whose volume is at most V and reach (which is the supremum of all r such that any point at a distance less than r has a unique nearest point on the manifold) is at least \tau. Let L(M,P) denote mean-squared distance of a random point from the probability distribution P to M. We obtain an algorithm that tests the manifold hypothesis in the following sense. The algorithm takes i.i.d random samples from P as input, and determines which of the following two is true (at least one must be): (a) There exists M \in{} G(d,CV,\tau{}C) such that L(M,P) \leq{} C\epsilon{}. (b) There exists no M \in{} G(d,V/C,C\tau{}) such that L(M,P) \leq{} \epsilon{} C. The answer is correct with probability at least 1−\delta{}.}
}
@inproceedings{Fei-FeiaetalCVPR-04,
       author = {L. Fei-Fei and R. Fergus and P. Perona},
        title = {Learning generative visual models from few training examples: an incremental {B}ayesian approach tested on 101 object categories},
    booktitle = {Proceedings of CVPR-05 Workshop on Generative-Model Based Vision},
         year = 2004,
}  
@inproceedings{Fei-FeiandPeronaCVPR-05,
       author = {Fei-Fei Li and Pietro Perona},
        title = {A {B}ayesian Hierarchical Model for Learning Natural Scene Categories},
    booktitle = {Proceedings of the 2005 {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
       volume = 2,
        pages = {524-531},
         year = 2005,
}
@article{Fei-FeietalPAMI-06,
     author = {L. Fei-Fei, R. Fergus and P. Perona},
      title = {Learning generative visual models from few training examples: 
               an incremental {B}ayesian approach tested on 101 object categories},
    journal = {IEEE Transactions on Pattern Analysis \& Machine Intelligence},
      issue = 4,
     volume = 28,
      pages = {594-611},
       year = 2006,
}
@misc{FeinNOCICEPTORS-14,
        title = {Nociceptors and the Perception of Pain},
       author = {Alan Fein},
 howpublished = {University of Connecticut Health Center},
         year = 2015,
}
@article{FeinbergetalNEURON-08,
      title = {{GFP} Reconstitution Across Synaptic Partners ({GRASP}) Defines Cell Contacts and Synapses in Living Nervous Systems},
     author = {Feinberg, Evan H. and VanHoven, Miri K. and Bendesky, Andres and Wang, George and Fetter, Richard D. and Shen, Kang and Bargmann, Cornelia I.},
               
    journal = {Neuron},
  publisher = {Cell Press},
     volume = 57,
      issue = 3,
       year = 2008,
      pages = {353-363},
}
@article{FeinbergetalPLoS-10,
     title = {Multiplexed echo planar imaging for sub-second whole brain FMRI and fast diffusion imaging.},
    author = {Feinberg, D.A. and Moeller, S. and Smith, S.M. and Auerbach, E. and Ramanna, S. and Glasser, M.F. and Miller, K.L. and Ugurbil, K. and Yacoub, E.},
   journal = {{PLoS} ONE},
    volume = 5,
     issue = 12,
     pages = {e15710},
      year = 2010,
  abstract = {Echo planar imaging (EPI) is an MRI technique of particular value to neuroscience, with its use for virtually all functional MRI (fMRI) and diffusion imaging of fiber connections in the human brain. EPI generates a single 2D image in a fraction of a second; however, it requires 2-3 seconds to acquire multi-slice whole brain coverage for fMRI and even longer for diffusion imaging. Here we report on a large reduction in EPI whole brain scan time at 3 and 7 Tesla, without significantly sacrificing spatial resolution, and while gaining functional sensitivity. The multiplexed-EPI (M-EPI) pulse sequence combines two forms of multiplexing: temporal multiplexing (m) utilizing simultaneous echo refocused (SIR) EPI and spatial multiplexing (n) with multibanded RF pulses (MB) to achieve m\%n images in an EPI echo train instead of the normal single image. This resulted in an unprecedented reduction in EPI scan time for whole brain fMRI performed at 3 Tesla, permitting TRs of 400 ms and 800 ms compared to a more conventional 2.5 sec TR, and 2-4 times reductions in scan time for HARDI imaging of neuronal fibertracks. The simultaneous SE refocusing of SIR imaging at 7 Tesla advantageously reduced SAR by using fewer RF refocusing pulses and by shifting fat signal out of the image plane so that fat suppression pulses were not required. In preliminary studies of resting state functional networks identified through independent component analysis, the 6-fold higher sampling rate increased the peak functional sensitivity by 60\%. The novel M-EPI pulse sequence resulted in a significantly increased temporal resolution for whole brain fMRI, and as such, this new methodology can be used for studying non-stationarity in networks and generally for expanding and enriching the functional information.}
}
@article{FeinmanandLakeCoRR-18,
       author = {Reuben Feinman and Brenden M. Lake},
        title = {Learning Inductive Biases with Simple Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1802.02745},
          url = {http://arxiv.org/abs/1802.02745},
         year = {2018},
}
@article{FeizietalNATURE-BIOTECHNOLOGY-13,
       author = {Feizi, Soheil and Marbach, Daniel and Medard, Muriel and Kellis, Manolis},
        title = {Network deconvolution as a general method to distinguish direct dependencies in networks},
      journal = {Nature Biotechnology},
         year = 2013,
    publisher = {Nature Publishing Group},
       volume = 31,
        issue = 8,
        pages = {726-733},
     abstract = {Recognizing direct relationships between variables connected in a network is a pervasive problem in biological, social and information sciences as correlation-based networks contain numerous indirect relationships. Here we present a general method for inferring direct effects from an observed correlation matrix containing both direct and indirect effects. We formulate the problem as the inverse of network convolution, and introduce an algorithm that removes the combined effect of all indirect paths of arbitrary length in a closed-form solution by exploiting eigen-decomposition and infinite-series sums. We demonstrate the effectiveness of our approach in several network applications: distinguishing direct targets in gene expression regulatory networks; recognizing directly interacting amino-acid residues for protein structure prediction from sequence alignments; and distinguishing strong collaborations in co-authorship social networks using connectivity information alone. In addition to its theoretical impact as a foundational graph theoretic tool, our results suggest network deconvolution is widely applicable for computing direct dependencies in network science across diverse disciplines.},
}
@article{Felder93,
       author = {Richard Felder},
        title = {Reaching the Second Tier: Learning and Teaching Styles in College Science Education},
      journal = {Journal of College Science Teaching},
       volume = 23,
        issue = 5,
        pages = {286-290},
         year = 1993,
}
@article{FelderandSilverman88,
       author = {R. M. Felder and L. K. Silverman},
        title = {Learning and Teaching Styles in Engineering Education},
      journal = {Engineering Education},
       volume = 78,
        issue = 7,
        pages = {674-681},
         year = 1988,
}
@article{FeldmanCS-77,
       author = {Feldman, Jerome and Sproull, Robert},
        title = {Decision Theory and Artificial Intelligence II: the Hungry Machine},
      journal = {Cognitive Science},
       volume = 1,
         year = 1977,
        pages = {158-192},
}
@article{FeldmanandValiantNC-09,
     title = {Experience-Induced Neural Circuits That Achieve High Capacity},
    author = {Vitaly Feldman and Leslie G. Valiant},
   journal = {Neural Computation},
    volume = 21,
     issue = 12,
      year = 2009,
     pages = {2715-2754},
}
@article{FeldmeyeretalPiN-13,
        title = {Barrel cortex function},
      journal = {Progress in Neurobiology},
       volume = {103},
        pages = {3-27},
         year = {2013},
       author = {Dirk Feldmeyer and Michael Brecht and Fritjof Helmchen and Carl C.H. Petersen and James F.A. Poulet and Jochen F. Staiger and Heiko J. Luhmann and Cornelius Schwarz},
     abstract = {Neocortex, the neuronal structure at the base of the remarkable cognitive skills of mammals, is a layered sheet of neuronal tissue composed of juxtaposed and interconnected columns. A cortical column is considered the basic module of cortical processing present in all cortical areas. It is believed to contain a characteristic microcircuit composed of a few thousand neurons. The high degree of cortical segmentation into vertical columns and horizontal layers is a boon for scientific investigation because it eases the systematic dissection and functional analysis of intrinsic as well as extrinsic connections of the column. In this review we will argue that in order to understand neocortical function one needs to combine a microscopic view, elucidating the workings of the local columnar microcircuits, with a macroscopic view, which keeps track of the linkage of distant cortical modules in different behavioral contexts. We will exemplify this strategy using the model system of vibrissal touch in mice and rats. On the macroscopic level vibrissal touch is an important sense for the subterranean rodents and has been honed by evolution to serve an array of distinct behaviors. Importantly, the vibrissae are moved actively to touch - requiring intricate sensorimotor interactions. Vibrissal touch, therefore, offers ample opportunities to relate different behavioral contexts to specific interactions of distant columns. On the microscopic level, the cortical modules in primary somatosensory cortex process touch inputs at highest magnification and discreteness - each whisker is represented by its own so-called barrel column. The cellular composition, intrinsic connectivity and functional aspects of the barrel column have been studied in great detail. Building on the versatility of genetic tools available in rodents, new, highly selective and flexible cellular and molecular tools to monitor and manipulate neuronal activity have been devised. Researchers have started to combine these with advanced and highly precise behavioral methods, on par with the precision known from monkey preparations. Therefore, the vibrissal touch model system is exquisitely positioned to combine the microscopic with the macroscopic view and promises to be instrumental in our understanding of neocortical function.}
}
@book{FellbaumWORDNET-98,
       editor = {Fellbaum, Christiane},
        title = {WordNet: An Electronic Lexical Database},
    publisher = {The {MIT} Press},
         year = 1998,
}
@book{Felleisenetal01,
       author = {Matthias Felleisen and Robert Bruce Findler and Matthew Flatt and Shriram Krishnamurthi},
        title = {How to Design Programs: An Introduction to Programming and Computing},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2001,
}
@article{FellemanandVanEssenCC-91,
       author = {Felleman, D. J. and Van Essen, D. C.},
        title = {Distributed hierarchical processing in primate cerebral cortex},
      journal = {Cerebral Cortex},
       volume = 1,
        pages = {1-47},
         year = 1991,
	 Abstract={In recent years, many new cortical areas have been identified in the macaque monkey. The number of identified connections between areas has increased even more dramatically. We report here on (1) a summary of the layout of cortical areas associated with vision and with other modalities, (2) a computerized database for storing and representing large amounts of information on connectivity patterns, and (3) the application of these data to the analysis of hierarchical organization of the cerebral cortex. Our analysis concentrates on the visual system, which includes 25 neocortical areas that are predominantly or exclusively visual in function, plus an additional 7 areas that we regard as visual-association areas on the basis of their extensive visual inputs. A total of 305 connections among these 32 visual and visual-association areas have been reported. This represents 31\% of the possible number of pathways if each area were connected with all others. The actual degree of connectivity is likely to be closer to 40\%. The great majority of pathways involve reciprocal connections between areas. There are also extensive connections with cortical areas outside the visual system proper, including the somatosensory cortex, as well as neocortical, transitional, and archicortical regions in the temporal and frontal lobes. In the somatosensory/motor system, there are 62 identified pathways linking 13 cortical areas, suggesting an overall connectivity of about 40\%. Based on the laminar patterns of connections between areas, we propose a hierarchy of visual areas and of somatosensory/motor areas that is more comprehensive than those suggested in other recent studies. The current version of the visual hierarchy includes 10 levels of cortical processing. Altogether, it contains 14 levels if one includes the retina and lateral geniculate nucleus at the bottom as well as the entorhinal cortex and hippocampus at the top. Within this hierarchy, there are multiple, intertwined processing streams, which, at a low level, are related to the compartmental organization of areas V1 and V2 and, at a high level, are related to the distinction between processing centers in the temporal and parietal lobes. However, there are some pathways and relationships (about 10\% of the total) whose descriptions do not fit cleanly into this hierarchical scheme for one reason or another. In most instances, though, it is unclear whether these represent genuine exceptions to a strict hierarchy rather than inaccuracies or uncertainities in the reported assignment.}
}
@article{FelsenbergetalNATURE-17,
       author = {Felsenberg, Johannes and Barnstedt, Oliver and Cognigni, Paola and Lin, Suewei and Waddell, Scott},
        title = {Re-evaluation of learned information in Drosophila},
      journal = {Nature},
         year = {2017},
    publisher = {Macmillan Publishers Limited},
       volume = {VVVV},
        issue = {NNNN},
        pages = {PPPP},
     abstract = {Animals constantly assess the reliability of learned information to optimize their behaviour. On retrieval, consolidated long- term memory can be neutralized by extinction if the learned prediction was inaccurate1. Alternatively, retrieved memory can be maintained, following a period of reconsolidation during which it is labile. Although extinction and reconsolidation provide opportunities to alleviate problematic human memories lack a detailed mechanistic understanding of memory updating. Here we identify neural operations underpinning the re-evaluation of memory in Drosophila. Reactivation of reward-reinforced olfactory memory can lead to either extinction or reconsolidation, depending on prediction accuracy. Each process recruits activity in specific parts of the mushroom body output network and distinct subsets of reinforcing dopaminergic neurons. Memory extinction requires output neurons with dendrites in the ' and '′ lobes of the mushroom body, which drive negatively reinforcing dopaminergic neurons that innervate neighbouring zones. The aversive valence of these new extinction memories neutralizes previously learned odour preference. Memory reconsolidation requires the 21 mushroom body output neurons. This pathway recruits negatively reinforcing dopaminergic neurons innervating the same compartment and re-engages positively reinforcing dopaminergic neurons to reconsolidate the original reward memory. These data establish that recurrent and hierarchical connectivity between mushroom body output neurons and dopaminergic neurons enables memory re-evaluation driven by reward-prediction error.},
}
@article {FeltenetalJOURNAL_OF_IMMUNOLOGY-85,
       author = {Felten, DL and Felten, SY and Carlson, SL and Olschowka, JA and Livnat, S},
        title = {Noradrenergic and peptidergic innervation of lymphoid tissue},
      journal = {The Journal of Immunology},
       volume = {135},
       number = {2},
         year = {1985},
        pages = {755-765},
    publisher = {American Association of Immunologists},
}
@article{FelzenszwalbandHuttenlocherIJCV-04,
     author = {Pedro F. Felzenszwalb and Daniel P. Huttenlocher},
      title = {Efficient Graph-Based Image Segmentation},
    journal = {International Journal of Computer Vision},
  publisher = {Kluwer Academic Publishers},
     volume = 59,
      issue = 2,
      pages = {167-181},
       year = 2004,
}
@article{FelzenszwalbandHuttenlocherIJCV-04b,
     author = {Pedro F. Felzenszwalb and Daniel P. Huttenlocher},
      title = {Pictorial Structures for Object Recognition},
    journal = {International Journal of Computer Vision},
  publisher = {Kluwer Academic Publishers},
     volume = 61,
      issue = 1,
      pages = {55-79},
       year = 2004,
}
@inproceedings{FelzenszwalbetalCVPR-08,
    author = {Pedro F. Felzenszwalb and David A. McAllester and Deva Ramanan},
     title = {A discriminatively trained, multiscale, deformable part model},
 booktitle = {{IEEE} Computer Society Conference on Computer Vision and Pattern Recognition},
  location = {Alaska, USA},
 publisher = {IEEE Computer Society},
     pages = {1-8},
      year = 2008,
}
@inproceedings{FelzenszwalbetalCVPR-10,
      title = {Cascade object detection with deformable part models},
     author = {Felzenszwalb, P. F. and Girshick, R. B. and McAllester, D.},
  booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
  publisher = {IEEE},
       year = 2010,
      pages = {2241-2248},
  abstract = {We describe a general method for building cascade classifiers from part-based deformable models such as pictorial structures. We focus primarily on the case of star-structured models and show how a simple algorithm based on partial hypothesis pruning can speed up object detection by more than one order of magnitude without sacrificing detection accuracy. In our algorithm, partial hypotheses are pruned with a sequence of thresholds. In analogy to probably approximately correct (PAC) learning, we introduce the notion of probably approximately admissible (PAA) thresholds. Such thresholds provide theoretical guarantees on the performance of the cascade method and can be computed from a small sample of positive examples. Finally, we outline a cascade detection algorithm for a general class of models defined by a grammar formalism. This class includes not only tree-structured pictorial structures but also richer models that can represent each part recursively as a mixture of other parts.},
  comment = {This work builds an efficient part-based object detector2 which is organized in a hierarchical way. If the model has n parts, an arrangement of n models is required, where the i-th model contains i parts. A model is simpler if it has less parts, so if a simple model makes a weak detection, then the procedure rejects the detection, whereas if a simple model makes a strong detection, more complex models are activated. This work uses dynamic programming strategies (in which the work done in time t-1 is useful to complete the work in time t) to improve speed. Also, the pruning thresholds are modeled using the notion of probably approximately admissible thresholds, to make it more efficient. It also uses general grammar models to evaluate the consistency of part models. This work has been a successful approach in the VOC object detection challenge, and recently obtained a prize as the life-long contribution (or something like that) from PASCAL.}, 
}
@article{FelzenszwalbetalPAMI-10,
     author = {P. Felzenszwalb and R. Girshick and D. McAllester and D. Ramanan},
      title = {Object Detection with Discriminatively Trained Part Based Models},
    journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
     volume = 32,
      issue = 9,
      pages = {1627-1645},
       year = 2010,
}
@article{FerenczetalIJCV-07,
     author = {Andras Ferencz and Erik Learned-Miller and Jitendra Malik},
      title = {Learning to Locate Informative Features for Visual Identification},
    journal = {International Journal of Computer Vision: Special Issue on Learning and Vision},
  publisher = {Kluwer Academic Publishers},
       year = 2007,
}
@inproceedings{FergusetalCVPR-03,
    author = {R. Fergus and P. Perona and A. Zisserman},
     title = {Object class recognition by unsupervised scale-invariant learning},
 booktitle = {Proceedings {IEEE} Conference on Computer Vision and Pattern Recognition, 2003},
     pages = {264-271},
    volume = 2,
       url = {http://cs.nyu.edu/~fergus/papers/fergus03.pdf},
      year = 2003,
}
@inproceedings{FergusetalNIPS-09,
     author = {Rob Fergus, Y. Weiss, Antonio Torralba},
      title = {Semi-Supervised Learning in Gigantic Image Collections},
  booktitle = {Advances in Neural Information Processing Systems 22, Proceedings of the
               Twenty-Third Annual Conference on Neural Information Processing Systems},
       year = 2009,
}
@article{FergusonAS-73,
    author = {Ferguson, T.},
     title = {A {B}ayesian analysis of some nonparametric problems},
   journal = {Annals of Statistics},
    volume = 1,
     issue = 2,
     pages = {209-230},
      year = 1973,
}
@inproceedings{FernandezandVelosoAAMAS-06,
       author = {Fern\'{a}ndez, Fernando and Veloso, Manuela},
        title = {Probabilistic Policy Reuse in a Reinforcement Learning Agent},
    booktitle = {Proceedings of the Fifth International Joint Conference on Autonomous Agents and Multiagent Systems},
         year = {2006},
     location = {Hakodate, Japan},
        pages = {720-727},
    publisher = {ACM},
      address = {New York, NY, USA},
}
@inproceedings{FernandezetalIJCAI-07,
       author = {Santiago Fern\'{a}ndez and Alex Graves and J\"{u}rgen Schmidhuber},
        title = {Sequence Labelling in Structured Domains with Hierarchical Recurrent Neural Networks},
    booktitle = {Proceedings of the 20th International Joint Conference on Artificial Intelligence},
         year = 2007,
     abstract = {Modelling data in structured domains requires establishing the relations among patterns at multiple scales. When these patterns arise from sequential data, the multiscale structure also contains a dynamic component that must be modelled, particularly, as is often the case, if the data is unsegmented. Probabilistic graphical models are the pre- dominant framework for labelling unsegmented sequential data in structured domains. Their use requires a certain degree of a priori knowledge about the relations among patterns and about the patterns themselves. This paper presents a hierarchical system, based on the connectionist temporal classification algorithm, for labelling unsegmented sequential data at multiple scales with recurrent neural networks only. Experiments on the recognition of sequences of spoken digits show that the system outperforms hidden Markov models, while making fewer assumptions about the domain.},
}
@article{FernandoetalCoRR-17,
       author = {Chrisantha Fernando and Dylan Banarse and Charles Blundell and Yori Zwols and David Ha and Andrei A. Rusu and Alexander Pritzel and Daan Wierstra},
        title = {{PathNet: E}volution Channels Gradient Descent in Super Neural Networks},
      journal = {CoRR},
       volume = {arXiv:/1701.08734},
         year = {2017},
     abstract = {For artificial general intelligence (AGI) it would be efficient if multiple users trained the same giant neural network, permitting parameter reuse, without catastrophic forgetting. PathNet is a first step in this direction. It is a neural network algorithm that uses agents embedded in the neural network whose task is to discover which parts of the network to re-use for new tasks. Agents are pathways (views) through the network which determine the subset of parameters that are used and updated by the forwards and backwards passes of the backpropogation algorithm. During learning, a tournament selection genetic algorithm is used to select pathways through the neural network for replication and mutation. Pathway fitness is the performance of that pathway measured according to a cost function. We demonstrate successful transfer learning; fixing the parameters along a path learned on task A and re-evolving a new population of paths for task B, allows task B to be learned faster than it could be learned from scratch or after fine-tuning. Paths evolved on task B re-use parts of the optimal path evolved on task A. Positive transfer was demonstrated for binary MNIST, CIFAR, and SVHN supervised learning classification tasks, and a set of Atari and Labyrinth reinforcement learning tasks, suggesting PathNets have general applicability for neural network training. Finally, PathNet also significantly improves the robustness to hyperparameter choices of a parallel asynchronous reinforcement learning algorithm (A3C).}
}
@book{Fernyhough2016voices,
        title = {The Voices Within: {T}he History and Science of How We Talk to Ourselves},
       author = {Fernyhough, Charles},
         year = {2016},
    publisher = {Basic Books}
}
@incollection{FernyhoughandMcCarthy-Jones2013hallucination,
       author = {Fernyhough, Charles and McCarthy-Jones, Simon},
        title = {Thinking Aloud about Mental Voices},
    booktitle = {Hallucination: Philosophy and Psychology},
       editor = {Fiona Macpherson and Dimitris Platchias},
    publisher = {The MIT Press},
         year = {2013},
     abstract = {It is commonly believed that auditory verbal hallucinations (AVHs) come from a misattribution of inner speech to an external agency. This chapter analyzes whether a developmental view of inner speech can deal with some of the problems associated with inner-speech theories. The chapter examines neurophysiological and phenomenological evidence relevant to the issue and looks up some key issues for future research.}
}
@inproceedings{FerranetalPMLR-18,
        title = {Modular meta-learning},
       author = {Alet, Ferran and Lozano-Perez, Tomas and Kaelbling, Leslie P.},
    booktitle = {Proceedings of The 2nd Conference on Robot Learning},
       editor = {Billard, Aude and Dragan, Anca and Peters, Jan and Morimoto, Jun},
    publisher = {PMLR},
       volume = {87},
         year = {2018},
        pages = {856-868},
       series = {Proceedings of Machine Learning Research},
     abstract = {Many prediction problems, such as those that arise in the context of robotics, have a simplifying underlying structure that, if known, could accelerate learning. In this paper, we present a strategy for learning a set of neural network modules that can be combined in different ways. We train different modular structures on a set of related tasks and generalize to new tasks by composing the learned modules in new ways. By reusing modules to generalize we achieve combinatorial generalization, akin to the "infinite use of finite means" displayed in language. Finally, we show this improves performance in two robotics-related problems.}
}
@inproceedings{Fesqetal96,
       author = {Lorraine Fesq and Abdullah Aljabri and ChristineAnderson and Robert Connerton and Richard Doyle and Mark Hoffman and Guy Man},
        title = {Spacecraft Autonomy in the New Millenium},
    booktitle = {NASA Workship},
     location = {Breckenridge, Colorado},
 organization = {NASA},
         year = 1996,
}
@book{Festinger1962,
       author = {Festinger, Leon},
        title = {A Theory of Cognitive Dissonance},
    publisher = {Stanford University Press},
         year = {1962},
}
@inproceedings{FidlerandLeonardisCVPR-07,
       author = {Sanja Fidler and Ales Leonardis},
        title = {Towards Scalable Representations of Object Categories: Learning a Hierarchy of Parts},
    booktitle = {Proceedings 2007 of the {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = {2007},
}
@inproceedings{FiechterCOLT-94,
       author = {Claude-Nicolas Fiechter},
        title = {Efficient Reinforcement Learning},
    booktitle = {Proceedings of the Seventh Annual {ACM} Conference on Computational Learning Theory},                
         year = 1994,
        pages = {88-97},
}
@article{FieldetalNATURE-10,
        title = {Functional connectivity in the retina at the resolution of photoreceptors},
       author = {Field, G.D. and Gauthier, J.L. and Sher, A. and Greschner, M. and Machado, T.A. and Jepson, L.H. and Shlens, J. and Gunning, D.E. and Mathieson, K. and Dabrowski, W. and Paninski, L. and Litke, A.M. and Chichilnisky, E.J.},
      journal = {Nature},
       volume = 467,
        issue = 7316,
        pages = {673-7},
         year = 2010,
     abstract = {To understand a neural circuit requires knowledge of its connectivity. Here we report measurements of functional connectivity between the input and ouput layers of the macaque retina at single-cell resolution and the implications of these for colour vision. Multi-electrode technology was used to record simultaneously from complete populations of the retinal ganglion cell types (midget, parasol and small bistratified) that transmit high-resolution visual signals to the brain. Fine-grained visual stimulation was used to identify the location, type and strength of the functional input of each cone photoreceptor to each ganglion cell. The populations of ON and OFF midget and parasol cells each sampled the complete population of long- and middle-wavelength-sensitive cones. However, only OFF midget cells frequently received strong input from short-wavelength-sensitive cones. ON and OFF midget cells showed a small non-random tendency to selectively sample from either long- or middle-wavelength-sensitive cones to a degree not explained by clumping in the cone mosaic. These measurements reveal computations in a neural circuit at the elementary resolution of individual neurons.}
}
@article{FieldsNATURE-15,
       author = {Fields, R. Douglas},
        title = {A new mechanism of nervous system plasticity: activity-dependent myelination},
      journal = {Nat Rev Neurosci},
    publisher = {Nature Publishing Group},
       volume = {16},
        issue = {12},
         year = {2015},
        pages = {756-767},
     abstract = {The synapse is the focus of experimental research and theory on the cellular mechanisms of nervous system plasticity and learning, but recent research is expanding the consideration of plasticity into new mechanisms beyond the synapse, notably including the possibility that conduction velocity could be modifiable through changes in myelin to optimize the timing of information transmission through neural circuits. This concept emerges from a confluence of brain imaging that reveals changes in white matter in the human brain during learning, together with cellular studies showing that the process of myelination can be influenced by action potential firing in axons. This Opinion article summarizes the new research on activity-dependent myelination, explores the possible implications of these studies and outlines the potential for new research.},
}
@phdthesis{FietePhD-03,
       author = {Ila Rani Fiete},
        title = {Learning and coding in biological neural networks},
       school = {Harvard University},
         year = 2003,
}
@incollection{FieteandSeungBIRDSONG-09,
       author = {Fiete, I.R. and Seung, H.S},
        title = {Birdsong Learning},
    booktitle = {Encyclopedia of Neuroscience},
    publisher = {Elsevier, New York},
         year = 2009,
        pages = {227-239},
}
@incollection{FieteetalBIRDSONG-07,
       author = {Fiete, Ila and Seung, Hyunjune},
        title = {Neural network models of birdsong production, learning, and coding},
    booktitle = {New Encyclopedia of Neuroscience},
       editor = {L. Squire and T.Albright and F. Bloom and F.Gage and N. Spitzer},
    publisher = {Elsevier},
         year = {2019},
}
@article{FieteetalNEURON-10,
        title = {Spike-Time-Dependent Plasticity and Heterosynaptic Competition Organize Networks to Produce Long Scale-Free Sequences of Neural Activity},
       author = {Ila R. Fiete and Walter Senn and Claude Z.H. Wang and Richard H.R. Hahnloser},
      journal = {Neuron},
       volume = 65,
        issue = 4,
        pages = {563-576},
         year = 2010,
     abstract = {Summary Sequential neural activity patterns are as ubiquitous as the outputs they drive, which include motor gestures and sequential cognitive processes. Neural sequences are long, compared to the activation durations of participating neurons, and sequence coding is sparse. Numerous studies demonstrate that spike-time-dependent plasticity (STDP), the primary known mechanism for temporal order learning in neurons, cannot organize networks to generate long sequences, raising the question of how such networks are formed. We show that heterosynaptic competition within single neurons, when combined with STDP, organizes networks to generate long unary activity sequences even without sequential training inputs. The network produces a diversity of sequences with a power law length distribution and exponent, independent of cellular time constants. We show evidence for a similar distribution of sequence lengths in the recorded premotor song activity of songbirds. These results suggest that neural sequences may be shaped by synaptic constraints and network circuitry rather than cellular time constants.},
}
@article{FikesetalAIJ-71,
       author = {Fikes, Richard and Nilsson, Nils J.},
        title = {{STRIPS}: A new approach to the application of theorem proving to problem solving},
      journal = AIJ,
       volume = 2,
         year = 1971,
        pages = {189-208},
}
@article{FikesetalAIJ-72,
       author = {Fikes, Richard E. and Hart, Peter E. and Nilsson, Nils J.},
        title = {Learning and Executing Generalized Robot Plans},
      journal = AIJ,
       volume = 3,
        pages = {251-288},
         year = 1972,
}
@article{FilonovetalACI-12,
       author = {Filonov, Grigory S. and Krumholz, Arie and Xia, Jun and Yao, Junjie and Wang, Lihong V. and Verkhusha, Vladislav V.},
        title = {Deep-tissue photoacoustic tomography of genetically encoded {iRFP} probe()},
      journal = {Angewandte Chemie International Edition},
         year = {2012},
       volume = {51},
        issue = {6},
        pages = {1448-1451},
}
@book{Findeisenetal80,
        title = {Control and Coordination in Hierarchical Systems},
       author = {Findeisen, W. and Bailey, F. N. and Brdys, M. and Malinowski, K. and Tatjewski, K. and Wozniak, A.},
     location = {New York},
    publisher = {John Wiley and Sons},
         year = 1980,
}
@inproceedings{FindlerandChenIJCAI-71,
       author = {Findler, N. and Chen, D.},
        title = {On the problems of time retrieval, temporal relations, causality, and coexistence},
    booktitle = {Proceedings {IJCAI} 2},
     location = {London},
 organization = {IJCAII},
         year = 1971,
}
@article{FineetalML-98,
       author = {Shai Fine and Yoram Singer and Naftali Tishby},
        title = {The Hierarchical Hidden {M}arkov Model: Analysis and Applications},
      journal = {Machine Learning},
       volume = 32,
        issue = 1,
         year = 1998,
        pages = {41-62},
}
@techreport{FingerPhD,
       author = {Finger, Joseph Jeffrey},
        title = {Exploiting Constraints in Design Synthesis},
         type = {{Ph.D.} Thesis},
  institution = {Stanford University},
         year = 1987,
}
@incollection{FinlayENCYCNEURO-09,
    author = {Finlay, B. L.},
 booktitle = {Encyclopedia of Neuroscience, {V}olume 2},
    editor = {Squire, L. R.}, 
 publisher = {Oxford: Academic Press}, 
     title = {Brain Evolution: Developmental Constraints and Relative Developmental Growth},
      year = 2009,
     pages = {337-345|},
}  
@techreport{FinnPhD-18,
       author = {Chelsea B. Finn},
        title = {Learning to Learn with Gradients},
        issue = {Doctor of Philosophy, Computer Science},
  institution = {University of California, Berkeley},
         year = 2018,
}
@article{FinnandLevineCoRR-16,
       author = {Chelsea Finn and Sergey Levine},
        title = {Deep Visual Foresight for Planning Robot Motion},
      journal = {CoRR},
       volume = {arXiv:1610.00696},
         year = {2016},
     abstract = {A key challenge in scaling up robot learning to many skills and environments is removing the need for human supervision, so that robots can collect their own data and improve their own performance without being limited by the cost of requesting human feedback. Model-based reinforcement learning holds the promise of enabling an agent to learn to predict the effects of its actions, which could provide flexible predictive models for a wide range of tasks and environments, without detailed human supervision. We develop a method for combining deep action-conditioned video prediction models with model-predictive control that uses entirely unlabeled training data. Our approach does not require a calibrated camera, an instrumented training set-up, nor precise sensing and actuation. Our results show that our method enables a real robot to perform nonprehensile manipulation -- pushing objects -- and can handle novel objects not seen during training.}
}
@article{FinnetalCoRR-17,
       author = {Chelsea Finn and Pieter Abbeel and Sergey Levine},
        title = {Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks},
      journal = {CoRR},
       volume = {arXiv:1703.03400},
         year = {2017},
     abstract = {We propose an algorithm for meta-learning that is model-agnostic, in the sense that it is compatible with any model trained with gradient descent and applicable to a variety of different learning problems, including classification, regression, and reinforcement learning. The goal of meta-learning is to train a model on a variety of learning tasks, such that it can solve new learning tasks using only a small number of training samples. In our approach, the parameters of the model are explicitly trained such that a small number of gradient steps with a small amount of training data from a new task will produce good generalization performance on that task. In effect, our method trains the model to be easy to fine-tune. We demonstrate that this approach leads to state-of-the-art performance on two few-shot image classification benchmarks, produces good results on few-shot regression, and accelerates fine-tuning for policy gradient reinforcement learning with neural network policies.},
}
@inproceedings{FirbyAAAI-87,
       author = {Firby, R. James},
        title = {An Investigation in Reactive Planning in Complex Domains},
    booktitle = {Proceedings AAAI-87},
     location = {Seattle, Washington},
 organization = {AAAI},
         year = 1987,
        pages = {202-206},
}
@techreport{FirbyPhD,
       author = {Firby, R. James},
        title = {Adaptive Execution in Complex Dynamic Worlds},
        issue = 672,
  institution = {Yale University Computer Science Department},
         year = 1989,
}
@inproceedings{FirbyetalIASTED-85,
      author = {Firby, R. James and Dean, Thomas L. and Miller, David P.},
       title = {Efficient Robot Planning with Deadlines and Travel Time},
   booktitle = {Proceedings of the 6th International Symposium on Robotics and Automation},
    location = {Santa Barbara, California},
organization = {IASTED},
        year = 1985,
}
@article{FisacetalCoRR-18,
       author = {Jaime F. Fisac and Chang Liu and Jessica B. Hamrick and S. Shankar Sastry and J. Karl Hedrick and Thomas L. Griffiths and Anca D. Dragan},
        title = {Generating Plans that Predict Themselves},
      journal = {CoRR},
       volume = {arXiv:1802.05250},
         year = {2018},
     abstract = {Collaboration requires coordination, and we coordinate by anticipating our teammates' future actions and adapting to their plan. In some cases, our teammates' actions early on can give us a clear idea of what the remainder of their plan is, i.e. what action sequence we should expect. In others, they might leave us less confident, or even lead us to the wrong conclusion. Our goal is for robot actions to fall in the first category: we want to enable robots to select their actions in such a way that human collaborators can easily use them to correctly anticipate what will follow. While previous work has focused on finding initial plans that convey a set goal, here we focus on finding two portions of a plan such that the initial portion conveys the final one. We introduce t-\ACty{}: a measure that quantifies the accuracy and confidence with which human observers can predict the remaining robot plan from the overall task goal and the observed initial t actions in the plan. We contribute a method for generating t-predictable plans: we search for a full plan that accomplishes the task, but in which the first t actions make it as easy as possible to infer the remaining ones. The result is often different from the most efficient plan, in which the initial actions might leave a lot of ambiguity as to how the task will be completed. Through an online experiment and an in-person user study with physical robots, we find that our approach outperforms a traditional efficiency-based planner in objective and subjective collaboration metrics.}
}
@article{FischbachandDittrichCTR-89,
       author = {Fischbach, K. -F. and Dittrich, A. P. M.},
        title = {The optic lobe of {D}rosophila melanogaster. I. A {G}olgi analysis of wild-type structure},
      journal = {Cell and Tissue Research},
         year = {1989},
       volume = {258},
       number = {3},
        pages = {441-475},
     abstract = {Golgi studies of the neurons in the optic lobes of Drosophila melanogaster reveal a large number of neuronal cell types. These can be classified as either columnar or tangential. Columnar elements establish the retinotopic maps of the lamina, medulla, and lobula-complex neuropiles. They are classified according to the position of their cell bodies, the number, width, and level of their arborizations, and their projection areas. Tangential elements are oriented perpendicularly to the columns. The arborizations of different tangential neurons are restricted to different layers of the optic neuropiles, within such layers their dendritic fields may span the entire retinotopic field or only part of it. The abundance of cell types inside each of the columnar units of the optic lobe is discussed with regard to its possible functional significance. By means of their stratified arborizations the columnar neurons form what appear to be multiple sets of retinotopically organized parallel information processing networks. It is suggested that these parallel networks filter different kinds of visual information and thus represent structurally separated functional subunits of the optic lobe. Such a parallel organization of visual functions increases the sites for function-specific gene actions and may explain the behavioral phenotypes of recently isolated structural mutants of the optic lobe.},
}
@incollection{FischerCN-20,
       author = {Jason Fischer},
        title = {Na{\"{i}}ve physics: building a mental model of how the world behaves},
    booktitle = {{The Cognitive Neurosciences VI}},
    publisher = {The MIT Press},
       editor = {Poeppel, D. and Mangun, G. R. and Gazzaniga, M. S.},
         year = {2020}
}
@inproceedings{FischerandLadnerSTOC-77,
       author = {Fischer, Michael and Ladner, Richard},
        title = {Propositional Modal Logic of Programs},
    booktitle = {Proceedings of the Ninth Annual {ACM} Symposium on Theory of Computing},
     location = {Boulder, Colorado},
         year = 1977,
        pages = {286-294},
}
@article {FischeretalPNAS-16,
       author = {Fischer, Jason and Mikhael, John G. and Tenenbaum, Joshua B. and Kanwisher, Nancy},
        title = {Functional neuroanatomy of intuitive physical inference},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {113},
       number = {34},
         year = {2016},
        pages = {E5072-E5081},
     abstract = {Perceiving the physical structure of the world and predicting how physical events will unfold over time are central to our daily lives. Recent behavioral and computational research has suggested that our physical intuitions may be supported by a 'physics engine' in the brain akin to the physical simulation engines built into video games. However, to date, there has been almost no investigation of the brain areas involved in intuitive physical inference. Here, using fMRI, we show that a variety of physical inference tasks as well as simply viewing physically rich scenes engage a common brain network in frontal and parietal cortices. These findings open the door to the cognitive neuroscientific study of physical inference in the human brain. To engage with the world - to understand the scene in front of us, plan actions, and predict what will happen next - we must have an intuitive grasp of the world's physical structure and dynamics. How do the objects in front of us rest on and support each other, how much force would be required to move them, and how will they behave when they fall, roll, or collide? Despite the centrality of physical inferences in daily life, little is known about the brain mechanisms recruited to interpret the physical structure of a scene and predict how physical events will unfold. Here, in a series of fMRI experiments, we identified a set of cortical regions that are selectively engaged when people watch and predict the unfolding of physical events - a 'physics engine' in the brain. These brain regions are selective to physical inferences relative to nonphysical but otherwise highly similar scenes and tasks. However, these regions are not exclusively engaged in physical inferences per se or, indeed, even in scene understanding; they overlap with the domain-general 'multiple demand' system, especially the parts of that system involved in action planning and tool use, pointing to a close relationship between the cognitive and neural mechanisms involved in parsing the physical content of a scene and preparing an appropriate action.},
}
@article{FischeretalPNAS-16_Supplements,
       author = {Fischer, Jason and Mikhael, John G. and Tenenbaum, Joshua B. and Kanwisher, Nancy},
        title = {{Supplementary Materials and Methods for Functional Neuroanatomy of Intuitive Physical Inference}},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {113},
       number = {34},
         year = {2016}
}
@article{FischlerandElschlagerTC-73,
     author = {M.A. Fischler and R.A. Elschlager},
      title = {The Representation and Matching of Pictorial Structures},
    journal = {{IEEE} Transactions on Computers},
     volume = 22,
      issue = 1,
  publisher = {IEEE Computer Society},
      pages = {67-92},
       year = 1973,
}
@book{FischlerandFirschein87,
   author =    {Fischler, Martin and Firschein, Oscar},
   title =     {Intelligence: The Eye, the Brain, and the Computer},
   publisher = {Addison-Wesley},
   address =   {Reading, Massachusetts},
   year =      1987,
}
@incollection{FishburnandVickson,
   author =   {Fishburn, Peter C. and Vickson, Raymond G.},
   title =    {Theoretical foundations of stochastic dominance},
   crossref = {Whitmore},
}
@article{FishellandHeintzNEURON-13,
     title = {The Neuron Identity Problem: Form Meets Function},
    author = {Gord Fishell and Nathaniel Heintz},
   journal = {Neuron},
    volume = 80,
     issue = 3,
     pages = {602-612},
      year = 2013,
  abstract = {A complete understanding of nervous system function cannot be achieved without the identification of its component cell types. In this Perspective, we explore a series of related issues surrounding cell identity and how revolutionary methods for labeling and probing specific neuronal types have clarified this question. Specifically, we ask the following questions: what is the purpose of such diversity, how is it generated, how is it maintained, and, ultimately, how can one unambiguously identity one cell type from another? We suggest that each cell type can be defined by a unique and conserved molecular ground state that determines its capabilities. We believe that gaining an understanding of these molecular barcodes will advance our ability to explore brain function, enhance our understanding of the biochemical basis of \{CNS\} disorders, and aid in the development of novel therapeutic strategies.},
}
@article{FitamenetalSCIENTIFIC-REPORTS-19,
       author = {Fitamen, Christophe and Blaye, Agn{\'{e}}s and Camos, Val{\`{e}}rie},
        title = {Five-year-old children's working memory can be improved when children act on a transparent goal cue},
      journal = {Scientific Reports},
       volume = {9},
        issue = {1},
         year = {2019},
        pages = {15342},
     abstract = {Working memory is a key component of human cognition and its development throughout childhood a major predictor of cognitive development and school achievement. Noticeably, preschoolers exhibit poor performance in working memory tasks. The present study aimed at testing different means to improve working memory performance in preschoolers. To this aim, we tested the effect of abstract and transparent goal cues in a Brown-Peterson task performed by 4- and 5-year-old preschoolers. If the transparent goal cue helps to better maintain the instructions, it should lead to better memory performance. Moreover, preschoolers had to track, either visually or with their fingers, the goal cue during the retention interval. If the motor activity favors the active engagement of the children in the task, the finger tracking should lead to improvement in memory performance. Our findings were that 5-year-old children benefitted from a transparent goal cue when they acted on it, while 4-year-old children did not show any improvement. These results suggest that working memory performance can be improved in 5-year-old children when the task embeds elements that can scaffold the task goal.}
}
@article{FitchPBR-17,
       author = {Fitch, W. Tecumseh},
        title = {Empirical approaches to the study of language evolution},
      journal = {Psychonomic Bulletin \& Review},
         year = {2017},
       volume = {24},
        issue = {1},
        pages = {3-33},
     abstract = {The study of language evolution, and human cognitive evolution more generally, has often been ridiculed as unscientific, but in fact it differs little from many other disciplines that investigate past events, such as geology or cosmology. Well-crafted models of language evolution make numerous testable hypotheses, and if the principles of strong inference (simultaneous testing of multiple plausible hypotheses) are adopted, there is an increasing amount of relevant data allowing empirical evaluation of such models. The articles in this special issue provide a concise overview of current models of language evolution, emphasizing the testable predictions that they make, along with overviews of the many sources of data available to test them (emphasizing comparative, neural, and genetic data). The key challenge facing the study of language evolution is not a lack of data, but rather a weak commitment to hypothesis-testing approaches and strong inference, exacerbated by the broad and highly interdisciplinary nature of the relevant data. This introduction offers an overview of the field, and a summary of what needed to evolve to provide our species with language-ready brains. It then briefly discusses different contemporary models of language evolution, followed by an overview of different sources of data to test these models. I conclude with my own multistage model of how different components of language could have evolved.},
}
@article{FitzHughBJ-61,
       author = {Richard FitzHugh},
        title = {Impulses and physiological states in theoretical models of nerve membrane},
      journal = {Biophysical Journal},
         year = {1961},
        pages = {445-466},
}
@inproceedings{FitzgibbonetalICCV-03,
    author = {Andrew Fitzgibbon and Yonatan Wexler and Andrew Zisserman},
     title = {Image-based rendering using image-based priors},
 booktitle = ICCV,
      year = 2003,
}
@inproceedings{FleetetalCVPR-98,
     author = {David Fleet and Michael Black and Allan Jepson},
      title = {Motion feature detection using steerable flow fields},
  booktitle = {Proceedings of the 1998 {IEEE} Conference on Computer Vision and Pattern Recognition},
  publisher = {IEEE Computer Society},
      pages = {274-281},
       year = 1998,
}
@misc{Flet-BerliacTHE-GRADIENT-19,
        title = {On the Promise of Hierarchical Reinforcement Learning},
       author = {Yannis Flet-Berliac},
 howpublished = {The {\mathax{\nabla}} Gradient},
         year = {2019},
          url = {https://thegradient.pub/the-promise-of-hierarchical-reinforcement-learning/}
}
@misc{FletcherNEURO_ANATOMY-15,
       author = {T. H. Fletcher},
        title = {{Neurohistology II: Synapses, Meninges, \& Receptors}},
         year = {2015},
 howpublished = {University of Minnesota, Neurohistology Course Notes},
}
@article{FletcherandMullinsNATURE-10,
       author = {Fletcher, Daniel A. and Mullins, R. Dyche},
        title = {Cell mechanics and the cytoskeleton},
      journal = {Nature},
         year = 2010,
    publisher = {Nature Publishing Group},
       volume = 463,
        issue = 7280,
        pages = {485-492},
     abstract = {The ability of a eukaryotic cell to resist deformation, to transport intracellular cargo and to change shape during movement depends on the cytoskeleton, an interconnected network of filamentous polymers and regulatory proteins. Recent work has demonstrated that both internal and external physical forces can act through the cytoskeleton to affect local mechanical properties and cellular behaviour. Attention is now focused on how cytoskeletal networks generate, transmit and respond to mechanical signals over both short and long timescales. An important insight emerging from this work is that long-lived cytoskeletal structures may act as epigenetic determinants of cell shape, function and fate.}
}
@incollection{FletcherandRanganNIPS-14,
        title = {Scalable Inference for Neuronal Connectivity from Calcium Imaging},
       author = {Fletcher, Alyson K and Rangan, Sundeep},
    booktitle = {Advances in Neural Information Processing Systems 27},
       editor = {Zoubin Ghahramani and Max Welling and C. Cortes and N.D. Lawrence and K.Q. Weinberger},
         year = 2014,
        pages = {2843-2851},
    publisher = {Curran Associates, Inc.},
     abstract = {Fluorescent calcium imaging provides a potentially powerful tool for inferring connectivity in neural circuits with up to thousands of neurons. However, a key challenge in using calcium imaging for connectivity detection is that current systems often have a temporal response and frame rate that can be orders of magnitude slower than the underlying neural spiking process. Bayesian inference methods based on expectation-maximization (EM) have been proposed to overcome these limitations, but are often computationally demanding since the E-step in the EM procedure typically involves state estimation for a high-dimensional nonlinear dynamical system. In this work, we propose a computationally fast method for the state estimation based on a hybrid of loopy belief propagation and approximate message passing (AMP). The key insight is that a neural system as viewed through calcium imaging can be factorized into simple scalar dynamical systems for each neuron with linear interconnections between the neurons. Using the structure, the updates in the proposed hybrid AMP methodology can be computed by a set of one-dimensional state estimation procedures and linear transforms with the connectivity matrix. This yields a computationally scalable method for inferring connectivity of large neural circuits. Simulations of the method on realistic neural networks demonstrate good accuracy with computation times that are potentially significantly faster than current approaches based on Markov Chain Monte Carlo methods.},
}
@article{FleuretandGemanIJCV-01,
       author = {Fleuret, Fran\c{c}ois and Geman, Donald},
        title = {Coarse-to-fine Face Detection},
      journal = {International Journal of Computer Vision},
         year = 2001,
       volume = 41,
        issue = {1/2},
        pages = {85-107},
}
@article{FlightNATURE-11,
        title = {Gene expression: The dynamics of the brain transcriptome revealed},
       author = {Flight, Monica Hoyos},
      journal = {Nature Review Neuroscience},
       volume = 12,
        issue = 12,
         year = 2011,
        pages = {703-703},
     abstract = {The tight regulation of gene expression in space and time is key to understanding how the complexity and variation within and between organisms can arise from a relatively simple DNA blueprint. Until now, few studies had been able to characterize the temporal dynamics of gene transcription in the human brain with the depth reported in two recent papers published in Nature.},
}
@book{Fodor75,
      author = {Fodor, Jerry},
       title = {The Language of Thought},
   publisher = {Harvard University Press},
     address = {Cambridge, MA},
        year = 1975,
}
@book{Fodor84,
      author = {Fodor, Jerry},
       title = {Modularity of Mind},
   publisher = {{MIT} Press},
     address = {Cambridge, Massachusetts},
        year = 1984,
}
@inproceedings{FodorandHuertaSLTW-06,
       author = {Paul Fodor and Juan M. Huerta},
        title = {Planning and logic programming for dialogue management},
    booktitle = {IEEE Spoken Language Technology Workshop},
    publisher = {IEEE},
         year = {2006},
     abstract = {Dialog interaction in conversational applications is subordinated to the goal of completing a domain-specific task. In this paper we present a basic architecture, a knowledge representation system, and a planning algorithm for dialogue management that decouples the interaction process from the planning task. In our system, the interaction is driven by the planner. We use logic programming, automatic planning and problem solving algorithms for representing information states and performing interaction management in the dialogue system. Our approach leverages recent advances in formalisms, inference engines, planning and problem solving, and is particularly suitable when implementing negotiation-intensive conversational applications.}
}
@article{FodorandPylyshynCOGNITION-88,
        title = {Connectionism and Cognitive Architecture},
    publisher = {Elsevier},
         year = {1988},
      journal = {Cognition},
       volume = {28},
       author = {Jerry A. Fodor and Zenon W. Pylyshyn},
        pages = {3-71},
       number = {1-2},
     abstract = {This paper explores differences between Connectionist proposals for cognitive architecture and the sorts of models that have traditionally been assumed in cognitive science. We claim that the major distinction is that, while both Connectionist and Classical architectures postulate representational mental states, the latter but not the former are committed to a symbol-level of representation, or to a 'language of thought': i.e., to representational states that have combinatorial syntactic and semantic structure. Several arguments for combinatorial structure in mental representations are then reviewed. These include arguments based on the 'systematicity' of mental representation: i.e., on the fact that cognitive capacities always exhibit certain symmetries, so that the ability to entertain a given thought implies the ability to entertain thoughts with semantically related contents. We claim that such arguments make a powerful case that mind/brain architecture is not Connectionist at the cognitive level. We then consider the possibility that Connectionism may provide an account of the neural (or 'abstract neurological') structures in which Classical cognitive architecture is implemented. We survey a number of the standard arguments that have been offered in favor of Connectionism, and conclude that they are coherent only on this interpretation.},
}
@article{FoersteretalCoRR-17,
       author = {Jakob N. Foerster and Richard Y. Chen and Maruan Al{-}Shedivat and Shimon Whiteson and Pieter Abbeel and Igor Mordatch},
        title = {Learning with Opponent-Learning Awareness},
      journal = {CoRR},
       volume = {abs/1709.04326},
         year = {2017},
     abstract = {Multi-agent settings are quickly gathering importance in machine learning. Beyond a plethora of recent work on deep multi-agent reinforcement learning, hierarchical reinforcement learning, generative adversarial networks and decentralized optimization can all be seen as instances of this setting. However, the presence of multiple learning agents in these settings renders the training problem non-stationary and often leads to unstable training or undesired final results. We present Learning with Opponent-Learning Awareness (LOLA), a method that reasons about the anticipated learning of the other agents. The LOLA learning rule includes an additional term that accounts for the impact of the agent's policy on the anticipated parameter update of the other agents. We show that the LOLA update rule can be efficiently calculated using an extension of the likelihood ratio policy gradient update, making the method suitable for model-free RL. This method thus scales to large parameter and input spaces and nonlinear function approximators. Preliminary results show that the encounter of two LOLA agents leads to the emergence of tit-for-tat and therefore cooperation in the iterated prisoners' dilemma (IPD), while independent learning does not. In this domain, LOLA also receives higher payouts compared to a naive learner, and is robust against exploitation by higher order gradient-based methods. Applied to infinitely repeated matching pennies, LOLA agents converge to the Nash equilibrium. In a round robin tournament we show that LOLA agents can successfully shape the learning of a range of multi-agent learning algorithms from literature, resulting in the highest average returns on the IPD. We also apply LOLA to a grid world task with an embedded social dilemma using deep recurrent policies. Again, by considering the learning of the other agent, LOLA agents learn to cooperate out of selfish interests.}
}
@article{FoldiakBC-90,
       author = {P. F\"{o}ldi\'{a}k},
        title = {Forming sparse representations by local anti-Hebbian learning},
      journal = {Biological Cybernetics},
       volume = 64,
        issue = 2,
        pages = {165-170},
         year = 1990,
}
@article{FoldiakNC-91,
     author = {P. F\"{o}ldi\'{a}k},
      title = {Learning invariance from transformation sequences},
    journal = {Neural Computation},
     volume = 3,
      pages = {194-200},
    comment = {http://www.st-andrews.ac.uk/\%7Epf2/FoldiakInvarianceLearningNC91.pdf},
       year = 1991,
}
@article{FongetalCoRR-17,
        title = {Using Human Brain Activity to Guide Machine Learning},
       author = {Ruth Fong and Walter Scheirer and David Cox},
      journal = {CoRR},
       volume = {arXiv:1703.05463},
         year = {2017},
     abstract = {Machine learning is a field of computer science that builds algorithms that learn. In many cases, machine learning algorithms are used to recreate a human ability like adding a caption to a photo, driving a car, or playing a game. While the human brain has long served as a source of inspiration for machine learning, little effort has been made to directly use data collected from working brains as a guide for machine learning algorithms. Here we demonstrate a new paradigm of "neurally-weighted" machine learning, which takes fMRI measurements of human brain activity from subjects viewing images, and infuses these data into the training process of an object recognition learning algorithm to make it more consistent with the human brain. After training, these neurally-weighted classifiers are able to classify images without requiring any additional neural data. We show that our neural-weighting approach can lead to large performance gains when used with traditional machine vision features, as well as to significant improvements with already high-performing convolutional neural network features. The effectiveness of this approach points to a path forward for a new class of hybrid machine learning algorithms which take both inspiration and direct constraints from neuronal data.}
}
@inproceedings{ForbesetalIJCAI-95,
       author = {Forbes, J. and Huang, T. and Kanazawa, K. and Russell, S. J.},
        title = {The {BAT}mobile: Towards a {B}ayesian Automated Taxi},
    booktitle = {Proceedings {IJCAI} 14},
     location = {Montreal, Canada},
 organization = {IJCAII},
        pages = {1878-1885},
         year = 1995,
}
@article{ForbusAIJ-84,
       author = {Forbus, Kenneth},
        title = {Qualitative Process Theory},
      journal = AIJ,
       volume = 24,
         year = 1984,
        pages = {85-168},
}
@incollection{ForbusQR-85,
       author = {Forbus, Kenneth D.},
        title = {Qualitative Process Theory},
     crossref = {QR-85},
        pages = {85-168},
}
@book{ForbusanddeKleer94,
       editor = {Forbus, Ken and de Kleer, Johan},
        title = {Building Problem Solvers},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1994,
}
@article{ForbusetalCS-16,
       author = {Forbus Kenneth D. and Ferguson Ronald W. and Lovett Andrew and Gentner Dedre},
        title = {Extending {SME} to Handle Large-Scale Cognitive Modeling},
      journal = {Cognitive Science},
       volume = {41},
       number = {5},
         year = {2016},
        pages = {1152-1201},
     abstract = {Analogy and similarity are central phenomena in human cognition, involved in processes ranging from visual perception to conceptual change. To capture this centrality requires that a model of comparison must be able to integrate with other processes and handle the size and complexity of the representations required by the tasks being modeled. This paper describes extensions to Structure-Mapping Engine (SME) since its inception in 1986 that have increased its scope of operation. We first review the basic SME algorithm, describe psychological evidence for SME as a process model, and summarize its role in simulating similarity-based retrieval and generalization. Then we describe five techniques now incorporated into the SME that have enabled it to tackle large-scale modeling tasks: (a) Greedy merging rapidly constructs one or more best interpretations of a match in polynomial time: O(n2log(n)); (b) Incremental operation enables mappings to be extended as new information is retrieved or derived about the base or target, to model situations where information in a task is updated over time; (c) Ubiquitous predicates model the varying degrees to which items may suggest alignment; (d) Structural evaluation of analogical inferences models aspects of plausibility judgments; (e) Match filters enable large-scale task models to communicate constraints to SME to influence the mapping process. We illustrate via examples from published studies how these enable it to capture a broader range of psychological phenomena than before.}
}
@techreport{ForgyTR-81,
       author = {Forgy, Charles L.},
        title = {OPS5 User's Manual},
        issue = {CMU-CS-81-135},
  institution = {Carnegie-Mellon University Department of Computer Science},
         year = 1981,
}
@book{ForsythPonce02,
      author = {David A. Forsyth and Jean Ponce},
       title = {Computer Vision: A Modern Approach},
   publisher = {Prentice Hall},
     address = {Englewood Cliffs, New Jersey},
        year = 2002,
}
@article{ForsythetalPAMI-91,
     author = {Forsyth, David and Mundy, Joseph L. and Zisserman, Andrew and Coelho, Chris and Heller, Aaron and Rothwell, Charles},
      title = {Invariant Descriptors for {3D} Object Recognition and Pose},
    journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
     volume = 13,
      issue = 10,
      pages = {971-991},
       year = 1991,
}
@article{FortunatoetalPNAS-06,
      title = {S. Fortunato and A. Flammini and F. Menczer and A. Vespignani},
     author = {Topical interests and the mitigation of search engine bias},
    journal = {Proceedings of the National Academy of Science},
     volume = 103,
      issue = 34,
      pages = {12684-12689},
       year = 2006,
}
@book{Fosback93,
       author = {Fosback, Norman G.},
        title = {Stock Market Logic},
    publisher = {Dearborn Financial Publishing},
         year = 1993,
}
@article {FosqueetalSCIENCE-15,
       author = {Fosque, Benjamin F. and Sun, Yi and Dana, Hod and Yang, Chao-Tsung and Ohyama, Tomoko and Tadross, Michael R. and Patel, Ronak and Zlatic, Marta and Kim, Douglas S. and Ahrens, Misha B. and Jayaraman, Vivek and Looger, Loren L. and Schreiter, Eric R.},
        title = {Labeling of active neural circuits in vivo with designed calcium integrators},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {347},
       number = {6223},
         year = {2015},
        pages = {755-760},
     abstract = {The identification of active neurons and circuits in vivo is a fundamental challenge in understanding the neural basis of behavior. Genetically encoded calcium (Ca2+) indicators (GECIs) enable quantitative monitoring of cellular-resolution activity during behavior. However, such indicators require online monitoring within a limited field of view. Alternatively, post hoc staining of immediate early genes (IEGs) indicates highly active cells within the entire brain, albeit with poor temporal resolution. We designed a fluorescent sensor, CaMPARI, that combines the genetic targetability and quantitative link to neural activity of GECIs with the permanent, large-scale labeling of IEGs, allowing a temporally precise 'activity snapshot' of a large tissue volume. CaMPARI undergoes efficient and irreversible green-to-red conversion only when elevated intracellular Ca2+ and experimenter-controlled illumination coincide. We demonstrate the utility of CaMPARI in freely moving larvae of zebrafish and flies, and in head-fixed mice and adult flies.},
}     
@article{FosterandJonesCoRR-17,
       author = {James M. Foster and Matt Jones},
        title = {Reinforcement Learning with Analogical Similarity to Guide Schema Induction and Attention},
      journal = {CoRR},
       volume = {arXiv:1712.10070},
         year = {2017},
     abstract = {Research in analogical reasoning suggests that higher-order cognitive functions such as abstract reasoning, far transfer, and creativity are founded on recognizing structural similarities among relational systems. Here we integrate theories of analogy with the computational framework of reinforcement learning (RL). We propose a psychology theory that is a computational synergy between analogy and RL, in which analogical comparison provides the RL learning algorithm with a measure of relational similarity, and RL provides feedback signals that can drive analogical learning. Simulation results support the power of this approach.}
}
@article{FosterandWilsonNATURE-06,
        title = {Reverse replay of behavioural sequences in hippocampal place cells during the awake state},
       author = {Foster, David J and Wilson, Matthew A},
      journal = {Nature},
       volume = {440},
       number = {7084},
        pages = {680},
         year = {2006},
    publisher = {Nature Publishing Group}
}         
@inproceedings {FowlkesetalCVPR-01,
        title = {Efficient Spatiotemporal Grouping Using the Nystr{\"o}m Method},
       author = {Charless Fowlkes and Serge Belongie and Jitendra Malik},
    booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
         year = 2001,
        pages = {231-238},
}
@article {FowlkesetalPAMI-04,
       author = {Charless Fowlkes and Serge Belongie and Fan Chung and Jitendra Malik},
        title = {Spectral Grouping Using the Nystr{\"o}m Method},
      journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
    publisher = {IEEE Computer Society},
       volume = 26,
        issue = 2,
         year = 2004,
        pages = {214-225},
}
@inproceedings{Fox,
       author = {Fox, M. S. and Allen, B. and Strohm, G.},
        title = {Job-Shop Scheduling: An Investigation in Constraint-Directed Reasoning},
    booktitle = {Proceedings AAAI-82},
     location = {Pittsburgh. Pennsylvania},
 organization = {AAAI},
         year = 1982,
        pages = {155-158},
}
@inproceedings{FoxandKempfIEEE-85,
       author = {Fox, B. R. and Kempf, K. G.},
        title = {Opportunistic Scheduling for Robotics Assembly},
    booktitle = {{IEEE} International Conference on Robotics and Automation},
     location = {St. Louis, MO},
         year = 1985,
        pages = {880-889},
}
@article{FoxandSmithES-85,
       author = {Fox, M. S. and Smith, S.},
        title = {ISIS: A Knowledge-Based System for Factory Scheduling},
      journal = {Expert Systems},
       volume = 1,
        pages = {25-49},
         year = 1985,
}
@incollection{Foxetal00,
       author = {D. Fox and S. Thrun, W. Burgard and F. Dellaert},
        title = {Particle filters for mobile robot localization},
       editor = {A. Doucet and N. de Freitas and Gordon. N.},
    publisher = {Springer-Verlag},
    booktitle = {Sequential Monte Carlo Methods in Practice},
         year = 2000,
}
@inproceedings{FragkiadakietalCVPR-15,
       author = {Fragkiadaki, Katerina and Arbelaez, Pablo and Felsen, Panna and Malik, Jitendra},
        title = {Learning to segment moving objects in videos},
    booktitle = {Proceedings of the 2015 {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 2015,
        pages = {4083-4090},
     abstract = {We segment moving objects in videos by ranking spatiotemporal segment proposals according to "moving objectness"; how likely they are to contain a moving object. In each video frame, we compute segment proposals using multiple figure-ground segmentations on per frame motion boundaries. We rank them with a Moving Objectness Detector trained on image and motion fields to detect moving objects and discard over/under segmentations or background parts of the scene. We extend the top ranked segments into spatio-temporal tubes using random walkers on motion affinities of dense point trajectories. Our final tube ranking consistently outperforms previous segmentation methods in the two largest video segmentation benchmarks currently available, for any number of proposals. Further, our per frame moving object proposals increase the detection rate up to 7\% over previous state-of-the-art static proposal methods.},   
}
@article{FrankJoCN-05,
       author = {Frank, M. J.},
        title = {Dynamic dopamine modulation in the basal ganglia: a neurocomputational account of cognitive deficits in medicated and nonmedicated {P}arkinsonism},
      journal = {Journal of Cognitive Neuroscience},
       volume = {17},
       number = {1},
         year = {2005},
        pages = {51-72},
     abstract = {Dopamine (DA) depletion in the basal ganglia (BG) of Parkinson's patients gives rise to both frontal-like and implicit learning impairments. Dopaminergic medication alleviates some cognitive deficits but impairs those that depend on intact areas of the BG, apparently due to DA ''overdose.'' These findings are difficult to accommodate with verbal theories of BG/DA function, owing to complexity of system dynamics: DA dynamically modulates function in the BG, which is itself a modulatory system. This article presents a neural network model that instantiates key biological properties and provides insight into the underlying role of DA in the BG during learning and execution of cognitive tasks. Specifically, the BG modulates the execution of ''actions'' e.g., motor different parts of the frontal cortex. Phasic changes in DA, which occur during error feedback, dynamically modulate the BG threshold for facilitating/suppressing a cortical command in response to particular stimuli. Reduced dynamic range of DA explains Parkinson and DA overdose deficits with a single underlying dysfunction, despite overall differences in raw DA levels. Simulated Parkinsonism and medication effects provide a theoretical basis for behavioral data in probabilistic classification and reversal tasks. The model also provides novel testable predictions for neuropsychological and pharmacological studies, and motivates further investigation of BG/DA interactions with the prefrontal cortex in working memory.}
}
@article{FrankandBadreCEREBRAL-CORTEX-12,
       author = {Frank, M. J. and Badre, D.},
        title = {Mechanisms of hierarchical reinforcement learning in corticostriatal circuits {I}: computational analysis},
      journal = {Cerebral Cortex},
         year = {2012},
       volume = {22},
       number = {3},
        pages = {509-526},
      comment = {See~\cite{FrankandBadreCEREBRAL-CORTEX-12.Supplements.pdf} for detail on the model and experimental methods},
     abstract = {Growing evidence suggests that the prefrontal cortex (PFC) is organized hierarchically, with more anterior regions having increasingly abstract representations. How does this organization support hierarchical cognitive control and the rapid discovery of abstract action rules? We present computational models at different levels of description. A neural circuit model simulates interacting corticostriatal circuits organized hierarchically. In each circuit, the basal ganglia gate frontal actions, with some striatal units gating the inputs to PFC and others gating the outputs to influence response selection. Learning at all of these levels is accomplished via dopaminergic reward prediction error signals in each corticostriatal circuit. This functionality allows the system to exhibit conditional if-then hypothesis testing and to learn rapidly in environments with hierarchical structure. We also develop a hybrid Bayesian-reinforcement learning mixture of experts (MoE) model, which can estimate the most likely hypothesis state of individual participants based on their observed sequence of choices and rewards. This model yields accurate probabilistic estimates about which hypotheses are attended by manipulating attentional states in the generative neural model and recovering them with the MoE model. This 2-pronged modeling approach leads to multiple quantitative predictions that are tested with functional magnetic resonance imaging in the companion paper.}
}   
@misc{FrankandBadreCEREBRAL-CORTEX-12.Supplements,
        title = {Mechanisms of hierarchical reinforcement learning in corticostriatal circuits {I}: computational analysis},
       author = {Frank, M. J. and Badre, D.},
 howpublished = {Supplementary materials for Frank and Badre. Cerebral Cortex. 2012},
         year = {2012},
}
@article{FranketalCOGNITION-08,
       author = {Frank, M. C. and Everett, D. L. and Fedorenko, E. and Gibson, E.},
        title = {Number as a cognitive technology: evidence from {P}irah\~{a} language and cognition},
      journal = {Cognition},
         year = {2008},
       volume = {108},
       number = {3},
        pages = {819-824},
     abstract = {Does speaking a language without number words change the way speakers of that language perceive exact quantities? The Pirah\~{a} are an Amazonian tribe who have been previously studied for their limited numerical system [Gordon, P. (2004). Numerical cognition without words: Evidence from Amazonia. Science 306, 496-499]. We show that the Pirah have no linguistic method whatsoever for expressing exact quantity, not even "one." Despite this lack, when retested on the matching tasks used by Gordon, Pirah speakers were able to perform exact matches with large numbers of objects perfectly but, as previously reported, they were inaccurate on matching tasks involving memory. These results suggest that language for exact number is a cultural invention rather than a linguistic universal, and that number words do not change our underlying representations of number but instead are a cognitive technology for keeping track of the cardinality of large sets across time, space, and changes in modality.}
}
@book{FranketalCOGNITIVE-NEUROSCIENCE-12,
        title = {Computational Cognitive Neuroscience},
       author = {Michael Frank and Yuko Munakata and Thomas Hazy and Randall O'Reilly},
    publisher = {Amazon Digital Services},
         year = {2012}
}
@article{FranklinandFrankPLoS-18,
        title = {Compositional clustering in task structure learning},
       author = {Franklin, Nicholas T. and Frank, Michael J.},
      journal = {{PLoS} Computational Biology},
    publisher = {Cold Spring Harbor Laboratory},
       volume = 4,
        issue = 4,
         year = 2018,
        pages = {e1006116},
     abstract = {Humans are remarkably adept at generalizing knowledge between experiences in a way that can be difficult for computers. Often, this entails generalizing constituent pieces of experiences that do not fully overlap, but nonetheless share useful similarities with, previously acquired knowledge. However, it is often unclear how knowledge gained in one context should generalize to another. Previous computational models and data suggest that rather than learning about each individual context, humans build latent abstract structures and learn to link these structures to arbitrary contexts, facilitating generalization. In these models, task structures that are more popular across contexts are more likely to be revisited in new contexts. However, these models can only re-use policies as a whole and are unable to transfer knowledge about the transition structure of the environment even if only the goal has changed (or vice-versa). This contrasts with ecological settings, where some aspects of task structure, such as the transition function, will be shared between context separately from other aspects, such as the reward function. Here, we develop a novel non-parametric Bayesian agent that forms independent latent clusters for transition and reward functions, affording separable transfer of their constituent parts across contexts. We show that the relative performance of this agent compared to an agent that jointly clusters reward and transition functions depends environmental task statistics: the mutual information between transition and reward functions and the stochasticity of the observations. We formalize our analysis through an information theoretic account of the priors, and propose a meta learning agent that dynamically arbitrates between strategies across task domains to optimize a statistical tradeoff.}
}
@article{FranklinandGoslingNATURE-53,
       author = {Franklin, R., Gosling, R. G.},
        title = {Evidence for 2-Chain Helix in Crystalline Structure of Sodium Deoxyribonucleate},
      journal = {Nature},
         year = {1953},
       volume = {171},
        issue = {4356},
        pages = {156-157},
}
@article{FranklinandWolpertNEURON-11,
        title = {Computational mechanisms of sensorimotor control},
       author = {Franklin, D.W. and Wolpert, D.M.},
      journal = {Neuron},
       volume = 72,
        issue = 3,
         year = 2011,
        pages = {425-442},
     abstract = {In order to generate skilled and efficient actions, the motor system must find solutions to several problems inherent in sensorimotor control, including nonlinearity, nonstationarity, delays, redundancy, uncertainty, and noise. We review these problems and five computational mechanisms that the brain may use to limit their deleterious effects: optimal feedback control, impedance control, predictive control, Bayesian decision theory, and sensorimotor learning. Together, these computational mechanisms allow skilled and fluent sensorimotor behavior.}
}
@article{FrascaetalCoRR-20,
        title = {{SIGN: Scalable Inception Graph Neural Networks}},
       author = {Fabrizio Frasca and Emanuele Rossi and Davide Eynard and Ben Chamberlain and Michael Bronstein and Federico Monti},
      journal = {CoRR},
         year = {2020},
       volume = {arXiv:2004.11198},
     abstract = {Graph representation learning has recently been applied to a broad spectrum of problems ranging from computer graphics and chemistry to high energy physics and social media. The popularity of graph neural networks has sparked interest, both in academia and in industry, in developing methods that scale to very large graphs such as Facebook or Twitter social networks. In most of these approaches, the computational cost is alleviated by a sampling strategy retaining a subset of node neighbors or subgraphs at training time. In this paper we propose a new, efficient and scalable graph deep learning architecture which sidesteps the need for graph sampling by using graph convolutional filters of different size that are amenable to efficient precomputation, allowing extremely fast training and inference. Our architecture allows using different local graph operators (e.g. motif-induced adjacency matrices or Personalized Page Rank diffusion matrix) to best suit the task at hand. We conduct extensive experimental evaluation on various open benchmarks and show that our approach is competitive with other state-of-the-art architectures, while requiring a fraction of the training and inference time. Moreover, we obtain state-of-the-art results on ogbn-papers100M, the largest public graph dataset, with over 110 million nodes and 1.5 billion edges.}
}
@incollection{FraserandDimitriadis94,
       author = {Fraser, Andrew M. and Dimitriadis, Alexis},
        title = {Forecasting Probability Densities by Using Hidden {M}arkov Models with Mixed States},
    booktitle = {Time Series Prediction: Forecasting the Future and Understanding the Past},
       editor = {Weigend, Andreas S. and Gershenfeld, Neil A.},
    publisher = {Addison-Wesley},
         year = 1994,
}
@article{FreemanCOIN-15,
       author = {Jeremy Freeman},
        title = {Open source tools for large-scale neuroscience},
      journal = {Current Opinion in Neurobiology},
       volume = {32},
         year = {2015},
        pages = {156-163},
     abstract = {New technologies for monitoring and manipulating the nervous system promise exciting biology but pose challenges for analysis and computation. Solutions can be found in the form of modern approaches to distributed computing, machine learning, and interactive visualization. But embracing these new technologies will require a cultural shift: away from independent efforts and proprietary methods and toward an open source and collaborative neuroscience.}
}
@article{FreemanetalIJCV-00,
       author = {W. T. Freeman and E. C. Pasztor and O. T. Carmichael},
        title = {Learning Low-Level Vision},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
         year = 2000,
       volume = 40,
        issue = 1,
        pages = {24-57},
}
@article{FreemanetalNATURE-METHODS-14,
       author = {Freeman, Jeremy and Vladimirov, Nikita and Kawashima, Takashi and Mu, Yu and Sofroniew, Nicholas J. and Bennett, Davis V. and Rosen, Joshua and Yang, Chao-Tsung and Looger, Loren L. and Ahrens, Misha B.},
        title = {Mapping brain activity at scale with cluster computing},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = {11},
        issue = {9},
         year = {2014},
        pages = {941-950},
     abstract = {Understanding brain function requires monitoring and interpreting the activity of large networks of neurons during behavior. Advances in recording technology are greatly increasing the size and complexity of neural data. Analyzing such data will pose a fundamental bottleneck for neuroscience. We present a library of analytical tools called Thunder built on the open-source Apache Spark platform for large-scale distributed computing. The library implements a variety of univariate and multivariate analyses with a modular, extendable structure well-suited to interactive exploration and analysis development. We demonstrate how these analyses find structure in large-scale neural data, including whole-brain light-sheet imaging data from fictively behaving larval zebrafish, and two-photon imaging data from behaving mouse. The analyses relate neuronal responses to sensory input and behavior, run in minutes or less and can be used on a private cluster or in the cloud. Our open-source framework thus holds promise for turning brain activity mapping efforts into biological insights.},
}
@misc{Freenet,
       author = {Freenet},
        title = {The Free Network Project},
 howpublished = {\url{http://freenet.sourceforge.net/}},
         year = 2001,
}
@incollection{French84,
       author = {French, Simon},
        title = {Fuzzy decision analysis: Some Criticisms},
    booktitle = {Fuzzy Sets and Decision Analysis},
       editor = {Zimmermann, H.-J. and Zadeh, L. A. and Gaines, B. R.},
    publisher = {North-Holland},
 series       = {{TIMS} Studies in the Management Sciences},
       volume = 20,
         year = 1984,
        pages = {29-44},
}
@article{FrenchCONNECTION-SCIENCE-97,
       author = {Robert M. French},
        title = {Pseudo-recurrent Connectionist Networks: An Approach to the 'Sensitivity-Stability' Dilemma},
      journal = {Connection Science},
    publisher = {Taylor \& Francis},
       volume = {9},
       number = {4},
         year = {1997},
        pages = {353-380},
     abstract = {In order to solve the 'sensitivity-stability' problem-and its immediate correlate, the problem of sequential learning-it is crucial to develop connectionist architectures that are simultaneously sensitive to, but not excessively disrupted by, new input. French (1992) suggested that to alleviate a particularly severe form of this disruption, catastrophic forgetting, it was necessary for networks to separate dynamically their internal representations during learning. McClelland et al. (1995) went even further. They suggested that nature's way of implementing this obligatory separation was the evolution of two separate areas of the brain, the hippocampus and the neocortex. In keeping with this idea of radical separation, a 'pseudo-recurrent' memory model is presented here that partitions a connectionist network into two functionally distinct, but continually interacting areas. One area serves as a final-storage area for representations; the other is an early-processing area where new representations are first learned by the system. The final-storage area continually supplies internally generated patterns (pseudopatterns; Robins, 1995), which are approximations of its content, to the early-processing area, where they are interleaved with the new patterns to be learned. Transfer of the new learning is done either by weight-copying from the early-processing area to the final-storage area or by pseudopattern transfer. A number of experiments are presented that demonstrate the effectiveness of this approach, allowing, in particular, effective sequential learning with gradual forgetting in the presence of new input. Finally, it is shown that the two interacting areas automatically produce representational compaction and it is suggested that similar representational streamlining may exist in the brain.}
}
@article{FrenchTiCS-99,
       author = {French, Robert},
        title = {Catastrophic Forgetting in Connectionist Networks},
      journal = {Trends in Cognitive Sciences},
       volume = {3},
         year = {1999},
        pages = {128-135},
     abstract = {All natural cognitive systems, and, in particular, our own, gradually forget previously learned information. Plausible models of human cognition should therefore exhibit similar patterns of gradual forgetting of old information as new information is acquired. Only rarely does new learning in natural cognitive systems completely disrupt or erase previously learned information; that is, natural cognitive systems do not, in general, forget 'catastrophically'. Unfortunately, though, catastrophic forgetting does occur under certain circumstances in distributed connectionist networks. The very features that give these networks their remarkable abilities to generalize, to function in the presence of degraded input, and so on, are found to be the root cause of catastrophic forgetting. The challenge in this field is to discover how to keep the advantages of distributed connectionist networks while avoiding the problem of catastrophic forgetting. In this article the causes, consequences and numerous solutions to the problem of catastrophic forgetting in neural networks are examined. The review will consider how the brain might have overcome this problem and will also explore the consequences of this solution for distributed connectionist networks.},
}
@article{FreundIC-95,
       author = {Yoav Freund},
        title = {Boosting a weak learning algorithm by majority},
      journal = {Information and Computation},
       volume = 121,
        issue = 2,
        pages = {256-285},
         year = 1995,
}
@inproceedings{FreundandRonCOLT-95,
       author = {Yoav Freund and Dana Ron},
        title = {Learning to Model Sequences Generated by Switching Distributions},
    booktitle = {{COLT}: Proceedings of the Workshop on Computational Learning Theory},
    publisher = {Morgan Kaufmann Publishers},
        pages = {41-50},
         year = 1995,
}
@article{FreundandSchapireHEDGE,
       author = {Freund, Y. and Schapire, R.},
        title = {An efficient boosting algorithm for combining preferences},
      journal = {Journal of Computer and System Sciences},
       volume = 55,
        issue = 1,
        pages = {119-139},
         year = 1997,
}
@incollection{FreyandJojicNIPS-02,
       author = {B. Frey and N. Jojic},
        title = {Fast, large-scale transformation-invariant clustering},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 14,
      address = {Cambridge, MA},
    publisher = {MIT Press},
         year = 2002,
}
@inproceedings{FriedmanUAI-98,
       author = {Nir Friedman},
        title = {The {Bayesian} Structural {EM} Algorithm},
    booktitle = {Proceedings of the Conference on Uncertainty in Artificial Intelligence},
        pages = {129-138},
         year = 1998,
}
@inproceedings{FriedmanandGoldszmidtUAI-96,
       author = {Friedman, Nir and Goldszmidt, Moises},
        title = {Learning {B}ayesian Networks with Local Structure},
    booktitle = {Twelfth Conference on Uncertainty in Artificial Intelligence},
     location = {Portland, Oregon},
         year = 1996,
}
@incollection{FriedmanandSingerNIPS-99,
       author = {N. Friedman and Y. Singer},
        title = {Efficient {B}ayesian parameter estimation in large discrete domains},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 11,
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1999,
}
@article{FriedmanetalAAS-07,
     author = {Friedman, Jerome  and Hastie, Trevor and H\"{o}fling, Holger and Tibshirani, Robert},
    journal = {Annals of Applied Statistics},
     volume = 1,
      issue = 2,
      pages = {302-332},
      title = {Pathwise coordinate optimization},
       year = {2007},
   abstract = {We consider "one-at-a-time" coordinate-wise descent algorithms for a class of convex optimization problems. An algorithm of this kind has been proposed for the L1-penalized regression (lasso) in the literature, but it seems to have been largely ignored. Indeed, it seems that coordinate-wise algorithms are not often used in convex optimization. We show that this algorithm is very competitive with the well-known LARS (or homotopy) procedure in large lasso problems, and that it can be applied to related methods such as the garotte and elastic net. It turns out that coordinate-wise descent does not work in the "fused lasso," however, so we derive a generalized algorithm that yields the solution in much less time that a standard convex optimizer. Finally, we generalize the procedure to the two-dimensional fused lasso, and demonstrate its performance on some image smoothing problems.},
}
@article{FriedmanetalJASA-84,
    author = {J. Friedman and W. Stuetzele and A. Schroeder},
     title = {Projection pursuit density estimation},
   journal = {Journal American Statistical Association},
      year = 1984,
    volume = 79,
     pages = {599-608}
}
@article{FriedmanetalML-97,
    author = {Nir Friedman and Dan Geiger and Moises Goldszmidt},
     title = {Bayesian Network Classifiers},
   journal = {Machine Learning},
    volume = 29,
     issue = {2-3},
     pages = {131-163},
      year = 1997,
}
@inproceedings{FriedmanetalUAI-98,
     title = {Learning the structure of {DBN}s},
    author = {Nir Friedman and Kevin Murphy and Stuart Russell},
 booktitle = {Proceedings of the Conference on Uncertainty in Artificial Intelligence},
     pages = {138-147},
      year = 1998,
}
@article {FriedrichetalBIORXIV-16,
       author = {Friedrich, Johannes and Yang, Weijian and Soudry, Daniel and Mu, Yu and Ahrens, Misha B and Yuste, Rafael and Peterka, Darcy S and Paninski, Liam},
        title = {Multi-scale approaches for high-speed imaging and analysis of large neural populations},
      journal = {bioRxiv},
         year = {2016},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {Progress in modern neuroscience critically depends on our ability to observe the activity of large neuronal populations with cellular spatial and high temporal resolution. However, two bottlenecks constrain efforts towards fast imaging of large populations. First, the resulting large video data is challenging to analyze. Second, there is an explicit tradeoff between imaging speed, signal-to-noise, and field of view: with current recording technology we cannot image very large neuronal populations with simultaneously high spatial and temporal resolution. Here we describe multi-scale approaches for alleviating both of these bottlenecks. First, we show that spatial and temporal decimation techniques provide order-of-magnitude speedups in spatiotemporally demixing calcium video data into estimates of single-cell neural activity. Second, once the shapes of individual neurons have been identified (e.g., after an initial phase of conventional imaging with standard temporal and spatial resolution), we find that the spatial/temporal resolution tradeoff shifts dramatically: after demixing we can accurately recover neural activity from data that has been spatially decimated by an order of magnitude. This offers a cheap method for compressing this large video data, and also implies that it is possible to either speed up imaging significantly, or to "zoom out" by a corresponding factor to image order-of-magnitude larger neuronal populations with minimal loss in accuracy or temporal resolution.},
}
@article{FriedrichetalFiNC-13,
       author = {Friedrich, Rainer W. and Genoud, Christel and Wanner, Adrian A.},
        title = {Analyzing the structure and function of neuronal circuits in zebrafish},
      journal = {Frontiers in Neural Circuits},
    publisher = {Frontiers Media S.A.},
       volume = {7},
         year = {2013},
        pages = {71},
     abstract = {The clever choice of animal models has been instrumental for many breakthrough discoveries in life sciences. One of the outstanding challenges in neuroscience is the in-depth analysis of neuronal circuits to understand how interactions between large numbers of neurons give rise to the computational power of the brain. A promising model organism to address this challenge is the zebrafish, not only because it is cheap, transparent and accessible to sophisticated genetic manipulations but also because it offers unique advantages for quantitative analyses of circuit structure and function. One of the most important advantages of zebrafish is its small brain size, both at larval and adult stages. Small brains enable exhaustive measurements of neuronal activity patterns by optical imaging and facilitate large-scale reconstructions of wiring diagrams by electron microscopic approaches. Such information is important, and probably essential, to obtain mechanistic insights into neuronal computations underlying higher brain functions and dysfunctions. This review provides a brief overview over current methods and motivations for dense reconstructions of neuronal activity and connectivity patterns. It then discusses selective advantages of zebrafish and provides examples how these advantages are exploited to study neuronal computations in the olfactory bulb.},
}
@article{FriedrichetalPLoS-17,
       author = {Friedrich, Johannes and Zhou, Pengcheng and Paninski, Liam},
      journal = {{PLoS} Computational Biology},
    publisher = {Public Library of Science},
        title = {Fast online deconvolution of calcium imaging data},
         year = {2017},
       volume = {13},
       number = {3},
        pages = {1-26},
     abstract = {Calcium imaging methods enable simultaneous measurement of the activity of thousands of neighboring neurons, but come with major caveats: the slow decay of the fluorescence signal compared to the time course of the underlying neural activity, limitations in signal quality, and the large scale of the data all complicate the goal of efficiently extracting accurate estimates of neural activity from the observed video data. Further, current activity extraction methods are typically applied to imaging data after the experiment is complete. However, in many cases we would prefer to run closed-loop experiments--analyzing data on-the-fly to guide the next experimental steps or to control feedback--and this requires new methods for accurate real-time processing. Here we present a fast activity extraction algorithm addressing both issues. Our approach follows previous work in casting the activity extraction problem as a sparse nonnegative deconvolution problem. To solve this optimization problem, we introduce a new algorithm that is an order of magnitude faster than previous methods, and progresses through the data sequentially from beginning to end, thus enabling, in principle, real-time online estimation of neural activity during the imaging session. This computational advance thus opens the door to new closed-loop experiments.},
}
@article{FristonENTROPY-12,
       author = {Karl, Friston},
        title = {A Free Energy Principle for Biological Systems},
      journal = {Entropy (Basel)},
         year = {2012},
       volume = {14},
       number = {11},
        pages = {2100-2121},
     abstract = {This paper describes a free energy principle that tries to explain the ability of biological systems to resist a natural tendency to disorder. It appeals to circular causality of the sort found in synergetic formulations of self-organization (e.g., the slaving principle) and models of coupled dynamical systems, using nonlinear Fokker Planck equations. Here, circular causality is induced by separating the states of a random dynamical system into external and internal states, where external states are subject to random fluctuations and internal states are not. This reduces the problem to finding some (deterministic) dynamics of the internal states that ensure the system visits a limited number of external states; in other words, the measure of its (random) attracting set, or the Shannon entropy of the external states is small. We motivate a solution using a principle of least action based on variational free energy (from statistical physics) and establish the conditions under which it is formally equivalent to the information bottleneck method. This approach has proved useful in understanding the functional architecture of the brain. The generality of variational free energy minimisation and corresponding information theoretic formulations may speak to interesting applications beyond the neurosciences; e.g., in molecular or evolutionary biology.}
}
@article{FristonNATURE-NEUROSCIENCE-18,
       author = {Friston, Karl},
        title = {Does predictive coding have a future?},
      journal = {Nature Neuroscience},
       volume = {21},
        issue = {8},
         year = {2018},
     abstract = {In the 20th century we thought the brain extracted knowledge from sensations. The 21st century witnessed a 'strange inversion', in which the brain became an organ of inference, actively constructing explanations for what's going on 'out there', beyond its sensory epithelia. One paper played a key role in this paradigm shift.},
}
@article{FristonNATURE-REVIEWS-NEUROSCIENCE-10,
       author = {Friston, Karl},
        title = {The free-energy principle: a unified brain theory?},
      journal = {Nature Reviews Neuroscience},
         year = {2010},
    publisher = {Nature Publishing Group},
       volume = {11},
        issue = {2},
        pages = {127-138},
     abstract = {A free-energy principle has been proposed recently that accounts for action, perception and learning. This Review looks at some key brain theories in the biological (for example, neural Darwinism) and physical (for example, information theory and optimal control theory) sciences from the free-energy perspective. Crucially, one key theme runs through each of these theories -- optimization. Furthermore, if we look closely at what is optimized, the same quantity keeps emerging, namely value (expected reward, expected utility) or its complement, surprise (prediction error, expected cost). This is the quantity that is optimized under the free-energy principle, which suggests that several global brain theories might be unified within a free-energy framework.}
}
@article{FristonNEURON-11,
        title = {What Is Optimal about Motor Control?},
      journal = {Neuron},
       volume = {72},
       number = {3},
        pages = {488-498},
         year = {2011},
       author = {Karl Friston},
     abstract = {This article poses a controversial question: is optimal control theory useful for understanding motor behavior or is it a misdirection? This question is becoming acute as people start to conflate internal models in motor control and perception (Poeppel et al., 2008, Hickok et al., 2011). However, the forward models in motor control are not the generative models used in perceptual inference. This Perspective tries to highlight the differences between internal models in motor control and perception and asks whether optimal control is the right way to think about things. The issues considered here may have broader implications for optimal decision theory and Bayesian approaches to learning and behavior in general.}
}
@article{FristonNRN-10,
       author = {Friston, Karl},
        title = {The free-energy principle: a unified brain theory?},
      journal = {Nature Reviews Neuroscience},
       volume = {11},
         year = {2010},
        pages = {127-38},
     abstract = {A free-energy principle has been proposed recently that accounts for action, perception and learning. This Review looks at some key brain theories in the biological (for example, neural Darwinism) and physical (for example, information theory and optimal control theory) sciences from the free-energy perspective. Crucially, one key theme runs through each of these theories - optimization. Furthermore, if we look closely at what is optimized, the same quantity keeps emerging, namely value (expected reward, expected utility) or its complement, surprise (prediction error, expected cost). This is the quantity that is optimized under the free-energy principle, which suggests that several global brain theories might be unified within a free-energy framework.},
}
@article{FristonandKiebelPTRSLB-09,
       author = {Friston, Karl and Kiebel, Stefan},
        title = {Predictive coding under the free-energy principle},
      journal = {Philosphical Transactions Royal Society London B Biological Science},
         year = {2009},
    publisher = {The Royal Society},
      address = {London},
       volume = {364},
        issue = {1521},
        pages = {1211-1221},
     abstract = {This paper considers prediction and perceptual categorization as an inference problem that is solved by the brain. We assume that the brain models the world as a hierarchy or cascade of dynamical systems that encode causal structure in the sensorium. Perception is equated with the optimization or inversion of these internal models, to explain sensory data. Given a model of how sensory data are generated, we can invoke a generic approach to model inversion, based on a free energy bound on the model's evidence. The ensuing free-energy formulation furnishes equations that prescribe the process of recognition, i.e. the dynamics of neuronal activity that represent the causes of sensory input. Here, we focus on a very general model, whose hierarchical and dynamical structure enables simulated brains to recognize and predict trajectories or sequences of sensory states. We first review hierarchical dynamical models and their inversion. We then show that the brain has the necessary infrastructure to implement this inversion and illustrate this point using synthetic birds that can recognize and categorize birdsongs.},
}
@article{FristonetalFiP-12,
       author = {Friston, Karl and Adams, Rick and Perrinet, Laurent and Breakspear, Michael},
        title = {Perceptions as Hypotheses: Saccades as Experiments},
      journal = {Frontiers in Psychology},
       volume = {3},
        pages = {151},
         year = {2012},
     abstract = {If perception corresponds to hypothesis testing (Gregory,1980); then visual searches might be construed as experiments that generate sensory data. In this work, we explore the idea that saccadic eye movements are optimal experiments, in which data are gathered to test hypotheses or beliefs about how those data are caused. This provides a plausible model of visual search that can be motivated from the basic principles of self-organized behavior: namely, the imperative to minimize the entropy of hidden states of the world and their sensory consequences. This imperative is met if agents sample hidden states of the world efficiently. This efficient sampling of salient information can be derived in a fairly straightforward way, using approximate Bayesian inference and variational free-energy minimization. Simulations of the resulting active inference scheme reproduce sequential eye movements that are reminiscent of empirically observed saccades and provide some counterintuitive insights into the way that sensory evidence is accumulated or assimilated into beliefs about the world.}
}
@article{FristonetalPLOS-ONE-09,
       author = {Friston, Karl and Daunizeau, Jean and Kiebel, Stefan},
        title = {Reinforcement Learning or Active Inference?},
      journal = {PloS one},
         year = {2009},
       volume = {4},
        pages = {e6421},
     abstract = {This paper questions the need for reinforcement learning or control theory when optimising behaviour. We show that it is fairly simple to teach an agent complicated and adaptive behaviours using a free-energy formulation of perception. In this formulation, agents adjust their internal states and sampling of the environment to minimize their free-energy. Such agents learn causal structure in the environment and sample it in an adaptive and self-supervised fashion. This results in behavioural policies that reproduce those optimised by reinforcement learning and dynamic programming. Critically, we do not need to invoke the notion of reward, value or utility. We illustrate these points by solving a benchmark problem in dynamic programming; namely the mountain-car problem, using active perception or inference under the free-energy principle. The ensuing proof-of-concept may be important because the free-energy formulation furnishes a unified account of both action and perception and may speak to a reappraisal of the role of dopamine in the brain.},
}
@book{Frith2007,
     author = {Chris Frith},
      title = {Making Up the Mind: How the Brain Creates our Mental World},
  publisher = {Blackwell},
    address = {Malden, MA},
       year = 2007,
    comment = {\urlh{http://en.wikipedia.org/wiki/Chris_Frith}{Chris Frith}},
}
@book{Frith2009making,
        title = {Making up the Mind: How the Brain Creates Our Mental World},
       author = {Frith, Chris},
         year = {2009},
    publisher = {Wiley}
}
@inproceedings{FromeetalDeVISE-13,
    author = {Andrea Frome and Greg S. Corrado and Jonathon Shlens and Samy Bengio and Jeffrey Dean and Marc'Aurelio Ranzato and Tom\`{a}s Mikolov},
     title = {{DeViSE}: {A} Deep Visual-Semantic Embedding Model},
 booktitle = {Advances in Neural Information Processing Systems 26},
      year = 2013,
     pages = {2121-2129}
}
@inproceedings{FromeetalICCV-07,
    author = {Andrea Frome and Fei Sha and Yoram Singer and Jitendra Malik}, 
     title = {Learning Globally-Consistent Local Distance Functions for Shape-Based Image Retrieval and Classification},
 booktitle = {{IEEE} International Conference on Computer Vision},
     month = {October},
      year = 2007,
}
@inproceedings{FromeetalNIPS-13,
    author = {Andrea Frome and Greg S. Corrado and Jonathon Shlens and Samy Bengio and Jeffrey Dean and Marc'Aurelio Ranzato and Tom\`{a}s Mikolov},
     title = {{DeViSE}: {A} Deep Visual-Semantic Embedding Model},
 booktitle = {Advances in Neural Information Processing Systems 26},
      year = 2013,
     pages = {2121-2129}
}
@book{FrontendVision,
      author = {Bart M.ter Haar Romeny},
       title = {Front-End Vision and Multiscale Image Analysis},
   publisher = {Kluwer Academic Publishers},
     address = {Boston, Massachusetts},
        year = 2002,
}
@article{FryetalJNM-08,
       author = {Fry, S. N. and Rohrseitz, N. and Straw, A. D. and Dickinson, M. H.},
        title = {{T}rack{F}ly: virtual reality for a behavioral system analysis in free-flying fruit flies},
      journal = {Journal of Neuroscience Methods},
       volume = {171},
       number = {1},
         year = {2008},
        pages = {110-117},
     abstract = {Modern neuroscience and the interest in biomimetic control design demand increasingly sophisticated experimental techniques that can be applied in freely moving animals under realistic behavioral conditions. To explore sensorimotor flight control mechanisms in free-flying fruit flies (Drosophila melanogaster), we equipped a wind tunnel with a Virtual Reality (VR) display system based on standard digital hardware and a 3D path tracking system. We demonstrate the experimental power of this approach by example of a 'one-parameter open loop' testing paradigm. It provided (1) a straightforward measure of transient responses in presence of open loop visual stimulation; (2) high data throughput and standardized measurement conditions from process automation; and (3) simplified data analysis due to well-defined testing conditions. Being based on standard hardware and software techniques, our methods provide an affordable, easy to replicate and general solution for a broad range of behavioral applications in freely moving animals. Particular relevance for advanced behavioral research tools originates from the need to perform detailed behavioral analyses in genetically modified organisms and animal models for disease research.}
}
@article{FuIEEE-78,
   author =  {Fu, K. S.},
   title =   {Learning Control Systems and Intelligent Control Systems: {A}n
              Intersection of Artificial Intelligence and Automatic Control},
   journal = {{IEEE} Transactions on Automatic Control},
   volume =  16,
    issue =  1,
   pages =   {70-72},
   year =    1971,
}
@inproceedings{FuWSC-2018,
       author = {M. C. {Fu}},
    booktitle = {2018 Winter Simulation Conference (WSC)},
       editor = {M. Rabe and A. A. Juan and N. Mustafee and A. Skoogh and S. Jain and B. Johansson},
        title = {Monte Carlo Tree Search: A Tutorial},
         year = {2018},
        pages = {222-236},
     abstract = {Monte Carlo tree search (MCTS) is a general approach to solving game problems, playing a central role in Google DeepMind's AlphaZero and its predecessor AlphaGo, which famously defeated the (human) world Go champion Lee Sedol in 2016 and world #1 Go player Ke Jie in 2017. Starting from scratch without using any domain-specific knowledge (other than the game rules), AlphaZero defeated not only its ancestors in Go but also the best computer programs in chess (Stockfish) and shogi (Elmo). In this tutorial, we provide an introduction to MCTS, including a review of its history and relationship to a more general simulation-based algorithm for Markov decision processes (MDPs) published in a 2005 Operations Research article; a demonstration of the basic mechanics of the algorithms via decision trees and the game of tic-tac-toe; and its use in AlphaGo and AlphaZero.}
}
@article{FuchsandClevelandSCIENCE-98,
       author = {Fuchs, Elaine and Cleveland, Don W.},
        title = {A Structural Scaffolding of Intermediate Filaments in Health and Disease},
      journal = {Science},
       volume = {279},
        issue = {5350},
         year = {1998},
        pages = {514-519},
     abstract = {The cytoplasm of animal cells is structured by a scaffolding composed of actin microfilaments, microtubules, and intermediate filaments. Intermediate filaments, so named because their 10-nanometer diameter is intermediate between that of microfilaments (6 nanometers) and microtubules (23 nanometers), assemble into an anastomosed network within the cytoplasm. In combination with a recently identified class of cross-linking proteins that mediate interactions between intermediate filaments and the other cytoskeletal networks, evidence is reviewed here that intermediate filaments provide a flexible intracellular scaffolding whose function is to structure cytoplasm and to resist stresses externally applied to the cell. Mutations that weaken this structural framework increase the risk of cell rupture and cause a variety of human disorders.},
}
@article{FuetalCoRR-19,
       author = {Justin Fu and Anoop Korattikara and Sergey Levine and Sergio Guadarrama},
        title = {From Language to Goals: Inverse Reinforcement Learning for Vision-Based Instruction Following},
      journal = {CoRR},
       volume = {arXiv:1902.07742},
         year = {2019},
     abstract = {Solving grounded language tasks often requires reasoning about relationships between objects in the context of a given task. For example, to answer the question "What color is the mug on the plate?" we must check the color of the specific mug that satisfies the "on" relationship with respect to the plate. Recent work has proposed various methods capable of complex relational reasoning. However, most of their power is in the inference structure, while the scene is represented with simple local appearance features. In this paper, we take an alternate approach and build contextualized representations for objects in a visual scene to support relational reasoning. We propose a general framework of Language-Conditioned Graph Networks (LCGN), where each node represents an object, and is described by a context-aware representation from related objects through iterative message passing conditioned on the textual input. E.g., conditioning on the "on" relationship to the plate, the object "mug" gathers messages from the object "plate" to update its representation to "mug on the plate", which can be easily consumed by a simple classifier for answer prediction. We experimentally show that our LCGN approach effectively supports relational reasoning and improves performance across several tasks and datasets.},
}
@inproceedings{FuetalICLR-19,
        title = {From Language to Goals: Inverse Reinforcement Learning for Vision-Based Instruction Following},
       author = {Justin Fu and Anoop Korattikara and Sergey Levine and Sergio Guadarrama},
    booktitle = {International Conference on Learning Representations},
         year = {2019},
     abstract = {Reinforcement learning is a promising framework for solving control problems, but its use in practical situations is hampered by the fact that reward functions are often difficult to engineer. Specifying goals and tasks for autonomous machines, such as robots, is a significant challenge: conventionally, reward functions and goal states have been used to communicate objectives. But people can communicate objectives to each other simply by describing or demonstrating them. How can we build learning algorithms that will allow us to tell machines what we want them to do? In this work, we investigate the problem of grounding language commands as reward functions using inverse reinforcement learning, and argue that language-conditioned rewards are more transferable than language-conditioned policies to new environments. We propose language-conditioned reward learning (LC-RL), which grounds language commands as a reward function represented by a deep neural network. We demonstrate that our model learns rewards that transfer to novel tasks and environments on realistic, high-dimensional visual environments with natural language commands, whereas directly learning a language-conditioned policy leads to poor performance.}
}
@article{FuetalPNAS-17,
       author = {Fu, Tian-Ming and Hong, Guosong and Viveros, Robert D. and Zhou, Tao and Lieber, Charles M.},
        title = {Highly scalable multichannel mesh electronics for stable chronic brain electrophysiology},
      journal = {Proceedings of the National Academy of Sciences of the United States of America},
    publisher = {National Academy of Sciences},
       volume = {114},
        issue = {47},
         year = {2017},
	 pages= {10046-E10055},
     abstract = {Implantable electrical probes have led to advances in neuroscience, brain-machine interfaces, and treatment of neurological diseases, yet they remain limited in several key aspects. Ideally, an electrical probe should be capable of recording from large numbers of neurons across multiple local circuits and, importantly, allow stable tracking of the evolution of these neurons over the entire course of study. Silicon probes based on microfabrication can yield large-scale, high-density recording but face challenges of chronic gliosis and instability due to mechanical and structural mismatch with the brain. Ultraflexible mesh electronics, on the other hand, have demonstrated negligible chronic immune response and stable long-term brain monitoring at single-neuron level, although, to date, it has been limited to 16 channels. Here, we present a scalable scheme for highly multiplexed mesh electronics probes to bridge the gap between scalability and flexibility, where 32 to 128 channels per probe were implemented while the crucial brain-like structure and mechanics were maintained. Combining this mesh design with multisite injection, we demonstrate stable 128-channel local field potential and single-unit recordings from multiple brain regions in awake restrained mice over 4 mo. In addition, the newly integrated mesh is used to validate stable chronic recordings in freely behaving mice. This scalable scheme for mesh electronics together with demonstrated long-term stability represent important progress toward the realization of ideal implantable electrical probes allowing for mapping and tracking single-neuron level circuit changes associated with learning, aging, and neurodegenerative diseases.},
}
@incollection{FujiiHBN-08,
       author = {Toshikatsu Fujii},
       editor = {Dere, E. and Easton, A. and Nadel, L. and Huston, J.P.},
        title = {Chapter 3.6 The basal forebrain and episodic memory},
    booktitle = {Handbook of Behavioral Neuroscience},
    publisher = {Elsevier Science},
         year = {2008},
        pages = {357-377}
}
@article{FujiietalNMC-16,
       author = {Fujii, Masazumi and Maesawa, Satoshi and Ishiai, Sumio and Iwami, Kenichiro and Futamura, Miyako and Saito, Kiyoshi},
        title = {Neural Basis of Language: An Overview of An Evolving Model},
      journal = {Neurologia Medico-Chirurgica},
    publisher = {The Japan Neurosurgical Society},
       volume = {56},
        issue = {7},
        pages = {379-386},
         year = {2016},
     abstract = {The neural basis of language had been considered as a simple model consisting of the Broca's area, the Wernicke's area, and the arcuate fasciculus (AF) connecting the above two cortical areas. However, it has grown to a larger and more complex model based upon recent advancements in neuroscience such as precise imaging studies of aphasic patients, diffusion tensor imaging studies, functional magnetic resonance imaging studies, and electrophysiological studies with cortical and subcortical stimulation during awake surgery. In the present model, language is considered to be processed through two distinct pathways, the dorsal stream and the ventral stream. The core of the dorsal stream is the superior longitudinal fasciculus/AF, which is mainly associated with phonological processing. On the other hand, semantic processing is done mainly with the ventral stream consisting of the inferior fronto-occipital fasciculus and the intratemporal networks. The frontal aslant tract has recently been named the deep frontal tract connecting the supplementary motor area and the Broca's area and it plays an important role in driving and initiating speech. It is necessary for every neurosurgeon to have basic knowledge of the neural basis of language. This knowledge is essential to plan safer surgery and preserve the above neural structures during surgery.},
}
@article{FukudaNEUROSCIENCE-17,
       author = {Takaichi Fukuda},
        title = {Structural organization of the dendritic reticulum linked by gap junctions in layer 4 of the visual cortex },
      journal = {Neuroscience},
       volume = {340},
       number = {6},
         year = {2017},
        pages = {76-90},
     abstract = {Neuronal gap junctions are ubiquitous in the brain, but their precise positions in actual neuronal circuits have been uncertain, and their functional roles remain unclear. In this study, I visualized connexin36-immunoreactive gap junctions and examined the structural features of the interconnected dendrites arising from parvalbumin (PV)-positive interneurons in layer 4 of the feline visual cortex. I observed evidence for net-like dense linkages of dendrites among virtually all \{PV\} neurons (56/58 cells, 96.6\%) in the tissue. This dendritic reticulum established connections among clustered cells and further among remote cells. The latter connectivity exhibited a preference for vertical direction, including translaminar linkages, but was also characterized by lateral continuity. Measurement of the distances from each dendritic gap junction back to the two connected somata revealed that at least one of two somata was within 50 $\mu$m from the junction in 77.5\% of the cases and within 75 $\mu$m in 91.2\% of the cases. Thus, distal gap junctions mediated morphologically asymmetrical connection where one soma was close to, but the other soma was far from the connecting junction. This connectivity was typically observed between neurons located apart in the same columnar space, where a long vertical dendrite bridged two neurons through an asymmetrically positioned gap junction. In contrast, gap junctions formed between nearby cells were close to both somata. Thalamocortical afferents established synapses densely on somata of layer 4 \{PV\} neurons, indicating the possible involvement of proximal gap junctions in visual stimulus-driven feedforward regulation. These findings provide a new structural basis for cortical investigations.},
}
@article{FukushimaBC-80,
       author = {Fukushima, K.},
        title = {Neocognitron: {A} self organizing neural network model for a mechanism of pattern recognition unaffected by shift in position},
      journal = {Biological Cybernnetics},
       volume = 36,
        issue = 4,
        pages = {93-202},
         year = 1980,
}
@incollection{Fukushimaetal88,
       author = {Kunihiko Fukushima and Sei Miyake and Takayuki Ito},
        title = {Neocognitron: {A} neural network model for a mechanism of visual pattern recognition},
    booktitle = {Artificial neural networks: theoretical concepts},
    publisher = {IEEE Computer Society Press},
        pages = {136-144},
         year = 1988,
}
@book{Fukuyama1992lastman,
        title = {{The End of History and the Last Man}},
       author = {Fukuyama, Francis},
       series = {Avon Book},
         year = {1992},
    publisher = {Free Press}
}
@article{FunckeetalMICCAI-14,
       author = {Funke, J. and Martel, J. N. and Gerhard, S. and Andres, B. and Cirean, D. C. and Giusti, A. and Gambardella, L. M. and Schmidhuber, J. and Pfister, H. and Cardona, A. and Cook, M.},
        title = {Candidate sampling for neuron reconstruction from anisotropic electron microscopy volumes},
      journal = {Medical Image Computing and Computer-assisted Intervention},
         year = {2014},
       volume = {17},
        issue = {1},
        pages = {17-24},
     abstract = {The automatic reconstruction of neurons from stacks of electron microscopy sections is an important computer vision problem in neuroscience. Recent advances are based on a two step approach: First, a set of possible 2D neuron candidates is generated for each section independently based on membrane predictions of a local classifier. Second, the candidates of all sections of the stack are fed to a neuron tracker that selects and connects them in {3D} to yield a reconstruction. The accuracy of the result is currently limited by the quality of the generated candidates. In this paper, we propose to replace the heuristic set of candidates used in previous methods with samples drawn from a conditional random field (CRF) that is trained to label sections of neural tissue. We show on a stack of Drosophila melanogaster neural tissue that neuron candidates generated with our method produce 30\% less reconstruction errors than current candidate generation methods. Two properties of our CRF are crucial for the accuracy and applicability of our method: (1) The CRF models the orientation of membranes to produce more plausible neuron candidates. (2) The interactions in the CRF are restricted to form a bipartite graph, which allows a great sampling speed-up without loss of accuracy.}
}
@inproceedings{FurlongetalPARCO-07,
     title = {Novel Brain-Derived Algorithms Scale Linearly with Number of Processing Elements},
    author = {Jeff Furlong and Andrew Felch and Jayram Moorkanikara Nageswaran and Nikil Dutt and
              Alex Nicolau and Alexander V. Veidenbaum and Ashok Chandrashekar and Richard Granger},
 booktitle = {Proceedings of the International Conference on Parallel Computing},
      year = 2007,
     pages = {767-776},
}
@book{Fuster2015prefrontal,
        title = {The Prefrontal Cortex},
       author = {Joaqu{\'{i}}n M. Fuster},
         year = {2015},
    publisher = {Elsevier Science}
}
@incollection{FusterBCN-08,
       author = {Joaqu{\'{i}}n M. Fuster},
       editor = {R.F. Thompson and J.L. McCelland}, 
        title = {Prefrontal Cortex},
    booktitle = {The International Encyclopedia of the Social \& Behavioral Sciences. Behavioral and Cognitive Neuroscience},
    publisher = {Elsevier, London},
         year = {2008},
        pages = {355-369}
}
@book{FusterPREFRONTAL-CORTEX-15,
       author = {Joaqu{\'{i}}n M. Fuster},
        title = {{Prefrontal Cortex, 5th Edition}},
    publisher = {Elsevier, London},
         year = {2015},
     abstract = {A functional model of the prefrontal cortex is developed from the material in the preceding chapters. The main general function of prefrontal cortex, to coordinate novel sequences of goal-directed behavior and speech, comprises several executive functions (planning, attentional set, working memory, decision-making, error monitoring and avoidance, and inhibitory control) performed in conjunction with other cortical and subcortical structures. These functions operate on a vast array of distributed, interactive cortical networks named cognits. Cognits are units of memory and knowledge, formed by experience through the synaptic association of neuronal assemblies representing simultaneous sensations and/or actions. Perceptual cognits are hierarchically self-organized in posterior cortex, executive cognits in frontal cortex. All cognitive functions serving the pursuit of new goal-directed behavior occur, under prefrontal control, in the perception-action cycle, the cybernetic adaptive interface between the organism and its environment. Since some functions are future oriented, the prefrontal cortex and perception-action cycle are both adaptive and preadaptive. From Chapter 8.},
}
@inbook{FusterPREFRONTAL-CORTEX-15-CHAPTER_8,
       author = {Joaqu{\'{i}}n M. Fuster},
        title = {{Chapter 8: An Overview of Prefrontal Functions}},
    booktitle = {{Prefrontal Cortex, 5th Edition}},
    publisher = {Elsevier, London},
         year = {2015},
        pages = {375-425},
}
@article{FusterTiCS-04,
       author = {Joaqu{\'{i}}n M. Fuster},
        title = {Upper processing stages of the perception-action cycle},
      journal = {Trends in Cognitive Sciences},
       volume = {8},
       number = {4},
         year = {2004},
        pages = {143-145},
     abstract = {The neural substrate for behavioral, cognitive and linguistic actions is hierarchically organized in the cortex of the frontal lobe. In their methodologically impeccable study, Koechlin et al. reveal the neural dynamics of the frontal hierarchy in behavioral action. Progressively higher areas control the performance of actions requiring the integration of progressively more complex and temporally dispersed information. The study substantiates the crucial role of the prefrontal cortex in the temporal organization of behavior.}
}
@article{FusterTiN-97,
       author = {Joaqu{\'{i}}n M. Fuster},
        title = {Network memory},
      journal = {Trends in Neurosciences},
       volume = {20},
       number = {10},
         year = {1997},
        pages = {451-459},
     abstract = {Our thinking on the cortical organization of primate memory is undergoing a copernican change, from a neuropsychology that localizes different memories in different areas to one that views memory as a distributed property of cortical systems. We are shifting our focus from `systems of memory' to the memory of systems. The same cortical systems that serve us to perceive and move in the world serve us to remember it. Our memories are networks of interconnected cortical neurons, formed by association, that contain our experiences in their connectional structure. Perceptual and motor memory networks are hierarchically organized in post-rolandic and pre-rolandic neocortex, respectively. Recall, recognition and working memory consist largely in their reactivation, also by association.}
}

@incollection{GaengelandMlodzikDROSOPHILA-08,
       author = {Gaengel, Konstantin and Mlodzik, Marek},
        title = {Microscopic Analysis of the Adult Drosophila Retina Using Semithin Plastic Sections},
    booktitle = {Drosophila},
    publisher = {Humana Press},
       series = {Methods in Molecular Biology},
       volume = 420,
       editor = {Dahmann, Christian},
         year = 2008,
        pages = {277-287},
}
@incollection{GageandBaarsARCUATE-FASCICULUS-18,
       author = {Nicole M. Gage and Bernard J. Baars},
        title = {Chapter 6 - Language and Thought},
       editor = {Nicole M. Gage and Bernard J. Baars},
    booktitle = {Fundamentals of Cognitive Neuroscience (Second Edition)},
    publisher = {Academic Press},
      address = {San Diego, CA},
         year = {2018},
        pages = {185-214},
}
@article{GaierandHaCoRR-19,
       author = {Adam Gaier and David Ha},
        title = {Weight Agnostic Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1906.04358},
         year = {2019},
     abstract = {Not all neural network architectures are created equal, some perform much better than others for certain tasks. But how important are the weight parameters of a neural network compared to its architecture? In this work, we question to what extent neural network architectures alone, without learning any weight parameters, can encode solutions for a given task. We propose a search method for neural network architectures that can already perform a task without any explicit weight training. To evaluate these networks, we populate the connections with a single shared weight parameter sampled from a uniform random distribution, and measure the expected performance. We demonstrate that our method can find minimal neural network architectures that can perform several reinforcement learning tasks without weight training. On a supervised learning domain, we find network architectures that achieve much higher than chance accuracy on MNIST using random weights.}
}
@article{GalanteetalPLoS-15,
       author = {Galante, Angelo and Sinibaldi, Raffaele and Conti, Allegra and De Luca, Cinzia and Catallo, Nadia and Sebastiani, Piero and Pizzella, Vittorio and Romani, Gian Luca and Sotgiu, Antonello and Della Penna, Stefania},
      journal = {PLoS ONE},
    publisher = {Public Library of Science},
        title = {Fast Room Temperature Very Low Field-Magnetic Resonance Imaging System Compatible with MagnetoEncephaloGraphy Environment},
         year = {2015},
       volume = {10},
       number = {12},
        pages = {1-21},
     abstract = {In recent years, ultra-low field (ULF)-MRI is being given more and more attention, due to the possibility of integrating ULF-MRI and Magnetoencephalography (MEG) in the same device. Despite the signal-to-noise ratio (SNR) reduction, there are several advantages to operating at ULF, including increased tissue contrast, reduced cost and weight of the scanners, the potential to image patients that are not compatible with clinical scanners, and the opportunity to integrate different imaging modalities. The majority of ULF-MRI systems are based, until now, on magnetic field pulsed techniques for increasing SNR, using SQUID based detectors with Larmor frequencies in the kHz range. Although promising results were recently obtained with such systems, it is an open question whether similar SNR and reduced acquisition time can be achieved with simpler devices. In this work a room-temperature, MEG-compatible very-low field (VLF)-MRI device working in the range of several hundred kHz without sample pre-polarization is presented. This preserves many advantages of ULF-MRI, but for equivalent imaging conditions and SNR we achieve reduced imaging time based on preliminary results using phantoms and <italic>ex-vivo</italic> rabbits heads.},
}
@article{GalarretaandHestrinNATURE-99,
       author = {Mario Galarreta and Shaul Hestrin},
        title = {A network of fast-spiking cells in the neocortex connected by electrical synapses},
      journal = {Nature},
       volume = 402,
         year = 1999,
        pages = {72-75},
     abstract = {Encoding of information in the cortex is thought to depend on synchronous firing of cortical neurons. Inhibitory neurons are known to be critical in the coordination of cortical activity, but how interaction among inhibitory cells promotes synchrony is not well understood. To address this issue directly, we have recorded simultaneously from pairs of fast-spiking (FS) cells, a type of \gamma{}-aminobutyric acid (GABA)-containing neocortical interneuron. Here we report a high occurrence of electrical coupling among FS cells. Electrical synapses were not found among pyramidal neurons or between FS cells and other cortical cells. Some FS cells were interconnected by both electrical and GABAergic synapses. We show that communication through electrical synapses allows excitatory signalling among inhibitory cells and promotes their synchronous spiking. These results indicate that electrical synapses establish a network of fast-spiking cells in the neocortex which may play a key role in coordinating cortical activity},
}
@article{GalarretaandHestrinNATURE-NEUROSCIENCE-98,
       author = {Mario Galarreta and Shaul Hestrin},
        title = {Frequency-dependent synaptic depression and the balance of excitation and inhibition in the neocortex},
      journal = {Nature Neuroscience},
       volume = 1,
        issue = 7,
         year = 1998,
        pages = {587-594},
     abstract = {The stability of cortical neuron activity in vivo suggests that the firing rates of both excitatory and inhibitory neurons are dynamically adjusted. Using dual recordings from excitatory pyramidal neurons and inhibitory fast-spiking neurons in neocortical slices, we report that sustained activation by trains of several hundred presynaptic spikes resulted in much stronger depression of synaptic currents at excitatory synapses than at inhibitory ones. The steady-state synaptic depression was frequency dependent and reflected presynaptic function. These results suggest that inhibitory terminals of fast-spiking cells are better equipped to support prolonged transmitter release at a high frequency compared with excitatory ones. This difference in frequency-dependent depression could produce a relative increase in the impact of inhibition during periods of high global activity and promote the stability of cortical circuits.},
}
@article{GalarretaandHestrinNATURE-REVIEWS-NEUROSCIENCE-01,
       author = {Mario Galarreta and Shaul Hestrin},
        title = {Electrical Synapses Between GABA-releasing Interneurons},
      journal = {Nature Reviews Neuroscience},
       volume = 2,
         year = 2001,
        pages = {225-433},
     abstract = {Although gap junctions were first demonstrated in the mammalian brain about 30 years ago, the distribution and role of electrical synapses have remained elusive. A series of recent reports has demonstrated that inhibitory interneurons in the cerebral cortex, thalamus, striatum and cerebellum are extensively interconnected by electrical synapses. Investigators have used paired recordings to reveal directly the presence of electrical synapses among identified cell types. These studies indicate that electrical coupling is a fundamental feature of local inhibitory circuits and suggest that electrical synapses define functionally diverse networks of GABA-releasing interneurons. Here, we discuss these results, their possible functional significance and the insights into neuronal circuit organization that have emerged from them.},
}
@article{GalarretaandHestrinPNAS-02,
       author = {Mario Galarreta and Shaul Hestrin},
        title = {Electrical and chemical synapses among parvalbumin fast-spiking GABAergic interneurons in adult mouse neocortex},
      journal = {PNAS},
       volume = 99,
        issue = 19,
         year = 2002,
        pages = {12438-12443},
     abstract = {Networks of \gamma{}-aminobutyric acid (GABA)ergic interneurons connected via electrical and chemical synapses are thought to play an important role in detecting and promoting synchronous activity in the cerebral cortex. Although the properties of electrical and chemical synaptic interactions among inhibitory interneurons are critical for their function as a network, they have only been studied systematically in juvenile animals. Here, we have used transgenic mice expressing the enhanced green fluorescent protein in cells containing parvalbumin (PV) to study the synaptic connectivity among fast-spiking (FS) cells in slices from adult animals (2-7 months old). We have recorded from pairs of PV-FS cells and found that the majority of them were electrically coupled (61\%, 14 of 23 pairs). In addition, 78\% of the pairs were connected via GABAergic chemical synapses, often reciprocally. The average coupling coef- ficient for step injections was 1.5\%, a smaller value than that reported in juvenile animals. GABA-mediated inhibitory postsynaptic currents and potentials decayed with exponential time constants of 2.6 and 5.9 ms, respectively, and exhibited paired-pulse depression (50-ms interval). The inhibitory synaptic responses in the adult were faster than those observed in young animals. Our results indicate that PV-FS cells are highly interconnected in the adult cerebral cortex by both electrical and chemical synapses, establishing networks that can have important implications for coordinating activity in cortical circuits.},
}
@article{GalarretaandHestrinSCIENCE-01,
       author = {Mario Galarreta and Shaul Hestrin},
        title = {Spike Transmission and Synchrony Detection in Networks of GABAergic Interneurons},
      journal = {Science},
       volume = 292,
         year = 2001,
        pages = {2295-2299},
     abstract = {The temporal pattern and relative timing of action potentials among neocortical neurons may carry important information. However, how cortical circuits detect or generate coherent activity remains unclear. Using paired recordings in rat neocortical slices, we found that the firing of fast-spiking cells can reflect the spiking pattern of single-axon pyramidal inputs. Moreover, this property allowed groups of fast-spiking cells interconnected by electrical and \gamma{}-aminobutyric acid (GABA)-releasing (GABAergic) synapses to detect the relative timing of their excitatory inputs. These results indicate that networks of fast- spiking cells may play a role in the detection and promotion of synchronous activity within the neocortex.},
}
@article{GalataetalCVIU-01,
    author = {Aphrodite Galata and Neil Johnson and David Hogg},
     title = {Learning Variable Length {M}arkov Models of Behaviour},
   journal = {Computer Vision and Image Understanding},
    volume = 81,
     issue = 3,
     pages = {398-413},
      year = 2001,
}
@article{GalesandYoungHMM-07,
       author = {Mark Gales and Steve Young},
        title = {The Application of Hidden Markov Models in Speech Recognition},
      journal = {Foundations and Trends in Signal Processing},
       volume = 1,
        issue = 3,
         year = 2007,
        pages = {195-304},
     abstract = {Hidden Markov Models (HMMs) provide a simple and effective framework for modelling time-varying spectral vector sequences. As a consequence, almost all present day large vocabulary continuous speech recognition (LVCSR) systems are based on HMMs. 
            Whereas the basic principles underlying HMM-based LVCSR are rather straightforward, the approximations and simplifying assumptions involved in a direct implementation of these principles would result in a system which has poor accuracy and unacceptable sensitivity to changes in operating environment. Thus, the practical application of HMMs in modern systems involves considerable sophistication.
            The aim of this review is first to present the core architecture of a HMM-based LVCSR system and then describe the various refinements which are needed to achieve state-of-the-art performance. These refinements include feature projection, improved covariance modelling, discriminative parameter estimation, adaptation and normalisation, noise compensation and multi-pass system combination. The review concludes with a case study of LVCSR for Broadcast News and Conversation transcription in order to illustrate the techniques described.}
}
@article{GaletalNATURE-NEUROSCIENCE-17,
       author = {Gal, Eyal and London, Michael and Globerson, Amir and Ramaswamy, Srikanth and Reimann, Michael W. and Muller, Eilif and Markram, Henry and Segev, Idan},
        title = {Rich cell-type-specific network topology in neocortical microcircuitry},
      journal = {Nature Neuroscience},
         year = {2017},
    publisher = {Nature Publishing Group},
     abstract = {Uncovering structural regularities and architectural topologies of cortical circuitry is vital for understanding neural computations. Recently, an experimentally constrained algorithm generated a dense network reconstruction of a [sim]0.3-mm3 volume from juvenile rat somatosensory neocortex, comprising [sim]31,000 cells and [sim]36 million synapses. Using this reconstruction, we found a small-world topology with an average of 2.5 synapses separating any two cells and multiple cell-type-specific wiring features. Amounts of excitatory and inhibitory innervations varied across cells, yet pyramidal neurons maintained relatively constant excitation/inhibition ratios. The circuit contained highly connected hub neurons belonging to a small subset of cell types and forming an interconnected cell-type-specific rich club. Certain three-neuron motifs were overrepresented, matching recent experimental results. Cell-type-specific network properties were even more striking when synaptic strength and sign were considered in generating a functional topology. Our systematic approach enables interpretation of microconnectomics 'big data' and provides several experimentally testable predictions.},
}
@book{GallagherMODELS-OF-THE-SELF-13,
        title = {Models of the Self},
       author = {Gallagher, Shaun},
         year = {2013},
    publisher = {Imprint Academic}
}
@article{GallagherandHollandPNAS-94,
       author = {Gallagher, M. and Holland, P. C.},
        title = {The amygdala complex: multiple roles in associative learning and attention},
      journal = {Proceedings of the National Academy of Science},
       volume = {91},
       number = {25},
         year = {1994},
        pages = {11771-11776},
     abstract = {Although certain neurophysiological functions of the amygdala complex in learning seem well established, the purpose of this review is to propose that an additional conceptualization of amygdala function is now needed. The research we review provides evidence that a subsystem within the amygdala provides a coordinated regulation of attentional processes. An important aspect of this additional neuropsychology of the amygdala is that it may aid in understanding the importance of connections between the amygdala and other neural systems in information processing.}
}
@article{GallandetalNATURE-13,
       author = {Galland, Remi and Leduc, Patrick and Guerin, Christophe and Peyrade, David and Blanchoin, Laurent and Thery, Manuel},
        title = {Fabrication of three-dimensional electrical connections by means of directed actin self-organization},
      journal = {Nature Materials},
    publisher = {Nature Publishing Group},
       volume = 12,
        issue = 5,
         year = 2013,
        pages = {416-421},
}
@article{GalleseetalTCS-04,
     title = {A unifying view of the basis of social cognition},
    author = {Vittorio Gallese and Christian Keysers and Giacomo Rizzolatti},
   journal = {Trends in Cognitive Sciences},
    volume = 8,
     issue = 9,
      year = 2004,
     pages = {396-403},
}
@incollection{Gallistel2012,
       author = {Charles R. Gallistel},
        title = {Machinery of Cognition},
    booktitle = {Evolution and the Mechanisms of Decision Making},
       editor = {Peter Hammerstein and Jeffrey R. Stevens},
    publisher = {MIT Press},
       volume = 2,
         year = 2012,
        pages = {39-52},
}
@article{GallistelBEHAVIOR_AND_PHILOSOPHY-10,
       author = {Charles Randy Gallistel},
        title = {Response to Donahoe Review},
      journal = {Behavior and Philosophy},
       volume = {38},
         year = {2010},
        pages = {103-110},
}
@incollection{GallistelLEARNING_AND_MEMORY-08,
        title = {Learning and representation},
       author = {Gallistel, C.R.},
    booktitle = {{Volume 1 of Learning and Memory: A Comprehensive Reference}},
       editor = {R. Menzel and J. Byrne},
    publisher = {Elsevier},
      address = {Oxford},
         year = {2008},
        pages = {227-242},
      comment = {https://ruccs.rutgers.edu/images/personal-charles-r-gallistel/publications/LearningAndRepresentation.pdf},
}
@book{Gallistel_and_King_Memory_and_the_Computational_Brain,
        title = {Memory and the Computational Brain: Why Cognitive Science will Transform Neuroscience},
       author = {Gallistel, C.R. and King, A.P.},
    publisher = {Wiley},
         year = {2009},
      comment = {http://www.ling.upenn.edu/~kroch/courses/lx400/Gallistel-King_chaps1-9.pdf},
}
@book{GallistelandKing2009computational,
        title = {{Memory and the Computational Brain: Why Cognitive Science will Transform Neuroscience}},
       author = {Gallistel, C.R. and King, A.P.},
    publisher = {Wiley},
         year = {2009},
}
@book{Gallwey2010,
        title = {The Inner Game of Tennis: The Classic Guide to the Mental Side of Peak Performance},
       author = {Gallwey, W. Timothy and Carroll, P. and Kleinman, Z.},
         year = {2010},
    publisher = {Random House Publishing Group}
}
@inproceedings{GalsworthyetalWCPG-01,
       author = {Galsworthy, M.J. and Paya-Cano, J.L. and Monleon, S. and Liu, L. and Plomin, R.},
        title = {The Puzzle Box: A New Task for Assessing Problem-Solving in Mice},
    booktitle = {{IX}th World Congress of Psychiatric Genetics},
     location = {Saint Louis, Missouri},
       volume = {105},
        issue = {7},
        pages = {559-649},
     abstract = {The Puzzle Box is a new intelligence task for mice, which has been developed as part of our ongoing research in the characterization of general cognitive ability (g) in mice. Puzzle tasks used to asse ss g in rats were adapted and extended to complement spatial and exploratory tasks in our battery. The apparatus is a box 73 \hmtimes{} 29 cm, divided by a removable barrier into a small dark goal box and a larger brightly lit start box. Mice are presented with a series of puzzles to solve in order to enter the goal box. These include filling an underpass with sawdust, where the mouse has to burrow through, blocking the entrance with a cardboard plug which the mouse must pull out, or presenting a set of mouse-size `cat-flap' doors. Individual differences analyses on our heterogeneous stock (HS) mice show that these mice can solve the problems with certain mice consistently outperforming others. Comparisons with other well-known cognitive tests also show that these tasks are good measures of g. Further analyses have identified difficulty levels, consistency and g-loadings of the task items. Task design has been re®ned on this basis to allow quicker and easier testing with tasks proven to be reliable. The Puzzle Box is therefore recommended as a useful addition to a cognitive battery in mice.}
}
@article{GambleNS-2013,
     title = {The sleep squeeze},
    author = {Jessa Gamble}, 
   journal = {New Scientist},
    volume = 217,
     issue = 2902,
      year = 2013,
     pages = {34-37}
}
@inproceedings{GandheetalASC-04,
    author = {Sudeep Gandhe and Andrew Gordon and Anton Leuski and David Traum and Douglas W. Oard},
     title = {First steps toward linking dialogues: Mediating between free-text questions and pre-recorded video answers},
 booktitle = {Proceedings of the Army Science Conference (24th)},
  location = {Orlando, Florida},
      year = 2004,
}
@inproceedings{GandheetalTACQ-08,
      title = {From Domain Specification to Virtual Humans: An integrated approach to authoring tactical questioning characters},
     author = {Sudeep Gandhe and David DeVault and Antonio Roque and Bilyana Martinovski and Ron Artstein and Anton Leuski and Jillian Gerten and David Traum},
  booktitle = {Proceedings of Interspeech},
       year = 2008,
      pages = {000-000},
}
@misc{GanguliMCGOVERN-16,
        title = {A theory of neural dimensionality, dynamics and measurement},
       author = {Surya Ganguli},
         note = {McGovern Institute Annual Symposium},
         year = {2016},
    publisher = {McGovern Institute for Brain Research at MIT},
}
@article{GanguliandCarmenaPLoS-09,
       author = {Ganguly, Karunesh and Carmena, Jose M.},
      journal = {{PLOS Biology}},
    publisher = {Public Library of Science},
        title = {Emergence of a Stable Cortical Map for Neuroprosthetic Control},
         year = 2009,
       number = 7,
       volume = 7,
        pages = {1-13},
     abstract = {In this article, the authors show that the neural representation for control of a neuroprosthetic device undergoes a process of consolidation, after which it is stable, readily recalled, and resistant to interference.},
}
@article{GanguliandSompolinskyARN-12,
       author = {Ganguli, S. and Sompolinsky, H.},
        title = {{C}ompressed sensing, sparsity, and dimensionality in neuronal information processing and data analysis},
      journal = {Annual Review of Neuroscience},
         year = 2012,
       volume = 35,
        pages = {485-508},
     abstract = {The curse of dimensionality poses severe challenges to both technical and conceptual progress in neuroscience. In particular, it plagues our ability to acquire, process, and model high-dimensional data sets. Moreover, neural systems must cope with the challenge of processing data in high dimensions to learn and operate successfully within a complex world. We review recent mathematical advances that provide ways to combat dimensionality in specific situations. These advances shed light on two dual questions in neuroscience. First, how can we as neuroscientists rapidly acquire high-dimensional data from the brain and subsequently extract meaningful models from limited amounts of these data? And second, how do brains themselves process information in their intrinsically high-dimensional patterns of neural activity as well as learn meaningful, generalizable models of the external world from limited experience?}
}
@article{GangulyetalJCB-15,
       author = {Ganguly, Archan and Tang, Yong and Wang, Lina and Ladt, Kelsey and Loi, Jonathan and Dargent, B\'{e}n\'{e}dicte and Leterrier, Christophe and Roy, Subhojit},
        title = {A dynamic formin-dependent deep F-actin network in axons},
      journal = {The Journal of Cell Biology},
       volume = 210,
        issue = 3,
         year = 2015,
        pages = {401-417},
     abstract = {Although actin at neuronal growth cones is well-studied, much less is known about actin organization and dynamics along axon shafts and presynaptic boutons. Using probes that selectively label filamentous-actin (F-actin), we found focal "actin hotspots" along axons--spaced \approx{}3-4 \hmmu{}m apart--where actin undergoes continuous assembly/disassembly. These foci are a nidus for vigorous actin polymerization, generating long filaments spurting bidirectionally along axons--a phenomenon we call "actin trails." Super-resolution microscopy reveals intra-axonal deep actin filaments in addition to the subplasmalemmal "actin rings" described recently. F-actin hotspots colocalize with stationary axonal endosomes, and blocking vesicle transport diminishes the actin trails, suggesting mechanistic links between vesicles and F-actin kinetics. Actin trails are formin--but not Arp2/3--dependent and help enrich actin at presynaptic boutons. Finally, formin inhibition dramatically disrupts synaptic recycling. Collectively, available data suggest a two-tier F-actin organization in axons, with stable "actin rings" providing mechanical support to the plasma membrane and dynamic "actin trails" generating a flexible cytoskeletal network with putative physiological roles.},
}
@article{GaninetalCoRR-18,
       author = {Yaroslav Ganin and Tejas Kulkarni and Igor Babuschkin and S. M. Ali Eslami and Oriol Vinyals},
        title = {Synthesizing Programs for Images using Reinforced Adversarial Learning},
      journal = {CoRR},
       volume = {arXiv:1804.01118},
         year = {2018},
     abstract = {Advances in deep generative networks have led to impressive results in recent years. Nevertheless, such models can often waste their capacity on the minutiae of datasets, presumably due to weak inductive biases in their decoders. This is where graphics engines may come in handy since they abstract away low-level details and represent images as high-level programs. Current methods that combine deep learning and renderers are limited by hand-crafted likelihood or distance functions, a need for large amounts of supervision, or difficulties in scaling their inference algorithms to richer datasets. To mitigate these issues, we present SPIRAL, an adversarially trained agent that generates a program which is executed by a graphics engine to interpret and sample images. The goal of this agent is to fool a discriminator network that distinguishes between real and rendered data, trained with a distributed reinforcement learning setup without any supervision. A surprising finding is that using the discriminator's output as a reward signal is the key to allow the agent to make meaningful progress at matching the desired output rendering. To the best of our knowledge, this is the first demonstration of an end-to-end, unsupervised and adversarial inverse graphics agent on challenging real world (MNIST, Omniglot, CelebA) and synthetic 3D datasets.}
}
@inproceedings{GaninetalICML-18,
        title = {Learning to write programs that generate images},
       author = {Yaroslav Ganin and Tejas Kulkarni and Igor Babuschkin and S. M. Ali Eslami and Oriol Vinyals},
    booktitle = {Proceedings of the 35th International Conference on Machine Learning},
         year = {2018},
       editor = {Jennifer Dy and Andreas Krause},
       series = {Proceedings of Machine Learning Research},
    publisher = {PMLR},
     abstract = {Advances in deep generative networks have led to impressive results in recent years. Nevertheless, such models can often waste their capacity on the minutiae of datasets, presumably due to weak inductive biases in their decoders. This is where graphics engines may come in handy since they abstract away low-level details and represent images as high-level programs. Current methods that combine deep learning and renderers are limited by hand-crafted likelihood or distance functions, a need for large amounts of supervision, or difficulties in scaling their inference algorithms to richer datasets. To mitigate these issues, we present SPIRAL, an adversarially trained agent that generates a program which is executed by a graphics engine to interpret and sample images. The goal of this agent is to fool a discriminator network that distinguishes between real and rendered data, trained with a distributed reinforcement learning setup without any supervision. A surprising finding is that using the discriminator's output as a reward signal is the key to allow the agent to make meaningful progress at matching the desired output rendering. To the best of our knowledge, this is the first demonstration of an end-to-end, unsupervised and adversarial inverse graphics agent on challenging real world (MNIST, OMNIGLOT, CELEBA) and synthetic 3D datasets.}
}
@article{GaninetalJMLR-16,
       author = {Yaroslav Ganin and Evgeniya Ustinova and Hana Ajakan and Pascal Germain and Hugo Larochelle and Fran\c{c}ois Laviolette and Mario Marchand and Victor Lempitsky},
        title = {Domain-Adversarial Training of Neural Networks},
      journal = {Journal of Machine Learning Research},
         year = {2016},
       volume = {17},
       number = {59},
        pages = {1-35},
     abstract = {We introduce a new representation learning approach for domain adaptation, in which data at training and test time come from similar but different distributions. Our approach is directly inspired by the theory on domain adaptation suggesting that, for effective domain transfer to be achieved, predictions must be made based on features that cannot discriminate between the training (source) and test (target) domains. The approach implements this idea in the context of neural network architectures that are trained on labeled data from the source domain and unlabeled data from the target domain (no labeled target-domain data is necessary). As the training progresses, the approach promotes the emergence of features that are (i) discriminative for the main learning task on the source domain and (ii) indiscriminate with respect to the shift between the domains. We show that this adaptation behaviour can be achieved in almost any feed-forward model by augmenting it with few standard layers and a new gradient reversal layer. The resulting augmented architecture can be trained using standard backpropagation and stochastic gradient descent, and can thus be implemented with little effort using any of the deep learning packages. We demonstrate the success of our approach for two distinct classification problems (document sentiment analysis and image classification), where state-of-the-art domain adaptation performance on standard benchmarks is achieved. We also validate the approach for descriptor learning task in the context of person re-identification application.},
}
@article{GaoandGanguliCOIN-15,
        title = {On simplicity and complexity in the brave new world of large-scale neuroscience},
       author = {Peiran Gao and Surya Ganguli},
      journal = {Current Opinion in Neurobiology},
       volume = {32},
         year = {2015},
        pages = {148-155},
     abstract = {Technological advances have dramatically expanded our ability to probe multi-neuronal dynamics and connectivity in the brain. However, our ability to extract a simple conceptual understanding from complex data is increasingly hampered by the lack of theoretically principled data analytic procedures, as well as theoretical frameworks for how circuit connectivity and dynamics can conspire to generate emergent behavioral and cognitive functions. We review and outline potential avenues for progress, including new theories of high dimensional data analysis, the need to analyze complex artificial networks, and methods for analyzing entire spaces of circuit models, rather than one model at a time. Such interplay between experiments, data analysis and theory will be indispensable in catalyzing conceptual advances in the age of large-scale neuroscience.}
}
@article{GaoandGanguliCoRR-15,
       author = {Peiran Gao and Surya Ganguli},
        title = {On Simplicity and Complexity in the Brave New World of Large-Scale Neuroscience},
      journal = {CoRR},
       volume = {arXiv:1503.08779},
         year = 2015,
     abstract = {Technological advances have dramatically expanded our ability to probe multi-neuronal dynamics and connectivity in the brain. However, our ability to extract a simple conceptual understanding from complex data is increasingly hampered by the lack of theoretically principled data analytic procedures, as well as theoretical frameworks for how circuit connectivity and dynamics can conspire to generate emergent behavioral and cognitive functions. We review and outline potential avenues for progress, including new theories of high dimensional data analysis, the need to analyze complex artificial networks, and methods for analyzing entire spaces of circuit models, rather than one model at a time. Such interplay between experiments, data analysis and theory will be indispensable in catalyzing conceptual advances in the age of large-scale neuroscience.}
}
@inproceedings{GaoandHammerstromIJCNN-03,
     title = {{Platform Performance Comparison of PALM Network on Pentium 4 and FPGA}},
    author = {Changjian Gao and Dan Hammerstrom},
 booktitle = {Proceedings of the International Joint Conference on Neural Networks},
  location = {Portland, OR},
      year = 2003,
}
@inproceedings{GaoandKollerICCV-11,
     title = {Discriminative Learning of Relaxed Hierarchy for Large-scale Visual Recognition},
    author = {Tianshi Gao and Daphne Koller},
 booktitle = {{IEEE} International Conference on Computer Vision},
 publisher = {IEEE},
      year = 2011
} 
@article{GaoetalCST-10,
        title = {Large-scale fabrication and electrical properties of an anisotropic conductive polymer composite utilizing preferable location of carbon nanotubes in a polymer blend},
      journal = {Composites Science and Technology},
       volume = {70},
        issue = {13},
         year = {2010},
        pages = {1973-1979},
       author = {Jie-Feng Gao and Ding-Xiang Yan and Bo Yuan and Hua-Dong Huang and Zhong-Ming Li},
     abstract = {An anisotropically conductive polymer composite (ACPC) based on carbon nanotubes (CNTs) and polycarbonate (PC)/polyethylene (PE) blend was fabricated via a slit die extrusion-hot stretch process. Under the influence of the shear flow and hot stretch, the \{PC\} phase is in situ deformed into aligned conductive fibrils in the \{PE\} matrix, whose surface region holds the majority of CNTs. When the stretch ratio rises to a certain level, the resistivity of the \{ACPC\} shows a strong anisotropy of three orders of magnitude difference between the perpendicular and parallel stretch directions. The fibrillar \{ACPC\} shows a weak positive temperature coefficient (PTC) effect, and two-process negative temperature coefficient (NTC) effect caused by the reorganization of \{PC\} fibrils below but near 230 \Angstrom{}\hmdeg{}C, and the transformation from anisotropy to isotropy beyond 230 \Angstrom{}\hmdeg{}C. The obtained \{ACPC\} allows it to have such potential applications as switch and sensor materials.}
}
@inproceedings{GaoetalIJCAI-18,
        title = {Three-Head Neural Network Architecture for Monte Carlo Tree Search},
       author = {Chao Gao and Martin M{\"{u}}ller and Ryan Hayward},
    booktitle = {Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence, {IJCAI-18}},
    publisher = {International Joint Conferences on Artificial Intelligence Organization},
        pages = {3762-3768},
         year = {2018},
     abstract = {AlphaGo Zero pioneered the concept of two-head neural networks in Monte Carlo Tree Search (MCTS), where the policy output is used for prior action probability and the state-value estimate is used for leaf node evaluation. We propose a three-head neural net architecture with policy, state- and action-value outputs, which could lead to more efficient MCTS since neural leaf estimate can still be back-propagated in tree with delayed node expansion and evaluation. To effectively train the newly introduced action-value head on the same game dataset as for two-head nets, we exploit the optimal relations between parent and children nodes for data augmentation and regularization. In our experiments for the game of Hex, the action-value head learning achieves similar error as the state-value prediction of a two-head architecture. The resulting neural net models are then combined with the same Policy Value MCTS (PV-MCTS) implementation. We show that, due to more efficient use of neural net evaluations, PV-MCTS with three-head neural nets consistently performs better than the two-head ones, significantly outplaying the state-of-the-art player MoHex-CNN.}
}
@article{GaoetalNATURE-COMMUNICATIONS-18,
       author = {Gao, Quan Q. and Putzbach, William E. and Murmann, Andrea E. and Chen, Siquan and Sarshad, Aishe A. and Peter, Johannes M. and Bartom, Elizabeth T. and Hafner, Markus and Peter, Marcus E.},
        title = {6mer seed toxicity in tumor suppressive microRNAs},
      journal = {Nature Communications},
         year = {2018},
       volume = {9},
        issue = {1},
        pages = {4504},
     abstract = {Many small-interfering (si)RNAs are toxic to cancer cells through a 6mer seed sequence (positions 2' of the guide strand). Here we performed an siRNA screen with all 4096 6mer seeds revealing a preference for guanine in positions 1 and 2 and a high overall G or C content in the seed of the most toxic siRNAs for four tested human and mouse cell lines. Toxicity of these siRNAs stems from targeting survival genes with C-rich 3′UTRs. The master tumor suppressor miRNA miR-34a-5p is toxic through such a G-rich 6mer seed and is upregulated in cells subjected to genotoxic stress. An analysis of all mature miRNAs suggests that during evolution most miRNAs evolved to avoid guanine at the miR-34a-5p end of the 6mer seed sequence of the guide strand. In contrast, for certain tumor-suppressive miRNAs the guide strand contains a G-rich toxic 6mer seed, presumably to eliminate cancer cells.},
}
@article{Garcia-LopezetalFiN-10,
       author = {Pablo Garcia-Lopez and Virginia Garcia-Marin and Miguel Freire},
        title = {The histological slides and drawings of Cajal},
      journal = {Frontiers in Neuroanatomy},
       volume = 4,
        issue = 9,
         year = 2010,
        pages = {1-16},
     abstract = {Ram\`{o}n y Cajal's studies in the field of neuroscience provoked a radical change in the course of its history. For this reason he is considered as the father of modern neuroscience. Some of his original preparations are housed at the Cajal Museum (Cajal Institute, CSIC, Madrid, Spain). In this article, we catalogue and analyse more than 4,500 of Cajal's histological preparations, the same preparations he used during his scientific career. Furthermore, we catalogued Cajal's original correspondence, both manuscripts and personal letters, drawings and plates. This is the first time anyone has compiled an account of Cajal's enormous scientific production, offering some curious insights into his work and his legacy.},
}
@article{GardinerPTRS_B-01,
       author = {Gardiner, J. M.},
        title = {Episodic memory and autonoetic consciousness: a first-person approach.},
      journal = {Philosophical Transactions of the Royal Society London B Biological Science},
         year = {2001},
       volume = {356},
        issue = {1413},
        pages = {1351-1361},
     abstract = {Episodic memory is identified with autonoetic consciousness, which gives rise to remembering in the sense of self-recollection in the mental re-enactment of previous events at which one was present. Autonoetic consciousness is distinguished from noetic consciousness, which gives rise to awareness of the past that is limited to feelings of familiarity or knowing. Noetic consciousness is identified not with episodic but with semantic memory, which involves general knowledge. A recently developed approach to episodic memory makes use of 'first-person' reports of remembering and knowing. Studies using this approach have revealed many independent variables that selectively affect remembering and others that selectively affect knowing. These studies can also be interpreted in terms of distinctiveness and fluency of processing. Remembering and knowing do not correspond with degrees of confidence in memory. Nor does remembering always control the memory response. There is evidence that remembering is selectively impaired in various populations, including not only amnesic patients and older adults but also adults with Asperger's syndrome. This first-person approach to episodic memory represents one way in which that most elusive aspect of consciousness, its subjectivity, can be investigated scientifically. The two kinds of conscious experiences can be manipulated experimentally in ways that are systematic, replicable and intelligible theoretically.},
}
@book{Gardner1995,
        title = {Self-Renewal: The Individual and the Innovative Society},
       author = {Gardner, John W.},
         year = {1995},
    publisher = {Norton}
} 
@article{Gardner70,
      author = {Martin Gardner},
       title = {The Fantastic Combinations of {John Conway's} New Solitaire Game of `{Life}'},
     journal = {Scientific American},
      volume = 223,
       pages = {120-123},
        year = 1970,
}}
@book{Gardner89,
      author = {Martin Gardner},
       title = {Gardner's Whys and Wherefores},
   publisher = {University of Chicago Press},
     address = {Chicago, IL},
        year = 1989,
}
@article{GardnerandAndrewCOiN-19,
       author = {Elizabeth Gardner and Andrew Ellington},
        title = {Reprogramming the brain with synthetic neurobiology},
      journal = {Current Opinion in Biotechnology},
       volume = {58},
        pages = {37-44},
         year = {2019},
     abstract = {The mammalian brain is among the most complex organs known in biology. Historically, neuroscience techniques have consisted primarily of low-throughput microscopy and electrophysiological approaches. While these methods will continue to serve the community, the emerging field of synthetic neurobiology may be better equipped to scale with systems neuroscience. By using genetic techniques to achieve cell-type specificity, a map of the connectome, neural activation and recording, and ultimately to program neural development itself, we can begin to build a better framework with which to understand the brain's mechanisms.}
}
@book{GareyandJohnson79,
      author = {Michael Garey and David Johnson},
       title = {Computers and Intractability: A Guide to the Theory of NP-Completeness},
   publisher = {W. H. Freeman},
     address = {New York, NY},
        year = 1979,
}
@incollection{GarriguesandGhaouiNIPS-08,
     title = {An Homotopy Algorithm for the Lasso with Online Observations},
    author = {Pierre Garrigues and Laurent El El Ghaoui},
 booktitle = {Advances in Neural Information Processing Systems 21},
    editor = {Dale Schuurmans and Yoshua Bengio},
 publisher = {MIT Press},
   address = {Cambridge, MA},
      year = 2008
}
@incollection{GarriguesandOlshausenNIPS-07,
      title = {Learning Horizontal Connections in a Sparse Coding Model of Natural Images},
     author = {Pierre Garrigues and Bruno Olshausen},
  booktitle = {Advances in Neural Information Processing Systems 20},
     editor = {J.C. Platt and D. Koller and Y. Singer and S. Roweis},
  publisher = {MIT Press},
    address = {Cambridge, MA},
      pages = {505-512},
       year = 2008
}
@inproceedings{GartnerFOCS-92,
       author = {B. Gartner},
        title = {A subexponential algorithm for abstract optimization problems},
    booktitle = {33rd Annual Symposium on Foundations of Computer Science},
        pages = {464-472},
         year = 1992,
}
@article{GartneretalFiP-13,
       author = {G\"{a}rtner, H. and Minnerop, M. and Pieperhoff, P. and Schleicher, A. and Zilles, K. and Altenm\"{u}ller, E. and Amunts, K.},
        title = {Brain morphometry shows effects of long-term musical practice in middle-aged keyboard players},
      journal = {Frontiers in Psychology},
         year = {2013},
       volume = {4},
        pages = {636-636},
     abstract = {To what extent does musical practice change the structure of the brain? In order to understand how long-lasting musical training changes brain structure, 20 male right-handed, middle-aged professional musicians and 19 matched controls were investigated. Among the musicians, 13 were pianists or organists with intensive practice regimes. The others were either music teachers at schools or string instrumentalists, who had studied the piano at least as a subsidiary subject, and practiced less intensively. The study was based on T1-weighted MR images, which were analyzed using deformation-based morphometry. Cytoarchitectonic probabilistic maps of cortical areas and subcortical nuclei as well as myeloarchitectonic maps of fiber tracts were used as regions of interest to compare volume differences in the brains of musicians and controls. In addition, maps of voxel-wise volume differences were computed and analyzed. Musicians showed a significantly better symmetric motor performance as well as a greater capability of controlling hand independence than controls. Structural MRI-data revealed significant volumetric differences between the brains of keyboard players, who practiced intensively and controls in right sensorimotor areas and the corticospinal tract as well as in the entorhinal cortex and the left superior parietal lobule. Moreover, they showed also larger volumes in a comparable set of regions than the less intensively practicing musicians. The structural changes in the sensory and motor systems correspond well to the behavioral results, and can be interpreted in terms of plasticity as a result of intensive motor training. Areas of the superior parietal lobule and the entorhinal cortex might be enlarged in musicians due to their special skills in sight-playing and memorizing of scores. In conclusion, intensive and specific musical training seems to have an impact on brain structure, not only during the sensitive period of childhood but throughout life.}
}
@article {GarwiczetalPNAS-09,
       author = {Garwicz, Martin and Christensson, Maria and Psouni, Elia},
        title = {A unifying model for timing of walking onset in humans and other mammals},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {106},
       number = {51},
         year = {2009},
        pages = {21889-21893},
     abstract = {The onset of walking is a fundamental milestone in motor development of humans and other mammals, yet little is known about what factors determine its timing. Hoofed animals start walking within hours after birth, rodents and small carnivores require days or weeks, and nonhuman primates take months and humans approximately a year to achieve this locomotor skill. Here we show that a key to the explanation for these differences is that time to the onset of walking counts from conception and not from birth, indicating that mechanisms underlying motor development constitute a functional continuum from pre- to postnatal life. In a multiple-regression model encompassing 24 species representative of 11 extant orders of placental mammals that habitually walk on the ground, including humans, adult brain mass accounted for 94\% of variance in time to walking onset postconception. A dichotomous variable reflecting species differences in functional limb anatomy accounted for another 3.8\% of variance. The model predicted the timing of walking onset in humans with high accuracy, showing that this milestone in human motor development occurs no later than expected given the mass of the adult human brain, which in turn reflects the duration of its ontogenetic development. The timing of motor development appears to be highly conserved in mammalian evolution as the ancestors of some of the species in the sample presented here diverged in phylogenesis as long as 100 million years ago. Fundamental patterns of early human life history may therefore have evolved before the evolution of primates.},
}
@article{Gasparetti03,
       author = {Fabio Gasparetti and Alessandro Micarelli},
        title = {Adaptive Web Search Based on a Colony of Cooperative Distributed Agents},
      journal = {Lecture Notes in Computer Science: Cooperative Information Agents},
       volume = 2782,
    publisher = {Springer-Verlag},
        pages = {168-183},
         year = 2003,
}
@article{GathercoleetalDEVELOPMENTAL-PSYCHOLOGY-04,
       author = {Gathercole, S. E. and Pickering, S. J. and Ambridge, B. and Wearing, H.},
        title = {The structure of working memory from 4 to 15 years of age},
      journal = {Developmental Psychology},
         year = {2004},
       volume = {40},
       number = {2},
        pages = {177-190},
     abstract = {The structure of working memory and its development across the childhood years were investigated in children 4-15 years of age. The children were given multiple assessments of each component of the A. D. Baddeley and G. Hitch (1974) working memory model. Broadly similar linear functions characterized performance on all measures as a function of age. From 6 years onward, a model consisting of 3 distinct but correlated factors corresponding to the working memory model provided a good fit to the data. The results indicate that the basic modular structure of working memory is present from 6 years of age and possibly earlier, with each component undergoing sizable expansion in functional capacity throughout the early and middle school years to adolescence.}
}
@article{GattoandKendal-11,
       author = {Gatto, Cheryl L. and Broadie, Kendal},
        title = {Drosophila Modeling of Heritable Neurodevelopmental Disorders},
      journal = {Current Opinion in Neurobiology},
       volume = {21},
        issue = {6},
         year = {2011},
        pages = {834-841},
     abstract = {Heritable neurodevelopmental disorders are multifaceted disease conditions encompassing a wide range of symptoms including intellectual disability, cognitive dysfunction, autism and myriad other behavioral impairments. In cases where single, causative genetic defects have been identified, such as Angelman syndrome, Rett syndrome, Neurofibromatosis Type 1 and Fragile X syndrome, the classical Drosophila genetic system has provided fruitful disease models. Recent Drosophila studies have advanced our understanding of UBE3A, MECP2, NF1 and FMR1 function, respectively, in genetic, biochemical, anatomical, physiological and behavioral contexts. Investigations in Drosophila continue to provide the essential mechanistic understanding required to facilitate the conception of rational therapeutic treatments.}
}
@book{GauglerandBilgrami2004,
        title = {Nematode Behaviour},
       author = {Gaugler, R. and Bilgrami, A.L.},
         year = 2004,
    publisher = {CABI}
}
@article{GauntetalCoRR-16,
       author = {Alexander L. Gaunt and Marc Brockschmidt and Rishabh Singh and Nate Kushman and Pushmeet Kohli and Jonathan Taylor and Daniel Tarlow},
        title = {{TerpreT}: {A} Probabilistic Programming Language for Program Induction},
      journal = {CoRR},
       volume = {arXiv:1612.00817},
         year = {2016},
     abstract = {We study machine learning formulations of inductive program synthesis; that is, given input-output examples, synthesize source code that maps inputs to corresponding outputs. Our key contribution is TerpreT, a domain-specific language for expressing program synthesis problems. A TerpreT model is composed of a specification of a program representation and an interpreter that describes how programs map inputs to outputs. The inference task is to observe a set of input-output examples and infer the underlying program. From a TerpreT model we automatically perform inference using four different back-ends: gradient descent (thus each TerpreT model can be seen as defining a differentiable interpreter), linear program (LP) relaxations for graphical models, discrete satisfiability solving, and the Sketch program synthesis system. TerpreT has two main benefits. First, it enables rapid exploration of a range of domains, program representations, and interpreter models. Second, it separates the model specification from the inference algorithm, allowing proper comparisons between different approaches to inference. We illustrate the value of TerpreT by developing several interpreter models and performing an extensive empirical comparison between alternative inference algorithms on a variety of program models. To our knowledge, this is the first work to compare gradient-based search over program space to traditional search-based alternatives. Our key empirical finding is that constraint solvers dominate the gradient descent and LP-based formulations.}
}
@article{GavranovicCoRR-19,
       author = {Bruno Gavranovic},
        title = {Compositional Deep Learning},
      journal = {CoRR},
       volume = {arXiv:1907.08292},
         year = {2019},
     abstract = {Neural networks have become an increasingly popular tool for solving many real-world problems. They are a general framework for differentiable optimization which includes many other machine learning approaches as special cases. In this thesis we build a category-theoretic formalism around a class of neural networks exemplified by CycleGAN. CycleGAN is a collection of neural networks, closed under composition, whose inductive bias is increased by enforcing composition invariants, i.e. cycle-consistencies. Inspired by Functorial Data Migration, we specify the interconnection of these networks using a categorical schema, and network instances as set-valued functors on this schema. We also frame neural network architectures, datasets, models, and a number of other concepts in a categorical setting and thus show a special class of functors, rather than functions, can be learned using gradient descent. We use the category-theoretic framework to conceive a novel neural network architecture whose goal is to learn the task of object insertion and object deletion in images with unpaired data. We test the architecture on three different datasets and obtain promising results.}
}
@book{GazdarandMellish89,
       author = {Gazdar, Gerald and Mellish, Chris},
        title = {Natural Language Processing in LISP},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1989,
}
@book{Gazzaniga2008,
        title = {Human},
       author = {Gazzaniga, Michael S.},
    publisher = {HarperCollins Publishers},
      address = {New York},
         year = 2008,
}
@book{Gazzaniga2009,
        title = {The Cognitive Neurosciences (Third Edition)},
       author = {Gazzaniga, Michael S.},
       series = {Bradford Books},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2009,
}
@inproceedings{GeertzenACL-09,
       author = {Geertzen, Jeroen},
        title = {Dialogue act prediction using stochastic context-free grammar induction},
    booktitle = {{Proceedings of the EACL 2009 Workshop on Computational Linguistic Aspects of Grammatical Inference}},
     location = {Athens, Greece},
         year = {2009},
        pages = {7-15},
    publisher = {Association for Computational Linguistics},
      address = {Stroudsburg, PA, USA},
}
@article{GeetalNATURE-06,
   Author="Ge, S.  and Goh, E. L.  and Sailor, K. A.  and Kitabatake, Y.  and Ming, G. L.  and Song, H. ",
   Title="{GABA} regulates synaptic integration of newly generated neurons in the adult brain",
   Journal="Nature",
   Number="7076",
   Volume="439",
   Year="2006",
   Pages="589-593",
   Abstract={Adult neurogenesis, the birth and integration of new neurons from adult neural stem cells, is a striking form of structural plasticity and highlights the regenerative capacity of the adult mammalian brain. Accumulating evidence suggests that neuronal activity regulates adult neurogenesis and that new neurons contribute to specific brain functions. The mechanism that regulates the integration of newly generated neurons into the pre-existing functional circuitry in the adult brain is unknown. Here we show that newborn granule cells in the dentate gyrus of the adult hippocampus are tonically activated by ambient GABA (gamma-aminobutyric acid) before being sequentially innervated by GABA- and glutamate-mediated synaptic inputs. GABA, the major inhibitory neurotransmitter in the adult brain, initially exerts an excitatory action on newborn neurons owing to their high cytoplasmic chloride ion content. Conversion of GABA-induced depolarization (excitation) into hyperpolarization (inhibition) in newborn neurons leads to marked defects in their synapse formation and dendritic development in vivo. Our study identifies an essential role for GABA in the synaptic integration of newly generated neurons in the adult brain, and suggests an unexpected mechanism for activity-dependent regulation of adult neurogenesis, in which newborn neurons may sense neuronal network activity through tonic and phasic GABA activation.}
}
@article{GehringetalCoRR-17,
       author = {Jonas Gehring and Michael Auli and David Grangier and Denis Yarats and Yann N. Dauphin},
        title = {Convolutional Sequence to Sequence Learning},
      journal = {CoRR},
       volume = {axXiv:1705.03122},
         year = {2017},
     abstract = {The prevalent approach to sequence to sequence learning maps an input sequence to a variable length output sequence via recurrent neural networks. We introduce an architecture based entirely on convolutional neural networks. Compared to recurrent models, computations over all elements can be fully parallelized during training and optimization is easier since the number of non-linearities is fixed and independent of the input length. Our use of gated linear units eases gradient propagation and we equip each decoder layer with a separate attention module. We outperform the accuracy of the deep LSTM setup of Wu et al. (2016) on both WMT'14 English-German and WMT'14 English-French translation at an order of magnitude faster speed, both on GPU and CPU.}
}
@inproceedings{GeibandSteidmanIJCAI-07,
       author = {Geib, Christopher W. and Steedman, Mark},
        title = {On Natural Language Processing and Plan Recognition},
    booktitle = {Proceedings of the 20th International Joint Conference on Artifical Intelligence},
    publisher = {Morgan Kaufmann Publishers Inc.},
      address = {San Francisco, CA, USA},
     location = {Hyderabad, India},
         year = {2007},
        pages = {1612-1617},
}
@inproceedings{GeigerandHeckermanUAI-95,
       author = {Dan Geiger and David Heckerman},
        title = {A Characterization of the {Dirichlet} Distribution with Application to Learning {Bayesian} Networks},
    publisher = {Morgan and Kaufmann},
    booktitle = {Proceedings of Eleventh Conference on Uncertainty in Artificial Intelligence},
     location = {Montreal, Quebec},
        pages = {196-207},
         year = 1995,
}
@inproceedings{GeigeretalUAI-96,
       author = {Geiger, Dan and Heckerman, David and Meek, Chris},
        title = {Asymptotic Model Selection for Directed Networks with Hidden Variables},
    booktitle = {Twelfth Conference on Uncertainty in Artificial Intelligence},
     location = {Portland, Oregon},
        pages = {283-290},
         year = 1996,
}
@book{Gelb74,
       editor = {Gelb, A.},
        title = {Applied Optimal Estimation},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1974,
}
@book{Gelernter92,
     author = {David Gelernter},
      title = {Mirror Worlds: or the Day Software Puts the Universe in a Shoebox ... How It Will Happen and What It Will Mean}, 
  publisher = {Oxford University Press},
    address = {Oxford, UK},
       year = 1992,
}
@book{Gelernter94,
       author = {David Gelernter},
        title = {The Muse in the Machine: Computerizing the Poetry of Human Thought},
    publisher = {The Free Press},
      address = {New York, NY},
         year = 1994,
}
@book{Gelfand63,
       author = {Gelfand, I. M. and Fomin, S. V.},
        title = {Calculus of Variations},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1963,
}
@incollection{Gelfond91,
       author = {M. Gelfond, V. Lifschitz, A. Rabinov},
        title = {What are the Limitations of Situational Calculus},
    booktitle = {Automated Reasoning: Essays in Honor of Woody Bledsoe},
       editor = {R. S. Boyer},
    publisher = {Kluwer},
         year = 1991,
}
@book{GellMann94,
       author = {Murray Gell-Mann},
        title = {Quark and the Jaguar: Adventures in the Simple and the Complex},
    publisher = {W. H. Freeman and Company},
      address = {New York},
         year = 1994,
}
@book{GelmanCarlinSternRubin95,
       author = {Andrew Gelman and John B. Carlin and Hal S. Stern and Donald B. Rubin},
        title = {Bayesian Data Analysis},
    publisher = {Chapman \& Hall},
         year = 1995,
}
@incollection{Geman87,
       author = {Geman, Stuart},
        title = {Stochastic Relaxation Methods for Image Restoration and Expert Systems},
       editor = {D.B. Cooper and R.L. Launer and D.E. McClure},
    booktitle = {Automated Image Analysis: Theory and Experiments},
    publisher = {Academic Press},
      address = {New York},
         year = 1987,
}
@incollection{GemanCTF-03,
     author = {Geman, Donald},
      title = {Coarse-to-fine classification and scene labeling},
     editor = {Denison, D.D. and Hansen, M.H. and Holmes, C.C. and Mallick, B. and Yu, B.},
  publisher = {Springer-Verlag},
  booktitle = {Nonlinear Estimation and Classification (Lecture Notes in Statistics Series)},
      pages = {31-48}, 
       year = 2003,
}
@misc{GemanCVPR-04,
  author = {Geman, Stuart},
   title = {Invited talk at {CVPR} 2004:
            {I}nvariance and {S}electivity in the {V}entral {V}isual {P}athway},
   month = {October},
 comment = {Invited talk at {CVPR} 2004},
 comment = {GemanCVPR-2004.ppt},
 comment = {GemanCommentary.pdf},
    year = 2004,
}
@article{GemanFacultyBulletin,
       author = {Geman, Stuart},
        title = {Compositionality},
      journal = {Brown University Faculty Bulletin},
       volume = {Spring},
      comment = {GemanFacultyBulletin.pdf},
         year = 1999,
}
@article{GemanJoP-06,
       author = {Geman, Stuart},
        title = {Invariance and Selectivity in the Ventral Visual Pathway},
      journal = {Journal of Physiology --- Paris},
       volume = 100,
        issue = 4,
        pages = {212-224},
         year = 2006,
}
@article{GemanandGemanPAMI-84,
       author = {S. Geman and D. Geman},
        title = {Stochastic relaxation, {G}ibbs distributions, and the {B}ayesian restoration of images},
      journal = PAMI,
         year = 1984,
       volume = 6,
        issue = 6,
        pages = {721-741},
}
@article{GemanandReynoldsPAMI-92,
   author = {Donald Geman and George Reynolds},
    title = {Constrained Restoration and the Recovery of Discontinuities},
  journal = PAMI,
     year = 1992,
   volume = 14,
    issue = 3,
    pages = {367-383},
}
@article{GemanetalQAM-02,
  author = {Geman, S. and Potter, D. and Chi, Z.},
   title = {Composition systems},
 journal = {Quarterly of Applied Mathematics},
  volume = {LX},
    year = 2002,
   pages = {707-736},
 comment = {GemanetalQAM-02.pdf},
}
@article{GemicietalCoRR-17,
       author = {Mevlana Gemici and Chia{-}Chun Hung and Adam Santoro and Greg Wayne and Shakir Mohamed and Danilo Jimenez Rezende and David Amos and Timothy P. Lillicrap},
        title = {Generative Temporal Models with Memory},
      journal = {CoRR},
       volume = {arXiv:1702.04649},
         year = {2017},
     abstract = {We consider the general problem of modeling temporal data with long-range dependencies, wherein new observations are fully or partially predictable based on temporally-distant, past observations. A sufficiently powerful temporal model should separate predictable elements of the sequence from unpredictable elements, express uncertainty about those unpredictable elements, and rapidly identify novel elements that may help to predict the future. To create such models, we introduce Generative Temporal Models augmented with external memory systems. They are developed within the variational inference framework, which provides both a practical training methodology and methods to gain insight into the models' operation. We show, on a range of problems with sparse, long-term temporal dependencies, that these models store information from early in a sequence, and reuse this stored information efficiently. This allows them to perform substantially better than existing models based on well-known recurrent neural networks, like LSTMs.}
}
@article{GeminianietalIJoNS-18,
       author = {Geminiani, Alice and Casellato, Claudia and Antonietti, Alberto and D'Angelo, Egidio and Pedrocchi, Alessandra},
        title = {A Multiple-Plasticity Spiking Neural Network Embedded in a Closed-Loop Control System to Model Cerebellar Pathologies},
      journal = {International Journal of Neural Systems},
       volume = {28},
       number = {05},
        pages = {1750017},
         year = {2018},
     abstract = {The cerebellum plays a crucial role in sensorimotor control and cerebellar disorders compromise adaptation and learning of motor responses. However, the link between alterations at network level and cerebellar dysfunction is still unclear. In principle, this understanding would benefit of the development of an artificial system embedding the salient neuronal and plastic properties of the cerebellum and operating in closed-loop. To this aim, we have exploited a realistic spiking computational model of the cerebellum to analyze the network correlates of cerebellar impairment. The model was modified to reproduce three different damages of the cerebellar cortex: (i) a loss of the main output neurons (Purkinje Cells), (ii) a lesion to the main cerebellar afferents (Mossy Fibers), and (iii) a damage to a major mechanism of synaptic plasticity (Long Term Depression). The modified network models were challenged with an Eye-Blink Classical Conditioning test, a standard learning paradigm used to evaluate cerebellar impairment, in which the outcome was compared to reference results obtained in human or animal experiments. In all cases, the model reproduced the partial and delayed conditioning typical of the pathologies, indicating that an intact cerebellar cortex functionality is required to accelerate learning by transferring acquired information to the cerebellar nuclei. Interestingly, depending on the type of lesion, the redistribution of synaptic plasticity and response timing varied greatly generating specific adaptation patterns. Thus, not only the present work extends the generalization capabilities of the cerebellar spiking model to pathological cases, but also predicts how changes at the neuronal level are distributed across the network, making it usable to infer cerebellar circuit alterations occurring in cerebellar pathologies.}
}
@article{GendronetalGERONTOLOGIST-16,
       author = {Gendron, Tracey L. and Welleford, E. Ayn and Inker, Jennifer and White, John T.},
        title = {The Language of Ageism: Why We Need to Use Words Carefully},
      journal = {The Gerontologist},
       volume = {56},
       number = {6},
         year = {2016},
        pages = {997-1006},
     abstract = {Language carries and conveys meaning which feeds assumptions and judgments that can lead to the development of stereotypes and discrimination. As a result, this study closely examined the specific language that is used to communicate attitudes and perceptions of aging and older adults. We conducted a qualitative study of a twitter assignment for 236 students participating in a senior mentoring program. Three hundred fifty-four tweets were qualitatively analyzed to explore language-based age discrimination using a thematic analytic approach. Twelve percent of the tweets (n = 43) were found to contain discriminatory language. Thematic analysis of the biased tweets identified 8 broad themes describing language-based age discrimination: assumptions and judgments, older people as different, uncharacteristic characteristics, old as negative, young as positive, infantilization, internalized ageism, and internalized microaggression.},
}
@inproceedings{GeneserethAAAI-83,
       author = {Genesereth, Michael R.},
        title = {An Overview of Metalevel Architecture},
    booktitle = {Proceedings AAAI-83},
     location = {Washington, DC, USA},
 organization = {AAAI},
         year = 1983,
        pages = {119-123},
}
@techreport{GeneserethTR-84,
       author = {Genesereth, Michael R.},
        title = {Partial Programs},
 type         = {Report No.},
        issue = {HPP-84-1},
  institution = {Stanford Heuristic Programming Project},
         year = 1984,
}
@inproceedings{GeneserethandHsuKR-91,
       author = {Genesereth, Michael R. and Hsu, Jane Yung-jen},
        title = {Partial programs},
    booktitle = {Principles of Knowledge Representation and Reasoning: Proceedings of the Second International Conference},
     location = {Cambridge, Massachusetts},
        pages = {238-249},
         year = 1991,
}
@book{GeneserethandNilsson87,
       author = {Genesereth, Michael R. and Nilsson, Nils J.},
        title = {Logical Foundations of Artificial Intelligence},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1987,
}
@article{GengetalACS-13,
       author = {Geng, Yanli and Pearson, Anthony C. and Gates, Elisabeth P. and Uprety, Bibek and Davis, Robert C. and Harb, John N. and Woolley, Adam T.},
        title = {Electrically Conductive Gold- and Copper-Metallized {DNA} Origami Nanostructures},
      journal = {Langmuir},
       volume = 29,
        issue = 10,
         year = 2013,
        pages = {3482-3490},
     abstract = {This work demonstrates the use of a circuit-like DNA origami structure as a template to fabricate conductive gold and copper nanostructures on Si surfaces. We improved over previous results by using multiple Pd seeding steps to increase seed uniformity and density. Our process has also been characterized through atomic force microscopy, particle size distribution analysis, and scanning electron microscopy. We found that four successive Pd seeding steps yielded the best results for electroless metal plating on DNA origami. Electrical resistance measurements were done on both Au- and Cu-metallized nanostructures, with each showing ohmic behavior. Gold-plated DNA origami structures made under optimal conditions had an average resistivity of 7.0 \% 10^{-5} \ohm \cdot m, whereas copper-metallized structures had a resistivity as low as 3.6 \% 10^{-4} \ohm \cdot m. Importantly, this is the first demonstration of electrically conductive Cu nanostructures fabricated on either DNA or DNA origami templates. Although resistivities for both gold and copper samples were larger than those of the bulk metal, these metal nanostructures have the potential for use in electrically connecting small structures. In addition, these metallized objects might find use in surface-enhanced Raman scattering experiments.}
}
@book{Gennaro2018,
        title = {The Routledge Handbook of Consciousness},
       author = {Gennaro, R.J.},
       series = {Routledge Handbooks in Philosophy},
    publisher = {Taylor \& Francis},
      address = {Oxford, UK},
         year = {2018},
}
@article{GentnerCS-83,
       author = {Dedre Gentner},
        title = {{Structure-mapping: A theoretical framework for analogy}},
      journal = {Cognitive Science},
       volume = {7},
       number = {2},
         year = {1983},
        pages = {155-170},
     abstract = {A theory of analogy must describe how the meaning of an analogy is derived from the meanings of its parts. In the structure-mapping theory, the interpretation rules are characterized as implicit rules for mapping knowledge about a base domain into a target domain. Two important features of the theory are (a) the rules depend only on syntactic properties of the knowledge representation, and not on the specific content of the domains; and (b) the theoretical framework allows analogies to be distinguished cleanly from literal similarity statements, applications of abstractions, and other kinds of comparisons. Two mapping principles are described: (a) Relations between objects, rather than attributes of objects, are mapped from base to target; and (b) The particular relations mapped are determined by systematicity, as defined by the existence of higher-order relations.}
}
@inproceedings{GenzelandCharniakACL-02,
    author = {Dmitriy Genzel and Eugene Charniak},
     title = {Entropy Rate Constancy in Text},
 booktitle = {Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics},
  location = {Philadelphia, PA},
     pages = {199-206},
      year = 2002,
}
@incollection{GeorgeandHawkinsIJCNN-05,
    author = {Dileep George and Jeff Hawkins},
     title = {A Hierarchical {B}ayesian Model of Invariant Pattern Recognition in the Visual Cortex},
 publisher = {IEEE},
    volume = 3,
 booktitle = {Proceedings of the International Joint Conference on Neural Networks},
  location = {Montreal, Canada},
     pages = {1812-1817},
      year = 2005,
}
@article{GeorgeandHawkinsPLoS-09,
        title = {Towards a mathematical theory of cortical micro-circuits},
       author = {George, Dileep and Hawkins, Jeff},
      journal = {{PLoS} Computational Biology},
       volume = 5,
        issue = 10,
         year = 2009,
    publisher = {Public Library of Science},
}
@incollection{Georgeff87,
       author = {Georgeff, Michael P.},
        title = {Planning},
       editor = {Traub, J. F.},
    booktitle = {Annual Review of Computer Science, Volume 2},
    publisher = {Annual Review Incorporated},
         year = 1987,
}
@inproceedings{GeorgeffandLanskyAAAI-87,
       author = {Georgeff, Michael P. and Amy L. Lansky},
        title = {Reactive Reasoning and Planning},
    booktitle = {Proceedings AAAI-87},
     location = {Seattle, Washington},
 organization = {AAAI},
         year = 1987,
        pages = {677-682},
}
@book{GeorgeffandLanskyPW-86,
       editor = {Georgeff, Michael P. and Lansky, Amy L.},
        title = {Reasoning About Actions and Plans},
    booktitle = {Reasoning About Actions and Plans},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1987,
}
@article{GepshteinandKubovyPNAS-01,
      title = {The emergence of visual objects in space-time},
     author = {Sergei Gepshtein and Michael Kubovy},
    journal = {Proceedings of the National Academy of Science},
     volume = 97,
      issue = 14,
      pages = {8186-8191},
        url = {http://www.pnas.org/cgi/content/full/97/14/8186},
       year = 2001,
}
@article{GerhardetalPLoS-13,
     title = {Successful Reconstruction of a Physiological Circuit with Known Connectivity from Spiking Activity Alone},
    author = {Gerhard, Felipe and Kispersky, Tilman and Gutierrez, Gabrielle J. and Marder, Eve and Kramer, Mark and Eden, Uri},
   journal = {{PLoS} Compututational Biology},
 publisher = {Public Library of Science},
    volume = 9,
     issue = 7,
      year = 2013,
     pages = {e1003138},
  abstract = {To appreciate how neural circuits control behaviors, we must understand two things. First, how the neurons comprising the circuit are connected, and second, how neurons and their connections change after learning or in response to neuromodulators. Neuronal connectivity is difficult to determine experimentally, whereas neuronal activity can often be readily measured. We describe a statistical model to estimate circuit connectivity directly from measured activity patterns. We use the timing relationships between observed spikes to predict synaptic interactions between simultaneously observed neurons. The model estimate provides each predicted connection with a curve that represents how strongly, and at which temporal delays, one circuit element effectively influences another. These curves are analogous to synaptic interactions of the level of the membrane potential of biological neurons and share some of their features such as being inhibitory or excitatory. We test our method on recordings from the pyloric circuit in the crab stomatogastric ganglion, a small circuit whose connectivity is completely known beforehand, and find that the predicted circuit matches the biological one --- a result other techniques failed to achieve. In addition, we show that drug manipulations impacting the circuit are revealed by this technique. These results illustrate the utility of our analysis approach for inferring connections from neural spiking activity.},
}
@misc{GersPhD-01,
        title = {Long Short-Term Memory in Recurrent Neural Networks},
       author = {Felix Gers},
         year = 2001,
 howpublished = {Diploma thesis. \'{E}colepolytechnique F\'{e}d\'{w}rale de Lausanne},
}
@article{GersetalJMLR-02,
       author = {Felix A. Gers and Nicol N. Schraudolph and J\"urgen Schmidhuber},
        title = {Learning Precise Timing with {LSTM} Recurrent Networks},
      journal = {Journal of Machine Learning Research},
       volume = 3,
         year = 2002,
        pages = {115-143},
     abstract = {The temporal distance between events conveys information essential for numerous sequential tasks such as motor control and rhythm detection. While Hidden Markov Models tend to ignore this information, recurrent neural networks (RNNs) can in principle learn to make use of it. We focus on Long Short-Term Memory (LSTM) because it has been shown to outperform other RNNs on tasks involving long time lags. We find that LSTM augmented by "peephole connections" from its internal cells to its multiplicative gates can learn the fine distinction between sequences of spikes spaced either 50 or 49 time steps apart without the help of any short training exemplars. Without external resets or teacher forcing, our LSTM variant also learns to generate stable streams of precisely timed spikes and other highly nonlinear periodic patterns. This makes LSTM a promising approach for tasks that require the accurate measurement or generation of time intervals.}
}
@book{Gershenfeld1999,
        title = {The Nature of Mathematical Modeling},
       author = {Gershenfeld, N.},
         year = {1999},
    publisher = {Cambridge University Press}
}
@book{Gershenfeld99,
      author = {Neil A. Gershenfeld},
       title = {When Things Start to Think},
   publisher = {Henry Holt},
     address = {New York, NY},
        year = 1999,
}
@article{GershmanCoRR-19,
        title = {What does the free energy principle tell us about the brain?},
       author = {Samuel J. Gershman},
      journal = {CoRR},
         year = {2019},
       volume = {arXiv:1901.07945},
     abstract = {The free energy principle has been proposed as a unifying account of brain function. It is closely related, and in some cases subsumes, earlier unifying ideas such as Bayesian inference, predictive coding, and active learning. This article clarifies these connections, teasing apart distinctive and shared predictions.}
}
@article{GershmanandDawANNUAL-REVIEWS-17,
        title = {Reinforcement Learning and Episodic Memory in Humans and Animals: {A}n Integrative Framework},
       author = {Samuel J. Gershman and Nathaniel D. Daw},
      journal = {Annual Reviews of Psychology},
         year = {2017},
       volume = {68},
        pages = {101-128},
	abstract = {We review the psychology and neuroscience of reinforcement learning (RL), which has experienced significant progress in the past two decades, enabled by the comprehensive experimental study of simple learning and decision-making tasks. However, one challenge in the study of RL is computational: The simplicity of these tasks ignores important aspects of reinforcement learning in the real world: (a) State spaces are high-dimensional, continuous, and partially observable; this implies that (b) data are relatively sparse and, indeed, precisely the same situation may never be encountered twice; furthermore, (c) rewards depend on the long-term consequences of actions in ways that violate the classical assumptions that make RL tractable. A seemingly distinct challenge is that, cognitively, theories of RL have largely involved procedural and semantic memory, the way in which knowledge about action values or world models extracted gradually from many experiences can drive choice. This focus on semantic memory leaves out many aspects of memory, such as episodic memory, related to the traces of individual events. We suggest that these two challenges are related. The computational challenge can be dealt with, in part, by endowing RL systems with episodic memory, allowing them to (a) efficiently approximate value functions over complex state spaces, (b) learn with very little data, and (c) bridge long-term dependencies between actions and rewards. We review the computational theory underlying this proposal and the empirical evidence to support it. Our proposal suggests that the ubiquitous and diverse roles of memory in RL may function as part of an integrated learning system.}
}
@article{GershmanetalPLoS-16,
       author = {Sam Gershman and Tobias Gerstenberg and Chris Baker and Fiery Cushman},
      journal = {PLoS ONE},
        title = {Plans, habits, and theory of mind},
         year = {2016},
     abstract = {Human success and even survival depends on our ability to predict what others will do by guessing what they are thinking. If I accelerate, will he yield? If I propose, will she accept? If I confess, will they forgive? Psychologists call this capacity "theory of mind." According to current theories, we solve this problem by assuming that others are rational actors. That is, we assume that others design and execute efficient plans to achieve their goals, given their knowledge. But if this view is correct, then our theory of mind is startlingly incomplete. Human action is not always a product of rational planning, and we would be mistaken to always interpret others' behaviors as such. A wealth of evidence indicates that we often act habitually--a form of behavioral control that depends not on rational planning, but rather on a history of reinforcement. We aim to test whether the human theory of mind includes a theory of habitual action and to assess when and how it is deployed. In a series of studies, we show that human theory of mind is sensitive to factors influencing the balance between habitual and planned behavior.},
}
@article{GestrietalDN-12,
       author = {Gestri, Gaia and Link, Brian A. and Neuhauss, Stephan C.F.},
        title = {The Visual System of Zebrafish and its Use to Model Human Ocular Diseases},
      journal = {Developmental Neurobiology},
         year = {2012},
       volume = {72},
        issue = {3},
        pages = {302-327},
     abstract = {Free swimming zebrafish larvae depend mainly on their sense of vision to evade predation and to catch prey. Hence there is strong selective pressure on the fast maturation of visual function and indeed the visual system already supports a number of visually-driven behaviors in the newly hatched larvae. The ability to exploit the genetic and embryonic accessibility of the zebrafish in combination with a behavioral assessment of visual system function has made the zebrafish a popular model to study vision and its diseases. Here, we review the anatomy, physiology and development of the zebrafish eye as the basis to relate the contributions of the zebrafish to our understanding of human ocular diseases.},
}
@book{GetoorandTaskar2006,
     title = {Introduction to Statistical Relational Learning},
    editor = {Lise Getoor and Ben Taskar},
 publisher = {MIT Press},
      year = 2006,
}
@inproceedings{GhahramaniandJordanNIPS-95,
       title = {Factorial Hidden {M}arkov Models},
      author = {Ghahramani, Zoubin and Jordan, Michael},
      editor = {Touretzky, D. S. and Leen, T. K.},
   booktitle = {Advances in Neural Information Processing 7},
   publisher = {{MIT} Press},
     address = {Cambridge, Massachusetts},
        year = 1995,
}
@techreport{GhahramaniandJordanTR-95,
       title = {Factorial Hidden {M}arkov Models},
      author = {Ghahramani, Zoubin and Jordan, Michael},
 institution = {{MIT} Department of Brain and Cognitive Sciences},
       issue = 9503,
        year = 1995,
}     
@inproceedings{GhahramaniandRoweisKFNN-04,
       author = {Ghahramani, Zoubin and Roweis, Sam T.},
        title = {Learning Nonlinear Dynamical Systems Using the Expectation-Maximization Algorithm},
    booktitle = {Kalman Filtering and Neural Networks},
       editor = {Simon Haykin},
    publisher = {John Wiley \& Sons},
         year = {2004},
        pages = {175-220},
}
@inproceedings{GhahramaniandRoweisNIPS-99,
       author = {Ghahramani, Zoubin and Roweis, Sam T.},
        title = {Learning Nonlinear Dynamical Systems Using an EM Algorithm},
    booktitle = {Proceedings of the 1998 Conference on Advances in Neural Information Processing Systems II},
    publisher = {MIT Press},
         year = {1999},
        pages = {431-437},
}
@inproceedings{GhemawatetalSOSP-03,
    author = {Sanjay Ghemawat and Howard Gobioff and Shun-Tak Leung},
     title = {The {Google} file system},
 booktitle = {Proceedings of the 19th ACM Symposium on Operating Systems Principles},
 publisher = {ACM},
      year = 2003,
     pages = {29-43},
}
@article{GhettiandBungeDEVELOPMENTAL-COGNITIVE_NEUROSCIENCE-12,
       author = {Simona Ghetti and Silvia A. Bunge},
        title = {Neural changes underlying the development of episodic memory during middle childhood},
      journal = {Developmental Cognitive Neuroscience},
       volume = {2},
       number = {4},
         year = {2012},
        pages = {381-395},
     abstract = {Episodic memory is central to the human experience. In typically developing children, episodic memory improves rapidly during middle childhood. While the developmental cognitive neuroscience of episodic memory remains largely uncharted, recent research has begun to provide important insights. It has long been assumed that hippocampus-dependent binding mechanisms are in place by early childhood, and that improvements in episodic memory observed during middle childhood result from the protracted development of the prefrontal cortex. We revisit the notion that binding mechanisms are age-invariant, and propose that changes in the hippocampus and its projections to cortical regions also contribute to the development of episodic memory. We further review the role of developmental changes in lateral prefrontal and parietal cortices in this development. Finally, we discuss changes in white matter tracts connecting brain regions that are critical for episodic memory. Overall, we argue that changes in episodic memory emerge from the concerted effort of a network of relevant brain structures.}
}
@article{GhoseTsoJN-97,
    author = {Geoffrey M. Ghose and Daniel Y. Ts'O},
     title = {Form Processing Modules in Primate Area {V4}},
   journal = {Journal of Neurophysiology},
    volume = 77,
     pages = {2191-2196},
      year = 1997, 
}
@article{GhoshetalCoRR-14a,
       author = {Shalini Ghosh and Daniel Elenius and Wenchao Li and Patrick Lincoln and Natarajan Shankar and Wilfried Steiner},
        title = {Automatically Extracting Requirements Specifications from Natural Language},
      journal = {CoRR},
       volume = {arXiv:1403.3142},
         year = 2014,
}
@article{GhoshetalCoRR-14b,
       author = {Shalini Ghosh and Patrick Lincoln and Christian Petersen and Alfonso Valdes},
        title = {Virus Detection in Multiplexed Nanowire Arrays using Hidden Semi-Markov models},
      journal = {CoRR},
       volume = {arXiv:1407.4490},
         year = 2014,
}
@article{GhoshetalCoRR-16,
        title = {Contextual {LSTM (CLSTM)} models for Large scale {NLP} tasks},
       author = {Shalini Ghosh and Oriol Vinyals and Brian Strope and Scott Roy and Tom Dean and Larry Heck},
      journal = {CoRR},
       volume = {arXiv:1602.06291},
         year = {2016},
     abstract = {Documents exhibit sequential structure at multiple levels of abstraction (e.g., sentences, paragraphs, sections). These abstractions constitute a natural hierarchy for representing the context in which to infer the meaning of words and larger fragments of text. In this paper, we present CLSTM (Contextual LSTM), an extension of the recurrent neural network LSTM (Long-Short Term Memory) model, where we incorporate contextual features (e.g., topics) into the model. We evaluate CLSTM on three specific NLP tasks: word prediction, next sentence selection, and sentence topic prediction. Results from experiments run on two corpora, English documents in Wikipedia and a subset of articles from a recent snapshot of English Google News, indicate that using both words andtopics as features improves performance of the CLSTM models over baseline LSTM models for these tasks. For example on the next sentence selection task, we get relative accuracy improvements of 21\% for the Wikipedia dataset and 18\% for the Google News dataset. This clearly demonstrates the significant benefit of using context appropriately in natural language (NL) tasks. This has implications for a wide variety of NL applications like question answering, sentence completion, paraphrase generation, and next utterance prediction in dialog systems.}
}
@article{GhoshetalNATURE-11,
     title = {Miniaturized integration of a fluorescence microscope},
    author = {Ghosh, K.K. and Burns, L.D. and Cocker, E.D. and Nimmerjahn, A. and Ziv, Y. and Gamal, A.E. and Schnitzer, M.J.},
   journal = {Nature Methods},
    volume = 8,
     issue = 10,
     pages = {871-8},
      year = 2011,
  abstract = {The light microscope is traditionally an instrument of substantial size and expense. Its miniaturized integration would enable many new applications based on mass-producible, tiny microscopes. Key prospective usages include brain imaging in behaving animals for relating cellular dynamics to animal behavior. Here we introduce a miniature (1.9 g) integrated fluorescence microscope made from mass-producible parts, including a semiconductor light source and sensor. This device enables high-speed cellular imaging across \~0.5 mm$^2$ areas in active mice. This capability allowed concurrent tracking of Ca2+ spiking in >200 Purkinje neurons across nine cerebellar microzones. During mouse locomotion, individual microzones exhibited large-scale, synchronized Ca2+ spiking. This is a mesoscopic neural dynamic missed by prior techniques for studying the brain at other length scales. Overall, the integrated microscope is a potentially transformative technology that permits distribution to many animals and enables diverse usages, such as portable diagnostics or microscope arrays for large-scale screens.}
}
@article{GhuloumetalITJ-07,
      title = {Future-Proof Data Parallel Algorithms and Software on {I}ntel Multi-Core Architecture}, 
     author = {A. Ghuloum and T. Smith and G. Wu and X. Zhou and J. Fang and 
               P. Guo and B. So and M. Rajagopalan and Y. Chen and B. Chen},
    journal = {Intel Technology Journal},
     volume = 11,
      issue = 4,
      pages = {333-347},
       year = 2007,
}}
@article{GibneyNATURE-15,
       author = {Gibney, Heidi},
        title = {Injectable brain implant spies on individual neurons: Electronic mesh has potential to unravel workings of mammalian brain.},
      journal = {Nature},
       volume = 522,
         year = 2015,
        pages = {137-138},
}
@book{Gibson63,
      author = {Gibson, John E.},
       title = {Nonlinear Automatic Control},
   publisher = {McGraw-Hill},
     address = {New York},
        year = 1963,
}
@book{GibsonAFFORDANCES-79,
      author = {Gibson, James J.},
       title = {The Ecological Approach to Visual Perception},
   publisher = {Houghton Mifflin}, 
     address = {Boston},
        year = 1979,
}
@book{GibsonPERCEPTION-50,
      author = {Gibson, James J.},
       title = {Perception of the Visual World},
   publisher = {Houghton Mifflin}, 
     address = {Boston},
        year = 1950,
}
@inproceedings{Gibsonetal98,
      author = {David Gibson and Jon M. Kleinberg and Prabhakar Raghavan},
       title = {Inferring {W}eb Communities from Link Topology},
   booktitle = {Proceedings of the 9th {ACM} Conference on Hypertext and Hypermedia},
       pages = {225-234},
       month = {June},
     address = {Pittsburgh, Pennsylvania},
         url = {http://citeseer.nj.nec.com/gibson98inferring.html},
        year = 1998,
}
@article{GickandHolyoakCOGPSYC-15,
       author = {Mary L. Gick and Keith J. Holyoak},
        title = {Schema induction and analogical transfer},
      journal = {Cognitive Psychology},
       volume = {15},
       number = {1},
         year = {1983},
        pages = {1-38},
     abstract = {An analysis of the process of analogical thinking predicts that analogies will be noticed on the basis of semantic retrieval cues and that the induction of a general schema from concrete analogs will facilitate analogical transfer. These predictions were tested in experiments in which subjects first read one or more stories illustrating problems and their solutions and then attempted to solve a disparate but analogous transfer problem. The studies in Part I attempted to foster the abstraction of a problem schema from a single story analog by means of summarization instructions, a verbal statement of the underlying principle, or a diagrammatic representation of it. None of these devices achieved a notable degree of success. In contrast, the experiments in Part II demonstrated that if two prior analogs were given, subjects often derived a problem schema as an incidental product of describing the similarities of the analogs. The quality of the induced schema was highly predictive of subsequent transfer performance. Furthermore, the verbal statements and diagrams that had failed to facilitate transfer from one analog proved highly beneficial when paired with two. The function of examples in learning was discussed in light of the present study.}
}
@article{GieddetalNN-99, 
    author = {Giedd, J. N. and Blumenthal, J. and Jeffries, N. O. and Castellanos, F. X. and Liu, H. and Zijdenbos, A. and Paus, T. and Evans, A. C. and Rapoport, J. L.}, 
     title = {Brain development during childhood and adolescence: a longitudinal {MRI} study}, 
    volume = 2, 
     issue = 10, 
   journal = {Nature Neuroscience}, 
 publisher = {Nature America}, 
      year = 1999, 
     pages = {861-863},
}
@article{GielowandZaborszkyCELL-17,
       author = {Gielow, Matthew R. and Zaborszky, Laszlo},
        title = {The Input-Output Relationship of the Cholinergic Basal Forebrain},
      journal = {Cell Reports},
         year = {2017},
    publisher = {Elsevier},
       volume = {18},
        issue = {7},
        pages = {1817-1830},
     abstract = {Basal forebrain cholinergic neurons influence cortical state, plasticity, learning, and attention. They collectively innervate the entire cerebral cortex, differentially controlling acetylcholine efflux across different cortical areas and timescales. Such control might be achieved by differential inputs driving separable cholinergic outputs, although no input-output relationship on a brain-wide level has ever been demonstrated. Here, we identify input neurons to cholinergic cells projecting to specific cortical regions by infecting cholinergic axon terminals with a monosynaptically restricted viral tracer. This approach revealed several circuit motifs, such as central amygdala neurons synapsing onto basolateral amygdala-projecting cholinergic neurons or strong somatosensory cortical input to motor cortex-projecting cholinergic neurons. The presence of input cells in the parasympathetic midbrain nuclei contacting frontally projecting cholinergic neurons suggest that the network regulating the inner eye muscles are additionally regulating cortical state via acetylcholine efflux. This dataset enables future circuit-level experiments to identify drivers of known cortical cholinergic functions.},
}
@article{GiepmansetalSCIENCE-06,
       author = {Giepmans, B. N. and Adams, S. R. and Ellisman, M. H. and Tsien, R. Y.},
        title = {The fluorescent toolbox for assessing protein location and function},
      journal = {Science},
         year = 2006,
       volume = 312,
        issue = 5771,
        pages = {217-224},
     abstract = {Advances in molecular biology, organic chemistry, and materials science have recently created several new classes of fluorescent probes for imaging in cell biology. Here we review the characteristic benefits and limitations of fluorescent probes to study proteins. The focus is on protein detection in live versus fixed cells: determination of protein expression, localization, activity state, and the possibility for combination of fluorescent light microscopy with electron microscopy. Small organic fluorescent dyes, nanocrystals ("quantum dots"), autofluorescent proteins, small genetic encoded tags that can be complexed with fluorochromes, and combinations of these probes are highlighted.},
}
@book{Gilbert2006,
     author = {Daniel Gilbert},
      title = {Stumbling On Happiness},
  publisher = {Albert A. Knopf},
    address = {New York, NY},
       year = 2006,
    comment = {http://en.wikipedia.org/wiki/Stumbling_on_Happiness},
}
@article{GilbertandLiNEURON-12,
       author = {Gilbert, Charles D. and Li, Wu},
        title = {Adult visual cortical plasticity},
      journal = {Neuron},
         year = {2012},
       volume = {75},
        issue = {2},
        pages = {250-264},
     abstract = {The visual cortex has the capacity for experience-dependent change, or cortical plasticity, that is retained throughout life. Plasticity is invoked for encoding information during perceptual learning, by internally representing the regularities of the visual environment, which is useful for facilitating intermediate-level vision--contour integration and surface segmentation. The same mechanisms have adaptive value for functional recovery after CNS damage, such as that associated with stroke or neurodegenerative disease. A common feature to plasticity in primary visual cortex (V1) is an association field that links contour elements across the visual field. The circuitry underlying the association field includes a plexus of long-range horizontal connections formed by cortical pyramidal cells. These connections undergo rapid and exuberant sprouting and pruning in response to removal of sensory input, which can account for the topographic reorganization following retinal lesions. Similar alterations in cortical circuitry may be involved in perceptual learning, and the changes observed in V1 may be representative of how learned information is encoded throughout the cerebral cortex.},
}
@article{GilbertetalPAMI-2010,
     author = {A. Gilbert and J. Illingworth and R. Bowden},
      title = {Action Recognition using Mined Hierarchical Compound Features},
    journal = {{IEEE} Transactions on Pattern Analysis and Machine Learning},
       year = 2010,
}
@article{GiljaetalNATURE-12,
     title = {A high-performance neural prosthesis enabled by control algorithm design},
    author = {Gilja, V. and Nuyujukian, P. and Chestek, C.A. and Cunningham, J.P. and Yu, B.M. and Fan, J.M. and Churchland, M.M. and Kaufman, M.T. and Kao, J.C. and Ryu, S.I. and Shenoy, K.V.},
   journal = {Nature Neuroscience},
    volume = 15,
     issue = 12,
      year = 2012,
     pages = {1752-1757},
  abstract = {Neural prostheses translate neural activity from the brain into control signals for guiding prosthetic devices, such as computer cursors and robotic limbs, and thus offer individuals with disabilities greater interaction with the world. However, relatively low performance remains a critical barrier to successful clinical translation; current neural prostheses are considerably slower, with less accurate control, than the native arm. Here we present a new control algorithm, the recalibrated feedback intention-trained Kalman filter (ReFIT-KF) that incorporates assumptions about the nature of closed-loop neural prosthetic control. When tested in rhesus monkeys implanted with motor cortical electrode arrays, the ReFIT-KF algorithm outperformed existing neural prosthetic algorithms in all measured domains and halved target acquisition time. This control algorithm permits sustained, uninterrupted use for hours and generalizes to more challenging tasks without retraining. Using this algorithm, we demonstrate repeatable high performance for years after implantation in two monkeys, thereby increasing the clinical viability of neural prostheses.}
}
@article{GillIC-61,
      author = {Gill, Arthur},
       title = {State-Identification Experiments in Finite Automata},
     journal = {Information and Computation},
      volume = 4,
       pages = {132-154},
        year = 1961,
}
@inproceedings{GilmeretalJMLR-17,
       author = {Justin Gilmer and Samuel S. Schoenholz and Patrick F. Riley and Oriol Vinyals and George E. Dahl},
        title = {Neural Message Passing for Quantum Chemistry},
      journal = {In Proceedings of the 34th International Conference on Machine Learning},
       volume = 70,
         year = 2017,
        pages = {1263-1272},
      comment = {arXiv:1704.01212},
     abstract = {Supervised learning on molecules has incredible potential to be useful in chemistry, drug discovery, and materials science. Luckily, several promising and closely related neural network models invariant to molecular symmetries have already been described in the literature. These models learn a message passing algorithm and aggregation procedure to compute a function of their entire input graph. At this point, the next step is to find a particularly effective variant of this general approach and apply it to chemical prediction benchmarks until we either solve them or reach the limits of the approach. In this paper, we reformulate existing models into a single common framework we call Message Passing Neural Networks (MPNNs) and explore additional novel variations within this framework. Using MPNNs we demonstrate state of the art results on an important molecular property prediction benchmark; these results are strong enough that we believe future work should focus on datasets with larger molecules or more accurate ground truth labels.},
}
@article{GimelfarbPAMI-96,
   author = {Gimelfarb, G.L.},
    title = {Texture modeling by multiple pairwise pixel interactions},
  journal = PAMI,
     year = 1996,
   volume = 18,
    issue = 11,
    pages = {1110-1114},
}
@book{Ginsberg93,
   author =    {Ginsberg, Matthew L.},
   title =     {Essentials of Artificial Intelligence},
   publisher = {Morgan Kaufmann},
   address =   {San Francisco, California},
   year =      1993,
}
@inproceedings{GinsbergAAAI-84,
   author =       {Ginsberg, Matthew L.},
   title =        {Non-Monotonic Reasoning Using {D}empster's Rule},
   booktitle =    {Proceedings of AAAI-84},
   location =     {Austin, Texas},
   organization = {AAAI},
   year =         1984,
}
@inproceedings{GinsbergAAAI-91,
   author =       {Ginsberg, Matthew L.},
   title =        {What is a modal truth criterion?},
   booktitle =    {Proceedings AAAI-91},
   location =     {Anaheim, California},
   organization = {AAAI},
   year =         1991,
}
@article{GinsbergAIM-89,
   author =  {Ginsberg, Matthew L.},
   title =   {Universal planning: An (almost) universally bad idea},
   journal = {AI Magazine},
   year =    1989,
   volume =  10,
    issue =  4,
   pages =   {40-44},
}
@inproceedings{GinsbergIJCAI-85,
   author =       {Ginsberg, M. L.},
   title =        {Does Probability Have a Place in Non-monotonic Reasoning?},
   booktitle =    {Proceedings {IJCAI} 9},
   location =     {Los Angeles, California},
   organization = {IJCAII},
   year =         1985,
}
@incollection{GinsbergPW-86,
   author =   {Ginsberg, M. L.},
   title =    {Possible Worlds Planning},
   crossref = {GeorgeffandLanskyPW-86},
   pages =    {213-243},
}      
@incollection{GinsbergUAI-85,
   author =   {Ginsberg, Matthew L.},
   title =    {Interpreting Probabilistic Reasoning},
   crossref = UAI-88,
   year =     1985,
}
@inproceedings{GinsbergandMcAllesterKR-94,
       author = {Ginsberg, Matthew L. and McAllester, David},
        title = {{GSAT} and Dynamic Backtracking},
    booktitle = {Proceedings of the 1994 International Conference on Principles of Knowledge Representation and Reasoning},
     location = {Bonn, Germany},
         year = 1994,
}
@article{GinsbergandSmithAIJ-88,
       author = {Ginsberg, Matthew L. and Smith, David E.},
        title = {Reasoning About Action {I}: a Possible Worlds Approach},
      journal = AIJ,
       volume = 35,
         year = 1988,
        pages = {165-195},
}
@techreport{GinsbergandSmithTR-86,
       author = {Ginsberg, Matthew L. and Smith, David E.},
        title = {Reasoning About Action {I}: A Possible Worlds Approach},
 type         = {Report No.},
        issue = {KSL-86-37},
  institution = {Stanford Knowledge Systems Laboratory},
         year = 1986,
}
@inproceedings{GionisandMannilaRECOMB-03,
       author = {Aristides Gionis and Heikki Mannila},
        title = {Finding recurrent sources in sequences},
    booktitle = {Proceedings of the seventh annual international conference on Research in computational molecular biology (RECOMB)},
        pages = {123-130},
     location = {Berlin, Germany},
    publisher = {ACM Press},
      address = {New York, NY},
         year = 2003,
}
@inproceedings{GionisetalVLDB-99,
       author = {Aristides Gionis and Piotr Indyk and Rajeev Motwani},
       editor = {Malcolm P. Atkinson and Maria E. Orlowska and Patrick Valduriez and Stanley B. Zdonik and Michael L. Brodie},
        title = {Similarity Search in High Dimensions via Hashing},
    booktitle = {Proceedings of the 25th International Conference on Very Large Data Bases},
    publisher = {Morgan Kaufmann},
        pages = {518-529},
         year = 1999,
}
@article{GirachetalCELL-13,
       author = {Girach, Fatima and Craig, Tim J. and Rocca, Daniel L. and Henley, Jeremy M.},
        title = {{RIM1}$\alpha$ {SUMO}ylation is Required for Fast Synaptic Vesicle Exocytosis},
      journal = {Cell},
 date         = {November, 27},
         year = 2013,
     abstract = {The rapid, activity-dependent quantal presynaptic release of neurotransmitter is vital for brain function. The complex process of vesicle priming, fusion, and retrieval is very precisely controlled and requires the spatiotemporal coordination of multiple protein-protein interactions. Here, we show that posttranslational modification of the active zone protein Rab3-interacting molecule 1$\alpha$ (RIM1$\alpha$) by the small ubiquitin-like modifier 1 (SUMO-1) functions as a molecular switch to direct these interactions and is essential for fast synaptic vesicle exocytosis. RIM1$\alpha$ SUMOylation at lysine residue K502 facilitates the clustering of Ca$_{2.1}$ calcium channels and enhances the Ca$^{2+}$ influx necessary for vesicular release, whereas non-SUMOylated RIM1\hmplusminus{} participates in the docking/priming of synaptic vesicles and maintenance of active zone structure. These results demonstrate that SUMOylation of RIM1$\alpha$\hmplusminus{} is a key determinant of rapid, synchronous neurotransmitter release, and the SUMO-mediated switching of RIM1$\alpha$ between binding proteins provides insight into the mechanisms underpinning synaptic function and dysfunction.},
}
@article{GiretetalPNAS-14,
       author = {Giret, Nicolas and Kornfeld, Joergen and Ganguli, Surya and Hahnloser, Richard H. R.},
        title = {Evidence for a causal inverse model in an avian cortico-basal ganglia circuit},
      journal = {Proceedings of the National Academy of Sciences USA},
         year = 2014,
    publisher = {National Academy of Sciences},
       volume = 111,
        issue = 16,
        pages = {6063-6068},
     abstract = {Auditory neural responses mirror motor activity in a songbird cortical area. The average temporal offset of mirrored responses is roughly equal to short sensorimotor loop delays. This correspondence between mirroring offsets and loop delays constitutes evidence for a causal inverse model. Causal inverse models can map a desired sensation into the required action.},
}
@article{GiustietalCoRR-16,
        title = {Two's company, three (or more) is a simplex: Algebraic-topological tools for understanding higher-order structure in neural data},
       author = {Chad Giusti and Robert Ghrist and Danielle S. Bassett},
      journal = {CoRR},
       volume = {arXiv:1601.01704},
         year = 2016,
     abstract = {The language of graph theory, or network science, has proven to be an exceptional tool for addressing myriad problems in neuroscience. Yet, the use of networks is predicated on a critical simplifying assumption: that the quintessential unit of interest in a brain is a dyad -- two nodes (neurons or brain regions) connected by an edge. While rarely mentioned, this fundamental assumption inherently limits the types of neural structure and function that graphs can be used to model. Here, we describe a generalization of graphs that overcomes these limitations, thereby offering a broad range of new possibilities in terms of modeling and measuring neural phenomena. Specifically, we explore the use of \emph{simplicial complexes}, a theoretical notion developed in the field of mathematics known as algebraic topology, which is now becoming applicable to real data due to a rapidly growing computational toolset. We review the underlying mathematical formalism as well as the budding literature applying simplicial complexes to neural data, from electrophysiological recordings in animal models to hemodynamic fluctuations in humans. Based on the exceptional flexibility of the tools and recent ground-breaking insights into neural function, we posit that this framework has the potential to eclipse graph theory in unraveling the fundamental mysteries of cognition.},
}
@article{GiustietalPNAS-15,
       author = {Giusti, Chad and Pastalkova, Eva and Curto, Carina and Itskov, Vladimir},
        title = {Clique topology reveals intrinsic geometric structure in neural correlations},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {112},
       number = {44},
         year = {2015},
        pages = {13455-13460},
     abstract = {Detecting meaningful structure in neural activity and connectivity data is challenging in the presence of hidden nonlinearities, where traditional eigenvalue-based methods may be misleading. We introduce a novel approach to matrix analysis, called clique topology, that extracts features of the data invariant under nonlinear monotone transformations. These features can be used to detect both random and geometric structure, and depend only on the relative ordering of matrix entries. We then analyzed the activity of pyramidal neurons in rat hippocampus, recorded while the animal was exploring a 2D environment, and confirmed that our method is able to detect geometric organization using only the intrinsic pattern of neural correlations. Remarkably, we found similar results during nonspatial behaviors such as wheel running and rapid eye movement (REM) sleep. This suggests that the geometric structure of correlations is shaped by the underlying hippocampal circuits and is not merely a consequence of position coding. We propose that clique topology is a powerful new tool for matrix analysis in biological settings, where the relationship of observed quantities to more meaningful variables is often nonlinear and unknown.},
}
@inproceedings{GivanandDeanIJCAI-97,
      author = {Givan, Robert and Dean, Thomas},
       title = {Model Minimization, Regression, and Propositional {STRIPS} Planning},
   booktitle = {Proceedings {IJCAI}-97},
   publisher = {Morgan Kaufmann Publishers},
     address = {San Francisco, California},
    location = {Nagoya, Japan},
organization = {IJCAI},
       pages = {1163-1168},
        year = 1997,
}
@article{GivanetalAIJ-00,
   author = {Givan, Robert and Leach, Sonia and Dean, Thomas},
    title = {Bounded Parameter {M}arkov Decision Processes},            
   volume = 122,
    issue = {1-2},
  journal = AIJ,
    pages = {71-109},
     year = 2000,
}
@article{GivanetalAIJ-03,
    title = {Equivalence Notions and Model Minimization 
             in {M}arkov Decision Processes},
   author = {Robert Givan and Thomas Dean and Matthew Greig},
  journal = AIJ,
   volume = 147,
    issue = {1-2},
    pages = {163-223},
     year = 2003,
}
@inproceedings{GivanetalECP-97,
      author = {Givan, Robert and Leach, Sonia and Dean, Thomas},
       title = {Bounded Parameter {M}arkov Decision Processes},
   booktitle = {Proceedings of the 4th European Conference on Planning},
       pages = {234-246},
        year = 1997,
}
@article{GlasseretalNATURE-16,
       author = {Glasser, Matthew F. and Coalson, Timothy S. and Robinson, Emma C. and Hacker, Carl D. and Harwell, John and Yacoub, Essa and Ugurbil, Kamil and Andersson, Jesper and Beckmann, Christian F. and Jenkinson, Mark and Smith, Stephen M. and Van Essen, David C.},
        title = {A multi-modal parcellation of human cerebral cortex},
      journal = {Nature},
    publisher = {Macmillan Publishers Limited},
         year = {2016},
       volume = {advance online publication},
     abstract = {Understanding the amazingly complex human cerebral cortex requires a map (or parcellation) of its major subdivisions, known as cortical areas. Making an accurate areal map has been a century-old objective in neuroscience. Using multi-modal magnetic resonance images from the Human Connectome Project (HCP) and an objective semi-automated neuroanatomical approach, we delineated 180 areas per hemisphere bounded by sharp changes in cortical architecture, function, connectivity, and/or topography in a precisely aligned group average of 210 healthy young adults. We characterized 97 new areas and 83 areas previously reported using post-mortem microscopy or other specialized study-specific approaches. To enable automated delineation and identification of these areas in new HCP subjects and in future studies, we trained a machine-learning classifier to recognize the multi-modal 'fingerprint' of each cortical area. This classifier detected the presence of 96.6\% of the cortical areas in new subjects, replicated the group parcellation, and could correctly locate areas in individuals with atypical parcellations. The freely available parcellation and classifier will enable substantially improved neuroanatomical precision for studies of the structural and functional organization of human cerebral cortex and its variation across individuals and in development, aging, and disease.},
}
@inproceedings{GlattandCostaAAAI-17,
        title = {Policy Reuse in Deep Reinforcement Learning},
       author = {Ruben Glatt and Anna Helena Reali Costa},
    booktitle = {Thirty-First AAAI Conference on Artificial Intelligence},
         year = {2017},
        pages = {4929-4930},
     abstract = {Driven by recent developments in Artificial Intelligence research, a promising new technology for building intelligent agents has evolved. The approach is termed Deep Reinforcement Learning and combines the classic field of Reinforcement Learning (RL) with the representational power of modern Deep Learning approaches. It is very well suited for single task learning but needs a long time to learn any new task. To speed up this process, we propose to extend the concept to multi-task learning by adapting Policy Reuse, a Transfer Learning approach from classic RL, to use with Deep Q-Networks.},
}
@book{Gleick02,
      author = {James Gleick},
       title = {What Just Happened: A Chronicle from the Information Frontier},
   publisher = {Pantheon Books},
     address = {New York, NY},
        year = 2002,
}
@book{Gleick99,
      author = {James Gleick},
       title = {Faster: The Acceleration of Just About Everything},
   publisher = {Pantheon Books},
     address = {New York, NY},
        year = 1999,
}
@article{GlicksteinCURRENT-BIOLOGY-06,
       author = {Glickstein, Mitch},
        title = {{Golgi and Cajal}: The neuron doctrine and the 100th anniversary of the 1906 {Nobel Prize}},
      journal = {Current Biology},
    publisher = {Elsevier},
       volume = {16},
        issue = {5},
         year = {2006},
        pages = {R147-R151}
}
@article{GloverPMC-11,
       author = {Glover, Gary H.},
        title = {Overview of functional magnetic resonance imaging},
      journal = {Neurosurgery clinics of North America},
         year = {2011},
       volume = {22},
        issue = {2},
        pages = {133-144},
     abstract = {Blood Oxygen Level Dependent (BOLD) functional magnetic resonance imaging (fMRI) depicts changes in deoxyhemoglobin concentration consequent to task-induced or spontaneous modulation of neural metabolism. Since its inception in 1990, this method has been widely employed in thousands of studies of cognition for clinical applications such as surgical planning, for monitoring treatment outcomes, and as a biomarker in pharmacologic and training programs. More recently, attention is turning to the use of pattern classification and other statistical methods to draw increasingly complex inferences about cognitive brain states from fMRI data. This article reviews the methods, challenges, and future of fMRI.},
}
@article{GluckandMyersHIPPOCAMPUS-93,
       author = {Gluck, M. A. and Myers, C. E.},
        title = {Hippocampal mediation of stimulus representation: a computational theory},
      journal = {Hippocampus},
         year = {1993},
       volume = {3},
       number = {4},
        pages = {491-516},
     abstract = {The authors propose a computational theory of the hippocampal region's function in mediating stimulus representations. The theory assumes that the hippocampal region develops new stimulus representations that enhance the discriminability of differentially predictive cues while compressing the representation of redundant cues. Other brain regions, including cerebral and cerebellar cortices, are presumed to use these hippocampal representations to recode their own stimulus representations. In the absence of an intact hippocampal region, the theory implies that other brain regions will attempt to learn associations using previously established fixed representations. Instantiated as a connectionist network model, the theory provides a simple and unified interpretation of the functional role of the hippocampal region in a wide range of conditioning paradigms, including stimulus discrimination, reversal learning, stimulus generalization, latent inhibition, sensory preconditioning, and contextual sensitivity. The theory makes novel predictions regarding the effects of hippocampal lesions on easy-hard transfer and compound preexposure. Several prior qualitative characterizations of hippocampal function--including stimulus selection, chunking, cue configuration, and contextual coding--are identified as task-specific special cases derivable from this more general theory. The theory suggests that a profitable direction for future empirical and theoretical research will be the study of learning tasks in which both intact and lesioned animals exhibit similar initial learning behaviors but differ on subsequent transfer and generalization tasks.}
}
@misc{Gnutella,
      author = {Gnutella},
       title = {Gnutella, a Peer-to-Peer Search System},
howpublished = {\url{http://gnutella.wego.com/}},
        year = 2001,
}
@article{GoardandYangNATURE-09,
    author = {Goard, Michael and Dan, Yang},
     title = {Basal forebrain activation enhances cortical coding of natural scenes},
   journal = {Nature Neuroscience},
    volume = 12,
     issue = 11,
      year = 2009,
     pages = {1444-1449},
  abstract = {The nucleus basalis of the basal forebrain is an essential component of the neuromodulatory system controlling the behavioral state of an animal and it is thought to be important in regulating arousal and attention. However, the effect of nucleus basalis activation on sensory processing remains poorly understood. Using polytrode recording in rat visual cortex, we found that nucleus basalis stimulation caused prominent decorrelation between neurons and marked improvement in the reliability of neuronal responses to natural scenes. The decorrelation depended on local activation of cortical muscarinic acetylcholine receptors, whereas the increased reliability involved distributed neural circuits, as evidenced by nucleus basalis-induced changes in thalamic responses. Further analysis showed that the decorrelation and increased reliability improved cortical representation of natural stimuli in a complementary manner. Thus, the basal forebrain neuromodulatory circuit, which is known to be activated during aroused and attentive states, acts through both local and distributed mechanisms to improve sensory coding.}
}
@incollection{GodingetalBIOINTERFACES-15,
       author = {Goding, Josef and Green, Rylie and Martens, Penny and Poole-Warren, Laura},
        title = {CHAPTER 8 Bioactive Conducting Polymers for Optimising the Neural Interface},
    booktitle = {Biointerfaces: Where Material Meets Biology},
         year = 2015,
        pages = {192-220},
    publisher = {The Royal Society of Chemistry},
     abstract = {Increasing demands on neuroprosthetic bioelectrodes have created a need for electrode materials that can support the safe and sustainable delivery of electrical stimulation to excitable tissues. Conducting polymers have become a focal point of research into next-generation electrode materials due to their superior electrical properties for charge transfer in biological environments. Perhaps the greatest potential of conducting polymers within neural applications is their ability to accommodate biofunctionality through the incorporation and controlled release of bioactive molecules. Incorporation of neurotrophic factors, cell adhesion molecules and various drug compounds within conducting polymers has been shown to assist in the development of higher quality tissue-electrode interfaces. However, limitations associated with drug loading and the impact of adding biological molecules on mechanical and electrical properties of biofunctionalised conducting polymers restrict their use in medical device applications. This chapter assesses the role of conducting polymers as bioelectrode materials, means of biofunctionalisation and the resultant challenges and limitations. Furthermore, this chapter evaluates the strategies developed to overcome these limitations in the pursuit of the development of high quality neural interfaces.}
}
@article{GoenseetalFiCN-16,
       author = {Glense, Jozien and Bohraus, Yvette and Logothetis, Nikos K.},
        title = {{fMRI} at High Spatial Resolution: Implications for {BOLD}-Models},
      journal = {Frontiers in computational neuroscience},
         year = {2016},
    publisher = {Frontiers Media S.A.},
       volume = {10},
        pages = {66-66},
     abstract = {As high-resolution functional magnetic resonance imaging (fMRI) and fMRI of cortical layers become more widely used, the question how well high-resolution fMRI signals reflect the underlying neural processing, and how to interpret laminar fMRI data becomes more and more relevant. High-resolution fMRI has shown laminar differences in cerebral blood flow (CBF), volume (CBV), and neurovascular coupling. Features and processes that were previously lumped into a single voxel become spatially distinct at high resolution. These features can be vascular compartments such as veins, arteries, and capillaries, or cortical layers and columns, which can have differences in metabolism. Mesoscopic models of the blood oxygenation level dependent (BOLD) response therefore need to be expanded, for instance, to incorporate laminar differences in the coupling between neural activity, metabolism and the hemodynamic response. Here we discuss biological and methodological factors that affect the modeling and interpretation of high-resolution fMRI data. We also illustrate with examples from neuropharmacology and the negative BOLD response how combining BOLD with CBF- and CBV-based fMRI methods can provide additional information about neurovascular coupling, and can aid modeling and interpretation of high-resolution fMRI.},
}
@article{GogtayetalPNAS-04,
    author = {Gogtay, Nitin and Giedd, Jay N. and Lusk, Leslie and Hayashi,Kiralee M. and Greenstein, Deanna and Vaituzis, A. Catherine and Nugent, Tom F. and Herman, David H. and Clasen, Liv S. and Toga, Arthur W. and Rapoport, Judith L. and Thompson, Paul M.}, 
     title = {Dynamic mapping of human cortical development during childhood through early adulthood}, 
   journal = {Proceedings of the National Academy of Sciences of the United States of America}, 
    volume = 101, 
     issue = 21, 
     pages = {8174-8179}, 
      year = 2004, 
  abstract = {We report the dynamic anatomical sequence of human cortical gray matter development between the age of 4-21 years using quantitative four-dimensional maps and time-lapse sequences. Thirteen healthy children for whom anatomic brain MRI scans were obtained every 2 years, for 8-10 years, were studied. By using models of the cortical surface and sulcal landmarks and a statistical model for gray matter density, human cortical development could be visualized across the age range in a spatiotemporally detailed time-lapse sequence. The resulting time-lapse "movies" reveal that (i) higher-order association cortices mature only after lower-order somatosensory and visual cortices, the functions of which they integrate, are developed, and (ii) phylogenetically older brain areas mature earlier than newer ones.},
}
@article{GokmenandVlasovCoRR-16,
       author = {Tayfun Gokmen and Yurii Vlasov},
        title = {Acceleration of Deep Neural Network Training with Resistive Cross-Point Devices},
      journal = {CoRR},
       volume = {arXiv:1603.07341},
         year = {2016},
     abstract = {In recent years, deep neural networks (DNN) have demonstrated significant business impact in large scale analysis and classification tasks such as speech recognition, visual object detection, pattern extraction, etc. Training of large DNNs, however, is universally considered as time consuming and computationally intensive task that demands datacenter-scale computational resources recruited for many days. Here we propose a concept of resistive processing unit (RPU) devices that can potentially accelerate DNN training by orders of magnitude while using much less power. The proposed RPU device can store and update the weight values locally thus minimizing data movement during training and allowing to fully exploit the locality and the parallelism of the training algorithm. We identify the RPU device and system specifications for implementation of an accelerator chip for DNN training in a realistic CMOS-compatible technology. For large DNNs with about 1 billion weights this massively parallel RPU architecture can achieve acceleration factors of 30,000X compared to state-of-the-art microprocessors while providing power efficiency of 84,000 GigaOps/s/W. Problems that currently require days of training on a datacenter-size cluster with thousands of machines can be addressed within hours on a single RPU accelerator. A system consisted of a cluster of RPU accelerators will be able to tackle Big Data problems with trillions of parameters that is impossible to address today like, for example, natural speech recognition and translation between all world languages, real-time analytics on large streams of business and scientific data, integration and analysis of multimodal sensory data flows from massive number of IoT (Internet of Things) sensors.}
}
@article{Gold72,
       author = {Gold, E. Mark},
        title = {System Identification via State Characterization},
      journal = {Automatica},
       volume = 8,
        pages = {621-636},
         year = 1972,
}
@article{GoldIC-78,
       author = {Gold, E. Mark},
        title = {Complexity of Automaton Identification from Given Sets},
      journal = {Information and Control},
       volume = 37,
        pages = {302-320},
         year = 1978,
}
@book{Goldberg89,
        title = {Genetic Algorithms in Search, Optimization, and Machine Learning},
       author = {David E. Goldberg},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1989,
}
@inproceedings{GoldbergeretalNIPS-05,
    author = {Goldberger, Jacob and Roweis, Sam and Hinton, Geoff and Salakhutdinov, Ruslan},
     title = {Neighborhood Component Analysis},
 booktitle = {Advances in Neural Information Processing Systems 17},
    editor = {Lawrence K. Saul and Yair Weiss and {L\'{e}on} Bottou},
 publisher = {MIT Press},
   address = {Cambridge, MA},
      year = 2005
}
@article{GoldetalJCN-07,
     title = {Using extracellular action potential recordings to constrain compartmental models},
    author = {Gold, C. and Henze, D.A. and Koch, C.},
   journal = {Journal Compututational Neuroscience},
    volume = 23,
     issue = 1,
     pages = {39-58},
      year = 2007,
  abstract = {We investigate the use of extracellular action potential (EAP) recordings for biophysically faithful compartmental models. We ask whether constraining a model to fit the EAP is superior to matching the intracellular action potential (IAP). In agreement with previous studies, we find that the IAP method under-constrains the parameters. As a result, significantly different sets of parameters can have virtually identical IAP's. In contrast, the EAP method results in a much tighter constraint. We find that the distinguishing characteristics of the waveform--but not its amplitude-resulting from the distribution of active conductances are fairly invariant to changes of electrode position and detailed cellular morphology. Based on these results, we conclude that EAP recordings are an excellent source of data for the purpose of constraining compartmental models.}
}
@article{Goldman-RakicARN-88,
       author = {Goldman-Rakic, Patricia S.},
        title = {Topography of Cognition: Parallel Distributed Networks in Primate Association Cortex},
      journal = {Annual Review of Neuroscience},
       volume = {11},
       number = {1},
        pages = {137-156},
         year = {1988},
}

@article{Goldman-RakicCHILD-DEVELOPMENT-87,
        title = {Development of Cortical Circuitry and Cognitive Function},
       author = {Patricia S. Goldman-Rakic},
      journal = {Child Development},
       number = {3},
    publisher = {[Wiley, Society for Research in Child Development]},
       volume = {58},
        pages = {601-622},
         year = {1987},
     abstract = {Recent functional and anatomical studies in nonhuman primates have elucidated the basic neural circuitry underlying delayed-response function in adult nonhuman primates. Thus circuitry includes connections of the principal sulcus with other areas of parietal association and limbic cortex and projections to the caudate nucleus, superior colliculus, and other premotor centers. Anatomical tracing in primate fetuses and in monkeys at various stages of postnatal development indicates that these various classes of cortical connections begin to form by the second trimester of pregnancy. Electromicroscopic studies of the principal sulcus and other areas of cerebral cortex show that the number and density of synapses in the cortex increase rapidly, reaching and maintaining higher than normal adult values between 2 and 4 months postnatally, before slowly declining over a period of years to stable adult levels. The capacity to perform delayed-response and/or AB̄ at short delays emerges around 4 months of age, coinciding with the end of the period of highest synaptic density in the principal sulcus. These findings suggest that a critical mass of cortical synapses is important for the emergence of this cognitive function, and that fully mature capacity may depend upon the elimination of excess synapses that occurs during adolescence and young adulthood. Knowledge of the neural basis of normal cognitive development may prove useful both to social and educational purposes as well as to understanding developmental disorders of cognition.},
}

@phdthesis{GoldmanPhD,
       author = {Goldman, Robert},
        title = {A Probabilistic Approach to Language Understanding},
       school = {Brown University},
         year = 1990,
      address = {Providence, RI},
       commen = {Available as Technical Report CS-TR-90-34},
}
@inproceedings{GoldmanandRivest86,
       author = {Goldman, Sally A. and Ronald Rivest},
        title = {Making Maximum Entropy Computations Easier By Adding Extra Constraints},
    booktitle = {Proceedings of the Sixth Annual Workshop on Maximum Entropy and {B}ayesian Methods in Applied Statistics},
         year = 1986,
}
@article{GoldmanetalCEREBRAL-CORTEX-03,
       author = {Goldman, Mark S. and Levine, Joseph H. and Major, Guy and Tank, David W. and Seung, H.S.},
        title = {Robust persistent neural activity in a model integrator with multiple hysteretic dendrites per neuron},
      journal = {Cerebral Cortex},
       volume = {13},
        issue = {11},
         year = {2003},
        pages = {1185-1195},
     abstract = {Short-term memory is often correlated with persistent changes in neuronal firing rates in response to transient inputs. We model the persistent maintenance of an analog eye position signal by an oculomotor neural integrator receiving transient eye movement commands. Previous models of this network rely on precisely tuned positive feedback with <1\% tolerance to mistuning, or use neurons that exhibit large discontinuities in firing rate with small changes in eye position. We show analytically how using neurons with multiple bistable dendritic compartments can enhance the robustness of eye fixations to mistuning while reproducing the approximately linear and continuous relationship between neuronal firing rates and eye position, and the dependence of neuron pair firing rate relationships on the direction of the previous saccade. The response of the model to continuously varying inputs makes testable predictions for the performance of the vestibuloocular reflex. Our results suggest that dendritic bistability could stabilize the persistent neural activity observed in working memory systems.},
}
@incollection{GoldmanetalNEN-07,
       author = {Mark Goldman and Albert Compte and Xiao-Jing Wang},
        title = {Neural integrators: recurrent mechanisms and models},
    booktitle = {New Encyclopedia of Neuroscience},
       editor = {Binder, Marc D. and Hirokawa, Nobutaka and Windhorst, Uwe},
         year = {2007},
    publisher = {Springer-Verlag New York} ,
          url = {http://wanglab.med.yale.edu/pdf_pub/ens.integrator.submitted.revised.pdf},
}
@inproceedings{GoldsteinIJCAI-75,
       author = {Goldstein, Ira P.},
        title = {NUDGE: A Knowledge-Based Scheduling Program},
    booktitle = {Proceedings {IJCAI} 4},
     location = {Tbilisi, Georgia, USSR},
 organization = {IJCAII},
         year = 1975,
}
@inproceedings{GoldszmidtandDarwicheSSS-94,
       author = {Goldszmidt, Moises and Darwiche, Adnan},
        title = {Action Networks: A Framework for Reasoning about Actions and Change Uncertainty},
    booktitle = {Proceedings of the {AAAI} Spring Symposium on Decision Theoretic Planning},
     location = {Stanford, California},
         year = 1994,
}
@book{Goleman1995,
       author = {Daniel Goleman},
        title = {Emotional Intelligence: Why It Can Matter More Than {IQ}},
    publisher = {Bantam Books},
      address = {New York, NY},
         year = 1995,
      comment = {http://en.wikipedia.org/wiki/Emotional_intelligence},
}
@book{Goleman2006,
     author = {Daniel Goleman},
      title = {Social Intelligence: The New Science of Human Relationships},
  publisher = {Bantam Books},
    address = {New York, NY},
       year = 2006,
    comment = {http://danielgoleman.info/topics/social-intelligence/},
}
@article{GomesetalCELL-13,
    author = {Gomes, Ana P. and Price, Nathan L. and Ling, Alvin J.Y. and Moslehi, Javid J. and Montgomery, Magdalene K. and u9Rajman, Luis and White, James P. and Teodoro, João S. and Wrann, Christiane D. and Hubbard, Basil P. and Mercken, Evi M. and Palmeira, Carlos M. and de Cabo, Rafael and Rolo, Anabela P. and Turner, Nigel and Bell, Eric L. and Sinclair, David A.},
     title = {Declining NAD+ Induces a Pseudohypoxic State Disrupting Nuclear-Mitochondrial Communication during Aging},
   journal = {Cell},
 publisher = {Cell Press},
    volume = {155},
     issue = {7},
      year = {2013},
     pages = {1624-1638},
  abstract = {Ever since eukaryotes subsumed the bacterial ancestor of mitochondria, the nuclear and mitochondrial genomes have had to closely coordinate their activities, as each encode different subunits of the oxidative phosphorylation (OXPHOS) system. Mitochondrial dysfunction is a hallmark of aging, but its causes are debated. We show that, during aging, there is a specific loss of mitochondrial, but not nuclear, encoded OXPHOS subunits. We trace the cause to an alternate PGC-1$\alpha/\beta$-independent pathway of nuclear-mitochondrial communication that is induced by a decline in nuclear NAD+ and the accumulation of HIF-1\hmplusminus{} under normoxic conditions, with parallels to Warburg reprogramming. Deleting SIRT1 accelerates this process, whereas raising NAD+ levels in old mice restores mitochondrial function to that of a young mouse in a SIRT1-dependent manner. Thus, a pseudohypoxic state that disrupts PGC-1$\alpha/\beta$-independent nuclear-mitochondrial communication contributes to the decline in mitochondrial function with age, a process that is apparently reversible. "A specific decline in mitochondrially encoded genes occurs during aging in muscle"Nuclear NAD+ levels regulate mitochondrial homeostasis independently of PGC-1$\alpha/\beta$"Declining NAD+ during aging causes pseudohypoxia, which disrupts OXPHOS function"Raising nuclear NAD+ in old mice reverses pseudohypoxia and metabolic dysfunction Aging disrupts an NAD+-dependent nuclear-mitochondrial communication pathway, causing a decline in the mitochondrially encoded oxidative phosphorylation components relative to nuclear-encoded components. Raising NAD+ levels in old mice alleviates the pseudohypoxic conditions that disrupt the communication pathway, reversing this mitochondrial defect.},
}
@article{Gomez-RoblesetalPNAS-15,
       author = {G{\'o}mez-Robles, Aida and Hopkins, William D. and Schapiro, Steven J. and Sherwood, Chet C.},
        title = {Relaxed genetic control of cortical organization in human brains compared with chimpanzees},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {112},
       number = {48},
         year = {2015},
        pages = {14799-14804},
     abstract = {Despite decades of research, we still have a very incomplete understanding of what is special about the human brain compared with the brains of our closest fossil and living relatives. Parsing the genetic versus environmental factors that govern the structure of the cerebral cortex in humans and chimpanzees may shed light on the evolution of behavioral flexibility in the human lineage. We show that the morphology of the human cerebral cortex is substantially less genetically heritable than in chimpanzees and therefore is more responsive to molding by environmental influences. This anatomical property of increased plasticity, which is likely related to the human pattern of development, may underlie our species{\textquoteright} capacity for cultural evolution.The study of hominin brain evolution has focused largely on the neocortical expansion and reorganization undergone by humans as inferred from the endocranial fossil record. Comparisons of modern human brains with those of chimpanzees provide an additional line of evidence to define key neural traits that have emerged in human evolution and that underlie our unique behavioral specializations. In an attempt to identify fundamental developmental differences, we have estimated the genetic bases of brain size and cortical organization in chimpanzees and humans by studying phenotypic similarities between individuals with known kinship relationships. We show that, although heritability for brain size and cortical organization is high in chimpanzees, cerebral cortical anatomy is substantially less genetically heritable than brain size in humans, indicating greater plasticity and increased environmental influence on neurodevelopment in our species. This relaxed genetic control on cortical organization is especially marked in association areas and likely is related to underlying microstructural changes in neural circuitry. A major result of increased plasticity is that the development of neural circuits that underlie behavior is shaped by the environmental, social, and cultural context more intensively in humans than in other primate species, thus providing an anatomical basis for behavioral and cognitive evolution.},
}
@misc{GonfalonieriUNDERSTANDING-INVERSE-REINFORCEMENT-LEARNING-18,
        title = {Inverse Reinforcement Learning},
       author = {Alexandre Gonfalonieri},
 howpublished = {https://towardsdatascience.com/inverse-reinforcement-learning-6453b7cdc90d},
         year = {2018},
}
@article{GongetalNATURE-16,
       author = {Gong, Hui and Xu, Dongli and Yuan, Jing and Li, Xiangning and Guo, Congdi and Peng, Jie and Li, Yuxin and Schwarz, Lindsay A. and Li, Anan and Hu, Bihe and Xiong, Benyi and Sun, Qingtao and Zhang, Yalun and Liu, Jiepeng and Zhong, Qiuyuan and Xu, Tonghui and Zeng, Shaoqun and Luo, Qingming},
        title = {High-throughput dual-colour precision imaging for brain-wide connectome with cytoarchitectonic landmarks at the cellular level},
      journal = {Nature Communications},
         year = {2016},
    publisher = {Nature Publishing Group},
       volume = {7},
        pages = {12142},
     abstract = {The precise annotation and accurate identification of neural structures are prerequisites for studying mammalian brain function. The orientation of neurons and neural circuits is usually determined by mapping brain images to coarse axial-sampling planar reference atlases. However, individual differences at the cellular level likely lead to position errors and an inability to orient neural projections at single-cell resolution. Here, we present a high-throughput precision imaging method that can acquire a co-localized brain-wide data set of both fluorescent-labelled neurons and counterstained cell bodies at a voxel size of 0.32 x 0.32 x 2.0 \micron{} in 3 days for a single mouse brain. We acquire mouse whole-brain imaging data sets of multiple types of neurons and projections with anatomical annotation at single-neuron resolution. The results show that the simultaneous acquisition of labelled neural structures and cytoarchitecture reference in the same brain greatly facilitates precise tracing of long-range projections and accurate locating of nuclei.},
}
@article{GongetalPOLYMER-15,
        title = {Anisotropic electrical conductivity of polymer composites with aligned carbon nanotubes},
      journal = {Polymer},
       volume = 56,
         year = 2015,
        pages = {498-506},
       author = {S. Gong and Z.H. Zhu and S.A. Meguid},
     abstract = {This paper studies the effect of carbon nanotube (CNT) alignment on anisotropy of electrical conductivity of \{CNT\} enhanced polymer composites. A new \{CNT\} percolation network model is developed by considering structural distortion of \{CNT\} walls at crossed junctions. The structural distortion is modeled by a pseudo-potential energy that consists of the Lennard-Jones potential of Van der Waals interaction between \{CNTs\} and the bending energy of \{CNT\} walls. Simulation results show better agreement with experimental data by the newly development. The current study reveals the anisotropy of electrical property of aligned CNT/polymer composites is mainly affected by the average conductive pathway density of CNTs, which depends heavily on \{CNT\} alignment structure. The study further identifies that the aligned \{CNT\} structure in polymer matrix effectively affects the contributions of each resistance components in the \{CNT\} networks to the total resistance. Finally, parametric studies are conducted to enable systematic exploration of designing multifunctional nanocomposites.}
}
@article {GongetalSCIENCE-15,
       author = {Gong, Yiyang and Huang, Cheng and Li, Jin Zhong and Grewe, Benjamin F. and Zhang, Yanping and Eismann, Stephan and Schnitzer, Mark J.},
        title = {High-speed recording of neural spikes in awake mice and flies with a fluorescent voltage sensor},
      journal = {Science},
         year = {2015},
    publisher = {American Association for the Advancement of Science},
     abstract = {Genetically encoded voltage indicators (GEVIs) are a promising technology for fluorescence readout of millisecond-scale neuronal dynamics. Prior GEVIs had insufficient signaling speed and dynamic range to resolve action potentials in live animals. We coupled fast voltage-sensing domains from a rhodopsin protein to bright fluorophores via resonance energy transfer. The resulting GEVIs are sufficiently bright and fast to report neuronal action potentials and membrane voltage dynamics in awake mice and flies, resolving fast spike trains with 0.2-millisecond timing precision at spike detection error rates orders of magnitude better than prior GEVIs. In vivo imaging revealed sensory-evoked responses, including somatic spiking, dendritic dynamics, and intracellular voltage propagation. These results empower in vivo optical studies of neuronal electrophysiology and coding and motivate further advancements in high-speed microscopy.},
}
@misc{GonzalezSIMULATED-ANNEALING-20,
        title = {Neural Network Pruning with simulated annealing - Medium, April 1, 2020},
       author = {Octavio Gonzalez-Lugo},
 howpublished = {\url{https://tavoglc.medium.com/neural-network-pruning-with-simulated-annealing-811b1a948223}},
         year = {2020},
}
@book{Good1983,
        title = {Good Thinking: The Foundations of Probability and Its Applications},
       author = {Good, I.J.},
         year = {1983},
    publisher = {University of Minnesota Press}
}
@incollection{GoodMI-68,
       author = {Good, I. J.},
        title = {A Five-Year Plan for Automatic Chess},
    booktitle = {Machine Intelligence 2},
    publisher = {Oliver and Boyd},
         year = 1968,
       editor = {Dale, E. and Michie, D.},
       volume = 2,
        pages = {89-118},
      address = {London},
}
@article{GoodaleandMilnerTiNS-92,
       author = {Melvyn A. Goodale and A.David Milner},
        title = {Separate visual pathways for perception and action},
      journal = {Trends in Neurosciences},
       volume = {15},
       number = {1},
        pages = {20-25},
         year = {1992},
     abstract = {Accumulating neuropsychological, electrophysiological and behavioural evidence suggests that the neural substrates of visual perception may be quite distinct from those underlying the visual control of actions. In other words, the set of object descriptions that permit identification and recognition may be computed independently of the set of descriptions that allow an observer to shape the hand appropriately to pick up an object. We propose that the ventral stream of projections from the striate cortex to the inferotemporal cortex plays the major role in the perceptual identification of objects, while the dorsal stream projecting from the striate cortex to the posterior parietal region mediates the required sensorimotor transformations for visually guided actions directed at such objects.}
}
@article{GoodaleandWestwood04,
   author = {M. A. Goodale and D. A. Westwood},
    title = {An evolving view of duplex vision: separate but 
             interacting cortical pathways for perception and action},
  journal = {Current Opinion in Neurobiology},
   volume = 14,
    pages = {203-211},
     year = 2004,
}
@article{GoodfellowGANT-16,
       author = {Ian Goodfellow},
        title = {NIPS 2016 Tutorial: Generative Adversarial Networks},
      journal = {CoRR},
       volume = {arXiv:1701.00160},
         year = {2016},
     abstract = {This report summarizes the tutorial presented by the author at NIPS 2016 on generative adversarial networks (GANs). The tutorial describes: (1) Why generative modeling is a topic worth studying, (2) how generative models work, and how GANs compare to other generative models, (3) the details of how GANs work, (4) research frontiers in GANs, and (5) state-of-the-art image models that combine GANs with other methods. Finally, the tutorial contains three exercises for readers to complete, and the solutions to these exercises.}
}
@book{Goodfellowetal2016deep,
        title = {Deep Learning},
       author = {Goodfellow, I. and Bengio, Y. and Courville, A.},
       series = {Adaptive Computation and Machine Learning series},
         year = {2016},
    publisher = {MIT Press}
}
@article{GoodfellowetalCoRR-13,
       author = {Ian J. Goodfellow and David Warde-Farley and Mehdi Mirza and Aaron Courville and Yoshua Bengio},
        title = {Maxout Networks},
      journal = {CoRR},
       volume = {arXiv:1302.4389},
         year = {2013},
     abstract = {We consider the problem of designing models to leverage a recently introduced approximate model averaging technique called dropout. We define a simple new model called maxout (so named because its output is the max of a set of inputs, and because it is a natural companion to dropout) designed to both facilitate optimization by dropout and improve the accuracy of dropout's fast approximate model averaging technique. We empirically verify that the model successfully accomplishes both of these tasks. We use maxout and dropout to demonstrate state of the art classification performance on four benchmark datasets: MNIST, CIFAR-10, CIFAR-100, and SVHN.},
}
@article{GoodfellowetalCoRR-14,
       author = {Generative Adversarial Networks},
        title = {Ian J. Goodfellow and Jean Pouget-Abadie and Mehdi Mirza and Bing Xu and David Warde-Farley and Sherjil Ozair and Aaron Courville and Yoshua Bengio},
      journal = {CoRR},
       volume = {arXiv:1406.2661},
         year = {2014},
     abstract = {We propose a new framework for estimating generative models via an adversarial process, in which we simultaneously train two models: a generative model G that captures the data distribution, and a discriminative model D that estimates the probability that a sample came from the training data rather than G. The training procedure for G is to maximize the probability of D making a mistake. This framework corresponds to a minimax two-player game. In the space of arbitrary functions G and D, a unique solution exists, with G recovering the training data distribution and D equal to 1/2 everywhere. In the case where G and D are defined by multilayer perceptrons, the entire system can be trained with backpropagation. There is no need for any Markov chains or unrolled approximate inference networks during either training or generation of samples. Experiments demonstrate the potential of the framework through qualitative and quantitative evaluation of the generated samples.}
}
@inproceedings{GoodfellowetalICLR-14,
       author = {Ian J. Goodfellow and  Mehdi Mirza and Xia Da and Aaron C. Courville and Yoshua Bengio},
        title = {An Empirical Investigation of Catastrophic Forgetting in Gradient-Based Neural Networks},
    booktitle = {2nd International Conference on Learning Representations},
      comment = {http://arxiv.org/abs/1312.6211},
         year = {2014},
     abstract = {Catastrophic forgetting is a problem faced by many machine learning models and algorithms. When trained on one task, then trained on a second task, many machine learning models "forget" how to perform the first task. This is widely believed to be a serious problem for neural networks. Here, we investigate the extent to which the catastrophic forgetting problem occurs for modern neural networks, comparing both established and recent gradient-based training algorithms and activation functions. We also examine the effect of the relationship between the first task and the second task on catastrophic forgetting. We find that it is always best to train using the dropout algorithm--the dropout algorithm is consistently best at adapting to the new task, remembering the old task, and has the best tradeoff curve between these two extremes. We find that different tasks and relationships between tasks result in very different rankings of activation function performance. This suggests the choice of activation function should always be cross-validated.},
}
@incollection{GoodfellowetalNIPS-09,
        title = {Measuring Invariances in Deep Networks},
       author = {Ian Goodfellow and Quoc Le and Andrew Saxe and Andrew Ng},
    booktitle = {Advances in Neural Information Processing Systems 22},
       editor = {Y. Bengio and D. Schuurmans and J. Lafferty and C. K. I. Williams and A. Culotta},
    publisher = {MIT Press},
         year = {2009},
        pages = {646-654},
}
@inproceedings{GoodfellowetalNIPS-14,
        title = {Generative Adversarial Nets},
       author = {Ian J. Goodfellow and Jean Pouget-Abadie and Mehdi Mirza and Bing Xu and David Warde-Farley and Sherjil Ozair and Aaron C. Courville and Yoshua Bengio},
    booktitle = {Proceedings of the 27th International Conference on Neural Information Processing Systems 27},
       editor = {Z. Ghahramani and M. Welling and C. Cortes and N. D. Lawrence and K. Q. Weinberger},
     location = {Montreal, Canada},
    publisher = {Curran Associates, Inc.},
         year = {2014},
        pages = {2672-2680},
     abstract = {We propose a new framework for estimating generative models via adversarial nets, in which we simultaneously train two models: a generative model G that captures the data distribution, and a discriminative model D that estimates the probability that a sample came from the training data rather than G. The training procedure for G is to maximize the probability of D making a mistake. This framework corresponds to a minimax two-player game. In the space of arbitrary functions G and D, a unique solution exists, with G recovering the training data distribution and D equal to 1/2 everywhere. In the case where G and D are defined by multilayer perceptrons, the entire system can be trained with backpropagation. There is no need for any Markov chains or unrolled approximate inference networks during either training or generation of samples. Experiments demonstrate the potential of the framework through qualitative and quantitatively evaluation of the generated samples.},
}
@article{GoodmanandFrankTiCS-16,
       author = {Goodman, N. D.  and Frank, M. C. },
        title = {{P}ragmatic {L}anguage {I}nterpretation as {P}robabilistic {I}nference},
      journal = {Trends in Cognitive Science},
         year = {2016},
       volume = {20},
       number = {11},
        pages = {818-829},
     abstract = {Understanding language requires more than the use of fixed conventions and more than decoding combinatorial structure. Instead, comprehenders make exquisitely sensitive inferences about what utterances mean given their knowledge of the speaker, language, and context. Building on developments in game theory and probabilistic modeling, we describe the rational speech act (RSA) framework for pragmatic reasoning. RSA models provide a principled way to formalize inferences about meaning in context; they have been used to make successful quantitative predictions about human behavior in a variety of different tasks and situations, and they explain why complex phenomena, such as hyperbole and vagueness, occur. More generally, they provide a computational framework for integrating linguistic structure, world knowledge, and context in pragmatic language understanding.}
}
@incollection{Goodmanetal2015,
        title = {Concepts in a probabilistic language of thought},
       author = {Goodman, Noah D. and Tenenbaum, Joshua B. and Gerstenberg, T.},
    booktitle = {The Conceptual Mind: New Directions in the Study of Concepts},
       editor = {Morgolis and Lawrence},
    publisher = {MIT Press},
         year = {2015}
}
@article{GoodmanetalNEURON-98,
       author = {Goodman, Miriam B. and Hall, David H. and Avery, Leon and Lockery, Shawn R.},
        title = {Active Currents Regulate Sensitivity and Dynamic Range in {\it{Caenorhabditis elegans}} Neurons},
      journal = {Neuron},
       volume = {20},
        issue = {4},
         year = {1998},
        pages = {763-772},
     abstract = {Little is known about the physiology of neurons in Caenorhabditis elegans. Using new techniques for in situ patch-clamp recording in C. elegans, we analyzed the electrical properties of an identified sensory neuron (ASER) across four developmental stages and 42 unidentified neurons at one stage. We find that ASER is nearly isopotential and fails to generate classical Na(+) action potentials. Rather, ASER displays a high sensitivity to input currents coupled to a depolarization-dependent reduction in sensitivity that may endow ASER with a wide dynamic range. Voltage clamp revealed depolarization-activated K(+) and Ca(2+) currents that contribute to high sensitivity near the zero-current potential. The depolarization-dependent reduction in sensitivity can be attributed to activation of K(+) current at voltages where it dominates the net membrane current. The voltage dependence of membrane current was similar in all neurons examined, suggesting that C. elegans neurons share a common mechanism of sensitivity and dynamic range.},
}
@book{GoodwinandSin84,
       author = {Goodwin, G. C. and Sin, K. S.},
        title = {Adaptive Filtering Prediction and Control},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1984,
}
@book{Goodwyn2012,
        title = {The Neurobiology of the Gods: {H}ow Brain Physiology Shapes the Recurrent Imagery of Myth and Dreams},
       author = {Goodwyn, Erik D.},
    publisher = {Taylor \& Francis},
         year = 2012,
}
@book{Gopal85,
       author = {Gopal, M.},
        title = {Modern Control System Theory},
    publisher = {Halsted Press},
      address = {New York},
         year = 1985,
}
@inproceedings{GopalanetalICAPS-17,
        title = {Planning with Abstract Markov Decision Processes},
       author = {Nakul Gopalan and Marie desJardins and Michael L. Littman and James MacGlashan and Shawn Squire and Stefanie Tellex and Robert John Winder and Lawson L. S. Wong},
    booktitle = {Proceedings of the Twenty-Seventh International Conference on Automated Planning and Scheduling},
         year = {2017},
        pages = {480-488},
     abstract = {Robots acting in human-scale environments must plan under uncertainty in large state-action spaces and face constantly changing reward functions as requirements and goals change. Planning under uncertainty in large state-action spaces requires hierarchical abstraction for efficient computation. We introduce a new hierarchical planning framework called Abstract Markov Decision Processes (AMDPs) that can plan in a fraction of the time needed for complex decision making in ordinary MDPs. AMDPs provide abstract states, actions, and transition dynamics in multiple layers above a base-level "flat" MDP. AMDPs decompose problems into a series of subtasks with both local reward and local transition functions used to create policies for subtasks. The resulting hierarchical planning method is independently optimal at each level of abstraction, and is recursively optimal when the local reward and transition functions are correct. We present empirical results showing significantly improved planning speed, while maintaining solution quality, in the Taxi domain and in a mobile-manipulation robotics problem. Furthermore, our approach allows specification of a decision-making model for a mobile-manipulation problem on a Turtlebot, spanning from low-level control actions operating on continuous variables all the way up through high-level object manipulation tasks.},
}
@inproceedings{Gopaletal02,
      author = {Gopal Pandurangan and Prabhakara Raghavan and Eli Upfal},
       title = {Using {P}age{R}ank to Characterize {W}eb structure},
   booktitle = {8th Annual International Computing and Combinatorics Conference ({COCOON})},
        year = 2002,
}
@inproceedings{GopaletalFOCS-01,
      author = {Gopal Pandurangan, Prabhakar Raghavan and Eli Upfal},
       title = {Building Low-diameter {P2P} Networks},
   booktitle = {Proceedings of the 42nd IEEE Symposium on theFoundations of Computer Science},
        year = 2001,
}
@article {GopnikSCIENCE-12,
       author = {Gopnik, Alison},
        title = {Scientific Thinking in Young Children: Theoretical Advances, Empirical Research, and Policy Implications},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {337},
       number = {6102},
        pages = {1623-1627},
         year = {2012},
     abstract = {New theoretical ideas and empirical research show that very young children{\textquoteright}s learning and thinking are strikingly similar to much learning and thinking in science. Preschoolers test hypotheses against data and make causal inferences; they learn from statistics and informal experimentation, and from watching and listening to others. The mathematical framework of probabilistic models and Bayesian inference can describe this learning in precise ways. These discoveries have implications for early childhood education and policy. In particular, they suggest both that early childhood experience is extremely important and that the trend toward more structured and academic early childhood programs is misguided.},
}
@book{GopnikandMeltzoff98,
       author = {Alison Gopnik and Andrew N. Meltzoff},
        title = {Words, Thoughts, and Theories},
         year = 1998,
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
}
@article{GordonAIJ-86,
       author = {Gordon, J. and E. H. Shortliffe},
        title = {A Method for Managing Evidential Reasoning in a Hierarchical Hypothesis Space},
      journal = AIJ,
       volume = 26,
         year = 1986,
        pages = {323-359},
}
@inproceedings{GordonandSchubertWTE-11,
       author = {Gordon, Jonathan and Schubert, Lenhart K.},
        title = {Discovering Commonsense Entailment Rules Implicit in Sentences},
    booktitle = {Proceedings of the TextInfer 2011 Workshop on Textual Entailment},
         year = 2011,
        pages = {59-63},
    publisher = {Association for Computational Linguistics},
      address = {Stroudsburg, PA, USA},
     abstract = {Reasoning about ordinary human situations and activities requires the availability of diverse types of knowledge, including expectations about the probable results of actions and the lexical entailments for many predicates. We describe initial work to acquire such a collection of conditional (if-then) knowledge by exploiting presuppositional discourse patterns (such as ones involving 'but', 'yet', and 'hoping to') and abstracting the matched material into general rules},
}
@article{GordonetalFiSN-11,
       author = {Gordon, Goren and Kaplan, David M. and Lankow, Benjamin and Little, Daniel Ying-Jeh and Sherwin, Jason and Suter, Benjamin A. and Thaler, Lore},
        title = {Toward an integrated approach to perception and action: conference report and future directions},
      journal = {Frontiers in Systems Neuroscience},
    publisher = {Frontiers Research Foundation},
         year = {2011},
       volume = {5},
        pages = {20-20},
     abstract = {This article was motivated by the conference entitled ``Perception \& Action - An Interdisciplinary Approach to Cognitive Systems Theory,'' which took place September 14-16, 2010 at the Santa Fe Institute, NM, USA. The goal of the conference was to bring together an interdisciplinary group of neuroscientists, roboticists, and theorists to discuss the extent and implications of action-perception integration in the brain. The motivation for the conference was the realization that it is a widespread approach in biological, theoretical, and computational neuroscience to investigate sensory and motor function of the brain in isolation from one another, while at the same time, it is generally appreciated that sensory and motor processing cannot be fully separated. Our article summarizes the key findings of the conference, provides a hypothetical model that integrates the major themes and concepts presented at the conference, and concludes with a perspective on future challenges in the field.},
}
@article{GorelicketalPAMI-07,
     author = {Lena Gorelick and Moshe Blank and Eli Shechtman and Michal Irani and Ronen Basri},
      title = {Actions as Space-Time Shapes},
    journal = {Transactions on Pattern Analysis and Machine Intelligence},
     volume = 29,
      issue = 12,
      pages = {2247-2253},
      month = {December},
       year = 2007,
}
@article{GornetandSchefferBIORXIV-17,
       author = {Gornet, Jonathan and Scheffer, Louis K},
        title = {Simulating Extracted Connectomes},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Labs Journals},
         year = {2017},
     abstract = {Connectomes derived from volume EM imaging of the brain can generate detailed physical models of every neuron, and simulators such as NEURON or GENESIS are designed to work with such models. In principal, combining these technologies, plus transmitter and channel models, should allow detailed and accurate simulation of real neural circuits. Here we experiment with this combination, using a well-studied system (motion detection in Drosophila. Since simulation requires both the physical geometry (which we have) and the models of the synapses (which are not currently available), we built approximate synapses corresponding to their known and estimated function. Once we did so, we reproduced direction selectivity in T4 cells, one of the main functions of this neural circuit. This verified the basic functionality of both extraction and simulations, and provided a biologically relevant computation we could use in further experiments. We then compared models with different degrees of physical realism, from full detailed models down to models consisting of a single node, to examine the tradeoff of simulation resources required versus accuracy achieved. Our results show that much simpler models may be adequate, at least in the case of medulla neurons in Drosophila. Such models can be easily derived from fully detailed models, and result in simulations that are much smaller, much faster, and accurate enough for many purposes. Biologically, we show that a lumped neuron model reproduces the main motion detector operation, confirming the result of Gruntman, that dendritic compution is not required for this function.},
}
@article{GornickandHoffmanNUCLEATION-66,
       author = {Fred. Gornick and John D. Hoffman},
        title = {Nucleation in Polymers},
      journal = {Industrial \& Engineering Chemistry},
    publisher = {American Chemistry Society},
       volume = 58,
        issue = 3,
         year = 66,
        pages = {41-53},
}
@inproceedings{GorryetalIRMA-07,
     author = {Benjamin Gorry and Zezhi Chen and Kevin Hammond and Andy Wallace and Greg Michaelson},
      title = {Using Mean-Shift Tracking Algorithms for Real-Time Tracking of Moving
               Images on an Autonomous Vehicle Testbed Platform},
  booktitle = {International Conference on Intelligent Robotics and Manufacturing Automation},
   location = {Venice, Italy},
       year = 2007,
        url = {http://www-fp.cs.st-andrews.ac.uk/embounded/pubs/papers/IRMA07.pdf},
}
@misc{GottmanRAISING-CHILDREN-15,
        title = {Fostering emotionally intelligent children, families, and communities},
       author = {John Gottman},
         year = 2015,
          url = {http://www.gottman.com/wp-content/uploads/EMOTIONALLY_INTELLIGENT_CHILDREN_Updated4.pdf},
 howpublished = {The Gottman Institute},
}
@article{GouldandVrbaPALEOBIOLOGY-82,
      title = {Exaptation --- A Missing Term in the Science of Form},
     author = {Gould, Stephen Jay and Vrba, Elisabeth S.},
    journal = {Paleobiology},
     volume = 8,
      issue = 1,
      pages = {4-15},
       year = 1982,
  publisher = {Paleontological Society},
   abstract = {Adaptation has been defined and recognized by two different criteria: historical genesis (features built by natural selection for their present role) and current utility (features now enhancing fitness no matter how they arose). Biologists have often failed to recognize the potential confusion between these different definitions because we have tended to view natural selection as so dominant among evolutionary mechanisms that historical process and current product become one. Yet if many features of organisms are non-adapted, but available for useful cooptation in descendants  then an important concept has no name in our lexicon (and unnamed ideas generally remain unconsidered): features that now enhance fitness but were not built by natural selection for their current role. We propose that such features be called exaptations and that adaptation be restricted, as Darwin suggested, to features built by selection for their current role. We present several examples of exaptation, indicating where a failure to conceptualize such an idea limited the range of hypotheses previously available. We explore several consequences of exaptation and propose a terminological solution to the problem of preadaptation.}, 
}
@inproceedings{GouldetalICCV-09,
     author = {S. Gould and R. Fulton and D. Koller},
      title = {Decomposing a Scene into Geometric and Semantically Consistent Regions},
  booktitle = {Proceedings of the International Conference on Computer Vision (ICCV 2009)}, 
       year = 2009
}
@incollection{GouldetalNIPS-09,
      title = {Region-based Segmentation and Object Detection},
     author = {Stephen Gould and Tianshi Gao and Daphne Koller},
  booktitle = {Advances in Neural Information Processing Systems 22},
  publisher = {MIT Press},
    address = {Cambridge, MA},
     editor = {Y. Bengio and D. Schuurmans and J. Lafferty and C. K. I. Williams and A. Culotta},
      pages = {655-663},
       year = 2009
}
@article{GoyaletalCoRR-19,
        title = {Recurrent Independent Mechanisms},
       author = {Anirudh Goyal and Alex Lamb and Jordan Hoffmann and Shagun Sodhani and Sergey Levine and Yoshua Bengio and Bernhard Sch{\"o}lkopf},
      journal = {ArXiv},
         year = {2019},
       volume = {arXiv:1909.10893},
     abstract = {Learning modular structures which reflect the dynamics of the environment can lead to better generalization and robustness to changes which only affect a few of the underlying causes. We propose Recurrent Independent Mechanisms (RIMs), a new recurrent architecture in which multiple groups of recurrent cells operate with nearly independent transition dynamics, communicate only sparingly through the bottleneck of attention, and are only updated at time steps where they are most relevant. We show that this leads to specialization amongst the RIMs, which in turn allows for dramatically improved generalization on tasks where some factors of variation differ systematically between training and evaluation.}
}
@inproceedings{Graham77,
       author = {Graham, R. L. and Lawler, E. L. and Lenstra, J. K. and Rinnooy Kan, A. H. G.},
        title = {Optimization and Approximation in Deterministic Sequencing and Scheduling: A Survey},
    booktitle = {Proceedings Discrete Optimization},
     location = {Vancouver, British Columbia},
         year = 1977,
}
@book{Grahametal89,
       author = {Graham, Ronald L. and Knuth, Donald E. and Patashnik Oren},
        title = {Concrete Mathematics},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1989,
}
@article{GrangerAIM-06,
       author = {Granger, Richard},
        title = {Engines of the brain: the computational instruction set of human cognition},
      journal = {{AI} Magazine},
       volume = 27,
        issue = 2,
    publisher = {American Association for Artificial Intelligence},
      address = {Menlo Park, CA},
         year = 2006,
        pages = {15-32},
}
@article{GrangerBBS-06,
        title = {The Evolution of Computation in Brain Circuitry [excerpt from review of Striedter 2005]},
       author = {Richard Granger},
      journal = {Behavioral and Brain Sciences},
       volume = 29,
        pages = {17-18},
         year = 2006
}
@article{GrangerECONOMETRICA-69,
       author = {C. W. J. Granger},
        title = {Investigating Causal Relations by Econometric Models and Cross-spectral Methods},
      journal = {Econometrica},
       number = {3},
    publisher = {Wiley, Econometric Society},
       volume = {37},
         year = {1969},
        pages = {424-438},
     abstract = {There occurs on some occasions a difficulty in deciding the direction of causality between two related variables and also whether or not feedback is occurring. Testable definitions of causality and feedback are proposed and illustrated by use of simple two-variable models. The important problem of apparent instantaneous causality is discussed and it is suggested that the problem often arises due to slowness in recording information or because a sufficiently wide class of possible causal variables has not been used. It can be shown that the cross spectrum between two variables can be decomposed into two parts, each relating to a single causal arm of a feedback situation. Measures of causal lag and causal strength can then be constructed. A generalisation of this result with the partial cross spectrum is suggested.},
}
@article{GrangierandBengioPAMI-08,
       author = {David Grangier and Samy Bengio},
        title = {A discriminative kernel-based model to rank images from text queries},
      journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
    publisher = {IEEE Computer Society},
         year = 2008,
}
@article{GrattarolaetalMBEC-95,
       author = {M. Grattarola and Marco Bove Sergio Martinoia and G. Massobrio},
      journal = {Medical \& Biological Engineering \& Computing},
        title = {Silicon neuron simulation with {SPICE}: tool for neurobiology and neural networks},
       volume = {33},
        issue = {4},
         year = {1995},
        pages = {533-536},
     abstract = {The paper deals with computer simulations of 'silicon neurons', which are assemblies of CMOS circuits that generate the equivalents of the ionic currents and of the action potentials of real (biological) neurons. The circuit simulation program SPICE is used to simulate the generation of action potentials by a silicon neuron. Moreover, the equivalent circuits of silicon synapses are described and the behaviours of simple two- and three-neuron networks are analysed. Implications for the areas of neurobiology and formal neural networks are briefly considered.}
}
@article{GraumanandDarrellJLMR-03,
       author = {Kristen Grauman and Trevor Darrell},
        title = {The Pyramid Match Kernel: Efficient Learning with Sets of Features},
      journal = {Journal of Machine Learning Research},
       volume = 8,
        pages = {725-760},
         year = 2007,
}
@article{GraveetalCoRR-13,
       author = {Alex Graves and {Abdel-Rahman} Mohamed and Geoffrey Hinton},
        title = {Speech Recognition with Deep Recurrent Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1303.5778},
         year = 2013,
     abstract = {Recurrent neural networks (RNNs) are a powerful model for sequential data. End-to-end training methods such as Connectionist Temporal Classification make it possible to train RNNs for sequence labelling problems where the input-output alignment is unknown. The combination of these methods with the Long Short-term Memory RNN architecture has proved particularly fruitful, delivering state-of-the-art results in cursive handwriting recognition. However RNN performance in speech recognition has so far been disappointing, with better results returned by deep feedforward networks. This paper investigates deep recurrent neural networks, which combine the multiple levels of representation that have proved so effective in deep networks with the flexible use of long range context that empowers RNNs. When trained end-to-end with suitable regularisation, we find that deep Long Short-term Memory RNNs achieve a test set error of 17.7\% on the TIMIT phoneme recognition benchmark, which to our knowledge is the best recorded score.}
}
@article{GravenCLINICAL-PERINATOLOGY-04,
       author = {Graven, S. N.},
        title = {{E}arly neurosensory visual development of the fetus and newborn},
      journal = {Clinical Perinatology},
         year = {2004},
       volume = {31},
       number = {2},
        pages = {199-216},
     abstract = {Neurosensory development of the visual system has its origins long before birth. The genetic processes of basic structure formation are followed by endogenous retinal ganglion cell activation in the form of spontaneous synchronous waves of stimulation. These waves of stimulation are required to establish the topographic relationship among retina, lateral geniculate nucleus, and visual cortex. This process prepares the visual system for visual experience. Visual experience ultimately stimulates creation of columns of neurons in the visual cortex, which are needed to see and interpret patterns, lines, movement, and color. Spontaneous synchronous retinal waves occur in preterm infants in the neonatal intensive care unit and must be protected, as they are critical for visual development.}
}
@article{GravesCoRR-12,
       author = {Alex Graves},
        title = {Sequence Transduction with Recurrent Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1211:3711},
         year = 2012,
}
@article{GravesCoRR-13,
       author = {Alex Graves},
        title = {Generating Sequences With Recurrent Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1308.0850},
         year = 2012,
     abstract = {This paper shows how Long Short-term Memory recurrent neural networks can be used to generate complex sequences with long-range structure, simply by predicting one data point at a time. The approach is demonstrated for text (where the data are discrete) and online handwriting (where the data are real-valued). It is then extended to handwriting synthesis by allowing the network to condition its predictions on a text sequence. The resulting system is able to generate highly realistic cursive handwriting in a wide variety of styles.},
} 
@article{GravesCoRR-16,
       author = {Alex Graves},
        title = {Adaptive Computation Time for Recurrent Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1603.08983},
         year = {2016},
          url = {http://arxiv.org/abs/1603.08983},
     abstract = {This paper introduces Adaptive Computation Time (ACT), an algorithm that allows recurrent neural networks to learn how many computational steps to take between receiving an input and emitting an output. ACT requires minimal changes to the network architecture, is deterministic and differentiable, and does not add any noise to the parameter gradients. Experimental results are provided for four synthetic problems: determining the parity of binary vectors, applying binary logic operations, adding integers, and sorting real numbers. Overall, performance is dramatically improved by the use of ACT, which successfully adapts the number of computational steps to the requirements of the problem. We also present character-level language modelling results on the Hutter prize Wikipedia dataset. In this case ACT does not yield large gains in performance; however it does provide intriguing insight into the structure of the data, with more computation allocated to harder-to-predict transitions, such as spaces between words and ends of sentences. This suggests that ACT or other adaptive computation methods could provide a generic method for inferring segment boundaries in sequence data.},
}
@misc{GravesPhD-09,
        title = {Supervised Sequence Labelling with Recurrent Neural Networks},
       author = {Alex Graves},
         year = 2009,
 howpublished = {Diploma thesis. Technische Universit\"{a}t M\"{u}nchen},
}
@inproceedings{GravesandJaitlyICML-14,
       author = {Alex Graves and Navdeep Jaitly},
        title = {Towards End-to-End Speech Recognition with Recurrent Neural Networks},
    booktitle = {Proceedings of the 31th International Conference on Machine Learning},
       volume = 32,
        pages = {1764-1772},
         year = 2014,
     abstract = {This paper presents a speech recognition system that directly transcribes audio data with text, without requiring an intermediate phonetic representation. The system is based on a combination of the deep bidirectional LSTM recurrent neural network architecture and the Connectionist Temporal Classification objective function. A modification to the objective function is introduced that trains the network to minimise the expectation of an arbitrary transcription loss function. This allows a direct optimisation of the word error rate, even in the absence of a lexicon or language model. The system achieves a word error rate of 27.3\% on the Wall Street Journal corpus with no prior linguistic information, 21.9\% with only a lexicon of allowed words, and 8.2\% with a trigram language model. Combining the network with a baseline system further reduces the error rate to 6.7\%.},
}
@inproceedings{GravesandSchmidhuberIJCNN-05,
       author = {Graves, Alex and Schmidhuber, J\"{u}rgen},
    booktitle = {Proceedings of {IEEE} International Joint Conference on Neural Networks},
        title = {Framewise phoneme classification with bidirectional {LSTM} networks},
         year = 2005,
       volume = 4,
        pages = {2047-2052},
     abstract = {In this paper, we apply bidirectional training to a long short term memory (LSTM) network for the first time. We also present a modified, full gradient version of the LSTM learning algorithm. We discuss the significance of framewise phoneme classification to continuous speech recognition, and the validity of using bidirectional networks for online causal tasks. On the TIMIT speech database, we measure the framewise phoneme classification scores of bidirectional and unidirectional variants of both LSTM and conventional recurrent neural networks (RNNs). We find that bidirectional LSTM outperforms both RNNs and unidirectional LSTMs.},
}
@article{GravesetalARXIV-13,
    author = {Alex Graves and Abdel-rahman Mohamed and Geoffrey E. Hinton},
     title = {Speech Recognition with Deep Recurrent Neural Networks},
   journal = {CoRR},
    volume = {arXiv:1303.5778},
        ee = {http://arxiv.org/abs/1303.5778},
      year = 2013,
}
@inproceedings{GravesetalASRU-13,
       author = {Graves, A. and Jaitly, N. and Mohamed, A.-R.},
    booktitle = {IEEE Workshop on Automatic Speech Recognition and Understanding},
        title = {Hybrid speech recognition with Deep Bidirectional {LSTM}},
         year = 2013,
        pages = {273-278},
     abstract = {Deep Bidirectional LSTM (DBLSTM) recurrent neural networks have recently been shown to give state-of-the-art performance on the TIMIT speech database. However, the results in that work relied on recurrent-neural-network-specific objective functions, which are difficult to integrate with existing large vocabulary speech recognition systems. This paper investigates the use of DBLSTM as an acoustic model in a standard neural network-HMM hybrid system. We find that a DBLSTM-HMM hybrid gives equally good results on TIMIT as the previous work. It also outperforms both GMM and deep network benchmarks on a subset of the Wall Street Journal corpus. However the improvement in word error rate over the deep network is modest, despite a great increase in framelevel accuracy. We conclude that the hybrid approach with DBLSTM appears to be well suited for tasks where acoustic modelling predominates. Further investigation needs to be conducted to understand how to better leverage the improvements in frame-level accuracy towards better word error rates.},
}
@article{GravesetalCoRR-07,
       author = {Alex Graves and Santiago Fern{\'{a}}ndez and J\"{u}rgen Schmidhuber},
        title = {Multi-Dimensional Recurrent Neural Networks},
      journal = {CoRR},
       volume = {arXiv:0705.2011},
         year = 2007,
          URL = {http://arxiv.org/abs/0705.2011},
     abstract = {Recurrent neural networks (RNNs) have proved effective at one dimensional sequence learning tasks, such as speech and online handwriting recognition. Some of the properties that make RNNs suitable for such tasks, for example robustness to input warping, and the ability to access contextual information, are also desirable in multidimensional domains. However, there has so far been no direct way of applying RNNs to data with more than one spatio-temporal dimension. This paper introduces multi-dimensional recurrent neural networks (MDRNNs), thereby extending the potential applicability of RNNs to vision, video processing, medical imaging and many other areas, while avoiding the scaling problems that have plagued other multi-dimensional models. Experimental results are provided for two image segmentation tasks.}
}
@article{GravesetalCoRR-14,
        title = {Neural {T}uring Machines},
       author = {Alex Graves and Greg Wayne and Ivo Danihelka},
      journal = {CoRR},
       volume = {arXiv:1410.5401},
         year = 2014,
          url = {http://arxiv.org/abs/1410.5401},
     abstract = {We extend the capabilities of neural networks by coupling them to external memory resources, which they can interact with by attentional processes. The combined system is analogous to a Turing Machine or Von Neumann architecture but is differentiable end-to-end, allowing it to be efficiently trained with gradient descent. Preliminary results demonstrate that Neural Turing Machines can infer simple algorithms such as copying, sorting, and associative recall from input and output examples.}
}
@inproceedings{GravesetalICML-06,
       author = {Graves, Alex and Fern\'{a}ndez, Santiago and Gomez, Faustino and Schmidhuber, J\"{u}rgen},
        title = {Connectionist Temporal Classification: Labelling Unsegmented Sequence Data with Recurrent Neural Networks},
    booktitle = {Proceedings of the 23rd International Conference on Machine Learning},
         year = 2006,
     location = {Pittsburgh, Pennsylvania},
        pages = {369-376},
    publisher = {ACM},
      address = {New York, NY, USA},
}
@article{GravesetalNATURE-16,
       author = {Graves, Alex and Wayne, Greg and Reynolds, Malcolm and Harley, Tim and Danihelka, Ivo and Grabska-Barwi\'{n}ska, Agnieszka and Colmenarejo, Sergio G\'{o}mez and Grefenstette, Edward and Ramalho, Tiago and Agapiou, John and Badia, Adri\`{a} Puigdom\'{e}nech and Hermann, Karl Moritz and Zwols, Yori and Ostrovski, Georg and Cain, Adam and King, Helen and Summerfield, Christopher and Blunsom, Phil and Kavukcuoglu, Koray and Hassabis, Demis},
        title = {Hybrid computing using a neural network with dynamic external memory},
      journal = {Nature},
         year = {2016},
    publisher = {Macmillan Publishers Limited},
       volume = {538},
        issue = {7626},
        pages = {471-476},
     abstract = {Artificial neural networks are remarkably adept at sensory processing, sequence learning and reinforcement learning, but are limited in their ability to represent variables and data structures and to store data over long timescales, owing to the lack of an external memory. Here we introduce a machine learning model called a differentiable neural computer (DNC), which consists of a neural network that can read from and write to an external memory matrix, analogous to the random-access memory in a conventional computer. Like a conventional computer, it can use its memory to represent and manipulate complex data structures, but, like a neural network, it can learn to do so from data. When trained with supervised learning, we demonstrate that a DNC can successfully answer synthetic questions designed to emulate reasoning and inference problems in natural language. We show that it can learn tasks such as finding the shortest path between specified points and inferring the missing links in randomly generated graphs, and then generalize these tasks to specific graphs such as transport networks and family trees. When trained with reinforcement learning, a DNC can complete a moving blocks puzzle in which changing goals are specified by sequences of symbols. Taken together, our results demonstrate that DNCs have the capacity to solve complex, structured tasks that are inaccessible to neural networks without external read/write memory.},
}
@book{Graziano2013,
        title = {Consciousness and the Social Brain},
       author = {Graziano, Michael},
    publisher = {Oxford University Press},
      address = {New York, NY},
         year = {2013},
}
@book{Graziano2015,
        title = {Consciousness and the Social Brain},
       author = {Graziano, Michael},
    publisher = {Oxford University Press},
      address = {New York, NY},
         year = {2013},
}
@article{GrazianoATLANTIC-16a,
       author = {Graziano, Michael},
        title = {Consciousness Is Not Mysterious: It's just the brain describing itself--to itself},
      journal = {The Atlantic Magazine},
       volume = {2016},
        issue = {January},
         year = {2016},
} 
@article{GrazianoATLANTIC-16b,
       author = {Graziano, Michael},
        title = {How Phantom Limbs Explain Consciousness: The brain's model of the body can tell us a lot about its model of attention},
      journal = {The Atlantic Magazine},
       volume = {2016},
        issue = {February},
         year = {2016},
}
@article{GrazianoATLANTIC-16c,
       author = {Graziano, Michael},
        title = {Most Popular Theories of Consciousness Are Worse Than Wrong: They play to our intuitions, but don't actually explain anything},
      journal = {The Atlantic Magazine},
       volume = {2016},
        issue = {March},
         year = {2016},
 howpublished = {Lisbon Machine Learning School: {LxMLS} 2014},
}
@article{GrazianoATLANTIC-16d,
       author = {Graziano, Michael},
        title = {Your Brain Sees Things That You Don't: Understanding the difference between awareness and attention might be the key to unlocking the mystery of human consciousness},
      journal = {The Atlantic Magazine},
       volume = {2016},
        issue = {April},
         year = {2016},
}
@article{GrazianoATLANTIC-16e,
       author = {Graziano, Michael},
        title = {How Consciousness Explains Ventriloquists and Religion: The brain projects its own qualities onto the world around it--for better or worse},
      journal = {The Atlantic Magazine},
       volume = {2016},
        issue = {May},
         year = {2016},
}
@article{GrazianoATLANTIC-16f,
       author = {Graziano, Michael},
        title = {A New Theory Explains How Consciousness Evolved: A neuroscientist on how we came to be aware of ourselves},
      journal = {The Atlantic Magazine},
       volume = {2016},
        issue = {June},
         year = {2016},
}
@article{GrazianoFiRAI-17,
       author = {Graziano, Michael S.},
        title = {The Attention Schema Theory: A Foundation for Engineering Artificial Consciousness},
      journal = {Frontiers in Robotics and AI},
       volume = {4},
        pages = {60},
         year = {2017},
     abstract = {The purpose of the attention schema theory is to explain how an information-processing device, the brain, arrives at the claim that it possesses a non-physical, subjective awareness, and assigns a high degree of certainty to that extraordinary claim. The theory does not address how the brain might actually possess a non-physical essence. It is not a theory that deals in the non-physical. It is about the computations that cause a machine to make a claim and to assign a high degree of certainty to the claim. The theory is offered as a possible starting point for building artificial consciousness. Given current technology, it should be possible to build a machine that contains a rich internal model of what consciousness is, attributes that property of consciousness to itself and to the people it interacts with, and uses that attribution to make predictions about human behavior. Such a machine would "believe" it is conscious and act like it is conscious, in the same sense that the human machine believes and acts.}
}
@incollection{GrazianoROUTLEDGE_HANDBOOK-18,
       author = {Michael S. Graziano},
        title = {The Attention Schema Theory of Consciousness},
       editor = {Gennaro, R.J.},
    booktitle = {The Routledge Handbook of Consciousness},
    publisher = {Taylor \& Francis},
      address = {Oxford, UK},
         year = {2018},
        pages = {174-187}
}
@incollection{GrazianoandBotvinick2002,
       author = {Graziano, Michael and Botvinick, Mathew},
       editor = {Prinz, W. and Hommel, B.},
        title = {How the brain represents the body: insights from neurophysiology and psychology},
    booktitle = {Common Mechanisms in Perception and Action. Attention and Performance},
    publisher = {Oxford University Press},
      address = {Oxford, UK},
         year = {2002},
        pages = {136-157},
}
@article{GrazianoandWebb-15,
       author = {Graziano, Michael S. and Webb, Taylor W.},
        title = {The attention schema theory: a mechanistic account of subjective awareness},
      journal = {Frontiers in Psychology},
       volume = {6},
        pages = {500},
         year = {2015},
     abstract = {We recently proposed the attention schema theory, a novel way to explain the brain basis of subjective awareness in a mechanistic and scientifically testable manner. The theory begins with attention, the process by which signals compete for the brain's limited computing resources. This internal signal competition is partly under a bottom-up influence and partly under top-down control. We propose that the top-down control of attention is improved when the brain has access to a simplified model of attention itself. The brain therefore constructs a schematic model of the process of attention, the 'attention schema', in much the same way that it constructs a schematic model of the body, the 'body schema'. The content of this internal model leads a brain to conclude that it has a subjective experience. One advantage of this theory is that it explains how awareness and attention can sometimes become dissociated; the brain's internal models are never perfect, and sometimes a model becomes dissociated from the object being modeled. A second advantage of this theory is that it explains how we can be aware of both internal and external events. The brain can apply attention to many types of information including external sensory information and internal information about emotions and cognitive states. If awareness is a model of attention, then this model should pertain to the same domains of information to which attention pertains. A third advantage of this theory is that it provides testable predictions. If awareness is the internal model of attention, used to help control attention, then without awareness, attention should still be possible but should suffer deficits in control. In this article, we review the existing literature on the relationship between attention and awareness, and suggest that at least some of the predictions of the theory are borne out by the evidence.}
}
@incollection{Grazianoetal2004,
        title = {A System of Multimodal Areas in the Primate Brain},
       author = {Graziano, Michael S. and Gross, Charles G. and Taylor, Charlotte S.R. and Moore, Tirin},
    booktitle = {{Crossmodal Space and Crossmodal Attention}},
       editor = {Spence and Driver},
    publisher = {Oxford University Press},
      address = {Oxford},
         year = {2004},
     abstract = {This chapter suggests that a set of interconnected areas in the primate brain monitors the location and movement of objects near the body, and control flinch and other defensive responses. This hypothesized 'defensive' system includes the ventral intraparietal area (VIP), parietal area 7b, the polysensory zone (PZ) in the precentral gyrus, and the putamen. The chapter reviews experimental results on this system of areas and discusses evidence that they are involved in representing visual, tactile, and auditory space near the body and in controlling defensive movements.}
} 
@article{GrecoetalACS-13,
       author = {Greco, Francesco and Fujie, Toshinori and Ricotti, Leonardo and Taccola, Silvia and Mazzolai, Barbara and Mattoli, Virgilio},
        title = {Microwrinkled Conducting Polymer Interface for Anisotropic Multicellular Alignment},
      journal = {ACS Applied Materials \& Interfaces},
       volume = {5},
        issue = {3},
         year = {2013},
        pages = {573-584},
     abstract = {Surfaces with controlled micro and nanoscale topographical cues are useful as smart scaffolds and biointerfaces for cell culture. Recently, use of thin-film and surface wrinkling is emerging as a rapid unconventional method for preparing topographically patterned surfaces, especially suited for the production of smart patterns over large area surfaces. On the other hand, there is an increasing interest in employing conducting polymers as soft, biocompatible, conductive biointerfaces or as parts of bioelectronic devices. A novel convenient and versatile method is presented for producing anisotropic topographical cues at the micro- and nanoscale on conducting polymer surfaces. Micro and nanowrinkles were formed during the heat-shrinking process of a thermo-retractable polystyrene substrate. Surface wrinkling was due to the mismatch between the mechanical properties of a conducting polymer ultrathin film and the substrate. Various geometries of wrinkled structures were prepared, demonstrating the tunability of topography depending on the thickness of the conductive film. A method for patterning the conductive properties of the wrinkled substrates was also presented. Such surfaces acted as smart scaffolds for the functional alignment of cells, envisioning their electrical stimulation. Cell adhesion and proliferation were evaluated, comparing different topographies, and a preferential anisotropic alignment of C2C12 murine skeletal muscle cells along wrinkles was demonstrated. The observed trends were also confirmed concerning the formation of aligned myotubes in C2C12 differentiation stage. Furthermore, improved results in terms of aligned and mature myotube formation were obtained by co-culturing C2C12 cells with a fibroblasts feeder layer. The combination of living cells and tunable conductive nanowrinkles will represent a unique tool for the development of innovative biomedical devices.}
}
@inproceedings{GreenIJCAI-69,
   author =       {Green, Cordell C.},
   title =        {Application of Theorem Proving to Problem Solving},
   booktitle =    {Proceedings {IJCAI} 1},
   location =     {Washington, DC, USA},
   organization = {IJCAII},
   year =         1969,
   pages =        {219-239},
}
@article{GreenMI-82,
       author = {Green, Cordell C. and Westfold, S. J.},
        title = {Knowledge-Based Programming and Deduction in Algorithm Design},
      journal = {Machine Intelligence},
    publisher = {Ellis Horwood Limited, Chichester, England},
       volume = 10,
         year = 1982,
        pages = {339-359},
}
@inproceedings{GreenPSI,
       author = {Green, Cordell C.},
        title = {The Design of the PSI Program Synthesizing System},
    booktitle = {Proceedings of the Second {IEEE} International Conference on Software Engineering},
 organization = {IEEE},
     location = {Long Beach, California},
         year = 1976,
        pages = {4-18},
}
@article{GreenandLabelleEBR-15,
       author = {Green, Andrea M. and Labelle, Jean-Philippe},
        title = {The influence of proprioceptive state on learning control of reach dynamics},
      journal = {Experimental Brain Research},
         year = {2015},
       volume = {233},
        issue = {10},
        pages = {2961-2975},
     abstract = {The motor system shows a remarkable capacity to generalize learned behavior to new contexts while simultaneously permitting learning of multiple and sometimes conflicting skills. To examine the influence of proprioceptive state on this capacity, we compared the effectiveness of changes in workspace location and limb orientation (horizontal vs. parasagittal plane posture) in facilitating learning of opposing dynamic force-field perturbations. When opposing fields were encountered in similar workspace positions and limb orientations, subjects failed to learn the two tasks. In contrast, differences in initial limb proprioceptive state were sufficient for significant learning to take place. The extent of learning was similar when the two fields were encountered in different arm orientations in a similar workspace location as compared to when learning took place in spatially separated workspace locations, consistent with the generalization of learning mainly in intrinsic joint coordinates. In keeping with these observations, examination of how trial-to-trial adaptation generalized showed that generalization tended to be greater across similar limb postures. However, when the two fields were encountered in distinct spatial locations, the extent of generalization of adaptation to one field depended on the limb orientation in which the other field was encountered. These results suggest that three-dimensional proprioceptive limb state plays an important role in modulating generalization patterns so as to permit the best compromise between broad generalization and the simultaneous learning of conflicting skills.}
}
@book{Greenblatt2011,
     title = {The Swerve: How the World Became Modern},
    author = {Stephen Greenblatt},
 publisher = {W.W. Norton \& Co.},
      year = 2011,
}
@article{GreeneandOlivaPS-09,
     title = {{The Briefest of Glances: The Time Course of Natural Scene Understanding}},
    author = {Greene, Michelle R. and Oliva, Aude},
   journal = {Psychological Science},
    volume = 20,
     issue = 4,
      year = 2009,
     pages = {464-472},
}
@article{GreenetalJEPLMC-12,
       author = {Green, A. E.  and Kraemer, D. J.  and Fugelsang, J. A.  and Gray, J. R.  and Dunbar, K. N. },
        title = {{{N}eural correlates of creativity in analogical reasoning}},
      journal = {Journal of Experimental Psychology, Learning, Memory and Cognition},
       volume = {38},
       number = {2},
         year = {2012},
        pages = {264-272},
     abstract = {Brain-based evidence has implicated the frontal pole of the brain as important for analogical mapping. Separately, cognitive research has identified semantic distance as a key determinant of the creativity of analogical mapping (i.e., more distant analogies are generally more creative). Here, we used functional magnetic resonance imaging to assess brain activity during an analogy generation task in which we varied the semantic distance of analogical mapping (as derived quantitatively from a latent semantic analysis). Data indicated that activity within an a priori region of interest in left frontopolar cortex covaried parametrically with increasing semantic distance, even after removing effects of task difficulty. Results implicate increased recruitment of frontopolar cortex as a mechanism for integrating semantically distant information to generate solutions in creative analogical reasoning.}
}
@article{GreengardandNestlerARPT-13,
       author = {Greengard, Paul and Nestler, Eric J.},
        title = {A Conversation with Paul Greengard},
      journal = {Annual Review of Pharmacology and Toxicology},
       volume = {53},
       number = {1},
        pages = {1-16},
         year = {2013},
     abstract = {Paul Greengard was born in New York City in 1925. After completing high school, he served three years in the US Navy during World War II and then completed his bachelor's degree at Hamilton College where he majored in physics and mathematics. He obtained a PhD in biophysics from Johns Hopkins University in 1953 and pursued postdoctoral training with Wilhelm Feldberg at the National Institute for Medical Research in England. After eight years as head of biochemistry at Geigy, and sabbaticals at Albert Einstein College of Medicine and Vanderbilt University, he joined the Yale University faculty as a full professor of pharmacology in 1968. While he was at Yale, Greengard's laboratory performed groundbreaking research, which demonstrated a role for cyclic nucleotides, protein kinases and protein phosphatases, and their protein substrates in the regulation of synaptic transmission. In 1983, Greengard moved to The Rockefeller University, where he has since served as the Vincent Astor Professor and Head of the Laboratory of Molecular and Cellular Neuroscience. Greengard's paradigm-shifting research has continued at Rockefeller and has informed our understanding and possible treatment of a host of brain disorders, including schizophrenia, Alzheimer's disease, Parkinson's disease, and depression. He is the author of more than 950 research articles and reviews. Greengard has received numerous awards and honors, including the Nobel Prize in Physiology or Medicine in 2000, the Metropolitan Life Foundation Award for Medical Research, The National Academy of Sciences Award in Neuroscience, the Ralph W. Gerard Prize in Neuroscience for the Society for Neuroscience, and the Karolinska Institutet's Bicentennial Gold Medal. He is a member of the US National Academy of Sciences and the Institute of Medicine of the National Academies. The following interview was conducted on May 29, 2012.}
}
@article{GreenspanCURRENT-BIOLOGY-07,
       author = {Ralph J. Greenspan},
        title = {Seymour Benzer (1921-2007)},
      journal = {Current Biology},
       volume = {18},
        issue = {3},
         year = {2007},
        pages = {106-110},
}
@article{GreenspanetalTMI-16,
       author = {H. Greenspan and B. van Ginneken and R. M. Summers},
      journal = {IEEE Transactions on Medical Imaging},
        title = {Guest Editorial Deep Learning in Medical Imaging: Overview and Future Promise of an Exciting New Technique},
         year = {2016},
       volume = {35},
       number = {5},
        pages = {1153-1159},
     abstract = {The papers in this special section focus on the technology and applications supported by deep learning. Deep learning is a growing trend in general data analysis and has been termed one of the 10 breakthrough technologies of 2013. Deep learning is an improvement of artificial neural networks, consisting of more layers that permit higher levels of abstraction and improved predictions from data. To date, it is emerging as the leading machine-learning tool in the general imaging and computer vision domains. In particular, convolutional neural networks (CNNs) have proven to be powerful tools for a broad range of computer vision tasks. Deep CNNs automatically learn mid-level and high-level abstractions obtained from raw data (e.g., images). Recent results indicate that the generic descriptors extracted from CNNs are extremely effective in object recognition and localization in natural images. Medical image analysis groups across the world are quickly entering the field and applying CNNs and other deep learning methodologies to a wide variety of applications.},
}
@inproceedings{GreenwaldandDeanAIPS-98,
      author = {Greenwald, Lloyd and Dean, Thomas},
       title = {A conditional scheduling approach to designing real-time systems},
   booktitle = {Proceedings of the 4th International Conference on Artificial Intelligence Planning Systems ({ICAPS}-98)},
    location = {Pittsburgh, Pennsylvania},
       pages = {224-231},
        year = 1998,
}
@article{GreenwaldandDeanNETWORKS-95,
   author = {Greenwald, Lloyd and Dean, Thomas},
    title = {Package Routing in Transportation Networks with Fixed Vehicle Schedules: Formulation, Complexity Results and Approximation Algorithms},
  journal = {Networks},
   volume = 27,
    issue = 1,
     year = 1996,
    pages = {81-93},
}
@article{GrefenstetteetalNIPS-15,
       author = {Edward Grefenstette and Karl Moritz Hermann and Mustafa Suleyman and Phil Blunsom},
        title = {Learning to Transduce with Unbounded Memory},
      journal = {CoRR},
       volume = {arXiv:1506.02516},
         year = {2015},
     abstract = {Recently, strong results have been demonstrated by Deep Recurrent Neural Networks on natural language transduction problems. In this paper we explore the representational power of these models using synthetic grammars designed to exhibit phenomena similar to those found in real transduction problems such as machine translation. These experiments lead us to propose new memory-based recurrent networks that implement continuously differentiable analogues of traditional data structures such as Stacks, Queues, and DeQues. We show that these architectures exhibit superior generalisation performance to Deep RNNs and are often able to learn the underlying generating algorithms in our transduction experiments.}
}
@book{GregerHOW_NOT_TO_DIE-15,
        title = {How Not to Die: Discover the Foods Scientifically Proven to Prevent and Reverse Disease},
       author = {Greger M.D., Michael and Stone, Gene},
    publisher = {Flatiron Books},
         year = {2015},
}
@inproceedings{GregorandLeCunICML-10,
        title = {Learning Fast Approximations of Sparse Coding},
       author = {Gregor, Karol and LeCun, Yann},
    booktitle = {Proceedings of International Conference on Machine learning},
         year = 2010,
}
@article{GregorandLeCunNEC-10,
        title = {Emergence of Complex-Like Cells in a Temporal Product Network with Local Receptive Fields},
       author = {Gregor, Karol and LeCun, Yann},
      journal = {Neural and Evolutionary Computing},
         year = 2010,
      comment = {Gregor and LeCun take the appraoch described in the work of Koray Kavukcuoglu [KavukcuogluetalTR-08, KavukcuogluetalCVPR-09] and extend it by separating the "what" and "where" of visual stimuli into separate channels, as is done in the mammalian ventral and dorsal visual streams, to address the tradeoff between invariance and selectivity [GemanJoP-06].},
     abstract = {We introduce a new neural architecture and an unsupervised algorithm for learning invariant representations from temporal sequence of images. The system uses two groups of complex cells whose outputs are combined multiplicatively: one that represents the content of the image, constrained to be constant over several consecutive frames, and one that represents the precise location of features, which is allowed to vary over time but constrained to be sparse. The architecture uses an encoder to extract features, and a decoder to reconstruct the input from the features. The method was applied to patches extracted from consecutive movie frames and produces orientation and frequency selective units analogous to the complex cells in V1. An extension of the method is proposed to train a network composed of units with local receptive field spread over a large image of arbitrary size. A layer of complex cells, subject to sparsity constraints, pool feature units over overlapping local neighborhoods, which causes the feature units to organize themselves into pinwheel patterns of orientation-selective receptive fields, similar to those observed in the mammalian visual cortex. A feed-forward encoder efficiently computes the feature representation of full images.}
}
@article{GregoretalCoRR-15,
       author = {Karol Gregor and Ivo Danihelka and Alex Graves and Daan Wierstra},
        title = {{DRAW}: A Recurrent Neural Network For Image Generation},
      journal = {CoRR},
       volume = {arXiv:1502.04623},
          url = {http://arxiv.org/abs/1502.04623},
         year = 2015,
     abstract = {This paper introduces the Deep Recurrent Attentive Writer (DRAW) neural network architecture for image generation. DRAW networks combine a novel spatial attention mechanism that mimics the foveation of the human eye, with a sequential variational auto-encoding framework that allows for the iterative construction of complex images. The system substantially improves on the state of the art for generative models on MNIST, and, when trained on the Street View House Numbers dataset, it generates images that cannot be distinguished from real data with the naked eye.}
}
@article{GregoryBMJ-98,
       author = {Richard Gregory},
        title = {Brainy mind},
      journal = {British Medical Journal},
       volume = 317,
        pages = {1693-1695},
         year = 1998,
}
@article{GregoryPERCEPTION-83,
       author = {Gregory R. L.},
        title = {Forty years on: Kenneth Craik's The Nature of Explanation (1943)},
      journal = {Perception},
       volume = {12},
        issue = {3},
         year = {1983},
        pages = {233-237}
}
@article{GregoryetalPTRS_B-80,
       author = {Gregory, Richard Langton and Longuet-Higgins, Hugh Christopher and Sutherland, N. S.},
        title = {Perceptions as hypotheses},
      journal = {Philosophical Transactions of the Royal Society of London. B, Biological Sciences},
       volume = {290},
       number = {1038},
        pages = {181-197},
         year = {1980},
     abstract = {Perceptions may be compared with hypotheses in science. The methods of acquiring scientific knowledge provide a working paradigm for investigating processes of perception. Much as the information channels of instruments, such as radio telescopes, transmit signals which are processed according to various assumptions to give useful data, so neural signals are processed to give data for perception. To understand perception, the signal codes and the stored knowledge or assumptions used for deriving perceptual hypotheses must be discovered. Systematic perceptual errors are important clues for appreciating signal channel limitations, and for discovering hypothesis-generating procedures. Although this distinction between 'physiological' and 'cognitive' aspects of perception may be logically clear, it is in practice surprisingly difficult to establish which are responsible even for clearly established phenomena such as the classical distortion illusions. Experimental results are presented, aimed at distinguishing between and discovering what happens when there is mismatch with the neural signal channel, and when neural signals are processed inappropriately for the current situation. This leads us to make some distinctions between perceptual and scientific hypotheses, which raise in a new form the problem: What are 'objects'?}
}
@article{GreigetalNATURE-13,
       author = {Greig, L. C. and Woodworth, M. B. and Galazo, M. J. and Padmanabhan, H. and Macklis, J. D.},
        title = {Molecular logic of neocortical projection neuron specification, development and diversity},
      journal = {Nature Reviews Neuroscience},
       volume = 14,
        issue = 11,
         year = 2013,
        pages = {755-769},
     abstract = {The sophisticated circuitry of the neocortex is assembled from a diverse repertoire of neuronal subtypes generated during development under precise molecular regulation. In recent years, several key controls over the specification and differentiation of neocortical projection neurons have been identified. This work provides substantial insight into the 'molecular logic' underlying cortical development and increasingly supports a model in which individual progenitor-stage and postmitotic regulators are embedded within highly interconnected networks that gate sequential developmental decisions. Here, we provide an integrative account of the molecular controls that direct the progressive development and delineation of subtype and area identity of neocortical projection neurons.},
}
@book{Grenander93,
       author = {Ulf Grenander},
        title = {General Pattern Theory: A Mathematical Study of Regular Structures},
    publisher = {Oxford University Press},
      address = {New York, NY},
         year = 1993,
}
@book{Grenander96,
       author = {Ulf Grenander},
        title = {Elements of Pattern Theory},
    publisher = {Johns Hopkins University Press},
      address = {Baltimore, Maryland},
         year = 1996,
}
@misc{GrettonRKHS-18,
       author = {Arthur Gretton},
        title = {Introduction to {RKHS} and some simple kernel algorithms},
         year = {2018},
 howpublished = {{\tt{http://www.gatsby.ucl.ac.uk/\textasciitilde{}gretton/coursefiles/lecture4_introToRKHS.pdf}}},
     abstract = {In this document, we give a nontechical introduction to reproducing kernel {Hilbert} spaces ({RKHS}s), and describe some basic algorithms in {RHKS}: [1] What is a kernel, how do we construct it? [2] Operations on kernels that allow us to combine them. [3] The reproducing kernel Hilbert space. [4] Application 1: Difference in means in feature space. [5] Application 2: Kernel PCA. [6] Application 3: Ridge regression},
}
@article{GreweetalNATURE-10,
       author = {Benjamin F. Grewe and Dominik Langer and Hansj\"{o}rg Kasper and Bj\"{o}rn M. Kampa and Fritjof Helmchen},
        title = {High-speed in vivo calcium imaging reveals neuronal network activity with near-millisecond precision},
      journal = {Nature Methods},
       volume = 2,
        issue = 7,
         year = 2010,
        pages = {399-405},
     abstract = {Two-photon calcium imaging of neuronal populations enables optical recording of spiking activity in living animals, but standard laser scanners are too slow to accurately determine spike times. Here we report in vivo imaging in mouse neocortex with greatly improved temporal resolution using random-access scanning with acousto-optic deflectors. We obtained fluorescence measurements from 34-91 layer 2/3 neurons at a 180-490 Hz sampling rate. We detected single action potential-evoked calcium transients with signal-to-noise ratios of 2-5 and determined spike times with near-millisecond precision and 5-15 ms confidence intervals. An automated 'peeling' algorithm enabled reconstruction of complex spike trains from fluorescence traces up to 20-30 Hz frequency, uncovering spatiotemporal trial-to-trial variability of sensory responses in barrel cortex and visual cortex. By revealing spike sequences in neuronal populations on a fast time scale, high-speed calcium imaging will facilitate optical studies of information processing in brain microcircuits.},
}
@article{GrienbergerandKonnerthNEURON-12,
        title = {Imaging Calcium in Neurons},
       author = {Christine Grienberger and Arthur Konnerth},
      journal = {Neuron},
       volume = 73,
        issue = 5,
         year = 2012,
        pages = {862-885},
     abstract = {Calcium ions generate versatile intracellular signals that control key functions in all types of neurons. Imaging calcium in neurons is particularly important because calcium signals exert their highly specific functions in well-defined cellular subcompartments. In this Primer, we briefly review the general mechanisms of neuronal calcium signaling. We then introduce the calcium imaging devices, including confocal and two-photon microscopy as well as miniaturized devices that are used in freely moving animals. We provide an overview of the classical chemical fluorescent calcium indicators and of the protein-based genetically encoded calcium indicators. Using application examples, we introduce new developments in the field, such as calcium imaging in awake, behaving animals and the use of calcium imaging for mapping single spine sensory inputs in cortical neurons in vivo. We conclude by providing an outlook on the prospects of calcium imaging for the analysis of neuronal signaling and plasticity in various animal models.}
}
@misc{GriffithsBBC-FUTURE-19,
        title = {Can you trust your earliest childhood memories?},
       author = {Sarah Griffiths},
 howpublished = {{BBC} Future Magazine},
         year = {2019},
}
@article{GriffithsetalTSC-10,
       author = {Thomas L. Griffiths and Nick Chater and Charles Kemp and Amy Perfors and Joshua B. Tenenbaum},
        title = {Probabilistic models of cognition: exploring representations and inductive biases},
    publisher = {Elsevier Science},
         year = 2010,
      journal = {Trends in Cognitive Sciences},
       volume = 14,
        issue = 8,
        pages = {357-364},
     abstract = {Cognitive science aims to reverse-engineer the mind, and many of the engineering challenges the mind faces involve induction. The probabilistic approach to modeling cognition begins by identifying ideal solutions to these inductive problems. Mental processes are then modeled using algorithms for approximating these solutions, and neural processes are viewed as mechanisms for implementing these algorithms, with the result being a top-down analysis of cognition starting with the function of cognitive processes. Typical connectionist models, by contrast, follow a bottom-up approach, beginning with a characterization of neural mechanisms and exploring what macro-level functional phenomena might emerge. We argue that the top-down approach yields greater flexibility for exploring the representations and inductive biases that underlie human cognition.}
}
@article{GrigorescuetalIVC-04,
       author = {Cosmin Grigorescu and Nicolai Petkov and Michel A. Westenberg},
        title = {Contour and boundary detection improved by surround suppression of texture edges},
      journal = {Image and Vision Computing},
       volume = 22,
        pages = {609-622},
         year = 2004,
}
@book{GrillnerandGraybiel2006,
        title = {Microcircuits: {T}he Interface Between Neurons and Global Brain Function},
       author = {Grillner, Sten and Graybiel, Ann M.},
    publisher = {MIT Press},
         year = {2006},
       series = {Dahlem Workshop Reports},
}
@article{GrillneretalTiN-05,
       author = {Sten Grillner and Henry Markram and Erik De Schutter and Gilad Silberberg and Fiona E. N. LeBeau},
        title = {Microcircuits in action -- from {CPGs} to neocortex},
      journal = {Trends in Neurosciences},
       volume = {28},
       number = {10},
        pages = {525-533},
         year = {2005},
     abstract = {To understand the interface between global brain function and molecular neuroscience - that is, the microcircuit level - a major challenge. Such understanding is prerequisite if we are to account for neural function in cellular terms. Very few vertebrate microcircuits are yet understood because their analysis is demanding technically. In this review of the \{TINS\} Microcircuits Special Feature, we attempt to shed light on the problem by comparing the operation of four types of microcircuit, to identify common molecular and cellular components. Central pattern generator (CPG) networks underlying rhythmic movements and hippocampal microcircuits that generate gamma and theta rhythms are compared with the neocortical microcircuits used in cognitive tasks and a cerebellar network. The long-term goal is to identify the components of a molecular and synaptic tool kit for the design of different microcircuits.}
}
@techreport{GrinvaldetalTR-01,
     title = {{In-vivo optical imaging of cortical architecture and dynamics}},
    author = {Grinvald, A. and Shoham, D. and Shmuel, A. and Glaser, D. and Vanzetta, I. and 
              Shtoyerman, E. and Slovin, H. and Sterkin, A. and Wijnbergen, C. and Hildesheim, R. and Arieli, A.},
institution = {Grodetsky Center for Research of Higher Brain Functions at the Weizmann Institute of Science, Israel},
     issue = {BC-AG/99-6},
      year = 2001
}
@article{GrioletalACL-13,
     title = {Domain-Independent Statistical Methodology for Dialog Management in Spoken Dialog Systems},
   journal = {Computer Speech \& Language},
    volume = 27,
     issue = 7,
      year = 2013,
     pages = {000-000},
    author = {David Griol and Zoraida Callejas and Ramon Lopez-Cozar and Giuseppe Riccardi},
  abstract = {This paper proposes a domain-independent statistical methodology to develop dialog managers for spoken dialog systems. Our methodology employs a data-driven classification procedure to generate abstract representations of system turns taking into account the previous history of the dialog. A statistical framework is also introduced for the development and evaluation of dialog systems created using the methodology, which is based on a dialog simulation technique. The benefits and flexibility of the proposed methodology have been validated by developing statistical dialog managers for four spoken dialog systems of different complexity, designed for different languages (English, Italian, and Spanish) and application domains (from transactional to problem-solving tasks). The evaluation results show that the proposed methodology allows rapid development of new dialog managers as well as to explore new dialog strategies, which permit developing new enhanced versions of already existing systems.}
}
@inproceedings{GrioletalCIARP-08,
    author = {Griol, D. and Hurtado, L. F. and Segarra, E. and Sanchis, E.},
     title = {A Dialog Management Methodology Based on Neural Networks and Its Application to Different Domains},
 booktitle = {Proceedings of the 13th Iberoamerican congress on Pattern Recognition: Progress in Pattern Recognition, Image Analysis and Applications},
  location = {Havana, Cuba},
 publisher = {Springer-Verlag},
   address = {Berlin, Heidelberg},
      year = 2008,
     pages = {643-650},
  abstract = {In this paper, we present a statistical approach for dialog management within the framework of two different domains. The dialog model, that is automatically learned from a data corpus, is based on the use of a classification process to generate the next system answer. A neural network classifier is used for the selection process. This methodology has been applied in a spoken dialog system that provides railway information. The definition of an extended methodology that takes into account new system functionalities and its application for developing a dialog system for booking sports facilities is also described.},
}
@article{GriponandBerrouTNN-11,
       author = {Gripon, V. and Berrou, C.},
      journal = {{IEEE} Transactions on Neural Networks},
        title = {Sparse Neural Networks With Large Learning Diversity},
         year = 2011,
       volume = 22,
        issue = 7,
        pages = {1087-1096},
     abstract = {Coded recurrent neural networks with three levels of sparsity are introduced. The first level is related to the size of messages that are much smaller than the number of available neurons. The second one is provided by a particular coding rule, acting as a local constraint in the neural activity. The third one is a characteristic of the low final connection density of the network after the learning phase. Though the proposed network is very simple since it is based on binary neurons and binary connections, it is able to learn a large number of messages and recall them, even in presence of strong erasures. The performance of the network is assessed as a classifier and as an associative memory.},
}
@article {GrosenicketalBIORXIV-17,
       author = {Grosenick, Logan M. and Broxton, Michael and Kim, Christina K. and Liston, Conor and Poole, Ben and Yang, Samuel and Andalman, Aaron S. and Scharff, Edward and Cohen, Noy and Yizhar, Ofer and Ramakrishnan, Charu and Ganguli, Surya and Suppes, Patrick and Levoy, Marc and Deisseroth, Karl},
        title = {Identification Of Cellular-Activity Dynamics Across Large Tissue Volumes In The Mammalian Brain},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Labs Journals},
          url = {http://biorxiv.org/content/early/2017/05/01/132688},
         year = {2017},
     abstract = {Tracking the coordinated activity of cellular events through volumes of intact tissue is a major challenge in biology that has inspired significant technological innovation. Yet scanless measurement of the high-speed activity of individual neurons across three dimensions in scattering mammalian tissue remains an open problem. Here we develop and validate a computational imaging approach (SWIFT) that integrates high-dimensional, structured statistics with light field microscopy to allow the synchronous acquisition of single-neuron resolution activity throughout intact tissue volumes as fast as a camera can capture images (currently up to 100 Hz at full camera resolution), attaining rates needed to keep pace with emerging fast calcium and voltage sensors. We demonstrate that this large field-of-view, single-snapshot volume acquisition method {\textemdash} which requires only a simple and inexpensive modification to a standard fluorescence microscope {\textemdash} enables scanless capture of coordinated activity patterns throughout mammalian neural volumes. Further, the volumetric nature of SWIFT also allows fast in vivo imaging, motion correction, and cell identification throughout curved subcortical structures like the dorsal hippocampus, where cellular-resolution dynamics spanning hippocampal subfields can be simultaneously observed during a virtual context learning task in a behaving animal. SWIFT{\textquoteright}s ability to rapidly and easily record from volumes of many cells across layers opens the door to widespread identification of dynamical motifs and timing dependencies among coordinated cell assemblies during adaptive, modulated, or maladaptive physiological processes in neural systems.},
}
@article{GroshevetalCoRR-17,
       author = {Edward Groshev and Aviv Tamar and Siddharth Srivastava and Pieter Abbeel},
        title = {Learning Generalized Reactive Policies using Deep Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1708.07280},
         year = {2017},
     abstract = {We consider the problem of learning for planning, where knowledge acquired while planning is reused to plan faster in new problem instances. For robotic tasks, among others, plan execution can be captured as a sequence of visual images. For such domains, we propose to use deep neural networks in learning for planning, based on learning a reactive policy that imitates execution traces produced by a planner. We investigate architectural properties of deep networks that are suitable for learning long-horizon planning behavior, and explore how to learn, in addition to the policy, a heuristic function that can be used with classical planners or search algorithms such as A*. Our results on the challenging Sokoban domain show that, with a suitable network design, complex decision making policies and powerful heuristic functions can be learned through imitation.}
}
@inproceedings{GrossCCSW-14,
       author = {Gross, Thomas R.},
        title = {Efficient Certification and Zero-Knowledge Proofs of Knowledge on Infrastructure Topology Graphs},
    booktitle = {Proceedings of the 6th Edition of the ACM Workshop on Cloud Computing Security},
    publisher = {ACM},
         year = {2014},
        pages = {69-80},
     abstract = {Digital signature schemes are a foundational cryptographic building block in certification and the projection of trust. Based on a signature scheme on committed graphs, we propose a framework of certification and proof methods to sign topology graphs and to prove properties of their certificates in zero-knowledge. This framework allows an issuer, such as an auditing system, to sign the topology representation of an infrastructure. The prover, such as an infrastructure provider, can then convince a verifier of topology properties including connectivity and isolation without disclosing the blueprint of the topology itself. By that, we can certify the structure of critical systems while still maintaining confidentiality. We offer zero-knowledge proofs of knowledge for a general specification language of security goals for virtualized infrastructures such that high-level security goals can be proven over topology certificates. We offer an efficient and practical construction, built upon the Camenisch-Lysyanskaya signature scheme, honest-verifier proofs and the strong RSA assumption.}
}
@article{GrossbergFiNC-17,
       author = {Grossberg, Stephen},
        title = {Acetylcholine Neuromodulation in Normal and Abnormal Learning and Memory: Vigilance Control in Waking, Sleep, Autism, Amnesia and Alzheimer's Disease},
      journal = {Frontiers in Neural Circuits},
       volume = {11},
        pages = {82},
         year = {2017},
     abstract = {Adaptive Resonance Theory, or ART, is a neural model that explains how normal and abnormal brains may learn to categorize and recognize objects and events in a changing world, and how these learned categories may be remembered for a long time. This article uses ART to propose and unify the explanation of diverse data about normal and abnormal modulation of learning and memory by acetylcholine (ACh). In ART, vigilance control determines whether learned categories will be general and abstract, or specific and concrete. ART models how vigilance may be regulated by ACh release in layer 5 neocortical cells by influencing after-hyperpolarization (AHP) currents. This phasic ACh release is mediated by cells in the nucleus basalis (NB) of Meynert that are activated by unexpected events. The article additionally discusses data about ACh-mediated tonic control of vigilance. ART proposes that there are often dynamic breakdowns of tonic control in mental disorders such as autism, where vigilance remains high, and medial temporal amnesia, where vigilance remains low. Tonic control also occurs during sleep-wake cycles. Properties of Up and Down states during slow wave sleep arise in ACh-modulated laminar cortical ART circuits that carry out processes in awake individuals of contrast normalization, attentional modulation, decision-making, activity-dependent habituation, and mismatch-mediated reset. These slow wave sleep circuits interact with circuits that control circadian rhythms and memory consolidation. Tonic control properties also clarify how Alzheimer's disease symptoms follow from a massive structural degeneration that includes undermining vigilance control by ACh in cortical layers 3 and 5. Sleep disruptions before and during Alzheimer's disease, and how they contribute to a vicious cycle of plaque formation in layers 3 and 5, are also clarified from this perspective.}
}
@article{GrossbergNEURAL-NETWORKS-13,
       author = {Grossberg, Stephen},
        title = {Adaptive {R}esonance {T}heory: how a brain learns to consciously attend, learn, and recognize a changing world},
      journal = {Neural Networks},
         year = {2013},
       volume = {37},
        pages = {1-47},
     abstract = {Adaptive Resonance Theory, or ART, is a cognitive and neural theory of how the brain autonomously learns to categorize, recognize, and predict objects and events in a changing world. This article reviews classical and recent developments of ART, and provides a synthesis of concepts, principles, mechanisms, architectures, and the interdisciplinary data bases that they have helped to explain and predict. The review illustrates that ART is currently the most highly developed cognitive and neural theory available, with the broadest explanatory and predictive range. Central to ART's predictive power is its ability to carry out fast, incremental, and stable unsupervised and supervised learning in response to a changing world. ART specifies mechanistic links between processes of consciousness, learning, expectation, attention, resonance, and synchrony during both unsupervised and supervised learning. ART provides functional and mechanistic explanations of such diverse topics as laminar cortical circuitry; invariant object and scenic gist learning and recognition; prototype, surface, and boundary attention; gamma and beta oscillations; learning of entorhinal grid cells and hippocampal place cells; computation of homologous spatial and temporal mechanisms in the entorhinal-hippocampal system; vigilance breakdowns during autism and medial temporal amnesia; cognitive-emotional interactions that focus attention on valued objects in an adaptively timed way; item-order-rank working memories and learned list chunks for the planning and control of sequences of linguistic, spatial, and motor information; conscious speech percepts that are influenced by future context; auditory streaming in noise during source segregation; and speaker normalization. Brain regions that are functionally described include visual and auditory neocortex; specific and nonspecific thalamic nuclei; inferotemporal, parietal, prefrontal, entorhinal, hippocampal, parahippocampal, perirhinal, and motor cortices; frontal eye fields; supplementary eye fields; amygdala; basal ganglia: cerebellum; and superior colliculus. Due to the complementary organization of the brain, ART does not describe many spatial and motor behaviors whose matching and learning laws differ from those of ART. ART algorithms for engineering and technology are listed, as are comparisons with other types of models.}
}
@article{GrossbergandWilliamsonCORTEX-01,
       author = {Grossberg, Stephen and Williamson, James R.},
        title = {A Neural Model of how Horizontal and Interlaminar Connections of Visual Cortex Develop into Adult Circuits that Carry Out Perceptual Grouping and Learning},
      journal = {Cerebral Cortex},
       volume = 11,
        issue = 1,
        pages = {37-58},
         year = 2001,
     abstract = {A neural model suggests how horizontal and interlaminar connections in visual cortical areas V1 and V2 develop within a laminar cortical architecture and give rise to adult visual percepts. The model suggests how mechanisms that control cortical development in the infant lead to properties of adult cortical anatomy, neurophysiology and visual perception. The model clarifies how excitatory and inhibitory connections can develop stably by maintaining a balance between excitation and inhibition. The growth of long-range excitatory horizontal connections between layer 2/3 pyramidal cells is balanced against that of short-range disynaptic interneuronal connections. The growth of excitatory on-center connections from layer 6-to-4 is balanced against that of inhibitory interneuronal off-surround connections. These balanced connections interact via intracortical and intercortical feedback to realize properties of perceptual grouping, attention and perceptual learning in the adult, and help to explain the observed variability in the number and temporal distribution of spikes emitted by cortical neurons. The model replicates cortical point spread functions and psychophysical data on the strength of real and illusory contours. The on-center, off-surround layer 6-to-4 circuit enables top-down attentional signals from area V2 to modulate, or attentionally prime, layer 4 cells in area V1 without fully activating them. This modulatory circuit also enables adult perceptual learning within cortical area V1 and V2 to proceed in a stable way.},
}
@article{GrosseetalPRE-02,
       author = {Ivo Grosse and Pedro Bernaola-Galv\'{a}n and Pedro Carpena and Ram\'{o}n Rom\'{a}n-Rold\'{a}n and Jose Oliver and H. Eugene Stanley},
        title = {Analysis of symbolic sequences using the Jensen-Shannon divergence},
      journal = {Physical Review E},
       volume = 65,
        issue = 041905,
        pages = {1-16},
         year = 2002,
}
@article{GrossmanF-HN-13,
       author = {Grossmann, Tobias},
        title = {The role of medial prefrontal cortex in early social cognition},
      journal = {Frontiers in Human Neuroscience},
       volume = {7},
         year = {2013},
          url = {https://www.frontiersin.org/article/10.3389/fnhum.2013.00340},
     abstract = {One major function of our brain is to enable us to behave with respect to socially relevant information. Much research on how the adult human brain processes the social world has shown that there is a network of specific brain areas, also called the social brain, preferentially involved during social cognition. Among the specific brain areas involved in the adult social brain, functional activity in prefrontal cortex (PFC), particularly the medial prefrontal cortex (mPFC), is of special importance for human social cognition and behavior. However, from a developmental perspective, it has long been thought that PFC is functionally silent during infancy (first year of life), and until recently, little was known about the role of PFC in the early development of social cognition. I shall present an emerging body of recent neuroimaging studies with infants that provide evidence that mPFC exhibits functional activation much earlier than previously thought, suggesting that the mPFC is involved in social information processing from early in life. This review will highlight work examining infant mPFC function across a range of social contexts. The reviewed findings will illustrate that the human brain is fundamentally adapted to develop within a social context.}
}
@article{GrossmanetalCELL-17,
       author = {Grossman, Nir and Bono, David and Dedic, Nina and Kodandaramaiah, Suhasa B. and Rudenko, Andrii and Suk, Ho-Jun and Cassara, Antonino M. and Neufeld, Esra and Kuster, Niels and Tsai, Li-Huei and Pascual-Leone, Alvaro and Boyden, Edward S.},
        title = {Noninvasive deep brain stimulation via temporally interfering electric fields},
      journal = {Cell},
       volume = {169},
        issue = {6},
         year = {2017},
        pages = {1029-1041},
     abstract = {We report a noninvasive strategy for electrically stimulating neurons at depth. By delivering to the brain multiple electric fields at frequencies too high to recruit neural firing, but which differ by a frequency within the dynamic range of neural firing, we can electrically stimulate neurons throughout a region where interference between the multiple fields results in a prominent electric field envelope modulated at the difference frequency. We validated this temporal interference (TI) concept via modeling and physics experiments, and verified that neurons in the living mouse brain could follow the electric field envelope. We demonstrate the utility of TI stimulation by stimulating neurons in the hippocampus of living mice without recruiting neurons of the overlying cortex. Finally, we show that by altering the currents delivered to a set of immobile electrodes, we can steerably evoke different motor patterns in living mice.},
}
@book{GroszetalNLP,
       editor = {Grosz, Barbara and Sparck Jones, Karen and Webber, Bonnie},
        title = {Readings in Natural Language Processing},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1986,
}
@article{GrotheeretalNATURE-COMMUNICATIONS-19,
       author = {Grotheer, Mareike and Zhen, Zonglei and Lerma-Usabiaga, Garikoitz and Grill-Spector, Kalanit},
        title = {Separate lanes for adding and reading in the white matter highways of the human brain},
      journal = {Nature Communications},
         year = {2019},
       volume = {10},
        issue = {1},
        pages = {3675},
     abstract = {Math and reading involve distributed brain networks and have both shared (e.g. encoding of visual stimuli) and dissociated (e.g. quantity processing) cognitive components. Yet, to date, the shared vs. dissociated gray and white matter substrates of the math and reading networks are unknown. Here, we define these networks and evaluate the structural properties of their fascicles using functional MRI, diffusion MRI, and quantitative MRI. Our results reveal that there are distinct gray matter regions which are preferentially engaged in either math (adding) or reading, and that the superior longitudinal and arcuate fascicles are shared across the math and reading networks. Strikingly, within these fascicles, reading- and math-related tracts are segregated into parallel sub-bundles and show structural differences related to myelination. These findings open a new avenue of research that examines the contribution of sub-bundles within fascicles to specific behaviors.},
}
@inproceedings{GrundmannetalCVPR-10,
    author = {M. Grundmann and V. Kwatra and M. Han and I. Essa},
     title = {Efficient Hierarchical Graph Based Video Segmentation},
 booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
      year = 2010,
}
@inproceedings{GrundmannetalCVPR-11,
    author = {M. Grundmann and V. Kwatra and I. Essa},
     title = {Auto-Directed Video Stabilization with Robust L1 Optimal Camera Paths},
 booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
      year = 2011,
}
@inproceedings{GuadarramaetalICRAS-13,
       author = {Sergio Guadarrama and Lorenzo Riano and Dave Golland and Daniel Gouhring and Yangqing Jia and Dan Klein and Pieter Abbeel and Trevor Darrell},
        title = {Grounding spatial relations for human-robot interaction},
    booktitle = {2013 {IEEE/RSJ} International Conference on Intelligent Robots and Systems},
         year = 2013,
        pages = {1640-1647},
     abstract = {We propose a system for human-robot interaction that learns both models for spatial prepositions and for object recognition. Our system grounds the meaning of an input sentence in terms of visual percepts coming from the robot's sensors in order to send an appropriate command to the PR2 or respond to spatial queries. To perform this grounding, the system recognizes the objects in the scene, determines which spatial relations hold between those objects, and semantically parses the input sentence. The proposed system uses the visual and spatial information in conjunction with the semantic parse to interpret statements that refer to objects (nouns), their spatial relationships (prepositions), and to execute commands (actions). The semantic parse is inherently compositional, allowing the robot to understand complex commands that refer to multiple objects and relations such as: "Move the cup close to the robot to the area in front of the plate and behind the tea box". Our system correctly parses 94\% of the 210 online test sentences, correctly interprets 91\% of the correctly parsed sentences, and correctly executes 89\% of the correctly interpreted sentences.}
}
@article{Guanetal97,
       author = {Guan, L. and Anderson, J.A. and Sutton, J.P.},
        title = {A network of networks processing model for image regularlization},
      journal = {IEEE Transactions on Neural Networks},
    publisher = {IEEE Neural Network Society},
      comment = {now called the Computational Intelligence Society},
       volume = 8,
        pages = {1-6},
         year = 1997,
}
@book{GuckenheimerandHolmes83,
       author = {Guckenheimer, J. and Holmes, P.},
        title = {Nonlinear Oscillations, Dynamical Systems, and Bifurcations of Vector Fields},
    publisher = {Springer-Verlag},
      address = {New York},
         year = 1983,
}
@article{GuestrinetalJAIR-03,
       author = {Carlos Guestrin and Daphne Koller and Ronald Parr and Shobha Venkataraman},
        title = {Efficient Solution Algorithms for Factored {MDP}s},
      journal = {Journal of Artificial Intelligence Research},
       volume = {19},
         year = {2003},
        pages = {399-468}
}
@article{GuetalCoRR-17,
       author = {Jiuxiang Gu and Zhenhua Wang and Jason Kuen and Lianyang Ma and Amir Shahroudy and Bing Shuai and Ting Liu and Xingxing Wang and Gang Wang},
        title = {Recent Advances in Convolutional Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1512.07108},
         year = {2017},
     abstract = {In the last few years, deep learning has led to very good performance on a variety of problems, such as visual recognition, speech recognition and natural language processing. Among different types of deep neural networks, convolutional neural networks have been most extensively studied. Leveraging on the rapid growth in the amount of the annotated data and the great improvements in the strengths of graphics processor units, the research on convolutional neural networks has been emerged swiftly and achieved state-of-the-art results on various tasks. In this paper, we provide a broad survey of the recent advances in convolutional neural networks. We detailize the improvements of CNN on different aspects, including layer design, activation function, loss function, regularization, optimization and fast computation. Besides, we also introduce various applications of convolutional neural networks in computer vision, speech and natural language processing.}
}
@inproceedings{GuetalECCV-12,
     title = {Multi-component Models for Object Detection},
    author = {Chunhui Gu and Pablo Andr{\'e}s Arbel{\'a}ez and Yuanqing Lin and Kai Yu and Jitendra Malik},
 booktitle = {Proceedings of the European Conference on Computer Vision},
      year = 2012,
     pages = {445-458},
}
@article{GuevaraetalABB-13,
       author = {Marta L\'{o}pez-Santib\'{a}\~{n}ez Guevara and Eileen Uribe-Querol and Alma Lilia Fuentes Far\'{i}as and Esperanza Mel\'{e}ndez-Herrera and Agustine Joseph D'Ercole and Gabriel Guti\'{e}rrez-Ospina},
        title = {Cortical columns (barrels) display normal size in the brain's primary somatosensory cortex of mice carrying null mutations of the insulin receptor substrate 1 gene: A preliminary report},
      journal = {Advances in Bioscience and Biotechnology},
       volume = {4},
         year = {2013},
        pages = {945-948},
}
@article{GuezetalCoRR-18,
       author = {Arthur Guez and Th{\'{e}}ophane Weber and Ioannis Antonoglou and Karen Simonyan and Oriol Vinyals and Daan Wierstra and R{\'{e}}mi Munos and David Silver},
        title = {Learning to Search with {MCTS}nets},
      journal = {CoRR},
       volume = {arXiv:1802.04697},
         year = {2018},
     abstract = {Planning problems are among the most important and well-studied problems in artificial intelligence. They are most typically solved by tree search algorithms that simulate ahead into the future, evaluate future states, and back-up those evaluations to the root of a search tree. Among these algorithms, Monte-Carlo tree search (MCTS) is one of the most general, powerful and widely used. A typical implementation of MCTS uses cleverly designed rules, optimized to the particular characteristics of the domain. These rules control where the simulation traverses, what to evaluate in the states that are reached, and how to back-up those evaluations. In this paper we instead learn where, what and how to search. Our architecture, which we call an MCTSnet, incorporates simulation-based search inside a neural network, by expanding, evaluating and backing-up a vector embedding. The parameters of the network are trained end-to-end using gradient-based optimisation. When applied to small searches in the well known planning problem Sokoban, the learned search algorithm significantly outperformed MCTS baselines.}
}
@article{GuggenmosetalPNAS-13,
    author = {Guggenmos, David J. and Azin, Meysam and Barbay, Scott and Mahnken, Jonathan D. and Dunham, Caleb and Mohseni, Pedram and Nudo, Randolph J.}, 
     title = {Restoration of function after brain damage using a neural prosthesis},
   journal = {Proceedings of the National Academy of Sciences}, 
      year = 2013, 
  abstract = {Neural interface systems are becoming increasingly more feasible for brain repair strategies. This paper tests the hypothesis that recovery after brain injury can be facilitated by a neural prosthesis serving as a communication link between distant locations in the cerebral cortex. The primary motor area in the cerebral cortex was injured in a rat model of focal brain injury, disrupting communication between motor and somatosensory areas and resulting in impaired reaching and grasping abilities. After implantation of microelectrodes in cerebral cortex, a neural prosthesis discriminated action potentials (spikes) in premotor cortex that triggered electrical stimulation in somatosensory cortex continuously over subsequent weeks. Within 1 wk, while receiving spike-triggered stimulation, rats showed substantially improved reaching and grasping functions that were indistinguishable from prelesion levels by 2 wk. Post hoc analysis of the spikes evoked by the stimulation provides compelling evidence that the neural prosthesis enhanced functional connectivity between the two target areas. This proof-of-concept study demonstrates that neural interface systems can be used effectively to bridge damaged neural pathways functionally and promote recovery after brain injury.}, 
}
@incollection{GuilleryPBR-05,
       author = {R.W. Guillery},
        title = {Anatomical pathways that link perception and action},
       series = {Progress in Brain Research},
    publisher = {Elsevier},
       volume = {149},
        pages = {235-256},
         year = {2005},
    booktitle = {Cortical Function: a View from the Thalamus},
     abstract = {Pathways linking action to perception are generally presented as passing from sensory pathways, through the thalamus, and then to a putative hierarchy of corticocortical links to motor outputs or to memory. Evidence for more direct sensorimotor links is now presented to show that cerebral cortex rarely, if ever, receives messages representing receptor activity only; thalamic inputs to cortex also carry copies of current motor instructions. Pathways afferent to the thalamus represent the primary input to neocortex. Generally they are made up of branching axons that send one branch to the thalamus and another to output centers of the brain stem or spinal cord. The information transmitted through the classical 'sensory' pathways to the thalamus represents not only information about the environment and the body, but also about instructions currently on their way to motor centers. The proposed hierarchy of direct corticocortical connections of the sensory pathways is not the only possible hierarchy of cortical connections. There is also a hierarchy of the corticofugal pathways to motor centers in the midbrain, and there are transthalamic corticocortical pathways that may show a comparable hierarchy. The extent to which these hierarchies may match each other, and relate to early developmental changes are poorly defined at present, but are important for understanding mechanisms that can link action and perception in the developing brain.}
}
@article{GuilleryandShermanBRR-11,
       author = {Guillery, R. W. and Sherman, S. Murray},
        title = {Branched thalamic afferents: {W}hat are the messages that they relay to the cortex?},
      journal = {Brain research reviews},
       volume = {66},
        issue = {1-2},
         year = {2011},
        pages = {205-219},
     abstract = {Many of the axons that carry messages to the thalamus for relay to the cerebral cortex are branched in a pattern long known from Golgi preparations. They send one branch to the thalamus and the other to motor centers of the brainstem or spinal cord. Because the thalamic branches necessarily carry copies of the motor instructions their messages have the properties of efference copies. That is, they can be regarded as providing reliable information about impending instructions contributing to movements that will produce changes in inputs to receptors, thus allowing neural centers to compensate for these changes of input. We consider how a sensory pathway like the medial lemniscus, the spinothalamic tract or the optic tract can also be seen to act as a pathway for an efference copy. The direct connections that ascending and cortical inputs to the thalamus also establish to motor outputs create sensorimotor relationships that provide cortex with a model of activity in lower circuits and link the sensory and the motor sides of behavior more tightly than can be expected from motor outputs with a single, central origin. These transthalamic connectional patterns differ from classical models of separate neural pathways for carrying efference copies of actions generated at higher levels, and introduce some different functional possibilities.},
}
@misc{GujarGENERATIVE_ADVERSARIAL-IMITATION-LEARNING-18,
        title = {Generative Adversarial Imitation Learning},
       author = {Sanket Gujar},
 howpublished = {https://medium.com/@sanketgujar95/generative-adversarial-imitation-learning-266f45634e60},
         year = {2018},
}
@article{GulcehreetalCoRR-16,
       author = {Caglar G\"{u}lcehre and Marcin Moczulski and Francesco Visin and Yoshua Bengio},
        title = {Mollifying Networks},
      journal = {CoRR},
       volume = {arXiv:1608.04980},
         year = {2016},
     abstract = {The optimization of deep neural networks can be more challenging than traditional convex optimization problems due to the highly non-convex nature of the loss function, e.g. it can involve pathological landscapes such as saddle-surfaces that can be difficult to escape for algorithms based on simple gradient descent. In this paper, we attack the problem of optimization of highly non-convex neural networks by starting with a smoothed -- or {\it{mollified}} -- objective function that gradually has a more non-convex energy landscape during the training. Our proposition is inspired by the recent studies in continuation methods: similar to curriculum methods, we begin learning an easier (possibly convex) objective function and let it evolve during the training, until it eventually goes back to being the original, difficult to optimize, objective function. The complexity of the mollified networks is controlled by a single hyperparameter which is annealed during the training. We show improvements on various difficult optimization tasks and establish a relationship with recent works on continuation methods for neural networks and mollifiers.},
}
@article{GulchereandBengioJMLR-16,
       author = {Caglar G\"{u}lcehre and Yoshua Bengio},
        title = {Knowledge Matters: Importance of Prior Information for Optimization},
      journal = {Journal of Machine Learning Research},
         year = {2016},
       volume = {17},
       number = {8},
        pages = {1-32},
     abstract = {We explored the effect of introducing prior knowledge into the intermediate level of deep supervised neural networks on two tasks. On a task we designed, all black-box state-of-the-art machine learning algorithms which we tested, failed to generalize well. We motivate our work from the hypothesis that, there is a training barrier involved in the nature of such tasks, and that humans learn useful intermediate concepts from other individuals by using a form of supervision or guidance using a curriculum. Our results provide a positive evidence in favor of this hypothesis. In our experiments, we trained a two-tiered MLP architecture on a dataset for which each input image contains three sprites, and the binary target class is 1 if all of three shapes belong to the same category and otherwise the class is 0. In terms of generalization, black-box machine learning algorithms could not perform better than chance on this task. Standard deep supervised neural networks also failed to generalize. However, using a particular structure and guiding the learner by providing intermediate targets in the form of intermediate concepts (the presence of each object) allowed us to solve the task efficiently. We obtained much better than chance, but imperfect results by exploring different architectures and optimization variants. This observation might be an indication of optimization difficulty when the neural network trained without hints on this task. We hypothesize that the learning difficulty is due to the composition of two highly non-linear tasks. Our findings are also consistent with the hypotheses on cultural learning inspired by the observations of training of neural networks sometimes getting stuck, even though good solutions exist, both in terms of training and generalization error.}
}
@article{GulwanietalFaTiPL-17,
       author = {Sumit Gulwani and Oleksandr Polozov and Rishabh Singh},
        title = {Program Synthesis},
      journal = {Foundations and Trends in Programming Languages},
       volume = {4},
       number = {1-2},
         year = {2017},
        pages = {1-119},
     abstract = {Program synthesis is the task of automatically finding a program in the underlying programming language that satisfies the user intent expressed in the form of some specification. Since the inception of AI in the 1950s, this problem has been considered the holy grail of Computer Science. Despite inherent challenges in the problem such as ambiguity of user intent and a typically enormous search space of programs, the field of program synthesis has developed many different techniques that enable program synthesis in different real-life application domains. It is now used successfully in software engineering, biological discovery, computeraided education, end-user programming, and data cleaning. In the last decade, several applications of synthesis in the field of programming by examples have been deployed in mass-market industrial products. This survey is a general overview of the state-of-the-art approaches to program synthesis, its applications, and subfields. We discuss the general principles common to all modern synthesis approaches such as syntactic bias, oracle-guided inductive search, and optimization techniques. We then present a literature review covering the four most common state-of-the-art techniques in program synthesis: enumerative search, constraint solving, stochastic search, and deduction-based programming by examples. We conclude with a brief list of future horizons for the field.}
}     
@article{GunningDARPA-18,
       author = {David Gunning},
        title = {Machine Common Sense Concept Paper},
      journal = {CoRR},
       volume = {arXiv:1810.07528},
         year = {2018},
     abstract = {This paper summarizes some of the technical background, research ideas, and possible development strategies for achieving machine common sense. Machine common sense has long been a critical-but-missing component of Artificial Intelligence (AI). Recent advances in machine learning have resulted in new AI capabilities, but in all of these applications, machine reasoning is narrow and highly specialized. Developers must carefully train or program systems for every situation. General commonsense reasoning remains elusive. The absence of common sense prevents intelligent systems from understanding their world, behaving reasonably in unforeseen situations, communicating naturally with people, and learning from new experiences. Its absence is perhaps the most significant barrier between the narrowly focused AI applications we have today and the more general, human-like AI systems we would like to build in the future. Machine common sense remains a broad, potentially unbounded problem in AI. There are a wide range of strategies that could be employed to make progress on this difficult challenge. This paper discusses two diverse strategies for focusing development on two different machine commonsense services: (1) a service that learns from experience, like a child, to construct computational models that mimic the core domains of child cognition for objects (intuitive physics), agents (intentional actors), and places (spatial navigation); and (2) service that learns from reading the Web, like a research librarian, to construct a commonsense knowledge repository capable of answering natural language and image-based questions about commonsense phenomena.}
}
@inproceedings{GuptaandNauAAAI-91,
       author = {Gupta, Naresh and Nau, Dana S.},
        title = {Complexity Results for Blocks-World Planning},
    booktitle = {Proceedings AAAI-91},
     location = {Anaheim, California},
 organization = {AAAI},
         year = 1991,
        pages = {629-633},
}
@article{GuptaetalCELL-11,
     author = {Gupta, P. B. and Fillmore, C. M. and Jiang, G. and Shapira, S. D. and Tao, K. and Kuperwasser, C. and Lander, E. S.},
      title = {{S}tochastic state transitions give rise to phenotypic equilibrium in populations of cancer cells},
    journal = {Cell},
     volume = 146,
      issue = 4,
       year = 2011,
      pages = {633-644},
}
@article{GuptaetalCoRR-15,
       author = {Isha Gupta and Alexantrou Serb and Ali Khiat and Ralf Zeitler and Stefano Vassanelli and Themistoklis Prodromakis},
        title = {Memristive integrative sensors for neuronal activity},
      journal = {CoRR},
       volume = {arXiv:1507.06832},
         year = 2015,
}
@inproceedings{GuptaetalICVLSI-10,
       author = {N. Gupta and S. K. Mandal and J. Malave and A. Mandal and R. N. Mahapatra},
    booktitle = {2010 23rd International Conference on VLSI Design},
        title = {A Hardware Scheduler for Real Time Multiprocessor System on Chip},
         year = {2010},
        pages = {264-269},
     abstract = {This paper presents the design and implementation of a low-power hardware scheduler for multiprocessor system-on-chips. The Pfair scheduling algorithm is considered with three different implementation schemes: replicated software scheduler running on each processor, single software scheduler running on a dedicated processor and the proposed hardware scheduler. Experimental evaluation with benchmarks shows that the hardware scheduler outperforms the other two schemes in terms of energy consumption by an order of magnitude of $10^5$ and scheduling delay by an order of magnitude of $10^3$.},
}
@article{GuptaetalNEURON-10,
       author = {Gupta, Anoopum S. and van der Meer, Matthijs A. A. and Touretzky, David S. and Redish, A. David},
        title = {Hippocampal replay is not a simple function of experience},
      journal = {Neuron},
         year = {2010},
       volume = {65},
        issue = {5},
        pages = {695-705},
     abstract = {Replay of behavioral sequences in the hippocampus during sharp wave ripple complexes (SWRs) provides a potential mechanism for memory consolidation and the learning of knowledge structures. Current hypotheses imply that replay should straightforwardly reflect recent experience. However, we find these hypotheses to be incompatible with the content of replay on a task with two distinct behavioral sequences (A and B). We observed forward and backward replay of B even when rats had been performing A for >10 min. Furthermore, replay of nonlocal sequence B occurred more often when B was infrequently experienced. Neither forward nor backward sequences preferentially represented highly experienced trajectories within a session. Additionally, we observed the construction of never-experienced novel-path sequences. These observations challenge the idea that sequence activation during SWRs is a simple replay of recent experience. Instead, replay reflected all physically available trajectories within the environment, suggesting a potential role in active learning and maintenance of the cognitive map.},
}
@misc{GuptaetalPATENT-06,
        title = {System and method for matching a plurality of ordered sequences with applications to call stack analysis to identify known software problems},
       author = {Rajeev Gupta and Guy Maring Lohman and Tanveer Fathima Mahmood and Laurent Sebastien Mignet and Natwar Modani and Mark Francis Wilding},
 howpublished = {https://patents.google.com/patent/US7840946},
         year = {2006},
}
@article{GureletalJCN-10,
       author = {Gurel, T. and Rotter, S. and Egert, U.},
        title = {Functional identification of biological neural networks using reservoir adaptation for point processes},
      journal = {Journal Computational Neuroscience},
         year = {2010},
       volume = {29},
       number = {1-2},
        pages = {279-299},
     abstract = {The complexity of biological neural networks does not allow to directly relate their biophysical properties to the dynamics of their electrical activity. We present a reservoir computing approach for functionally identifying a biological neural network, i.e. for building an artificial system that is functionally equivalent to the reference biological network. Employing feed-forward and recurrent networks with fading memory, i.e. reservoirs, we propose a point process based learning algorithm to train the internal parameters of the reservoir and the connectivity between the reservoir and the memoryless readout neurons. Specifically, the model is an Echo State Network (ESN) with leaky integrator neurons, whose individual leakage time constants are also adapted. The proposed ESN algorithm learns a predictive model of stimulus-response relations in in vitro and simulated networks, i.e. it models their response dynamics. Receiver Operating Characteristic (ROC) curve analysis indicates that these ESNs can imitate the response signal of a reference biological network. Reservoir adaptation improved the performance of an ESN over readout-only training methods in many cases. This also held for adaptive feed-forward reservoirs, which had no recurrent dynamics. We demonstrate the predictive power of these ESNs on various tasks with cultured and simulated biological neural networks.},
}
@article{GurneyetalBC-01a,
       author = {Gurney, K. and Prescott, T. J.  and Redgrave, P.},
        title = {A computational model of action selection in the basal ganglia. {I}. {A} new functional anatomy},
      journal = {Biological Cybernetics},
         year = {2001},
       volume = {84},
       number = {6},
        pages = {401-410},
     abstract = {We present a biologically plausible model of processing intrinsic to the basal ganglia based on the computational premise that action selection is a primary role of these central brain structures. By encoding the propensity for selecting a given action in a scalar value (the salience), it is shown that action selection may be recast in terms of signal selection. The generic properties of signal selection are defined and neural networks for this type of computation examined. A comparison between these networks and basal ganglia anatomy leads to a novel functional decomposition of the basal ganglia architecture into 'selection' and 'control' pathways. The former pathway performs the selection per se via a feedforward off-centre on-surround network. The control pathway regulates the action of the selection pathway to ensure its effective operation, and synergistically complements its dopaminergic modulation. The model contrasts with the prevailing functional segregation of basal ganglia into 'direct' and 'indirect' pathways.},
}
@article{GurneyetalBC-01b,
    author = {Gurney, K. and Prescott, T. J. and Redgrave, P.},
     title = {A computational model of action selection in the basal ganglia. {II}. {A}nalysis and simulation of behaviour},
   journal = {Biological Cybernetics},
    volume = 84,
     issue = 6,
      year = 2001,
     pages = {411-423},
  abstract = {In a companion paper a new functional architecture was proposed for the basal ganglia based on the premise that these brain structures play a central role in behavioural action selection. The current paper quantitatively describes the properties of the model using analysis and simulation. The decomposition of the basal ganglia into selection and control pathways is supported in several ways. First, several elegant features are exposed--capacity scaling, enhanced selectivity and synergistic dopamine modulation--which might be expected to exist in a well designed action selection mechanism. The discovery of these features also lends support to the computational premise of selection that underpins our model. Second, good matches between model globus pallidus external segment output and globus pallidus internal segment and substantia nigra reticulata area output, and neurophysiological data, have been found which are indicative of common architectural features in the model and biological basal ganglia. Third, the behaviour of the model as a signal selection mechanism has parallels with some kinds of action selection observed in animals under various levels of dopaminergic modulation.},
}
@article{GurneyetalNETWORK-04,
       author = {Gurney, K. N.  and Humphries, M.  and Wood, R.  and Prescott, T. J.  and Redgrave, P.},
        title = {Testing computational hypotheses of brain systems function: a case study with the basal ganglia},
      journal = {Network},
         year = {2004},
       volume = {15},
       number = {4},
        pages = {263-290},
     abstract = {We develop a methodology for testing computational hypotheses about neural functionality articulated in models at the systems level of description. In this approach, the first step is to attempt the construction of a model of the underlying brain system which is consistent with the known anatomy and physiology, but which is also able to exhibit functional properties consistent with a putative computational hypothesis. If this is successful, the second step consists of including additional known pathways into the model and testing the new models to see whether they show an improvement in functional performance (using appropriate performance metrics). A positive outcome is taken as evidence in support of the hypothesis. A final step is to construct 'control' models by including pathways that are not consistent with biological data. In this case a performance detriment is taken as support for the hypothesis. The methodology is applied to the basal ganglia, and builds on a previously published model of this system (Gurney et al 2001 Biol. Cybern. 84 401-23) which was based on the hypothesis that the basal ganglia perform action selection. The realistically constrained models show a selection benefit, while control models show a decrement in selection ability. These results, taken together, provide further validation of our selection hypothesis of basal ganglia function.},
}
@article{GutkinetalJoP-03,
       author = {Boris Gutkin and David Pinto and Bard Ermentrout},
        title = {Mathematical Neuroscience: From Neurons to Circuits to Systems},
      journal = {Journal of Physiology Paris},
         year = {2003},
       volume = {97},
        pages = {209-219},
     abstract = {Applications of mathematics and computational techniques to our understanding of neuronal systems are provided. Reduction of membrane models to simplified canonical models demonstrates how neuronal spike-time statistics follow from simple properties ofneurons. Averaging over space allows one to derive a simple model for the whisker barrel circuit and use this to explain and suggest several experiments. Spatio-temporal pattern formation methods are applied to explain the patterns seen in the early stages of drug-induced visual hallucinations.}
}
@incollection{GvozdenoivicetalELECTROPOLYMERIZATION-11,
        title = {Electrochemical Polymerization of Aniline},
       author = {Milica M. Gvozdenovi\'{c} and Branimir Z. Jugovi\'{c} and Jasmina S. Stevanovi\'{c} and Tomislav Lj. Tris\v{o}vi\'{c} and Branimir N. Grgur},
    booktitle = {Electropolymerization},
       editor = {Ewa Schab-Balcerzak},
    publisher = {InTech},
         year = 2011,
}
@article{GyglietalCoRR-17,
        title = {Deep Value Networks Learn to Evaluate and Iteratively Refine Structured Outputs},
       author = {Michael Gygli and Mohammad Norouzi and Anelia Angelova},
      journal = {CoRR},
       volume = {arXiv:1703.04363},
         year = {2017},
     abstract = {We approach structured output prediction by optimizing a deep value network (DVN) to precisely estimate the task loss on different output configurations for a given input. Once the model is trained, we perform inference by gradient descent on the continuous relaxations of the output variables to find outputs with promising scores from the value network. When applied to image segmentation, the value network takes an image and a segmentation mask as inputs and predicts a scalar estimating the intersection over union between the input and ground truth masks. For multi-label classification, the DVN's objective is to correctly predict the F1 score for any potential label configuration. The DVN framework achieves the state-of-the-art results on multi-label prediction and image segmentation benchmarks.}
}
@inproceedings{GyglietalICML-17,
        title = {Deep Value Networks Learn to Evaluate and Iteratively Refine Structured Outputs},
       author = {Michael Gygli and Mohammad Norouzi and Anelia Angelova},
    booktitle = {Proceedings of the 34th International Conference on Machine Learning},
        pages = {1341-1351},
         year = {2017},
       editor = {Doina Precup and Yee Whye Teh},
       volume = {70},
       series = {Proceedings of Machine Learning Research},
    publisher = {PMLR},
     abstract = {We approach structured output prediction by optimizing a deep value network (DVN) to precisely estimate the task loss on different output configurations for a given input. Once the model is trained, we perform inference by gradient descent on the continuous relaxations of the output variables to find outputs with promising scores from the value network. When applied to image segmentation, the value network takes an image and a segmentation mask as inputs and predicts a scalar estimating the intersection over union between the input and ground truth masks. For multi-label classification, the DVN's objective is to correctly predict the F1 score for any potential label configuration. The DVN framework achieves the state-of-the-art results on multi-label prediction and image segmentation benchmarks.}
}
@article{GyohdaetalJoB-00,
       author = {Gyohda, Atsuko and Komano, Teruya},
        title = {Purification and Characterization of the R64 Shufflon-Specific Recombinase},
      journal = {Journal of Bacteriology},
       volume = 182,
        issue = 10,
         year = 2000,
        pages = {2787-2792},
     abstract = {The shufflon, a multiple DNA inversion system in plasmid R64, consists of four invertible DNA segments which are separated and flanked by seven 19-bp repeat sequences. The product of a site-specific recombinase gene, RCI, promotes site-specific recombination between any two of the inverted 19-bp repeat sequences of the shufflon. To analyze the molecular mechanism of this recombination reaction, RCI protein was overproduced and purified. The purified RCI protein promoted the in vitro recombination reaction between the inverted 19-bp repeats of supercoiled DNA of a plasmid carrying segment A of the R64 shufflon. The recombination reaction was enhanced by the bacterial host factor HU. Gel electrophoretic analysis indicated that the RCI protein specifically binds to the DNA segments carrying the 19-bp sequences. The binding affinity of the RCI protein to the four shufflon segments as well as four synthetic 19-bp sequences differed greatly: among the four 19-bp repeat sequences, the repeat-a and -d sequences displayed higher affinity to RCI protein. These results suggest that the differences in the affinity of RCI protein for the 19-bp repeat sequences determine the inversion frequencies of the four segments.},
}
@article{HaandSchmidhuberCoRR-18,
       author = {David Ha and J{\"{u}}rgen Schmidhuber},
        title = {World Models},
      journal = {CoRR},
       volume = {arXiv:1803.10122},
         year = {2018},
     abstract = {We explore building generative neural network models of popular reinforcement learning environments. Our world model can be trained quickly in an unsupervised manner to learn a compressed spatial and temporal representation of the environment. By using features extracted from the world model as inputs to an agent, we can train a very compact and simple policy that can solve the required task. We can even train our agent entirely inside of its own hallucinated dream generated by its world model, and transfer this policy back into the actual environment. An interactive version of this paper is available at this {\urlh{https://worldmodels.github.io/}{URL}}},
}
@article{HaarandCoussiosHIFU-07,
       author = {Gail ter Haar and Constantin Coussios},
        title = {High Intensity Focused Ultrasound (HIFU): Physical principles and devices},
      journal = {International Journal of Hyperthermia},
       volume = 23,
        issue = 2,
         year = 2007,
        pages = {89-104},
          url = {http://informahealthcare.com/doi/pdf/10.1080/02656730601186138},
} 
@inproceedings{HaberetalICWIBSN-14,
       author = {D. Haber and A. A. C. Thomik and A. A. Faisal},
    booktitle = {2014 11th International Conference on Wearable and Implantable Body Sensor Networks},
        title = {Unsupervised Time Series Segmentation for High-Dimensional Body Sensor Network Data Streams},
         year = {2014},
        pages = {121-126},
     abstract = {The vast amounts of data which can be collected using body-sensor networks with high temporal and spatial resolution require a novel analysis approach. In this context, state-of-the-art Bayesian approaches based on variational, non-parametric or MCMC derived methods often become computationally intractable when faced with several million data points. Here, we present how the simple combination of PCA, approximate Bayesian segmentation and temporal correlation processing can achieve reliable time series segmentation. We use our method, which relies on simple iterative covariance, correlation and maximum likelihood operations, to perform complex behavioural time series segmentation over millions of samples in 18 dimensions in linear time and space. Our approach is suitable for even higher dimensional data streams as performance scales near constantly with the dimensionality of the time series samples. We validate this novel approach on an artificially-generated time series and demonstrate that our method is very robust to noise and achieves a segmentation accuracy of over 86\% of matching segments against ground-truth. We conclude that our approach makes Big Data driven approaches to stream processing Body Sensor Network (BSN) data tractable, and is required for BSN-driven Neurotechnology applications in Brain-Machine Interfacing and Neuroprosthetics.}
}
@article{HackamandRedelmeirJAMA-06,
       author = {Daniel G. Hackam and Donald A. Redelmeier},
        title = {Translation of Research Evidence From Animals to Humans},
      journal = {Journal of the American Medical Association},
       volume = 296,
        issue = 14,
         year = 2006,
        pages = {1731-1732},
}
@book{Hadamard1945,
        title = {The Mathematician's Mind: The Psychology of Invention in the Mathematical Field},
       author = {Hadamard, Jacques},
         year = {1945},
    publisher = {Princeton University Press}
}
@inproceedings{Haddawy,
       author = {Haddawy, Peter},
        title = {Time, Chance, and Action},
    booktitle = {Proceedings of the Sixth Workshop on Uncertainty in {AI}},
     location = {Cambridge, Massachusetts},
         year = 1990,
        pages = {147-154},
}
@inproceedings{Haddawy91,
       author = {Haddawy, Peter},
        title = {A temporal probability logic for representing actions},
    booktitle = {Principles of Knowledge Representation and Reasoning: Proceedings of the Second International Conference},
     location = {Cambridge, Massachusetts},
        pages = {313-324},
         year = 1991,
}
@inproceedings{HaddawyDARPA-90,
       author = {Haddawy, Peter and Hanks, Steve},
        title = {Issues in Decision-Theoretic Planning: Symbolic Goals and Numeric Utilities},
    booktitle = {Proceedings of the DARPA Workshop on Innovative Approaches to Planning, Scheduling, and Control},
 organization = {DARPA},
        pages = {48-58},
         year = 1990,
}
@inproceedings{HaddawyandSuwandiSSS-94,
       author = {Haddawy, Peter and Suwandi, Meliani},
        title = {Decision-Theoretic Planning using Inheritance Abstraction},
    booktitle = {Proceedings of the {AAAI} Spring Symposium on Decision Theoretic Planning},
     location = {Stanford, California},
         year = 1994,
}
@inproceedings{HadselletalCVPR-06,
       author = {Raia Hadsell and Sumit Chopra and Yann LeCun},
        title = {Dimensionality Reduction by Learning an Invariant Mapping},
    booktitle = {Proceedings of {IEEE} Computer Vision and Pattern Recognition},
    publisher = {{IEEE} Computer Society},
       volume = {2},
        pages = {1735-1742},
         year = 2006
}
@book{Haeckel1883,
        title = {The History of Creation},
      comment = {E. Ray Lankester},
       author = {Ernst Haeckel},
    publisher = {Kegan Paul, Trench, Tr\"{u}bner \& Co.},
      address = {London},
         year = 1883,
}
@article{HaesemeyeretalCELL-15,
       author = {Haesemeyer, Martin and Robson, Drew N. and Li, Jennifer M. and Schier, Alexander F. and Engert, Florian},
        title = {The Structure and Timescales of Heat Perception in Larval Zebrafish},
      journal = {Cell Systems},
    publisher = {Elsevier},
       volume = {1},
        issue = {5},
         year = {2015},
        pages = {338-348},
     abstract = {Avoiding temperatures outside the physiological range is critical for animal survival, but how temperature dynamics are transformed into behavioral output is largely not understood. Here, we used an infrared laser to challenge freely swimming larval zebrafish with ?white noise? heat stimuli and built quantitative models relating external sensory information and internal state to behavioral output. These models revealed that larval zebrafish integrate temperature information over a time-window of 400 ms preceding a swim bout and that swimming is suppressed right after the end of a bout. Our results suggest that larval zebrafish compute both an integral and a derivative across heat in time to guide their next movement. Our models put important constraints on the type of computations that occur in the nervous system and reveal principles of how somatosensory temperature information is processed to guide behavioral decisions such as sensitivity to both absolute levels and changes in stimulation.}
}
@article{HafneretalCoRR-19,
        title = {Learning Latent Dynamics for Planning from Pixels},
       author = {Hafner, Danijar and Lillicrap, Timothy and Fischer, Ian and Villegas, Ruben and Ha, David and Lee, Honglak and Davidson, James},
      journal = {CoRR},
       volume = {arXiv:1811.04551},
         year = {2019},
     abstract = {Planning has been very successful for control tasks with known environment dynamics. To leverage planning in unknown environments, the agent needs to learn the dynamics from interactions with the world. However, learning dynamics models that are accurate enough for planning has been a long-standing challenge, especially in image-based domains. We propose the Deep Planning Network (PlaNet), a purely model-based agent that learns the environment dynamics from images and chooses actions through fast online planning in latent space. To achieve high performance, the dynamics model must accurately predict the rewards ahead for multiple time steps. We approach this problem using a latent dynamics model with both deterministic and stochastic transition components and a multi-step variational inference objective that we call latent overshooting. Using only pixel observations, our agent solves continuous control tasks with contact dynamics, partial observability, and sparse rewards, which exceed the difficulty of tasks that were previously solved by planning with learned models. PlaNet uses substantially fewer episodes and reaches final performance close to and sometimes higher than strong model-free algorithms.}
}
@book{HagemanandYoung81,
       author = {Hageman, L. A. and Young, D. M.},
        title = {Applied Iterative Methods},
    publisher = {Academic Press},
         year = 1981,
}
@book{Hager,
       author = {Hager, Gregory D.},
        title = {Task-Directed Sensor Fusion and Planning: A Computational Approach},
    publisher = {Kluwer},
      address = {Boston, Massachusetts},
         year = 1990,
}
@techreport{HagerPhD,
       author = {Hager, Gregory D.},
        title = {Active Reduction of Uncertainty in Multi-Sensor Systems},
         type = {{Ph.D.} Thesis},
  institution = {University of Pennsylvania, Department of Computer and Information Science},
         year = 1988,
}
@article{HagmannetalPLoS-08,
       author = {Hagmann, Patric and Cammoun, Leila and Gigandet, Xavier and Meuli, Reto and Honey, Christopher J. and Wedeen, Van J. and Sporns, Olaf},
      journal = {{PLoS} Biology},
    publisher = {Public Library of Science},
        title = {Mapping the Structural Core of Human Cerebral Cortex},
       volume = 6,
        pages = {e159},
     abstract = {Mapping of major structural connections of the human cortex reveals a core of brain regions that are highly interconnected and highly central with respect to the rest of the brain.},
        issue = 7,
         year = 2008,
}
@article{HahnloseretalNATURE-00,
       author = {Hahnloser, Richard H. R. and Sarpeshkar, Rahul and Mahowald, Misha A. and Douglas, Rodney J. and Seung, H. Sebastian},
        title = {Digital selection and analogue amplification coexist in a cortex-inspired silicon circuit},
      journal = {Nature},
    publisher = {Macmillian Magazines Ltd.},
       volume = {405},
        issue = {6789},
         year = {2000},
        pages = {947-951},
     abstract = {Digital circuits such as the flip-flop use feedback to achieve multi-stability and nonlinearity to restore signals to logical levels, for example 0 and 1. Analogue feedback circuits are generally designed to operate linearly, so that signals are over a range, and the response is unique. By contrast, the response of cortical circuits to sensory stimulation can be both multistable and graded. We propose that the neocortex combines digital selection of an active set of neurons with analogue response by dynamically varying the positive feedback inherent in its recurrent connections. Strong positive feedback causes differential instabilities that drive the selection of a set of active neurons under the constraints embedded in the synaptic weights. Once selected, the active neurons generate weaker, stable feedback that provides analogue amplification of the input. Here we present our model of cortical processing as an electronic circuit that emulates this hybrid operation, and so is able to perform computations that are similar to stimulus selection, gain modulation and spatiotemporal pattern generation in the neocortex.},
}
@unpublished {Hakkani-TurandHeckICASP-14,
         note = {Interspeech 2014 tutorial slides},
       author = {Dilek Hakkani-Tur and Larry Heck},
        title = {Semantic Web and Linked Big Data Resources for Spoken Language Processing},
          url = {http://research.microsoft.com/apps/pubs/default.aspx?id=228809},
         year = 2014,
        month = {September},
}
@misc{HalassaCBMM-18,
       author = {Michael Halasa},
        title = {The Thalamus in Cognitive Control and Flexibility},
 howpublished = {https://cbmm.mit.edu/video/thalamus-cognitive-control-and-flexibility-11921},
         year = {2018}
}
@article{HalassaandShermanNEURON-19,
       author = {Michael M. Halassa and S. Murray Sherman},
        title = {Thalamocortical Circuit Motifs: A General Framework},
      journal = {Neuron},
       volume = {103},
       number = {5},
         year = {2019},
        pages = {762-770},
     abstract = {The role of the thalamus in cortical sensory transmission is well known, but its broader role in cognition is less appreciated. Recent studies have shown thalamic engagement in dynamic regulation of cortical activity in attention, executive control, and perceptual decision-making, but the circuit mechanisms underlying such functionality are unknown. Because the thalamus is composed of excitatory neurons that are devoid of local recurrent excitatory connectivity, delineating long-range, input-output connectivity patterns of single thalamic neurons is critical for building functional models. We discuss this need in relation to existing organizational schemes such as core versus matrix and first-order versus higher-order relay nuclei. We propose that a new classification is needed based on thalamocortical motifs, where structure naturally informs function. Overall, our synthesis puts understanding thalamic organization at the forefront of existing research in systems and computational neuroscience, with both basic and translational applications.}
}
@article{HalassaetalCELL-14,
       author = {Halassa, M.M. and Chen, Z. and Wimmer, R.D. and Brunetti, P.M. and Zhao, S. and Zikopoulos, B. and Wang, F. and Brown, E.N. and Wilson, M.A.},
        title = {State-dependent architecture of thalamic reticular subnetworks},
      journal = {Cell},
       volume = {158},
       number = {4},
         year = {2014},
        pages = {808-821},
     abstract = {Behavioral state is known to influence interactions between thalamus and cortex, which are important for sensation, action, and cognition. The thalamic reticular nucleus (TRN) is hypothesized to regulate thalamo-cortical interactions, but the underlying functional architecture of this process and its state dependence are unknown. By combining the first TRN ensemble recording with psychophysics and connectivity-based optogenetic tagging, we found reticular circuits to be composed of distinct subnetworks. While activity of limbic-projecting TRN neurons positively correlates with arousal, sensory-projecting neurons participate in spindles and show elevated synchrony by slow waves during sleep. Sensory-projecting neurons are suppressed by attentional states, demonstrating that their gating of thalamo-cortical interactions is matched to behavioral state. Bidirectional manipulation of attentional performance was achieved through subnetwork-specific optogenetic stimulation. Together, our findings provide evidence for differential inhibition of thalamic nuclei across brain states, where the TRN separately controls external sensory and internal limbic processing facilitating normal cognitive function. PAPERFLICK:}
} 
@book{Hale73,
       author = {Hale, Francis J.},
        title = {Introduction to Control System Analysis and Design},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1973,
}
@article{HalleandStevensIRE-62,
       author = {M. Halle and K. Stevens},
      journal = {{IRE} Transactions on Information Theory},
        title = {Speech recognition: A model and a program for research},
         year = {1962},
       volume = {8},
       number = {2},
        pages = {155-159},
     abstract = {A speech recognition model is proposed in which the transformation from an input speech signal into a sequence of phonemes is carried out largely through an active or feedback process. In this process, patterns are generated internally in the analyzer according to an adaptable sequence of instructions until a best match with the input signal is obtained. Details of the process are given, and the areas where further research is needed are indicated.}
}
@article{HallemandCarlsonNEURON-04,
       author = {Elissa A Hallem and John R Carlson},
        title = {The Spatial Code for Odors Is Changed by Conditioning},
      journal = {Neuron},
       volume = {42},
       number = {3},
         year = {2004},
        pages = {359-361},
     abstract = {Drosophila displays robust olfactory learning. In this issue of Neuron, Davis and colleagues use optical imaging to demonstrate that olfactory conditioning alters odor representations among the glomeruli of the fly antennal lobe. New synaptic activity is recruited into the representation of an odor and is localized to projection neurons (PNs).}
}
@article{HallworthandBlandEXPERIMENTAL-NEUROLOGY-04,
       author = {Hallworth, N. E. and Bland, B. H.},
        title = {Basal ganglia-hippocampal interactions support the role of the hippocampal formation in sensorimotor integration},
      journal = {Experimental Neurology},
       volume = {188},
       number = {2},
         year = {2004},
        pages = {430-443},
     abstract = {Experiments were carried out to evaluate whether neural activity in the basal ganglia is functionally related to the neural activity underlying mechanisms of theta band oscillation and synchrony in the hippocampal formation. Experiment 1 demonstrated that electrical stimulation administered to the substantia nigra, globus pallidus (GP) and caudate-putamen (CPu) in urethane anesthetized rats elicited theta field activity in the hippocampal formation. Subsequent microinfusion of the local anesthetic procaine hydrochloride into the medial septum reversibly abolished this effect. In Experiment 2, single cell discharge profiles established for 152 cells recorded in nuclei of the basal ganglia resulted in 101 (66\%) being classified as theta-related and 51 (34\%) classified as nonrelated. Theta-related cells were further subclassified as tonic theta-ON cells (n = 79) and tonic theta-OFF (n = 22). Tonic theta-ON and tonic theta-OFF cells displayed irregular or regular (tonic) discharge patterns. Rhythmic discharge patterns did not occur in any theta-related cells in the nuclei of the basal ganglia. However, analyses using Kaneoke and Vitek's [J. Neurosci. Methods 68, (1996) 211] algorithms revealed that 51/101 (50\%) theta-related cells displayed periodicity in their discharge patterns whereas 27/51 (53\%) of the nonrelated cells displayed periodicity in their discharge patterns. The periodicities in the majority of cells were in frequency ranges above that of theta band oscillation and synchrony. The results support the following conclusions: (1) the cellular activity of the basal ganglia, composed of nuclei traditionally associated with motor functions, is functionally connected with the neural circuitry involved in the generation of theta band oscillation and synchrony in the hippocampal formation; (2) the observed functional connectivity provides support for the role of the hippocampal formation in sensorimotor integration.}
}
@article{HalpernAIJ-90,
       author = {Halpern, Joseph Y.},
        title = {An analysis of first-order logics of probability},
      journal = AIJ,
       volume = 46,
         year = 1990,
        pages = {311-350},
}
@inproceedings{HalpernIJCAI-89,
       author = {Halpern, Joseph},
        title = {An Analysis of First-Order Logics of Probability},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
         year = 1989,
        pages = {1375-1381},
}
@inproceedings{HalpernSTOC-84,
       author = {Halpern, J. Y. and M. Rabin},
        title = {A Logic to Reason About Likelihood},
    booktitle = {Proceedings of the {ACM} Fifteenth Annual Symposium on Theory of Computing},
         year = 1984,
}
@book{HalpernTARK-86a,
       editor = {Halpern, Joseph Y.},
        title = {Theoretical Aspects of Reasoning About Knowledge: Proceedings of the 1986 Conference},
    booktitle = {Theoretical Aspects of Reasoning About Knowledge: Proceedings ofthe 1986 Conference},
         year = 1986,
    publisher = {Morgan Kaufmann},
}
@inproceedings{HalpernTARK-86b,
       author = {Halpern, Joseph Y.},
        title = {Reasoning About Knowledge: An Overview},
     crossref = {HalpernTARK-86a},
        pages = {1-17},
}
@inproceedings{HalpernandMosesIJCAI-85,
       author = {Halpern, Joseph and Moses, Yoram},
        title = {A Guide to the Modal Logics of Knowledge and Belief},
    booktitle = {Proceedings {IJCAI} 9},
     location = {Los Angeles, California},
 organization = {IJCAII},
         year = 1985,
        pages = {480-490},
}
@inproceedings{HalpernandShoham,
       author = {Halpern, Joseph and Shoham, Yoav},
        title = {A Propositional Modal Logic of Time Intervals},
    booktitle = {Proceedings of the Conference on Logic in Computer Science},
     location = {Cambridge, Massachusetts},
         year = 1986,
        pages = {279-292},
}
@article{HameletalNEURON-15,
       author = {Hamel, E. J. and Grewe, B. F. and Parker, J. G. and Schnitzer, M. J.},
        title = {Cellular level brain imaging in behaving mammals: an engineering approach},
      journal = {Neuron},
         year = {2015},
       volume = {86},
       number = {1},
        pages = {140-159},
     abstract = {Fluorescence imaging offers expanding capabilities for recording neural dynamics in behaving mammals, including the means to monitor hundreds of cells targeted by genetic type or connectivity, track cells over weeks, densely sample neurons within local microcircuits, study cells too inactive to isolate in extracellular electrical recordings, and visualize activity in dendrites, axons, or dendritic spines. We discuss recent progress and future directions for imaging in behaving mammals from a systems engineering perspective, which seeks holistic consideration of fluorescent indicators, optical instrumentation, and computational analyses. Today, genetically encoded indicators of neural Ca(2+) dynamics are widely used, and those of trans-membrane voltage are rapidly improving. Two complementary imaging paradigms involve conventional microscopes for studying head-restrained animals and head-mounted miniature microscopes for imaging in freely behaving animals. Overall, the field has attained sufficient sophistication that increased cooperation between those designing new indicators, light sources, microscopes, and computational analyses would greatly benefit future progress.}
}
@inproceedings{HametalICML-04,
       author = {Ham, Jihun and Lee, Daniel D. and Mika, Sebastian and Sch\"{o}lkopf, Bernhard},
        title = {A Kernel View of the Dimensionality Reduction of Manifolds},
    booktitle = {Proceedings of the Twenty-first International Conference on Machine Learning},
    publisher = {ACM},
         year = 2004,
     location = {Banff, Alberta, Canada},
        pages = {369-376},
     abstract = {We interpret several well-known algorithms for dimensionality reduction of manifolds as kernel methods. Isomap, graph Laplacian eigenmap, and locally linear embedding (LLE) all utilize local neighborhood information to construct a global embedding of the manifold. We show how all three algorithms can be described as kernel PCA on specially constructed Gram matrices, and illustrate the similarities and differences between the algorithms with representative examples.},
}
@article{HamiltonAAD-11,
        title = {Memory skills of deaf learners: Implications and applications},
       author = {Hamilton, Harley},
      journal = {American Annals of the Deaf},
       volume = 156,
         year = 4,
        pages = 2011,
}
@inproceedings{HamiltonetalNIPS-17,
        title = {Inductive Representation Learning on Large Graphs},
       author = {Hamilton, William L. and Ying, Rex and Leskovec, Jure},
    booktitle = {Proceedings of the 31st International Conference on Neural Information Processing Systems},
    publisher = {Curran Associates Inc.},
      address = {Red Hook, NY, USA},
         year = {2017},
        pages = {1025-1035},
     abstract = {Low-dimensional embeddings of nodes in large graphs have proved extremely useful in a variety of prediction tasks, from content recommendation to identifying protein functions. However, most existing approaches require that all nodes in the graph are present during training of the embeddings; these previous approaches are inherently transductive and do not naturally generalize to unseen nodes. Here we present GraphSAGE, a general inductive framework that leverages node feature information (e.g., text attributes) to efficiently generate node embeddings for previously unseen data. Instead of training individual embeddings for each node, we learn a function that generates embeddings by sampling and aggregating features from a node's local neighborhood. Our algorithm outperforms strong baselines on three inductive node-classification benchmarks: we classify the category of unseen nodes in evolving information graphs based on citation and Reddit post data, and we show that our algorithm generalizes to completely unseen graphs using a multi-graph dataset of protein-protein interactions.},
}
@book{Hammond89,
       author = {Hammond, Kristian J.},
        title = {Case-Based Planning},
    publisher = {Academic Press},
      address = {New York},
         year = 1989,
}
@inproceedings{HammondAAAI-86,
       author = {Hammond, Kris},
        title = {CHEF: A Model of Case-based Planning},
    booktitle = {Proceedings AAAI-86},
     location = {Philadelphia, Pennsylvania},
 organization = {AAAI},
         year = 1986,
        pages = {267-271},
}
@article{HamrickCURRENT-OPINION-19,
        title = {Analogues of mental simulation and imagination in deep learning},
       author = {Jessica B Hamrick},
      journal = {Current Opinion in Behavioral Sciences},
       volume = {29},
        pages = {8-16},
         year = {2019},
     abstract = {Mental simulation-the capacity to imagine what will or what could be-is a salient feature of human cognition, playing a key role in a wide range of cognitive abilities. In artificial intelligence, the last few years have seen the development of methods which are analogous to mental models and mental simulation. This paper outlines recent methods in deep learning for constructing such models from data and learning to use them via reinforcement learning, and compares such approaches to human mental simulation. Model-based methods in deep learning can serve as powerful tools for building and scaling cognitive models. However, a number of challenges remain in matching the capacity of human mental simulation for efficiency, compositionality, generalization, and creativity.}
}
@inproceedings{HamricketalCSS-18,
       author = {Jessica B. Hamrick and Kelsey R. Allen and Victor Bapst and Tina Zhu and Kevin R. McKee and Joshua B. Tenenbaum and Peter W. Battaglia},
        title = {Relational inductive bias for physical construction in humans and machines},
    booktitle = {Proceedings of the 40th Annual Conference of the Cognitive Science Society},
         year = {2018},
        pages = {1773-1779},
     abstract = {While current deep learning systems excel at tasks such as object classification, language processing, and gameplay, few can construct or modify a complex system such as a tower of blocks. We hypothesize that what these systems lack is a "relational inductive bias": a capacity for reasoning about inter-object relations and making choices over a structured description of a scene. To test this hypothesis, we focus on a task that involves gluing pairs of blocks together to stabilize a tower, and quantify how well humans perform. We then introduce a deep reinforcement learning agent which uses object- and relation-centric scene and policy representations and apply it to the task. Our results show that these structured representations allow the agent to outperform both humans and more naive approaches, suggesting that relational inductive bias is an important component in solving structured reasoning problems and for building more intelligent, flexible machines.}
}
@article{HamricketalCoRR-17,
       author = {Jessica B. Hamrick and Andrew J. Ballard and Razvan Pascanu and Oriol Vinyals and Nicolas Heess and Peter W. Battaglia},
        title = {Metacontrol for Adaptive Imagination-Based Optimization},
      journal = {CoRR},
       volume = {arXiv:1705.02670},
         year = {2017},
     abstract = {Many machine learning systems are built to solve the hardest examples of a particular task, which often makes them large and expensive to run---especially with respect to the easier examples, which might require much less computation. For an agent with a limited computational budget, this "one-size-fits-all" approach may result in the agent wasting valuable computation on easy examples, while not spending enough on hard examples. Rather than learning a single, fixed policy for solving all instances of a task, we introduce a metacontroller which learns to optimize a sequence of "imagined" internal simulations over predictive models of the world in order to construct a more informed, and more economical, solution. The metacontroller component is a model-free reinforcement learning agent, which decides both how many iterations of the optimization procedure to run, as well as which model to consult on each iteration. The models (which we call "experts") can be state transition models, action-value functions, or any other mechanism that provides information useful for solving the task, and can be learned on-policy or off-policy in parallel with the metacontroller. When the metacontroller, controller, and experts were trained with "interaction networks" (Battaglia et al., 2016) as expert models, our approach was able to solve a challenging decision-making problem under complex non-linear dynamics. The metacontroller learned to adapt the amount of computation it performed to the difficulty of the task, and learned how to choose which experts to consult by factoring in both their reliability and individual computational resource costs. This allowed the metacontroller to achieve a lower overall cost (task loss plus computational cost) than more traditional fixed policy approaches.}
}
@article{HamricketalCoRR-18,
       author = {Jessica B. Hamrick and Kelsey R. Allen and Victor Bapst and Tina Zhu and Kevin R. McKee and Joshua B. Tenenbaum and Peter W. Battaglia},
        title = {Relational inductive bias for physical construction in humans and machines},
      journal = {CoRR},
      comment = {Proceedings of the 40th Annual Conference of the Cognitive Science Society},
         year = {2018},
       volume = {abs/1806.01203},
}
@article{HamzeietalCC-16,
       author = {Hamzei, Farsin and Vry, Magnus-Sebastian and Saur, Dorothee and Glauche, Volkmar and Hoeren, Markus and Mader, Irina and Weiller, Cornelius and Rijntjes, Michel},
        title = {The Dual-Loop Model and the Human Mirror Neuron System: an Exploratory Combined fMRI and DTI Study of the Inferior Frontal Gyrus},
      journal = {Cerebral Cortex},
       volume = {26},
       number = {5},
        pages = {2215-2224},
         year = {2016},
     abstract = {The inferior frontal gyrus (IFG) is active during both goal-directed action and while observing the same motor act, leading to the idea that also the meaning of a motor act (action understanding) is represented in this "mirror neuron system" (MNS). However, in the dual-loop model, based on dorsal and ventral visual streams, the MNS is thought to be a function of the dorsal steam, projecting to pars opercularis (BA44) of IFG, while recent studies suggest that conceptual meaning and semantic analysis are a function of ventral connections, projecting mainly to pars triangularis (BA45) of IFG. To resolve this discrepancy, we investigated action observation (AO) and imitation (IMI) using fMRI in a large group of subjects. A grasping task (GR) assessed the contribution from movement without AO. We analyzed connections of the MNS-related areas within IFG with postrolandic areas with the use of activation-based DTI. We found that action observation with imitation are mainly a function of the dorsal stream centered on dorsal part of BA44, but also involve BA45, which is dorsally and ventrally connected to the same postrolandic regions. The current finding suggests that BA45 is the crucial part where the MNS and the dual-loop system interact.}
}
@book{Handbook_of_Brain_Microcircuits2010,
        title = {Handbook of Brain Microcircuits},
       author = {Shepherd, G.M. and Grillner, S.},
         year = {2018},
    publisher = {Oxford University Press}
}
@article{HaneschetalCTR-89,
       author = {Hanesch, U. and Fischbach, K. F. and Heisenberg, M.},
        title = {Neuronal architecture of the central complex in {D}rosophila melanogaster},
      journal = {Cell and Tissue Research},
         year = {1989},
       volume = {257},
       number = {2},
        pages = {343-366},
     abstract = {On the basis of 1200 Golgi-impregnated brains the structure of the central complex of Drosophila melanogaster was analyzed at the cellular level. The four substructures of the central complex --- the ellipsoid body, the fanshaped body, the noduli, and the protocerebral bridge --- are composed of (a) columnar small-field elements linking different substructures or regions in the same substructure and (b) tangential large-field neurons forming strata perpendicular to the columns. At least some small-field neurons belong to isomorphic sets, which follow various regular projection patterns. Assuming that the blebs of a neuron are presynaptic and the spines are postsynaptic, the Golgi preparations indicate that small-field neurons projecting to the ventral bodies (accessory area) are the main output from the central complex and that its main input is through the large-field neurons. These in turn are presumed to receive input in various neuropils of the brain including the ventral bodies. Transmitters can be attributed immunocytochemically to some neuron types. For example, GABA is confined to the R1--R4 neurons of the ellipsoid body, whereas these cells are devoid of choline acetyltransferase-like immunore-activity. It is proposed that the central complex is an elaboration of the interhemispheric commissure serving the fast exchange of data between the two brain hemispheres in the control of behavioral activity.},
}
@article{HanetalCELL-13,
        title = {Forebrain Engraftment by Human Glial Progenitor Cells Enhances Synaptic Plasticity and Learning in Adult Mice},
       author = {Han, Xiaoning and Chen, Michael and Wang, Fushun and Windrem, Martha and Wang, Su and Shanz, Steven and Xu, Qiwu and Oberheim, Nancy Ann and Bekar, Lane and Betstadt, Sarah and Silva, Alcino J. and Takano, Takahiro and Goldman, Steven A. and Nedergaard, Maiken},
      journal = {Cell Stem Cell},
    publisher = {Cell Press},
       volume = 12,
        issue = 3,
         year = 2013,
        pages = {342-353},
}
@article{HanetalNATURE-01,
        title = {Quantum-dot-tagged microbeads for multiplexed optical coding of biomolecules},
       author = {Han, M. and Gao, X. and Su, J.Z. and Nie, S.},
      journal = {Nature Biotechnology},
       volume = 19,
        issue = 7,
        pages = {631-635},
         year = 2001,
     abstract = {Multicolor optical coding for biological assays has been achieved by embedding different-sized quantum dots (zinc sulfide-capped cadmium selenide nanocrystals) into polymeric microbeads at precisely controlled ratios. Their novel optical properties (e.g., size-tunable emission and simultaneous excitation) render these highly luminescent quantum dots (QDs) ideal fluorophores for wavelength-and-intensity multiplexing. The use of 10 intensity levels and 6 colors could theoretically code one million nucleic acid or protein sequences. Imaging and spectroscopic measurements indicate that the QD-tagged beads are highly uniform and reproducible, yielding bead identification accuracies as high as 99.99\% under favorable conditions. DNA hybridization studies demonstrate that the coding and target signals can be simultaneously read at the single-bead level. This spectral coding technology is expected to open new opportunities in gene expression studies, high-throughput screening, and medical diagnostics.}
}
@article{HanetalNATURE-18,
       author = {Han, Yunyun and Kebschull, Justus M. and Campbell, Robert A. A. and Cowan, Devon and Imhof, Fabia and Zador, Anthony M. and Mrsic-Flogel, Thomas D.},
        title = {The logic of single-cell projections from visual cortex},
      journal = {Nature},
         year = {2018},
    publisher = {Macmillan Publishers Limited},
       volume = {556},
        pages = {51-56},
     abstract = {Neocortical areas communicate through extensive axonal projections, but the logic of information transfer remains poorly understood, because the projections of individual neurons have not been systematically characterized. It is not known whether individual neurons send projections only to single cortical areas or distribute signals across multiple targets. Here we determine the projection patterns of 591 individual neurons in the mouse primary visual cortex using whole-brain fluorescence-based axonal tracing and high-throughput DNA sequencing of genetically barcoded neurons (MAPseq). Projections were highly diverse and divergent, collectively targeting at least 18 cortical and subcortical areas. Most neurons targeted multiple cortical areas, often in non-random combinations, suggesting that sub-classes of intracortical projection neurons exist. Our results indicate that the dominant mode of intracortical information transfer is not based on 'one neuron-one target area' mapping. Instead, signals carried by individual cortical neurons are shared across subsets
of target areas, and thus concurrently contribute to multiple functional pathways}
}
@inproceedings{HangaletalUIST-11,
        title = {Muse: Reviving memories using email archives},
       author = {Sudheendra Hangal, Monica S. Lam, Jeffrey Heer},
    booktitle = {Proceedings of the 24th Symposium on User Interface Software and Technology},
         year = 2011,
        pages = {75-84},
}
@inproceedings{HanksAAAI-86,
       author = {Hanks, Steve and Drew V. McDermott},
        title = {Default Reasoning, Nonmonotonic Logics, and the Frame Problem},
    booktitle = {Proceedings AAAI-86},
     location = {Philadelphia, Pennsylvania},
 organization = {AAAI},
         year = 1986,
        pages = {328-333},
}
@inproceedings{HanksAAAI-88,
       author = {Hanks, Steve},
        title = {Representing and Computing Temporally Scoped Beliefs},
    booktitle = {Proceedings AAAI-88},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
        pages = {501-505},
         year = 1988,
}
@article{HanksAIJ-87,
       author = {Hanks, Steve and Drew V. McDermott},
        title = {Nonmonotonic Logic and Temporal Projection},
      journal = AIJ,
       volume = 33,
         year = 1987,
        pages = {379-412},
}
@article{HanksandMcDermottAIJ-94,
       author = {Hanks, Steve and Drew V. McDermott},
        title = {Modeling a Dynamic and Uncertain World {I}: Symbolic and Probabilistic Reasoning About Change},
      journal = AIJ,
       volume = 66,
        pages = {1-55},
        issue = 1,
         year = 1994,
}
@inproceedings{HanksandWeldAIPS-92,
       author = {Hanks, Steven and Weld, Daniel S.},
        title = {Systematic Adaptation for Case-Based Planning},
    booktitle = {Proceedings of the First International Conference on Artificial Intelligence Planning Systems},
        pages = {96-105},
         year = 1992,
}
@article{HannaetalCELL-10,
    author = {Hanna, Jacob H. and Saha, Krishanu and Jaenisch, Rudolf},
     title = {Pluripotency and Cellular Reprogramming: Facts, Hypotheses, Unresolved Issues},
   journal = {Cell},
 publisher = {Cell Press},
    volume = 143,
     issue = 4,
     pages = {508-525},
      year = 2010,
  abstract = {Direct reprogramming of somatic cells to induced pluripotent stem cells by ectopic expression of defined transcription factors has raised fundamental questions regarding the epigenetic stability of the differentiated cell state. In addition, evidence has accumulated that distinct states of pluripotency can interconvert through the modulation of both cell-intrinsic and exogenous factors. To fully realize the potential of in vitro reprogrammed cells, we need to understand the molecular and epigenetic determinants that convert one cell type into another. Here we review recent advances in this rapidly moving field and emphasize unresolved and controversial questions.}
} 
@inproceedings{HansenCUAI-98,
       author = {Hansen, E. A.},
        title = {Solving {POMDP}s by Searching in Policy Space},
    booktitle = {Proceedings of the Fourteenth Conference on Uncertainty in Artificial Intelligence},
        pages = {211-219},
      address = {Madison, WI},
         year = 1998,
}
@inproceedings{HanssonandMayerUAI-88,
       author = {Hansson, Othar and Mayer, Andrew},
        title = {The Optimality of Satisficing Solutions},
    booktitle = UAI-88,
     location = {Minneapolis, Minnesota},
        pages = {148-157},
         year = 1988,
}
@inproceedings{HanssonandMayerUAI-89,
       author = {Hansson, Othar and Mayer, Andrew},
        title = {Heuristic Search as Evidential Reasoning},
    booktitle = {Proceedings of the Fifth Workshop on Uncertainty in {AI}},
     location = {Windsor, Ontario},
         year = 1989,
        pages = {152-161},
}
@inproceedings{Hanssonetal,
       author = {Hansson, Othar and Mayer, Andrew and Russell, Stuart},
        title = {Decision-Theoretic Planning in BPS},
    booktitle = {Position Papers for the {AAAI} Spring Symposium on Planning in Uncertain Environments},
     location = {Palo Alto, California},
         year = 1989,
}
@article{HanuschkinetalNEUROSCIENCE-11,
       author = {Hanuschkin, Alexander and Diesmann, Markus and Morrison, Abigail},
        title = {A Reafferent and Feed-forward Model of Song Syntax Generation in the Bengalese Finch},
      journal = {Journal Compututational Neuroscience},
       volume = 31,
        issue = 3,
         year = 2011,
        pages = {509-532},
    publisher = {Springer-Verlag New York, Inc.},
      address = {Secaucus, NJ, USA},
     abstract = {Adult Bengalese finches generate a variable song that obeys a distinct and individual syntax. The syntax is gradually lost over a period of days after deafening and is recovered when hearing is restored. We present a spiking neuronal network model of the song syntax generation and its loss, based on the assumption that the syntax is stored in reafferent connections from the auditory to the motor control area. Propagating synfire activity in the HVC codes for individual syllables of the song and priming signals from the auditory network reduce the competition between syllables to allow only those transitions that are permitted by the syntax. Both imprinting of song syntax within HVC and the interaction of the reafferent signal with an efference copy of the motor command are sufficient to explain the gradual loss of syntax in the absence of auditory feedback. The model also reproduces for the first time experimental findings on the influence of altered auditory feedback on the song syntax generation, and predicts song- and species-specific low frequency components in the LFP. This study illustrates how sequential compositionality following a defined syntax can be realized in networks of spiking neurons.},
}
@article{HaraetalCoRR-17,
       author = {Kota Hara and Ming-Yu Liu and Oncel Tuzel and Amir-massoud Farahmand},
        title = {Attentional Network for Visual Object Detection},
      journal = {CoRR},
       volume = {ArXiv:1702.01478},
         year = {2017},
     abstract = {We propose augmenting deep neural networks with an attention mechanism for the visual object detection task. As perceiving a scene, humans have the capability of multiple fixation points, each attended to scene content at different locations and scales. However, such a mechanism is missing in the current state-of-the-art visual object detection methods. Inspired by the human vision system, we propose a novel deep network architecture that imitates this attention mechanism. As detecting objects in an image, the network adaptively places a sequence of glimpses of different shapes at different locations in the image. Evidences of the presence of an object and its location are extracted from these glimpses, which are then fused for estimating the object class and bounding box coordinates. Due to lacks of ground truth annotations of the visual attention mechanism, we train our network using a reinforcement learning algorithm with policy gradients. Experiment results on standard object detection benchmarks show that the proposed network consistently outperforms the baseline networks that does not model the attention mechanism.}
}
@book{Harel00,
       author = {David Harel},
        title = {Computers Ltd: What They Really Can't Do},
    publisher = {Oxford University Press},
      address = {Oxford},
         year = 2000,
}
@book{Harel87,
       author = {David Harel},
        title = {Algorithmics: The Spirit of Computing},
    publisher = {Addison-Wesley},
      address = {Reading, MA},
         year = 1987,
}
@book{HarelALGORITHMICS,
       author = {Harel, David},
        title = {ALGORITHMICS: The Spirit of Computing},
    publisher = {Addison-Wesley},
         year = 1987,
}
@book{HarelDL,
       author = {Harel, David},
        title = {First-Order Dynamic Logic},
    publisher = {Springer-Verlag},
      address = {New York},
         year = 1979,
}
@inproceedings{HariharanetalICCV-11,
  author    = {Bharath Hariharan and Pablo Arbelaez and Lubomir Bourdev and Subhransu Maji and Jitendra Malik},
  title     = {Semantic Contours from Inverse Detectors},
  booktitle = {International Conference on Computer Vision},
  year      = 2011,
}
@article{HarkeretalJMIV-07,
       author = {S. R. Harker and C. R. Vogel and T. Gedeon},
        title = {Analysis of Constrained Optimization Variants of the Map-Seeking Circuit Algorithm},
      journal = {Journal Mathematical Imaging and Vision},
       volume = 29,
        issue = 1,
        pages = {49-62},
         year = 2007,
}
@book{Harman1967,
       author = {Harman, H. H.},
        title = {Modern Factor Analysis},
    publisher = {University of Chicago Press},
      address = {Chicago, Illinois},
         year = 1967,
}
@article{HarrisPLoS-08,
       author = {Harris, William A.},
        title = {Seymour Benzer 1921-2007 The Man Who Took Us from Genes to Behaviour},
      journal = {PLoS Biology},
    publisher = {Public Library of Science},
         year = {2008},
       volume = {6},
       number = {2},
        pages = {1-3},
     abstract = {Seymour Benzer died suddenly in December 2007. This piece pays tribute to Seymour Benzer's pioneering work in solid state physics, molecular biology, and neurogenetics.},
}
@article{HarrisandGordonNATURE-NEUROSCIENCE-15,
       author = {Harris, Kenneth D. and Shepherd, Gordon M. G.},
        title = {The neocortical circuit: themes and variations},
      journal = {Nature Neuroscience},
         year = {2015},
       volume = {18},
        issue = {2},
        pages = {170-181},
     abstract = {Similarities in neocortical circuit organization across areas and species suggest a common strategy to process diverse types of information, including sensation from diverse modalities, motor control and higher cognitive processes. Cortical neurons belong to a small number of main classes. The properties of these classes, including their local and long-range connectivity, developmental history, gene expression, intrinsic physiology and in vivo activity patterns, are remarkably similar across areas. Each class contains subclasses; for a rapidly growing number of these, conserved patterns of input and output connections are also becoming evident. The ensemble of circuit connections constitutes a basic circuit pattern that appears to be repeated across neocortical areas, with area- and species-specific modifications. Such 'serially homologous' organization may adapt individual neocortical regions to the type of information each must process.},
}
@inproceedings{HarrisandStephensAVC-88,
     author = {C. Harris and M.J. Stephens},
      title = {A combined corner and edge detector},
  booktitle = {Alvey Vision Conference},
      pages = {147-152},
       year = 1988,
}
@article{HarrisetalNEURON-12,
      title = {Synaptic Energy Use and Supply},
     author = {Julia J. Harris and Renaud Jolivet and David Attwell},
    journal = {Neuron},
    volume = 75,
     issue = 5,
     pages = {762-777},
      year = 2012,
  abstract = {Neuronal computation is energetically expensive. Consequently, the brain's limited energy supply imposes constraints on its information processing capability. Most brain energy is used on synaptic transmission, making it important to understand how energy is provided to and used by synapses. We describe how information transmission through presynaptic terminals and postsynaptic spines is related to their energy consumption, assess which mechanisms normally ensure an adequate supply of \{ATP\} to these structures, consider the influence of synaptic plasticity and changing brain state on synaptic energy use, and explain how disruption of the energy supply to synapses leads to neuropathology.}
}
@article{HartetalIEEE-68,
       author = {Hart, Peter E. and Nilsson, Nils J. and Raphael, B.},
        title = {A Formal Basis for the Heuristic Determination of Minimum Cost paths},
      journal = {{IEEE} Transactions on Systems Science and Cybernetics},
       volume = 4,
         year = 1968,
}
@article{HartetalNLM-14,
       author = {Hart, G. and Leung, B. K. and Balleine, B. W. },
        title = {Dorsal and ventral streams: the distinct role of striatal subregions in the acquisition and performance of goal-directed actions},
      journal = {Neurobiology Learning and Memory},
         year = {2014},
       volume = {108},
        pages = {104-118},
     abstract = {Considerable evidence suggests that distinct neural processes mediate the acquisition and performance of goal-directed instrumental actions. Whereas a cortical-dorsomedial striatal circuit appears critical for the acquisition of goal-directed actions, a cortical-ventral striatal circuit appears to mediate instrumental performance, particularly the motivational control of performance. Here we review evidence that these distinct mechanisms of learning and performance constitute two distinct 'streams' controlling instrumental conditioning. From this perspective, the regulation of the interaction between these 'streams' becomes a matter of considerable importance. We describe evidence that the basolateral amygdala, which is heavily interconnected with both the dorsal and ventral subregions of the striatum, coordinates this interaction providing input to the final common path to action as a critical component of the limbic-motor interface.}
}
@article{HartetalSCIENTIFIC-REPORTS-18,
       author = {Hart, Yuval and Dillon, Moira R. and Marantan, Andrew and Cardenas, Anna L. and Spelke, Elizabeth and Mahadevan, L.},
        title = {The statistical shape of geometric reasoning},
      journal = {Scientific Reports},
         year = {2018},
       volume = {8},
        issue = {1},
        pages = {12906},
     abstract = {Geometric reasoning has an inherent dissonance: its abstract axioms and propositions refer to perfect, idealized entities, whereas its use in the physical world relies on dynamic perception of objects. How do abstract Euclidean concepts, dynamics, and statistics come together to support our intuitive geometric reasoning? Here, we address this question using a simple geometric task - planar triangle completion. An analysis of the distribution of participants errors in localizing a fragmented triangle's missing corner reveals scale-dependent deviations from a deterministic Euclidean representation of planar triangles. By considering the statistical physics of the process characterized via a correlated random walk with a natural length scale, we explain these results and further predict participants' estimates of the missing angle, measured in a second task. Our model also predicts the results of a categorical reasoning task about changes in the triangle size and shape even when such completion strategies need not be invoked. Taken together, our findings suggest a critical role for noisy physical processes in our reasoning about elementary Euclidean geometry.},
}
@book{HartmanisandStearns66,
       author = {Hartmanis, J. and Stearns, R. E.},
        title = {Algebraic Structure Theory of Sequential Machines},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, N.J.},
         year = 1966,
}
@misc{Harvest03,
       title = {Harvest: A Distributed Search System},
      author = {Harvest},
howpublished = {\url{http://harvest.sourceforge.net/}},
        year = 2003,
}
@article{HarveyetalNATURE-09,
     title = {Intracellular dynamics of hippocampal place cells during virtual navigation},
    author = {Harvey, C.D. and Collman, F. and Dombeck, D.A. and Tank, D.W.},
   journal = {Nature},
    volume = 461,
     issue = 7266,
     pages = {941-946},
      year = 2009,
  abstract = {Hippocampal place cells encode spatial information in rate and temporal codes. To examine the mechanisms underlying hippocampal coding, here we measured the intracellular dynamics of place cells by combining in vivo whole-cell recordings with a virtual-reality system. Head-restrained mice, running on a spherical treadmill, interacted with a computer-generated visual environment to perform spatial behaviours. Robust place-cell activity was present during movement along a virtual linear track. From whole-cell recordings, we identified three subthreshold signatures of place fields: an asymmetric ramp-like depolarization of the baseline membrane potential, an increase in the amplitude of intracellular theta oscillations, and a phase precession of the intracellular theta oscillation relative to the extracellularly recorded theta rhythm. These intracellular dynamics underlie the primary features of place-cell rate and temporal codes. The virtual-reality system developed here will enable new experimental approaches to study the neural circuits underlying navigation.}
}
@article{HarwoodetalJN-99,
     author = {M. R. Harwood and L. E. Mezey and C. M. Harris}, 
      title = {The spectral main sequence of human saccades},
    journal = {The Journal of Neuroscience},
     volume = 19, 
      pages = {9098-9106},
       year = 1999, 
}
@article{HashemietalCC-16,
       author = {Ezzat Hashemi and Jeanelle Ariza and Haille Rogers and Stephen C. Noctor and Ver\`{o}nica Mart\'{í}nez-Cerde\~{n}o},
        title = {The Number of Parvalbumin-Expressing Interneurons Is Decreased in the Prefrontal Cortex in Autism},
      journal = {Cerebral Cortex},
        pages = {1-13},
         year = 2016,
}
@article{HassabisandMaguirePTRS_B-09,
        title = {The construction system of the brain},
       author = {Hassabis, Demis and Maguire, Eleanor A.},
      journal = {Philosophical Transactions of the Royal Society London B Biological Science},
    publisher = {The Royal Society},
       volume = 364,
        issue = 1521,
         year = 2009,
        pages = {1263-1271},
     abstract = {The ability to construct a hypothetical situation in one's imagination prior to it actually occurring may afford greater accuracy in predicting its eventual outcome. The recollection of past experiences is also considered to be a reconstructive process with memories recreated from their component parts. Construction, therefore, plays a critical role in allowing us to plan for the future and remember the past. Conceptually, construction can be broken down into a number of constituent processes although little is known about their neural correlates. Moreover, it has been suggested that some of these processes may be shared by a number of other cognitive functions including spatial navigation and imagination. Recently, novel paradigms have been developed that allow for the isolation and characterization of these underlying processes and their associated neuroanatomy. Here, we selectively review this fast-growing literature and consider some implications for remembering the past and predicting the future.},
}
@article{HassabisandMaguireTiCS-07,
        title = {Deconstructing episodic memory with construction},
       author = {Hassabis, Demis and Maguire, Eleanor A.},
      journal = {Trends in Cognitive Science},
    publisher = {Elsevier Current Trends},
       volume = 11,
        issue = 7,
         year = 2007,
        pages = {299-306},
     abstract = {It has recently been observed that the brain network supporting recall of episodic memories shares much in common with other cognitive functions such as episodic future thinking, navigation and theory of mind. It has been speculated that 'self-projection' is the key common process. However, in this Opinion article, we note that other functions (e.g. imagining fictitious experiences) not explicitly connected to either the self or a subjective sense of time, activate a similar brain network. Hence, we argue that the process of 'scene construction' is better able to account for the commonalities in the brain areas engaged by an extended range of disparate functions. In light of this, we re-evaluate our understanding of episodic memory, the processes underpinning it and other related cognitive functions.} 
}
@article{HassabisetalNEURON-17,
       author = {Demis Hassabis and Dharshan Kumaran and Christopher Summerfield and Matthew Botvinick},
        title = {Neuroscience-Inspired Artificial Intelligence},
      journal = {Neuron},
       volume = {95},
       number = {2},
        pages = {245-258},
         year = {2017},
     abstract = {The fields of neuroscience and artificial intelligence (AI) have a long and intertwined history. In more recent times, however, communication and collaboration between the two fields has become less commonplace. In this article, we argue that better understanding biological brains could play a vital role in building intelligent machines. We survey historical interactions between the AI and neuroscience fields and emphasize current advances in AI that have been inspired by the study of neural computation in humans and other animals. We conclude by highlighting shared themes that may be key for advancing future research in both fields.},
}
@book{Hastieetal2001,
       author = {Hastie, Trevor  and Tibshirani, Robert and Friedman, Jerome},
        title = {The Elements of Statistical Learning: Data Mining, Inference, and Prediction},
    publisher = {Springer},
      address = {New York},
         year = 2001,
}
@book{Haugeland,
       author = {Haugeland, John},
        title = {Artificial Intelligence: The Very Idea},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1985,
}
@inproceedings{HaughAAAI-87,
       author = {Haugh, Brian},
        title = {Simple Causal Minimizations for Temporal Persistence and Projection},
    booktitle = {Proceedings AAAI-87},
     location = {Seattle, Washington},
 organization = {AAAI},
         year = 1987,
        pages = {218-223},
}
@incollection{HauserandSpelkeTCN-04,
       author = {Hauser, Marc D. and Spelke, Elizabeth},
        title = {Evolutionary and Developmental Foundations of Human Knowledge: A Case Study of Mathematics},
    booktitle = {The Cognitive Neurosciences, {III}},
       editor = {M. Gazzaniga and N. Logothetis},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = {2004},
        pages = {853-864},
     abstract = {Studies of human infants and of adult nonhuman primates provide evidence for two systems of representation that serve as building blocks for the number concepts of educated, enculturated humans. One is a system for representing small, precise numbers of objects; the other is a system for representing large, approximate numerical magnitudes. Studies of the signature limits of these systems reveal striking homologies across human and nonhuman primates and across infants and adults, suggesting that the systems emerged before the human species and arise in human development before the acquisition of language or culture. Children take a major step beyond these systems when they come to understand verbal counting and the natural number concepts that counting both defines and supports. Nevertheless, even the most sophisticated number concepts of educated humans rely in part on the core knowledge systems with deep roots in our ontogenetic and phylogenetic past. In this respect, numerical cognition may serve as a prototype for understanding other distinctively human, culturally supported concepts and cognitive skills.},
}
@inproceedings{HausknechtetalCIAI-GAMES-13,
    author = {Matthew Hausknecht, Joel Lehman, Risto Miikkulainen, and Peter Stone},
     title = {A Neuroevolution Approach to General Atari Game Playing},
 booktitle = {In {IEEE} Transactions on Computational Intelligence and {AI} in Games},
      year = 2013,
}
@inproceedings{HauskrechtandFraserAMIA-98,
       author = {Milos Hauskrecht and Hamish Fraser},
        title = {Modeling Treatment of Ischemic Heart Disease with Partially Observable {M}arkov Decision Processes},
    booktitle = {Proceedings of American Medical Informatics Association annual symposium on Computer Applications in Health Care},
        pages = {538-542},
      address = {Orlando, FL},
         year = 1998,
}
@inproceedings{HauskrechtandFraserDX-98,
       author = {Milos Hauskrecht and Hamish Fraser},
        title = {Planning medical therapy using partially observable {M}arkov decision processes},
    booktitle = {Proceedings of the Ninth International Workshop on Principles of Diagnosis (DX-98)},
        pages = {182-189},
      address = {Cape Cod, MA},
         year = 1998,
}
@inproceedings{HauskrechtetalIJCAI-99,
       author = {Milos Hauskrecht and Gopal Pandurangan and Eli Upfal},
        title = {Computing near optimal strategies for stochastic investment planning problems},
    booktitle = {Proceedings of the Sixteenth International Joint Conference on Artificial Intelligence},
      address = {Stockholm, Sweden},
         year = 1999,
}
@inproceedings{HauskrechtetalUAI-98,
      author = {Milos Hauskrecht and Nicolas Meuleau and Craig Boutilier and Leslie Pack Kaelbling and Thomas Dean},
       title = {Hierarchical solution of {M}arkov decision processes using macro-actions},
   booktitle = {Proceedings of the 14th Conference on Uncertainty in Artificial Intelligence ({UAI}-98)},
   publisher = {Morgan Kaufmann Publishers},
     address = {San Francisco, California},
organization = {AUAI},
       pages = {220-229},
    location = {Madison, Wisconsin},
        year = 1998,
}
@article{HausseretalSCIENCE-00,
       author = {H\"{a}usser, Michael and Spruston, Nelson and Stuart, Greg J.},
        title = {Diversity and Dynamics of Dendritic Signaling},
      journal = {Science},
       volume = {290},
       number = {5492},
        pages = {739-744},
         year = {2000},
     abstract = {Communication between neurons in the brain occurs primarily through synapses made onto elaborate treelike structures called dendrites. New electrical and optical recording techniques have led to tremendous advances in our understanding of how dendrites contribute to neuronal computation in the mammalian brain. The varied morphology and electrical and chemical properties of dendrites enable a spectrum of local and long-range signaling, defining the input-output relationship of neurons and the rules for induction of synaptic plasticity. In this way, diversity in dendritic signaling allows individual neurons to carry out specialized functions within their respective networks.},
}
@article{HausslerAIJ-88,
       author = {Haussler, David},
        title = {Quantifying Inductive Bias: {AI} Learning Algorithms and Valiant's Learning Framework},
      journal = AIJ,
       volume = 38,
         year = 1988,
        pages = {177-221},
}
@incollection{HausslerandWarmuthSFI-95,
       author = {Haussler, David and Warmuth, Manfred K.},
        title = {The Probably Approximately Correct ({PAC}) and Other Learning Models},
    booktitle = {The Mathematics of Generalization},
       editor = {Wolpert, David H.},
    publisher = {Addison-Wesley},
         year = 1995,
}
@article{HaussleretalIC-91,
       author = {Haussler, D. and Kearns, M. and Littlestone, N. and Warmuth, M. K.},
        title = {Equivalence of Models for Polynomial Time Learnability},
      journal = {Information and Computing},
       volume = 95,
        pages = {129-161},
         year = 1991,
}
@book{Hawkins04,
        title = {On Intelligence},
       author = {Jeff Hawkins and Sandra Blakeslee},
    publisher = {Henry Holt and Company},
      address = {New York},
         year = 2004,
}
@article{HawkinsandAhmadFiNC-16,
       author = {Jeff Hawkins and Subutai Ahmad},
        title = {Why Neurons Have Thousands of Synapses, a Theory of Sequence Memory in Neocortex},
      journal = {Frontiers in Neural Circuits},
    publisher = {Frontiers Media S.A.},
       volume = {10},
      article = {23},
         year = {2016},
     abstract = {Pyramidal neurons represent the majority of excitatory neurons in the neocortex. Each pyramidal neuron receives input from thousands of excitatory synapses that are segregated onto dendritic branches. The dendrites themselves are segregated into apical, basal, and proximal integration zones, which have different properties. It is a mystery how pyramidal neurons integrate the input from thousands of synapses, what role the different dendrites play in this integration, and what kind of network behavior this enables in cortical tissue. It has been previously proposed that non-linear properties of dendrites enable cortical neurons to recognize multiple independent patterns. In this paper we extend this idea in multiple ways. First we show that a neuron with several thousand synapses segregated on active dendrites can recognize hundreds of independent patterns of cellular activity even in the presence of large amounts of noise and pattern variation. We then propose a neuron model where patterns detected on proximal dendrites lead to action potentials, defining the classic receptive field of the neuron, and patterns detected on basal and apical dendrites act as predictions by slightly depolarizing the neuron without generating an action potential. By this mechanism, a neuron can predict its activation in hundreds of independent contexts. We then present a network model based on neurons with these properties that learns time-based sequences. The network relies on fast local inhibition to preferentially activate neurons that are slightly depolarized. Through simulation we show that the network scales well and operates robustly over a wide range of parameters as long as the network uses a sparse distributed code of cellular activations. We contrast the properties of the new network model with several other neural network models to illustrate the relative capabilities of each. We conclude that pyramidal neurons with thousands of synapses, active dendrites, and multiple integration zones create a robust and powerful sequence memory. Given the prevalence and similarity of excitatory neurons throughout the neocortex and the importance of sequence memory in inference and behavior, we propose that this form of sequence memory may be a universal property of neocortical tissue},
}
@misc{HawkinsandGeorgeHTM-06,
       title = {Hierarchical Temporal Memory: Concepts, Theory and Terminology},
      author = {Jeff Hawkins and Dileep George},
         url = {http://www.numenta.com/Numenta_HTM_Concepts.pdf},
        year = 2006,
}
@misc{HawkinsandGeorgeNUMENTA-06,
       title = {Hierarchical Temporal Memory: Concepts, Theory and Terminology},
      author = {Jeff Hawkins and Dileep George},
         url = {http://www.numenta.com/},
        year = 2006,
}
@article{HawkinsetalFiNC-19,
       author = {Hawkins, Jeff and Lewis, Marcus and Klukas, Mirko and Purdy, Scott and Ahmad, Subutai},
        title = {A Framework for Intelligence and Cortical Function Based on Grid Cells in the Neocortex},
      journal = {Frontiers in Neural Circuits},
       volume = {12},
        issue = {121},
        pages = {1-14},
         year = {2019},
     abstract = {How the neocortex works is a mystery. In this paper we propose a novel framework for understanding its function. Grid cells are neurons in the entorhinal cortex that represent the location of an animal in its environment. Recent evidence suggests that grid cell-like neurons may also be present in the neocortex. We propose that grid cells exist throughout the neocortex, in every region and in every cortical column. They define a location-based framework for how the neocortex functions. Whereas grid cells in the entorhinal cortex represent the location of one thing, the body relative to its environment, we propose that cortical grid cells simultaneously represent the location of many things. Cortical columns in somatosensory cortex track the location of tactile features relative to the object being touched and cortical columns in visual cortex track the location of visual features relative to the object being viewed. We propose that mechanisms in the entorhinal cortex and hippocampus that evolved for learning the structure of environments are now used by the neocortex to learn the structure of objects. Having a representation of location in each cortical column suggests mechanisms for how the neocortex represents object compositionality and object behaviors. It leads to the hypothesis that every part of the neocortex learns complete models of objects and that there are many models of each object distributed throughout the neocortex. The similarity of circuitry observed in all cortical regions is strong evidence that even high-level cognitive tasks are learned and represented in a location-based framework.}
}
@article{HawrylyczetalNATURE-12,
        title = {An anatomically comprehensive atlas of the adult human brain transcriptome},
       author = {Hawrylycz, Michael J. and Lein, Ed S. and Guillozet-Bongaarts, Angela L. and Shen, Elaine H. and Ng, Lydia and Miller, Jeremy A. and van de Lagemaat, Louie N. and Smith, Kimberly A. and Ebbert, Amanda and Riley, Zackery L. and Abajian, Chris and Beckmann, Christian F. and Bernard, Amy and Bertagnolli, Darren and Boe, Andrew F. and Cartagena, Preston M. and Chakravarty, M. Mallar and Chapin, Mike and Chong, Jimmy and Dalley, Rachel A. and Daly, Barry David and Dang, Chinh and Datta, Suvro and Dee, Nick and Dolbeare, Tim A. and Faber, Vance and Feng, David and Fowler, David R. and Goldy, Jeff and Gregor, Benjamin W. and Haradon, Zeb and Haynor, David R. and Hohmann, John G. and Horvath, Steve and Howard, Robert E. and Jeromin, Andreas and Jochim, Jayson M. and Kinnunen, Marty and Lau, Christopher and Lazarz, Evan T. and Lee, Changkyu and Lemon, Tracy A. and Li, Ling and Li, Yang and Morris, John A. and Overly, Caroline C. and Parker, Patrick D. and Parry, Sheana E. and Reding, Melissa and Royall, Joshua J. and Schulkin, Jay and Sequeira, Pedro Adolfo and Slaughterbeck, Clifford R. and Smith, Simon C. and Sodt, Andy J. and Sunkin, Susan M. and Swanson, Beryl E. and Vawter, Marquis P. and Williams, Derric and Wohnoutka, Paul and Zielke, H. Ronald and Geschwind, Daniel H.  and Hof, Patrick R. and Smith, Stephen M. and Koch, Christof and Grant, Seth G. N. and Jones, Allan R.},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = {489},
         year = {2012},
        pages = {391-399},
}
@article{HawrylyczetalNATURE-NEUROSCIENCE-15,
       author = {Hawrylycz, Michael and Miller, Jeremy A. and Menon, Vilas and Feng, David and Dolbeare, Tim and Guillozet-Bongaarts, Angela L. and Jegga, Anil G. and Aronow, Bruce J. and Lee, Chang-Kyu and Bernard, Amy and Glasser, Matthew F. and Dierker, Donna L. and Menche, J\"{o}rg and Szafer, Aaron and Collman, Forrest and Grange, Pascal and Berman, Kenneth A. and Mihalas, Stefan and Yao, Zizhen and Stewart, Lance and Barab\'{a}si, Albert-L\'{a}szl\'{o} and Schulkin, Jay and Phillips, John and Ng, Lydia and Dang, Chinh and Haynor, David R. and Jones, Allan and Van Essen, David C. and Koch, Christof and Lein, Ed},
        title = {Canonical genetic signatures of the adult human brain},
      journal = {Nature Neuroscience},
         year = {2015},
    publisher = {Nature Publishing Group},
       volume = {18},
        pages = {1832-1844},
     abstract = {The structure and function of the human brain are highly stereotyped, implying a conserved molecular program responsible for its development, cellular structure and function. We applied a correlation-based metric called differential stability to assess reproducibility of gene expression patterning across 132 structures in six individual brains, revealing mesoscale genetic organization. The genes with the highest differential stability are highly biologically relevant, with enrichment for brain-related annotations, disease associations, drug targets and literature citations. Using genes with high differential stability, we identified 32 anatomically diverse and reproducible gene expression signatures, which represent distinct cell types, intracellular components and/or associations with neurodevelopmental and neurodegenerative disorders. Genes in neuron-associated compared to non-neuronal networks showed higher preservation between human and mouse; however, many diversely patterned genes displayed marked shifts in regulation between species. Finally, highly consistent transcriptional architecture in neocortex is correlated with resting state functional connectivity, suggesting a link between conserved gene expression and functionally relevant circuitry}
}
@article{HawrylyczetalPL0S-11,
    author = {Michael Hawrylycz and Richard A. Baldock and Albert Burger and Tsutomu Hashikawa and G. Allan 
               Johnson and Maryann E. Martone and Lydia Ng and Christopher Lau and Stephen D. Larson and 
               Jonathan Nissanov and Luis Puelles and Seth Ruffins and Fons Verbeek and Ilya Zaslavsky and Jyl Boline},
     title = {Digital Atlasing and Standardization in the Mouse Brain},
   journal = {{PLoS} Computational Biology},
    volume = 7,
     issue = 2,
      year = 2011,
}
@article{HawrylyczetalPNAS-16,
       author = {Hawrylycz, Michael and Anastassiou, Costas and Arkhipov, Anton and Berg, Jim and Buice, Michael and Cain, Nicholas and Gouwens, Nathan W. and Gratiy, Sergey and Iyer, Ramakrishnan and Lee, Jung Hoon and Mihalas, Stefan and Mitelut, Catalin and Olsen, Shawn and Reid, R. Clay and Teeter, Corinne and de Vries, Saskia and Waters, Jack and Zeng, Hongkui and Koch, Christof and MindScope},
        title = {Inferring cortical function in the mouse visual system through large-scale systems neuroscience},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {113},
       number = {27},
        pages = {7337-7344},
         year = {2016},
     abstract = {The scientific mission of the Project MindScope is to understand neocortex, the part of the mammalian brain that gives rise to perception, memory, intelligence, and consciousness. We seek to quantitatively evaluate the hypothesis that neocortex is a relatively homogeneous tissue, with smaller functional modules that perform a common computational function replicated across regions. We here focus on the mouse as a mammalian model organism with genetics, physiology, and behavior that can be readily studied and manipulated in the laboratory. We seek to describe the operation of cortical circuitry at the computational level by comprehensively cataloging and characterizing its cellular building blocks along with their dynamics and their cell type-specific connectivities. The project is also building large-scale experimental platforms (i.e., brain observatories) to record the activity of large populations of cortical neurons in behaving mice subject to visual stimuli. A primary goal is to understand the series of operations from visual input in the retina to behavior by observing and modeling the physical transformations of signals in the corticothalamic system. We here focus on the contribution that computer modeling and theory make to this long-term effort.},
}
@inproceedings{HayashiandDeanIEEE-88,
      author = {Hayahsi, Akira and Dean, Thomas},
       title = {Locating a Mobile Robot Using Local Observations and a Global Satellite Map},
   booktitle = {Proceedings of the 3rd {IEEE} International Symposium on Intelligent Control},
    location = {Washington, DC},
organization = {IEEE},
       pages = {135-140},
        year = 1988,
}
@article{Hayes-RothAIJ-85,
       author = {Hayes-Roth, Barbara},
        title = {A Blackboard Architecture for Control},
      journal = AIJ,
       volume = 26,
         year = 1985,
        pages = {251-321},
}
@inproceedings{Hayes-RothIJCAI-89,
       author = {Hayes-Roth, Barbara and Washington, Richard and Hewett, Rattikorn and Hewett, Michael and Seiver, Adam},
        title = {Intelligent Monitoring and Control},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
         year = 1989,
        pages = {243-249},
}
@article{Hayes-RothPR-77,
       author = {Hayes-Roth, Barbara},
        title = {Evolution of cognitive structures and processe},
      journal = {Psychological Review},
       volume = 84,
        issue = 3,
         year = 1977,
        pages = {260-278},
}
@incollection{HayesLIQUIDS-85,
       author = {Hayes, Patrick J.},
        title = {Na\"{i}ve Physics {I}: Ontology for Liquids},
       editor = {Hobbs, Jerry E. and Moore, Robert C.},
    booktitle = {Formal Theories of the Common Sense World},
    publisher = {Ablex},
      address = {Norwood, New Jersey},
         year = 1985,
        pages = {71-107},
}
@incollection{HayesNPM-79,
       author = {Hayes, Patrick},
        title = {The Na\"{i}ve Physics Manifesto},
       editor = {Michie, Donald},
    booktitle = {Expert Systems in the Microelectronic Age},
    publisher = {Edinburgh University Press},
         year = 1979,
        pages = {242-270},
}
@inproceedings{HaykeletalIEE-09,
       author = {Haykel Ben Jamaa, M. and Carrara, S. and Georgiou, J. and Archontas, N. and De Micheli, G.},
    booktitle = {IEEE Conference on Nanotechnology},
        title = {Fabrication of memristors with poly-crystalline silicon nanowires},
         year = 2009,
        pages = {152-154},
     abstract = {Memristors are the two-terminal components that complete the symmetry between the fundamental circuit variables, and they are highly suitable for bioinspired and neural-network-based computational systems due to their inherent memory effect. In this paper we present a fabrication technique that uses only Complementary Metal Oxide Semiconductor (CMOS) processing steps and conventional photolithography, yielding poly-crystalline silicon nanowires that show a memristive behavior. Besides measurements, we performed numerical device simulations that address the observed memristive effect.},
}
@article{HayworthandMarblestoneBIORXIV-18,
       author = {Hayworth, Kenneth J. and Marblestone, Adam H.},
        title = {How thalamic relays might orchestrate supervised deep training and symbolic computation in the brain},
      journal = {bioRxiv},
         year = {2018},
    publisher = {Cold Spring Harbor Laboratory},
     abstract = {The thalamus appears to be involved in the flexible routing of information among cortical areas, yet the computational implications of such routing are only beginning to be explored. Here we create a connectionist model of how selectively gated cortico-thalamo-cortical relays could underpin both symbolic and sub-symbolic computations. We first show how gateable relays can be used to create a Dynamically Partitionable Auto-Associative Network (DPAAN) (Hayworth, 2012) consisting of a set of cross-connected cortical memory buffers. All buffers and relays in a DPAAN are trained simultaneously to have a common set of stable attractor states that become the symbol vocabulary of the DPAAN. We show via simulations that such a DPAAN can support operations necessary for syntactic rule-based computation, namely buffer-to-buffer copying and equality detection. We then provide each DPAAN module with a multilayer input network trained to map sensory inputs to the DPAAN{\textquoteright}s symbol vocabulary, and demonstrate how gateable thalamic relays can provide recall and clamping operations to train this input network by Contrastive Hebbian Learning (CHL) (Xie and Seung, 2003). We suggest that many such DPAAN modules may exist at the highest levels of the brain{\textquoteright}s sensory hierarchies and show how a joint snapshot of the contents of multiple DPAAN modules can be stored as a declarative memory in a simple model of the hippocampus. We speculate that such an architecture might first have been {\textquoteleft}discovered{\textquoteright} by evolution as a means to bootstrap learning of more meaningful cortical representations feeding the striatum, eventually leading to a system that could support symbolic computation. Our model serves as a bridging hypothesis for linking controllable thalamo-cortical information routing with computations that could underlie aspects of both learning and symbolic reasoning in the brain.},
}
@article{HayworthetalNATURE-METHODS-15,
       author = {Hayworth, Kenneth J. and Xu, C. Shan and Lu, Zhiyuan and Knott, Graham W. and Fetter, Richard D. and Tapia, Juan Carlos and Lichtman, Jeff W. and Hess, Harald F.},
        title = {Ultrastructurally smooth thick partitioning and volume stitching for large-scale connectomics},
      journal = {Nature Methods},
         year = 2015,
    publisher = {Nature Publishing Group},
       volume = 12,
        issue = 4,
        pages = {319-322},
     abstract = {Focused-ion-beam scanning electron microscopy (fib-sem) has become an essential tool for studying neural tissue at resolutions below 10 nm x 10 nm x 10 nm, producing data sets optimized for automatic connectome tracing. We present a technical advance, ultrathick sectioning, which reliably subdivides embedded tissue samples into chunks (20 mm thick) optimally sized and mounted for efficient, parallel fib-sem imaging. these chunks are imaged separately and then 'volume stitched' back together, producing a final three-dimensional data set suitable for connectome tracing.},
}
@article{HazyetalPTRS-07,
       author = {Hazy, T. E. and Frank, M. J. and O'Reilly, R. C.},
        title = {Towards an executive without a homunculus: computational models of the prefrontal cortex/basal ganglia system},
      journal = {Philosophical Transactions of the Royal Society London B, Biological Science},
       volume = {362},
       number = {1485},
         year = {2007},
        pages = {1601-1613},
     abstract = {The prefrontal cortex (PFC) has long been thought to serve as an 'executive' that controls the selection of actions and cognitive functions more generally. However, the mechanistic basis of this executive function has not been clearly specified often amounting to a homunculus. This paper reviews recent attempts to deconstruct this homunculus by elucidating the precise computational and neural mechanisms underlying the executive functions of the PFC. The overall approach builds upon existing mechanistic models of the basal ganglia (BG) and frontal systems known to play a critical role in motor control and action selection, where the BG provide a 'Go' versus 'NoGo' modulation of frontal action representations. In our model, the BG modulate working memory representations in prefrontal areas to support more abstract executive functions. We have developed a computational model of this system that is capable of developing human-like performance on working memory and executive control tasks through trial-and-error learning. This learning is based on reinforcement learning mechanisms associated with the midbrain dopaminergic system and its activation via the BG and amygdala. Finally, we briefly describe various empirical tests of this framework.}
}
@article{HeaveyetalCONSCIOUS_COGNITION-13,
       author = {Heavey C.L. and Hurlburt R.T. and Lefforge N.L.},
        title = {Toward a Phenomenology of Inner Speaking},
      journal = {Consciousness and Cognition},
       volume = {22},
       number = {4},
         year = {2013},
        pages = {1477-1493},
     abstract = {Inner speaking is a common and widely discussed phenomenon of inner experience. Based on our studies of inner experience using Descriptive Experience Sampling (a qualitative method designed to produce high fidelity descriptions of randomly selected pristine inner experience), we advance an initial phenomenology of inner speaking. Inner speaking does occur in many, though certainly not all, moments of pristine inner experience. Most commonly it is experienced by the person as speaking in his or her own naturally inflected voice but with no sound being produced. In addition to prototypical instances of inner speaking, there are wide-ranging variations that fit the broad category of inner speaking and large individual differences in the frequency with which individuals experience inner speaking. Our observations are discrepant from what many have said about inner speaking, which we attribute to the characteristics of the methods different researchers have used to examine inner speaking.},
}
@article{HeaveyetalEMOTION-12,
       author = {Heavey C.L. and Hurlburt R.T. and Lefforge N.L.},
        title = {Toward a phenomenology of feelings},
      journal = {Emotion},
       volume = {12},
       number = {4},
         year = {2012},
        pages = {763-777},
     abstract = {Our understanding of emotion cannot be complete without an understanding of feelings, the experiential aspect of emotion. Despite their importance, little effort has been devoted to the careful apprehension of feelings. Based on our apprehension of many randomly selected moments of pristine inner experience, we present a preliminary phenomenology of feelings. We begin by observing that often feelings occur as directly experienced phenomena of awareness; however, often no feelings are present in experience, or if they are present, they are too faint to be observed by a process intended to observe them. Feelings range from vague to distinct and sometimes do, but other times do not, include bodily sensations. When bodily sensations are present, there is a wide range of clarity and location of these sensations. Sometimes people experience multiple distinct feelings and sometimes people experience one feeling that is a mix or blend of different feelings. We also discuss what feelings are not, including instances when feelings do not appear to be present, despite evidence suggesting the presence of underlying emotional processes (e.g., behavioral evidence of emotion). These instances of emotion but not feeling lead us to speculate that experiencing feelings is a skill developed over time through an interaction of interpersonal and intrapersonal events.},
}
@book{Hebb1949organization,
        title = {The Organization of Behavior: A Neuropsychological Theory},
       author = {Hebb, Donald O.},
         year = {1949},
    publisher = {Wiley}
}
@book{Hebb49,
     author = {Donald Hebb},
      title = {The Organization of Behavior},
  publisher = {Wiley},
    address = {New York},
       year = 1949,
}
@book{Hebb_The_Organization_of_Behavior-1949,
       author = {Hebb, Donald O.},
        title = {The organization of behavior: {A} neuropsychological theory},
    publisher = {Wiley},
      address = {New York},
         year = 1949,
     abstract = {Donald Hebb pioneered many current themes in behavioural neuroscience. He saw psychology as a biological science, but one in which the organization of behaviour must remain the central concern. Through penetrating theoretical concepts, including the "cell assembly," "phase sequence," and "Hebb synapse," he offered a way to bridge the gap between cells, circuits and behaviour. He saw the brain as a dynamically organized system of multiple distributed parts, with roots that extend into foundations of development and evolutionary heritage. He understood that behaviour, as brain, can be sliced at various levels and that one of our challenges is to bring these levels into both conceptual and empirical register. He could move between theory and fact with an ease that continues to inspire both students and professional investigators. Although facts continue to accumulate at an accelerating rate in both psychology and neuroscience, and although these facts continue to force revision in the details of Hebb's earlier contributions, his overall insistence that we look at behaviour and brain together within a dynamic, relational and multilayered framework remains. His work touches upon current studies of population coding, contextual factors in brain representations, synaptic plasticity, developmental construction of brain/behaviour relations, clinical syndromes, deterioration of performance with age and disease, and the formal construction of connectionist models. The collection of papers in this volume represent these and related themes that Hebb inspired. We also acknowledge our appreciation for Don Hebb as teacher, colleague and friend.},
}
@inproceedings{HeckandHuangGlobalSIP-14,
       author = {L. Heck and H. Huang},
        title = {Deep Learning of Knowledge Graph Embeddings for Semantic Parsing of Twitter Dialogs},
    booktitle = {Proceedings of the {IEEE} Global Conference on Signal and Information Processing},
         year = 2014,
     abstract = {This paper presents a novel method to learn neural knowledge graph embeddings. The embeddings are used to compute semantic relatedness in a coherence-based semantic parser. The approach learns embeddings directly from structured knowledge representations. A deep neural network approach known as Deep Structured Semantic Modeling (DSSM) is used to scale the approach to learn neural embeddings for all of the concepts (pages) of Wikipedia. Experiments on Twitter dialogs show a 23.6\% reduction in semantic parsing errors compared to the state-of-the-art unsupervised approach.}
}
@techreport{HeckermanTR-95,
     author = {Heckerman, David},
      title = {A Tutorial on Learning {B}ayesian Networks},
institution = {Microsoft Research},
      issue = {MSR-95-06},
       year = 1995,
}
@incollection{HeckermanUAI-88,
     author = {Heckerman, David},
      title = {Probabilistic Interpretations for Mycin Certainty Factors},
   crossref = UAI-88,
}
@inproceedings{HeckermanUAI-89,
     author = {Heckerman, David E. and Breese, John S. and Horvitz, Eric J.},
      title = {The Compilation of Decision Models},
  booktitle = UAI-89,
   location = {Windsor, Ontario},
       year = 1989,
      pages = {162-173},
}
@article{HeckermanetalMIM-92,
     author = {D. Heckerman and E. Horvitz and B. Nathwani},
      title = {Toward normative expert systems: {P}art {I}. {T}he {P}athfinder project},
    journal = {Methods of Information in Medicine},
     volume = 31,
      pages = {90-105},
       year = 1992,
}
@inproceedings{HeetalACM-09,
       author = {He, Jingrui and Liu, Yan and Lawrence, Richard},
        title = {Graph-based Transfer Learning},
    booktitle = {Proceedings of the 18th ACM Conference on Information and Knowledge Management},
         year = {2009},
     location = {Hong Kong, China},
        pages = {937-946},
    publisher = {ACM},
      address = {New York, NY, USA},
     abstract = {Transfer learning is the task of leveraging the information from labeled examples in some domains to predict the labels for examples in another domain. It finds abundant practical applications, such as sentiment prediction, image classification and network intrusion detection. In this paper, we propose a graph-based transfer learning framework. It propagates the label information from the source domain to the target domain via the example-feature-example tripartite graph, and puts more emphasis on the labeled examples from the target domain via the example-example bi-partite graph. Our framework is semi-supervised and non-parametric in nature and thus more flexible. We also develop an iterative algorithm so that our framework is scalable to large-scale applications. It enjoys the theoretical property of convergence. Compared with existing transfer learning methods, the proposed framework propagates the label information to both the features irrelevant to the source domain and the unlabeled examples in the target omain via the common features in a principled way. Experimental results on 3 real data sets demonstrate the effectiveness of our algorithm.}
}
@article{HeetalCoRR-15,
       author = {Kaiming He and Xiangyu Zhang and Shaoqing Ren and Jian Sun},
        title = {Deep Residual Learning for Image Recognition},
      journal = {CoRR},
       volume = {arXiv:1512.03385},
         year = {2015},
     abstract = {Deeper neural networks are more difficult to train. We present a residual learning framework to ease the training of networks that are substantially deeper than those used previously. We explicitly reformulate the layers as learning residual functions with reference to the layer inputs, instead of learning unreferenced functions. We provide comprehensive empirical evidence showing that these residual networks are easier to optimize, and can gain accuracy from considerably increased depth. On the ImageNet dataset we evaluate residual nets with a depth of up to 152 layers---8x deeper than VGG nets but still having lower complexity. An ensemble of these residual nets achieves 3.57\% error on the ImageNet test set. This result won the 1st place on the ILSVRC 2015 classification task. We also present analysis on CIFAR-10 with 100 and layers. The depth of representations is of central importance for many visual recognition tasks. Solely due to our extremely deep representations, we obtain a 28\% relative improvement on the COCO object detection dataset. Deep residual nets are foundations of our submissions to ILSVRC \& COCO 2015 competitions, where we also won the 1st places on the tasks of ImageNet detection, ImageNet localization, COCO detection, and COCO segmentation.}
}
@article{HeisenbergLM-98,
       author = {Heisenberg, Martin},
        title = {What Do the Mushroom Bodies Do for the Insect Brain? An Introduction},
      journal = {Learning \& Memory},
       volume = {5},
       number = {1},
         year = {1998},
        pages = {1-10},
}
@article{HeisenbergNATURE-REVIEWS-NEUROSCIENCE-03,
       author = {Heisenberg, M.},
        title = {Mushroom body memoir: from maps to models},
      journal = {Nature Reviews Neuroscience},
       volume = {4},
       number = {4},
         year = {2003},
        pages = {266-275},
     abstract = {Genetic intervention in the fly Drosophila melanogaster has provided strong evidence that the mushroom bodies of the insect brain act as the seat of a memory trace for odours. This localization gives the mushroom bodies a place in a network model of olfactory memory that is based on the functional anatomy of the olfactory system. In the model, complex odour mixtures are assumed to be represented by activated sets of intrinsic mushroom body neurons. Conditioning renders an extrinsic mushroom-body output neuron specifically responsive to such a set. Mushroom bodies have a second, less understood function in the organization of the motor output. The development of a circuit model that also addresses this function might allow the mushroom bodies to throw light on the basic operating principles of the brain.},
} 
@article{HelmchenandDenkNATURE-05,
     title = {Deep tissue two-photon microscopy},
    author = {Helmchen, F. and Denk, W.},
   journal = {Nature Methods},
    volume = 2,
     issue = 12,
     pages = {932-940},
      year = 2005,
  abstract = {With few exceptions biological tissues strongly scatter light, making high-resolution deep imaging impossible for traditional-including confocal-fluorescence microscopy. Nonlinear optical microscopy, in particular two photon-excited fluorescence microscopy, has overcome this limitation, providing large depth penetration mainly because even multiply scattered signal photons can be assigned to their origin as the result of localized nonlinear signal generation. Two-photon microscopy thus allows cellular imaging several hundred microns deep in various organs of living animals. Here we review fundamental concepts of nonlinear microscopy and discuss conditions relevant for achieving large imaging depths in intact tissue.}
}
@article{HelmstaedteretalBRR-07,
        title = {Reconstruction of an average cortical column in silico},
       author = {M. Helmstaedter and C.P.J. de Kock and D. Feldmeyer and R.M. Bruno and B. Sakmann},
      journal = {Brain Research Reviews},
       volume = 55,
        issue = 2,
         year = 2007,
        pages = {193-203},
     abstract = {The characterization of individual neurons by Golgi and Cajal has been the basis of neuroanatomy for a century. A new challenge is to anatomically describe, at cellular resolution, complete local circuits that can drive behavior. In this essay, we review the possibilities to obtain a model cortical column by using in vitro and in vivo pair recordings, followed by anatomical reconstructions of the projecting and target cells. These pairs establish connection modules that eventually may be useful to synthesize an average cortical column in silico. Together with data on sensory evoked neuronal activity measured in vivo, this will allow to model the anatomical and functional cellular basis of behavior based on more realistic assumptions than previously attempted.}
}
@article{HelmstaedteretalNATURE-13,
    author = {Helmstaedter, Moritz and Briggman, Kevin L. and Turaga, Srinivas C. and Jain, Viren and Seung, H. Sebastian and Denk, Winfried},
     title = {Connectomic reconstruction of the inner plexiform layer in the mouse retina},
   journal = {Nature},
 publisher = {Nature Publishing Group},
    volume = 500,
     issue = 7461,
      year = 2013,
     pages = {168-174},
  abstract = {Comprehensive high-resolution structural maps are central to functional exploration and understanding in biology. For the nervous system, in which high resolution and large spatial extent are both needed, such maps are scarce as they challenge data acquisition and analysis capabilities. Here we present for the mouse inner plexiform layer--the main computational neuropil region in the mammalian retina--the dense reconstruction of 950 neurons and their mutual contacts. This was achieved by applying a combination of crowd-sourced manual annotation and machine-learning-based volume segmentation to serial block-face electron microscopy data. We characterize a new type of retinal bipolar interneuron and show that we can subdivide a known type based on connectivity. Circuit motifs that emerge from our data indicate a functional mechanism for a known cellular response in a ganglion cell that detects localized motion, and predict that another ganglion cell is motion sensitive.}
}
@book{HelprinFIRE,
       author = {Mark Helprin},
        title = {Refiner's Fire: the Life and Adventures of Marshall Pearl, a Foundling},
    publisher = {Alfred A. Knopf, Inc.},
      address = {New York},
         year = 1977,
}
@book{HelprinWINTER,
       author = {Mark Helprin},
        title = {Winter's Tale},
    publisher = {Harcourt, Brace, Jovanovitch},
      address = {New York},
         year = 1983,
}
@article{HenaffetalCoRR-17,
       author = {Mikael Henaff and William F. Whitney and Yann LeCun},
        title = {Model-Based Planning in Discrete Action Spaces},
      journal = {CoRR},
       volume = {arXiv:1705.07177},
         year = {2017},
     abstract = {Action planning using learned and differentiable forward models of the world is a general approach which has a number of desirable properties, including improved sample complexity over model-free RL methods, reuse of learned models across different tasks, and the ability to perform efficient gradient-based optimization in continuous action spaces. However, this approach does not apply straightforwardly when the action space is discrete. In this work, we show that it is in fact possible to effectively perform planning via backprop in discrete action spaces, using a simple paramaterization of the actions vectors on the simplex combined with input noise when training the forward model. Our experiments show that this approach can match or outperform model-free RL and discrete planning methods on gridworld navigation tasks in terms of performance and/or planning time while using limited environment interactions, and can additionally be used to perform model-based control in a challenging new task where the action space combines discrete and continuous actions. We furthermore propose a policy distillation approach which yields a fast policy network which can be used at inference time, removing the need for an iterative planning procedure.}
}
@phdthesis{HendersonPhD,
        title = {Recovering From Errors in Conversational Dialogue Systems},
       author = {Henderson, Matthew},
         year = 2011,
       school = {University of Edinburgh},
     abstract = {Spoken dialogue systems can encounter different types of errors, including non-understanding errors. This is where the system realises the user has spoken, but does not understand their utterance. Strategies for dealing with this kind of error have been proposed and tested in the context of slot-filling systems, for example by Dan Bohus with a system which helps reserve conference rooms [1]. However there has been little work into possible strategies for more conversational settings. This dissertation looks at how we could recover from non-understanding errors experienced by a robot tourguide, and tests the strategies in an experimental study. The main hypothesis of this study is that it is beneficial to use strategies which are designed to do something smarter than just asking the user to repeat themselves. The strategies implemented are motivated by the findings of work done on task-based dialogue systems [1,2,3], which suggest it is useful to move the user on through the dialogue instead of getting caught up with the non-understanding error.}
}
@article{HendersonTiCS-17,
       author = {John M. Henderson},
        title = {Gaze Control as Prediction},
      journal = {Trends in Cognitive Sciences},
       volume = {21},
       number = {1},
        pages = {15-23},
         year = {2017},
     abstract = {The recent study of overt attention during complex scene viewing has emphasized explaining gaze behavior in terms of image properties and image salience independently of the viewer's intentions and understanding of the scene. In this Opinion article, I outline an alternative approach proposing that gaze control in natural scenes can be characterized as the result of knowledge-driven prediction. This view provides a theoretical context for integrating and unifying many of the disparate phenomena observed in active scene viewing, offers the potential for integrating the behavioral study of gaze with the neurobiological study of eye movements, and provides a theoretical framework for bridging gaze control and other related areas of perception and cognition at both computational and neurobiological levels of analysis.}
}
@article{HendersonetalCoRR-17,
       author = {Matthew Henderson and Rami Al-Rfou and Brian Strope and Yun-Hsuan Sung and L{\'{a}}szl{\'{o}} Luk{\'{a}}cs and Ruiqi Guo and Sanjiv Kumar and Balint Miklos and Ray Kurzweil},
        title = {Efficient Natural Language Response Suggestion for Smart Reply},
      journal = {CoRR},
       volume = {arXiv:1705.00652},
         year = {2017},
     abstract = {This paper presents a computationally efficient machine-learned method for natural language response suggestion. Feed-forward neural networks using n-gram embedding features encode messages into vectors which are optimized to give message-response pairs a high dot-product value. An optimized search finds response suggestions. The method is evaluated in a large-scale commercial e-mail application, Inbox by Gmail. Compared to a sequence-to-sequence approach, the new system achieves the same quality at a small fraction of the computational requirements and latency.}
}
@book{HendlerAIPS-92,
       editor = {Hendler, James},
        title = {First International Conference on {AI} Planning Systems},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1992,
}
@inproceedings{HendlerDARPA-87,
       author = {Hendler, James and Sanborn, James},
        title = {Planning and Reaction in Dynamic Domains},
    booktitle = {Proceedings of the DARPA Knowledge-Based Planning Workshop},
     location = {Austin, Texas},
 organization = {DARPA},
         year = 1987,
}
@article{HendleretalAIM-90,
       author = {Hendler, James and Tate, Austin and Drummond, Mark},
        title = {{AI} Planning: Systems and techniques},
      journal = {AI Magazine},
       volume = 11,
        issue = 2,
        pages = {61-77},
         year = 1990,
}
@article{HendricksetalJoN-17,
       author = {Hendricks, William D. and Chen, Yang and Bensen, AeSoon L. and Westbrook, Gary L. and Schnell, Eric},
        title = {Short-Term Depression of Sprouted Mossy Fiber Synapses from Adult-Born Granule Cells},
      journal = {Journal of Neuroscience},
       volume = {37},
       number = {23},
        pages = {5722-5735},
         year = {2017},
    publisher = {Society for Neuroscience},
     abstract = {Epileptic seizures potently modulate hippocampal adult neurogenesis, and adult-born dentate granule cells contribute to the pathologic retrograde sprouting of mossy fiber axons, both hallmarks of temporal lobe epilepsy. The characteristics of these sprouted synapses, however, have been largely unexplored, and the specific contribution of adult-born granule cells to functional mossy fiber sprouting is unknown, primarily due to technical barriers in isolating sprouted mossy fiber synapses for analysis. Here, we used DcxCreERT2 transgenic mice to permanently pulse-label age-defined cohorts of granule cells born either before or after pilocarpine-induced status epilepticus (SE). Using optogenetics, we demonstrate that adult-born granule cells born before SE form functional recurrent monosynaptic excitatory connections with other granule cells. Surprisingly, however, although healthy mossy fiber synapses in CA3 are well characterized "detonator" synapses that potently drive postsynaptic cell firing through their profound frequency-dependent facilitation, sprouted mossy fiber synapses from adult-born cells exhibited profound frequency-dependent depression, despite possessing some of the morphological hallmarks of mossy fiber terminals. Mature granule cells also contributed to functional mossy fiber sprouting, but exhibited less synaptic depression. Interestingly, granule cells born shortly after SE did not form functional excitatory synapses, despite robust sprouting. Our results suggest that, although sprouted mossy fibers form recurrent excitatory circuits with some of the morphological characteristics of typical mossy fiber terminals, the functional characteristics of sprouted synapses would limit the contribution of adult-born granule cells to hippocampal hyperexcitability in the epileptic hippocampus.},
}
@article{Hendrix,
       author = {Hendrix, Gary},
        title = {Modeling simultaneous actions and continuous processes},
      journal = AIJ,
       volume = 4,
         year = 1973,
        pages = {145-180},
}
@incollection{HengstEMLDM-17,
       author = {Hengst, Bernhard},
       editor = {Sammut, Claude and Webb, Geoffrey I.},
        title = {Hierarchical Reinforcement Learning},
    booktitle = {Encyclopedia of Machine Learning and Data Mining},
    publisher = {Springer US},
      address = {Boston, MA},
         year = {2017},
        pages = {611-619},
     abstract = {Hierarchical reinforcement learning (HRL) decomposes a  reinforcement learning problem into a hierarchy of subproblems or subtasks such that higher-level parent-tasks invoke lower-level child tasks as if they were primitive actions. A decomposition may have multiple levels of hierarchy. Some or all of the subproblems can themselves be reinforcement learning problems. When a parent-task is formulated as a reinforcement learning problem it is commonly formalized as a semi-Markov decision problem because its actions are child-tasks that persist for an extended period of time. The advantage of hierarchical decomposition is a reduction in computational complexity if the overall problem can be represented more compactly and reusable subtasks learned or provided independently. While the solution to a HRL problem is optimal given the constraints of the hierarchy there are no guarantees in general that the decomposed solution is an optimal solution to the original reinforcement problem.}
}
@article{HenkeletalCoRR-18,
       author = {Jordan Henkel and Shuvendu Lahiri and Ben Liblit and Thomas W. Reps},
        title = {Code Vectors: Understanding Programs Through Embedded Abstracted Symbolic Traces},
      journal = {CoRR},
       volume = {arXiv:1803.06686},
         year = {2018},
     abstract = {With the rise of machine learning, there is a great deal of interest in treating programs as data to be fed to learning algorithms. However, programs do not start off in a form that is immediately amenable to most off-the-shelf learning techniques. Instead, it is necessary to transform the program to a suitable representation before a learning technique can be applied. In this paper, we use abstractions of traces obtained from symbolic execution of a program as a representation for learning word embeddings. We trained a variety of word embeddings under hundreds of parameterizations, and evaluated each learned embedding on a suite of different tasks. In our evaluation, we obtain 93\% top-1 accuracy on a benchmark consisting of over 19,000 API-usage analogies extracted from the Linux kernel. In addition, we show that embeddings learned from (mainly) semantic abstractions provide nearly triple the accuracy of those learned from (mainly) syntactic abstractions.},
}
@article{HenleyJVLVB-69,
       author = {Henley, Nancy M.},
        title = {A psychological study of the semantics of animal terms},
      journal = {Journal of Verbal Learning and Verbal Behauior},
       volume = 8,
         year = 1969,
        pages = {176-184},
}
@article {HenleyetalPHYSIOLOGICAL_REVIEWS-14,
       author = {Henley, Jeremy M. and Craig, Tim J. and Wilkinson, Kevin A.},
        title = {Neuronal {SUMO}ylation: Mechanisms, Physiology, and Roles in Neuronal Dysfunction},
      journal = {Physiological Reviews},
    publisher = {American Physiological Society},
       volume = {94},
       number = {4},
         year = {2014},
        pages = {1249-1285},
     abstract = {Protein SUMOylation is a critically important posttranslational protein modification that participates in nearly all aspects of cellular physiology. In the nearly 20 years since its discovery, SUMOylation has emerged as a major regulator of nuclear function, and more recently, it has become clear that SUMOylation has key roles in the regulation of protein trafficking and function outside of the nucleus. In neurons, SUMOylation participates in cellular processes ranging from neuronal differentiation and control of synapse formation to regulation of synaptic transmission and cell survival. It is a highly dynamic and usually transient modification that enhances or hinders interactions between proteins, and its consequences are extremely diverse. Hundreds of different proteins are SUMO substrates, and dysfunction of protein SUMOylation is implicated in a many different diseases. Here we briefly outline core aspects of the SUMO system and provide a detailed overview of the current understanding of the roles of SUMOylation in healthy and diseased neurons.},
}
@book{HennessyandPatterson2012,
    author = {John L. Hennessy and David A. Patterson},
     title = {Computer Architecture: A Quantitative Approach},
 publisher = {5th Edition, Morgan Kaufmann},
   address = {San Francisco, CA},
      year = 2012,
}
@book{HennessyandPatterson97,
     author = {John L. Hennessy and David A. Patterson},
      title = {Computer Organization and Design: The Hardware/Software Interface},
  publisher = {Second edition, Morgan Kaufmann},
    address = {San Francisco, CA},
       year = 1997,
}
@article {HenriksenetalELIFE-16,
       author = {Henriksen, Sid and Pang, Rich and Wronkiewicz, Mark},
       editor = {Johansen-Berg, Heidi},
        title = {A simple generative model of the mouse mesoscale connectome},
      journal = {eLife},
    publisher = {eLife Sciences Publications, Ltd},
       volume = {5},
         year = {2016},
        pages = {e12366},
     abstract = {Recent technological advances now allow for the collection of vast data sets detailing the intricate neural connectivity patterns of various organisms. Oh et al. (2014) recently published the most complete description of the mouse mesoscale connectome acquired to date. Here we give an in-depth characterization of this connectome and propose a generative network model which utilizes two elemental organizational principles: proximal attachment ‒ outgoing connections are more likely to attach to nearby nodes than to distant ones, and source growth ‒ nodes with many outgoing connections are likely to form new outgoing connections. We show that this model captures essential principles governing network organization at the mesoscale level in the mouse brain and is consistent with biologically plausible developmental processes.},
}
@inproceedings{Henrion87a,
       author = {Henrion, M. and Cooley, D.R.},
        title = {An experimental comparison of knowledge engineering for expert systems and for decision analysis},
    booktitle = {Proceedings AAAI-87 Sixth National Conference on Artificial Intelligence, Seattle, WA},
    publisher = {Morgan Kaufmann, San Mateo, CA},
        pages = {471-476},
         year = 1987,
}
@inproceedings{HenrionID-88,
       author = {Max Henrion},
        title = {Towards efficient probabilistic diagnosis in multiply connected belief networks},
    booktitle = {Proceedings of the Conference on Influence Diagrams},
      address = {Berkeley, California},
         year = 1988,
}
@incollection{HenrionUAI-88,
       author = {Max Henrion},
        title = {Propagating Uncertainty by Logic Sampling in {B}ayes Networks},
     crossref = UAI-88,
        pages = {149-163},
}
@article{HenrionetalAIM-91,
        title = {Decision Analysis and Expert Systems},
       author = {Henrion, Max and Breese, John S. and Horvitz, Eric J.},
      journal = {AI Magazine},
       volume = {Winter},
        pages = {64-91},
         year = 1991,
} 
@inproceedings{Henzingeretal00,
       author = {M. Henzinger and A. Heydon and M. Mitzenmacher and M. Najork},
        title = {On Near-Uniform {URL} Sampling},
    booktitle = {Proceedings of the 9th World Wide Web Conference},
        pages = {},
         year = 2000,
}
@inproceedings{Henzingeretal99,
       author = {M. Henzinger and A. Heydon and M. Mitzenmacher and M. Najork},
        title = {Measuring Index Quality Using Random Walks on the {W}eb},
    booktitle = {Proceedings of the 8th World Wide Web Conference},
        pages = {213-225},
         year = 1999,
}
@article{Herculano-HouzeletalPNAS-08,
       author = {Herculano-Houzel, Suzana and Collins, Christine E. and Wong, Peiyan and Kaas, Jon H. and Lent, Roberto},
        title = {The basic nonuniformity of the cerebral cortex},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {105},
       number = {34},
        pages = {12593-12598},
         year = {2008},
    publisher = {National Academy of Sciences},
     abstract = {Evolutionary changes in the size of the cerebral cortex, a columnar structure, often occur through the addition or subtraction of columnar modules with the same number of neurons underneath a unit area of cortical surface. This view is based on the work of Rockel et al. [Rockel AJ, Hiorns RW, Powell TP (1980) The basic uniformity in structure of the neocortex. Brain 103:221-244], who found a steady number of approximately 110 neurons underneath a surface area of 750 μm2 (147,000 underneath 1 mm2) of the cerebral cortex of five species from different mammalian orders. These results have since been either corroborated or disputed by different groups. Here, we show that the number of neurons underneath 1 mm2 of the cerebral cortical surface of nine primate species and the closely related Tupaia sp. is not constant and varies by three times across species. We found that cortical thickness is not inversely proportional to neuronal density across species and that total cortical surface area increases more slowly than, rather than linearly with, the number of neurons underneath it. The number of neurons beneath a unit area of cortical surface varies linearly with neuronal density, a parameter that is neither related to cortical size nor total number of neurons. Our finding of a variable number of neurons underneath a unit area of the cerebral cortex across primate species indicates that models of cortical organization cannot assume that cortical columns in different primates consist of invariant numbers of neurons.},
}
@article{Herculano-HouzeletalPNAS-10,
       author = {Herculano-Houzel, Suzana and Mota, Bruno and Wong, Peiyan and Kaas, Jon H.},
        title = {Connectivity-driven white matter scaling and folding in primate cerebral cortex},
      journal = {Proceedings of the National Academy of Sciences},
       volume = 107,
        issue = 44,
        pages = {19008-19013},
         year = 2010,
     abstract = {Larger brains have an increasingly folded cerebral cortex whose white matter scales up faster than the gray matter. Here we analyze the cellular composition of the subcortical white matter in 11 primate species, including humans, and one Scandentia, and show that the mass of the white matter scales linearly across species with its number of nonneuronal cells, which is expected to be proportional to the total length of myelinated axons in the white matter. This result implies that the average axonal cross-section area in the white matter, a, does not scale significantly with the number of neurons in the gray matter, N. The surface area of the white matter increases with N0.87, not N1.0. Because this surface can be defined as the product of N, a, and the fraction n of cortical neurons connected through the white matter, we deduce that connectivity decreases in larger cerebral cortices as a slowly diminishing fraction of neurons, which varies with N--0.16, sends myelinated axons into the white matter. Decreased connectivity is compatible with previous suggestions that neurons in the cerebral cortex are connected as a small-world network and should slow down the increase in global conduction delay in cortices with larger numbers of neurons. Further, a simple model shows that connectivity and cortical folding are directly related across species. We offer a white matter-based mechanism to account for increased cortical folding across species, which we propose to be driven by connectivity-related tension in the white matter, pulling down on the gray matter.},
}
@article{HerdetalCoRR-19,
       author = {Seth Herd and Kai Krueger and Ananta Nair and Jessica Mollick and Randall OReilly},
        title = {Neural Mechanisms of Human Decision-Making},
      journal = {CoRR},
       volume = {arXiv:1912.07660},
         year = {2019},
     abstract = {We present a computational and theoretical model of the neural mechanisms underlying human decision-making. We propose a detailed model of the interaction between brain regions, under a proposer-predictor-actor-critic framework. Task-relevant areas of cortex propose a candidate plan using fast, model-free, parallel constraint-satisfaction computations. Other areas of cortex and medial temporal lobe can then predict likely outcomes of that plan in this situation. This step is optional. This prediction-(or model-) based computation produces better accuracy and generalization, at the expense of speed. Next, linked regions of basal ganglia act to accept or reject the proposed plan based on its reward history in similar contexts. Finally the reward-prediction system acts as a critic to determine the value of the outcome relative to expectations, and produce dopamine as a training signal for cortex and basal ganglia. This model gains many constraints from the hypothesis that the mechanisms of complex human decision-making are closely analogous to those that have been empirically studied in detail for animal action-selection. We argue that by operating sequentially and hierarchically, these same mechanisms are responsible for the most complex human plans and decisions. Finally, we use the computational model to generate novel hypotheses on causes of human risky decision-making, and compare this to other theories of human decision-making.}
}
@article{HereldetalIEEE-04,
       author = {Hereld, M. and Stevens, R. L. and van Drongelen, W. and Lee, H. C.},
        title = {Developing a petascale neural simulation},
      journal = {Conference Proceedings {IEEE} Engineering in Medicine and Biology Society},
       volume = {6},
         year = {2004},
        pages = {3999-4002},
     abstract = {Simulations of large neural networks have the potential to contribute uniquely to the study of epilepsy, from the effects of extremely local changes in neuron environment and behavior, to the effects of large scale wiring anomalies. Currently, simulations with sufficient detail in the neuron model, however, are limited to cell counts that are far smaller than scales measured by typical probes. Furthermore, it is likely that future simulations will follow the path that large-scale simulations in other fields have and include hierarchically interacting components covering different scales and different biophysics. The resources needed for problem solving in this domain call for petascale computing--computing with supercomputers capable of 10^(15) operations a second and holding datasets of 10^(15) bytes in memory. We will lay out the structure of our simulation of epileptiform electrical activity in the neocortex, describe experiments and models of its scaling behavior in large cluster supercomputers, identify tight spots in this behavior, and project the performance onto a candidate next generation computing platform.}
}
@article{HereldetalIJBEM-05,
       author = {Mark Hereld and Rick Stevens and Justin Teller and Wim van Drongelen and Hyong Lee},
        title = {Large Neural Simulations on Large Parallel Computers},
      journal = {International Journal of Bioelectromagnetism},
       volume = {7},
        issue = {1},
         year = {2005},
     abstract = {Simulations of biologically realistic neurons in large densely connected networks pose many problems to application programmers, particularly on distributed memory computers. We discuss simulations of hundreds of thousands to millions of cells in a model of neocortex in the context of new computing platforms with many tens of thousands to hundreds of thousands of processing elements. We are developing a performance model for this simulation so that we can gauge its performance on these platforms in terms of memory usage, time to setup and execute the simulation, and to estimate the practical limits to the size and simulation timescale available to us in our simulation experiments. Recent results from runs on a BlueGene/L computer, which could ultimately scale to over one hundred thousand processors, are described.}
}
@article{HerffetalFiN-15,
       author = {Herff, Christian  and  Heger, Dominic  and  de Pesters, Adriana  and  Telaar, Dominic  and  Brunner, Peter  and  Schalk, Gerwin  and  Schultz, Tanja},
        title = {Brain-to-text: Decoding spoken phrases from phone representations in the brain},
      journal = {Frontiers in Neuroscience},
       volume = 9,
         year = 2015,
        issue = 217,
     abstract = {It has long been speculated whether communication between humans and machines based on natural speech related cortical activity is possible. Over the past decade, studies have suggested that it is feasible to recognize isolated aspects of speech from neural signals, such as auditory features, phones or one of a few isolated words. However, until now it remained an unsolved challenge to decode continuously spoken speech from the neural substrate associated with speech and language processing. Here, we show for the first time that continuously spoken speech can be decoded into the expressed words from intracranial electrocorticographic (ECoG) recordings. Specifically, we implemented a system, which we call Brain-To-Text that models single phones, employs techniques from automatic speech recognition (ASR), and thereby transforms brain activity while speaking into the corresponding textual representation. Our results demonstrate that our system can achieve word error rates as low as 25\% and phone error rates below 50\%. Additionally, our approach contributes to the current understanding of the neural basis of continuous speech production by identifying those cortical regions that hold substantial information about individual phones. In conclusion, the Brain-To-Text system described in this paper represents an important step toward human-machine communication based on imagined speech.}
}
@article{HeringandShengNATURE-01,
       author = {Hering, Heike and Sheng, Morgan},
        title = {Dentritic spines: structure, dynamics and regulation},
      journal = {Nature Reviews Neuroscience},
         year = {2001},
       volume = {2},
        issue = {12},
        pages = {880-888},
     abstract = {Dendritic spines are tiny protrusions that receive excitatory synaptic input and compartmentalize postsynaptic responses. Heterogeneous in size and shape, and modifiable by activity and experience, dendritic spines have long been thought to provide a morphological basis for synaptic plasticity. Although advanced imaging techniques have highlighted the rapid and regulated motility of spines in living neurons, the functional significance of spine plasticity remains elusive. Recent insights into the molecular mechanisms that regulate spine morphogenesis offer potential ways to manipulate dendritic spines in vivo and to explore their physiological roles in the brain.}
}
@article{HermannetalCoRR-15,
       author = {Karl Moritz Hermann and Tom\'{a}\v{s} Ko\v{c}isk\'{y} and Edward Grefenstette and Lasse Espeholt and Will Kay and Mustafa Suleyman and Phil Blunsom},
        title = {Teaching Machines to Read and Comprehend},
      journal = {CoRR},
       volume = {arXiv:1506.03340},
         year = 2015,
     abstract = {Teaching machines to read natural language documents remains an elusive challenge. Machine reading systems can be tested on their ability to answer questions posed on the contents of documents that they have seen, but until now large scale training and test datasets have been missing for this type of evaluation. In this work we define a new methodology that resolves this bottleneck and provides large scale supervised reading comprehension data. This allows us to develop a class of attention based deep neural networks that learn to read real documents and answer complex questions with minimal prior knowledge of language structure.}
}
@article{HermannetalCoRR-17,
       author = {Karl Moritz Hermann and Felix Hill and Simon Green and Fumin Wang and Ryan Faulkner and Hubert Soyer and David Szepesvari and Wojciech Marian Czarnecki and Max Jaderberg and Denis Teplyashin and Marcus Wainwright and Chris Apps and Demis Hassabis and Phil Blunsom},
        title = {Grounded Language Learning in a Simulated 3D World},
      journal = {CoRR},
       volume = {arXiv:1706.06551},
         year = {2017},
     abstract = {We are increasingly surrounded by artificially intelligent technology that takes decisions and executes actions on our behalf. This creates a pressing need for general means to communicate with, instruct and guide artificial agents, with human language the most compelling means for such communication. To achieve this in a scalable fashion, agents must be able to relate language to the world and to actions; that is, their understanding of language must be grounded and embodied. However, learning grounded language is a notoriously challenging problem in artificial intelligence research. Here we present an agent that learns to interpret language in a simulated 3D environment where it is rewarded for the successful execution of written instructions. Trained via a combination of reinforcement and unsupervised learning, and beginning with minimal prior knowledge, the agent learns to relate linguistic symbols to emergent perceptual representations of its physical surroundings and to pertinent sequences of actions. The agent's comprehension of language extends beyond its prior experience, enabling it to apply familiar language to unfamiliar situations and to interpret entirely novel instructions. Moreover, the speed with which this agent learns new words increases as its semantic knowledge grows. This facility for generalising and bootstrapping semantic knowledge indicates the potential of the present approach for reconciling ambiguous natural language with the complexity of the physical world.}
}
@misc{HermannetalLxMLS-14,
        title = {Natural Language Understanding in a Continuous Space},
       author = {Karl-Moritz Hermann and Nal Kalchbrenner and Edward Grefenstette and Phil Blunsom},
         year = 2014,
          url = {http://lxmls.it.pt/2014/blunsom-lxmls.pdf}}
@incollection{Hertz2002,
       author = {Hertz, J. A.},
        title = {Neurons, Networks, and Cognition: An Introduction to Neural Modeling},
    booktitle = {Models of Neural Networks {IV}: Early Vision and Attention},
       editor = {Hemmen, J. Leo and Cowan, Jack D. and Domany, Eytan},
    publisher = {Springer New York},
      address = {New York, NY},
         year = {2002},
        pages = {1-46},
     abstract = {This chapter deals with the modeling of neural systems at three levels: (1) single neurons, described by Hodgkin-Huxley equations and simpler integrate-and-fire models; (2) the dynamics of local cortical circuitry, in particular the observed irregularity of neuronal firing; and (3) cognitive computation, illustrated by the case of associative memory. The examples show how physically-based analysis, particularly methods from statistical mechanics, gives insight into the dynamics and function of the nervous system.}
}
@book{Hertz91,
       author = {Hertz, John and Krogh, Anders and Palmer, Richard G.},
        title = {Introduction to the Theory of Neural Computation},
    publisher = {Addison Wesley},
      address = {Redwood City, California},
         year = 1991,
}
@article{HertzbergetalMEDICAL_PHYSiCS-10,
    author = {Hertzberg, Y. and Volovick, A. and Zur, Y. and Medan, Y. and Vitek, S. and Navon G.},
     title = {Ultrasound focusing using magnetic resonance acoustic radiation force imaging: application to ultrasound transcranial therapy.},
   journal = {Medical Physics},
    volume = 37,
     issue = 6,
      year = 2010,
     pages = {2934-2942},
  abstract = {PURPOSE: Magnetic resonance guided ultrasonic therapy is a promising minimally invasive technology for constantly growing variety of clinical applications. Delivery of focused ultrasound (FUS) energy to the targeted point with optimal intensity is highly desired; however, due to tissue aberrations, optimal focal intensity is not always achieved. Especially in transcranial applications, the acoustic waves are shifted and distorted mainly by the skull. In order to verify that magnetic resonance acoustic radiation force imaging (MR-ARFI) can be used as a focusing tool in transcranial treatments, such an imaging was applied in vivo on a porcine brain via ex vivo human skull. Then, by the use of MR-ARFI technique, an improved ultrasound focusing algorithm is proposed and demonstrated for both transcranial and none brain applications.

METHODS: MR-ARFI images were acquired on a GE 1.5 T scanner equipped with InSightec FUS systems ExAblate 2000 and ExAblate 4000. Imaging was performed with MR-ARFI sequences of line-scan spin-echo and single-shot gradient-echo echo-planar. The in-plane resolution of both acquisitions was 0.9 x 0.9 mm2. The total acquisition time of MR-ARFI image was 31 s by the line-scan sequence and 1 s by the echo-planar sequence. An in vivo experiment was performed using FUS transducer, which is built out of 1024 ultrasound transmitting piezoelectric elements at 220 kHz frequency. The transducer was focused into the brain of a pig, which was wrapped in a human skull, in degassed water environment to resemble human treatments. The pig underwent a wide bilateral craniectomy to prevent a bone heating from the ultrasound beams. Two focusing experiments were performed in phantoms using 1 MHz and 710 kHz FUS transducers working with 208 and 225 elements, respectively. In the first experiment, aberration was added virtually to the apparatus by adding random phases to the phase map of the transducer. A simple focusing correction scheme was used, in which the corrected phase of a group of elements was chosen such that it maximizes the radiation force at the focal point. In the second experiment, aberrations made by a human skull were corrected using geometrical and phase based adjustments on segments of the transducer.

RESULTS: A maximum displacement of 10 microm was obtained using 1.4 kW acoustic power on a live pig's head that its skull was removed and replaced by ex vivo human skull. Aberration correction using MR-ARFI resulted in near optimal focus, as the radiation force was similar to the nonaberration case. Transcranial, MR-ARFI based aberration correction performed better than CT based aberration correction, a technique that is currently used in brain FUS treatments.

CONCLUSIONS: In the present work, the authors show for the first time a result of MR-ARFI in a live brain through ex vivo human skull. They have demonstrated that aberration correction could be done using MR-ARFI by measuring the radiation force at the focal point. Aberration correction using MR-ARFI is a promising noninvasive technique for transcranial focusing, which may result in near optimal focus and more reliable and safer brain FUS treatments.}
}
@article{HeskesNC-04,
       author = {Tom Heskes},
        title = {On the uniqueness of loopy belief propagation fixed points},
      journal = {Neural Computation},
    publisher = {MIT Press},
      address = {Cambridge, MA},
       volume = 16,
        issue = 11,
        pages = {2379-2413},
         year = 2004,
}
@inproceedings{HeskesetalUAI-03,
       author = {Tom Heskes and Kees Albers and Bert Kappen},
        title = {Approximate Inference and Constrained Optimization},
    booktitle = {Proceedings of the 19th Annual Conference on Uncertainty in Artificial Intelligence},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
        pages = {313-320},
         year = 2003,
}
@article{HesposandSpelkeNATURE-04,
       author = {Hespos, S. J. and Spelke, E. S.},
        title = {Conceptual precursors to language},
      journal = {Nature},
         year = {2004},
       volume = {430},
       number = {6998},
        pages = {453-456},
     abstract = {Because human languages vary in sound and meaning, children must learn which distinctions their language uses. For speech perception, this learning is selective: initially infants are sensitive to most acoustic distinctions used in any language, and this sensitivity reflects basic properties of the auditory system rather than mechanisms specific to language; however, infants' sensitivity to non-native sound distinctions declines over the course of the first year. Here we ask whether a similar process governs learning of word meanings. We investigated the sensitivity of 5-month-old infants in an English-speaking environment to a conceptual distinction that is marked in Korean but not English; that is, the distinction between 'tight' and 'loose' fit of one object to another. Like adult Korean speakers but unlike adult English speakers, these infants detected this distinction and divided a continuum of motion-into-contact actions into tight- and loose-fit categories. Infants' sensitivity to this distinction is linked to representations of object mechanics that are shared by non-human animals. Language learning therefore seems to develop by linking linguistic forms to universal, pre-existing representations of sound and meaning.}
}
@article{HewittAIJ-77,
       author = {Hewitt, Carl},
        title = {Viewing Control Structures as Patterns of Passing Messages},
      journal = AIJ,
       volume = 8,
         year = 1977,
        pages = {323-364},
}
@inproceedings{HewittPLANNER,
       author = {Hewitt, Carl},
        title = {PLANNER: A Language for Proving Theorems in Robots},
    booktitle = {Proceedings {IJCAI} 1},
     location = {Washington, DC, USA},
 organization = {IJCAII},
         year = 1969,
        pages = {295-301},
}
@article{HickokPLR-09,
       author = {Hickok, Gregory},
        title = {The functional neuroanatomy of language},
      journal = {Physics of Life Reviews},
       volume = {6},
        issue = {3},
         year = {2009},
        pages = {121-143},
     abstract = {There has been substantial progress over the last several years in understanding aspects of the functional neuroanatomy of language. These advances are summarized in this review. It will be argued that recognizing speech sounds is carried out in the superior temporal lobe bilaterally, that the superior temporal sulcus bilaterally is involved in phonological-level aspects of this process, that the frontal/motor system is not central to speech recognition although it may modulate auditory perception of speech, that conceptual access mechanisms are likely located in the lateral posterior temporal lobe (middle and inferior temporal gyri), that speech production involves sensory-related systems in the posterior superior temporal lobe in the left hemisphere, that the interface between perceptual and motor systems is supported by a sensory-motor circuit for vocal tract actions (not dedicated to speech) that is very similar to sensory-motor circuits found in primate parietal lobe, that verbal short-term memory can be understand as an emergent property of this sensory-motor circuit. These observations are understood within the context of a dual stream model of speech processing in which one pathway supports speech comprehension and the other supports sensory-motor integration. Additional topics of discussion include the functional organization of the planum temporale for spatial hearing and speech-related sensory-motor processes, the anatomical and functional basis of a form of acquired language disorder, conduction aphasia, the neural basis of vocabulary development, and sentence-level/grammatical processing.},
}
@article{HickokandPoeppelNATURE-07,
       author = {Hickok, Gregory and Poeppel, David},
        title = {The cortical organization of speech processing},
      journal = {Nature Reviews Neuroscience},
         year = {2007},
    publisher = {Nature Publishing Group},
       volume = {8},
        pages = {393},
     abstract = {Despite decades of research, the functional neuroanatomy of speech processing has been difficult to characterize. A major impediment to progress may have been the failure to consider task effects when mapping speech-related processing systems. We outline a dual-stream model of speech processing that remedies this situation. In this model, a ventral stream processes speech signals for comprehension, and a dorsal stream maps acoustic speech signals to frontal lobe articulatory networks. The model assumes that the ventral stream is largely bilaterally organized -- although there are important computational differences between the left- and right-hemisphere systems -- and that the dorsal stream is strongly left-hemisphere dominant.}
}
@book{HickokandSmall2015neurolanguage,
        title = {Neurobiology of Language},
       author = {Hickok, G. and Small, S.L.},
    publisher = {Elsevier},
         year = {2015},
     abstract = {Neurobiology of Language explores the study of language, a field that has seen tremendous progress in the last two decades. Key to this progress is the accelerating trend toward integration of neurobiological approaches with the more established understanding of language within cognitive psychology, computer science, and linguistics. This volume serves as the definitive reference on the neurobiology of language, bringing these various advances together into a single volume of 100 concise entries. The organization includes sections on the field's major subfields, with each section covering both empirical data and theoretical perspectives. "Foundational" neurobiological coverage is also provided, including neuroanatomy, neurophysiology, genetics, linguistic, and psycholinguistic data, and models.}
}
@article{HigginsetalICLR-17,
        title = {Towards a Definition of Disentangled Representations},
       author = {Irina Higgins and David Amos and David Pfau and Sebastien Racaniere and Loic Matthey and Danilo Rezende and Alexander Lerchner},
         year = {2017},
    booktitle = {International Conference on Learning Representations},
         year = 2017,
     abstract = {Learning an interpretable factorised representation of the independent data gen- erative factors of the world without supervision is an important precursor for the development of artificial intelligence that is able to learn and reason in the same way that humans do. We introduce \Beta-VAE, a new state-of-the-art framework for automated discovery of interpretable factorised latent representations from raw image data in a completely unsupervised manner. Our approach is a modification of the variational autoencoder (VAE) framework. We introduce an adjustable hy- perparameter \Beta that balances latent channel capacity and independence constraints with reconstruction accuracy. We demonstrate that \Beta-VAE with appropriately tuned \Beta > 1 qualitatively outperforms VAE (\Beta = 1), as well as state of the art unsu- pervised (InfoGAN) and semi-supervised (DC-IGN) approaches to disentangled factor learning on a variety of datasets (celebA, faces and chairs). Furthermore, we devise a protocol to quantitatively compare the degree of disentanglement learnt by different models, and show that our approach also significantly outperforms all baselines quantitatively. Unlike InfoGAN, \Beta-VAE is stable to train, makes few assumptions about the data and relies on tuning a single hyperparameter \Beta, which can be directly optimised through a hyperparameter search using weakly labelled data or through heuristic visual inspection for purely unsupervised data.}
}
@article{HigginsetalICLR-18,
        title = {{SCAN}: Learning Hierarchical Compositional Visual Concepts},
       author = {Irina Higgins and Nicolas Sonnerat and Loic Matthey and Arka Pal and Christopher P Burgess and Matko Bošnjak and Murray Shanahan and Matthew Botvinick and Demis Hassabis and Alexander Lerchner},
      journal = {International Conference on Learning Representations},
         year = {2018},
     abstract = {The seemingly infinite diversity of the natural world arises from a relatively small set of coherent rules, such as the laws of physics or chemistry. We conjecture that these rules give rise to regularities that can be discovered through primarily unsupervised experiences and represented as abstract concepts. If such representations are compositional and hierarchical, they can be recombined into an exponentially large set of new concepts. This paper describes SCAN (Symbol-Concept Association Network), a new framework for learning such abstractions in the visual domain. SCAN learns concepts through fast symbol association, grounding them in disentangled visual primitives that are discovered in an unsupervised manner. Unlike state of the art multimodal generative model baselines, our approach requires very few pairings between symbols and images and makes no assumptions about the form of symbol representations. Once trained, SCAN is capable of multimodal bi-directional inference, generating a diverse set of image samples from symbolic descriptions and vice versa. It also allows for traversal and manipulation of the implicit hierarchy of visual concepts through symbolic instructions and learnt logical recombination operations. Such manipulations enable SCAN to break away from its training data distribution and imagine novel visual concepts through symbolically instructed recombination of previously learnt concepts.}
}
@incollection{Higher_Cortical_Functions_Association,
       author = {Anthony Wright},
        title = {Higher Cortical Functions: Association and Executive Processing},
    booktitle = {Neuroscience Online: An electronic textbook for the neurosciences},
    publisher = {The University of Texas McGovern Medical School},
          url = {https://nba.uth.tmc.edu/neuroscience/m/s4/chapter09.html},
         year = {1997},
     abstract = {Higher-order integrative cortical areas, called association areas, intervene between the sensory inputs and motor outputs. Hierarchical organization of the cortex was proposed in the 1870s by the famous British neurologist, John Hughlings Jackson. These association areas are the site of cognitive processing and are the focal areas in the rapidly developing field of cognitive neuroscience.},
}
@inproceedings{HihiandBengioNIPS-96,
       author = {El Hihi, Salah and Bengio, Yoshua},
        title = {Hierarchical Recurrent Neural Networks for Long-Term Dependencies},
    booktitle = {Advances in Neural Information Processing Systems 8},
       editor = {Touretzky, D. S. and Mozer, M. and Hasselmo, M.E.},
         year = 1996,
    publisher = {MIT Press},
     abstract = {We have already shown that extracting long-term dependencies from sequential data is difficult, both for deterministic dynamical systems such as recurrent networks, and probabilistic models such as hidden Markov models (HMMs) or input/output hidden Markov models (IOHMMs). In practice, to avoid this problem, researchers have used domain specific a-priori knowledge to give meaning to the hidden or state variables representing past context. We propose to use a more general type of a-priori knowledge, namely that the temporal dependencies are structured hierarchically. This implies that long-term dependencies are represented by variables with a long time scale. This principle is applied to a recurrent network which includes delays and multiple time scales. Experiments confirm the advantages of such structures. A similar approach is proposed for HMMs and IOHMMs.},
}
@article{HildebrandetalNATURE-17,
       author = {Hildebrand, David Grant Colburn and Cicconet, Marcelo and Torres, Russel Miguel and Choi, Woohyuk and Quan, Tran Minh and Moon, Jungmin and Wetzel, Arthur Willis and Scott Champion, Andrew and Graham, Brett Jesse and Randlett, Owen and Plummer, George Scott and Portugues, Ruben and Bianco, Isaac Henry and Saalfeld, Stephan and Baden, Alexander David and Lillaney, Kunal and Burns, Randal and Vogelstein, Joshua Tzvi and Schier, Alexander Franz and Lee, Wei-Chung Allen and Jeong, Won-Ki and Lichtman, Jeff William and Engert, Florian},
        title = {Whole-brain serial-section electron microscopy in larval zebrafish},
      journal = {Nature},
       volume = {advance online publication},
         year = {2017},
    publisher = {Nature Publishing Group},
     abstract = {High-resolution serial-section electron microscopy (ssEM) makes it possible to investigate the dense meshwork of axons, dendrites, and synapses that form neuronal circuits1. However, the imaging scale required to comprehensively reconstruct these structures is more than ten orders of magnitude smaller than the spatial extents occupied by networks of interconnected neurons2, some of which span nearly the entire brain. Difficulties in generating and handling data for large volumes at nanoscale resolution have thus restricted vertebrate studies to fragments of circuits. These efforts were recently transformed by advances in computing, sample handling, and imaging techniques1, but high-resolution examination of entire brains remains a challenge. Here, we present ssEM data for the complete brain of a larval zebrafish (Danio rerio) at 5.5 days post-fertilization. Our approach utilizes multiple rounds of targeted imaging at different scales to reduce acquisition time and data management requirements. The resulting dataset can be analysed to reconstruct neuronal processes, permitting us to survey all myelinated axons (the projectome). These reconstructions enable precise investigations of neuronal morphology, which reveal remarkable bilateral symmetry in myelinated reticulospinal and lateral line afferent axons. We further set the stage for whole-brain structure-function comparisons by co-registering functional reference atlases and in vivo two-photon fluorescence microscopy data from the same specimen. All obtained images and reconstructions are provided as an open-access resource.}
}
@book{Hildreth84,
       author = {Ellen Catherine Hildreth},
        title = {Measurement of Visual Motion},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
         year = 1984,
}
@article{HilgetagandKaiserNI-04,
     title = {Clustered organization of cortical connectivity},
    author = {Claus Hilgetag and Marcus Kaiser},
   journal = {Neuroinformatics},
    volume = 2, 
     issue = 3, 
     pages = {353-360},
      year = 2004,
}
@misc{HillANALOGY-LIST-21,
        title = {Grounded Language Learning (With Analogy) and a Developmental Pathway to Human-Like Intelligence in Machines},
       author = {Felix Hill},
 howpublished = {\url{https://www.youtube.com/watch?v=U0wL5YKahP8}},
         year = {2021},
	 abstract = {Adult humans are capable of flexible conceptual abstract thought, which enables impressive feats of problem-solving and creativity. One view is that our ability to think in the abstract is the outcome of a process of cognitive development that begins with physical experiences in concrete situations; a process that is driven by analogy-making, and critically supported by language. Motivated by this perspective, I'll first present some recent work on embodied language learning at DeepMind that focuses on highly concrete situations. I'll then illustrate some preliminary work training neural networks to make analogies in more abstract domains. Finally, I hope to discuss how these two ingredients may in future be combined in a developmental pathway supporting the emergence of abstract and analogical reasoning, conceptual behaviour and language use in an embodied artificial agent.}
}
@article{HillandTononiNEUROPHYSIOLOGY-05,
     title = {Modeling sleep and wakefulness in the thalamocortical system},
    author = {Hill, S. and Tononi, G.},
   journal = {Journal of Neurophysiology},
    volume = 93,
     issue = 3,
     pages = {1671-98},
      year = 2005,
  abstract = {When the brain goes from wakefulness to sleep, cortical neurons begin to undergo slow oscillations in their membrane potential that are synchronized by thalamocortical circuits and reflected in EEG slow waves. To provide a self-consistent account of the transition from wakefulness to sleep and of the generation of sleep slow waves, we have constructed a large-scale computer model that encompasses portions of two visual areas and associated thalamic and reticular thalamic nuclei. Thousands of model neurons, incorporating several intrinsic currents, are interconnected with millions of thalamocortical, corticothalamic, and both intra- and interareal corticocortical connections. In the waking mode, the model exhibits irregular spontaneous firing and selective responses to visual stimuli. In the sleep mode, neuromodulatory changes lead to slow oscillations that closely resemble those observed in vivo and in vitro. A systematic exploration of the effects of intrinsic currents and network parameters on the initiation, maintenance, and termination of slow oscillations shows the following. 1) An increase in potassium leak conductances is sufficient to trigger the transition from wakefulness to sleep. 2) The activation of persistent sodium currents is sufficient to initiate the up-state of the slow oscillation. 3) A combination of intrinsic and synaptic currents is sufficient to maintain the up-state. 4) Depolarization-activated potassium currents and synaptic depression terminate the up-state. 5) Corticocortical connections synchronize the slow oscillation. The model is the first to integrate intrinsic neuronal properties with detailed thalamocortical anatomy and reproduce neural activity patterns in both wakefulness and sleep, thereby providing a powerful tool to investigate the role of sleep in information transmission and plasticity}
}
@article{HillarandSommerITIT-15,
       author = {C. J. Hillar and F. T. Sommer},
      journal = {IEEE Transactions on Information Theory},
        title = {When Can Dictionary Learning Uniquely Recover Sparse Data From Subsamples?},
         year = {2015},
       volume = {61},
       number = {11},
        pages = {6290-6297},
     abstract = {Sparse coding or sparse dictionary learning has been widely used to recover underlying structure in many kinds of natural data. Here, we provide conditions guaranteeing when this recovery is universal; that is, when sparse codes and dictionaries are unique (up to natural symmetries). Our main tool is a useful lemma in combinatorial matrix theory that allows us to derive bounds on the sample sizes guaranteeing such uniqueness under various assumptions for how training data are generated. Whenever the conditions to one of our theorems are met, any sparsity-constrained learning algorithm that succeeds in reconstructing the data recovers the original sparse codes and dictionary. We also discuss potential applications to neuroscience and data analysis.},
}
@inproceedings{HillelandWeinshallICML-07,
       author = {Aharon Bar Hillel and Daphna Weinshall},
        title = {Learning distance function by coding similarity},
    booktitle = {Proceedings of the 24th International Conference on Machine learning},
     location = {Corvalis, Oregon},
    publisher = {ACM Press},
      address = {New York, NY},
        pages = {65-72},
         year = 2007,
}
@inbook{HillertEncyclopedia_Psychology-16,
       author = {Hillert, Dieter},
        title = {Neurobiology of Language},
    booktitle = {Encyclopedia of Evolutionary Psychological Science},
       editor = {T.K. Shackelford and V.A. Weekes-Shackelford},
    publisher = {Springer},
         year = {2016},
        pages = {1-15},
     abstract = {The research area neurobiology of language aims to understand the genetic and neurological capacity that enables modern humans to acquire and use language. This language-related neurobiological capacity includes internalized syntactic, phonological, morphological, and semantic structures. Linguistic structures, which are externalized due to linguistic-cultural changes, are therefore not considered. Other relevant approaches comprise relevant genetic foundations, the examination of neural regions and circuits, and comparative studies mainly conducted with songbirds, cetaceans, and nonhuman primates. It is however controversially discussed what language is and how it might have emerged in the lineage of hominids.}
}
@article{HilletalCoRR-20,
       author = {Felix Hill and Olivier Tieleman and Tamara von Glehn and Nathaniel Wong and Hamza Merzic and Stephen Clark},
        title = {Grounded Language Learning Fast and Slow},
      journal = {CoRR},
       volume = {arXiv:2009.01719},
         year = {2020},
     abstract = {Recent work has shown that large text-based neural language models, trained with conventional supervised learning objectives, acquire a surprising propensity for few- and one-shot learning. Here, we show that an embodied agent situated in a simulated 3D world, and endowed with a novel dual-coding external memory, can exhibit similar one-shot word learning when trained with conventional reinforcement learning algorithms. After a single introduction to a novel object via continuous visual perception and a language prompt ("This is a dax"), the agent can re-identify the object and manipulate it as instructed ("Put the dax on the bed"). In doing so, it seamlessly integrates short-term, within-episode knowledge of the appropriate referent for the word "dax" with long-term lexical and motor knowledge acquired across episodes (i.e. "bed" and "putting"). We find that, under certain training conditions and with a particular memory writing mechanism, the agent's one-shot word-object binding generalizes to novel exemplars within the same ShapeNet category, and is effective in settings with unfamiliar numbers of objects. We further show how dual-coding memory can be exploited as a signal for intrinsic motivation, stimulating the agent to seek names for objects that may be useful for later executing instructions. Together, the results demonstrate that deep neural networks can exploit meta-learning, episodic memory and an explicitly multi-modal environment to account for 'fast-mapping', a fundamental pillar of human cognitive development and a potentially transformative capacity for agents that interact with human users.}
}
@inproceedings{HilletalICLR-19,
       author = {Felix Hill and Adam Santoro and David G. T. Barrett and Ari S. Morcos and Timothy P. Lillicrap},
        title = {Learning to Make Analogies by Contrasting Abstract Relational Structure},
    booktitle = {International Conference on Learning Representations},
      comment = {arXiv:1902.00120},
         year = {2019},
     abstract = {Analogical reasoning has been a principal focus of various waves of AI research. Analogy is particularly challenging for machines because it requires relational structures to be represented such that they can be flexibly applied across diverse domains of experience. Here, we study how analogical reasoning can be induced in neural networks that learn to perceive and reason about raw visual data. We find that the critical factor for inducing such a capacity is not an elaborate architecture, but rather, careful attention to the choice of data and the manner in which it is presented to the model. The most robust capacity for analogical reasoning is induced when networks learn analogies by contrasting abstract relational structures in their input domains, a training method that uses only the input data to force models to learn about important abstract features. Using this technique we demonstrate capacities for complex, visual and symbolic analogy making and generalisation in even the simplest neural network architectures.}
}
@inproceedings{HilletalICLR-20,
        title = {Environmental drivers of systematicity and generalization in a situated agent},
       author = {Felix Hill and Andrew Lampinen and Rosalia Schneider and Stephen Clark and Matthew Botvinick and James L. McClelland and Adam Santoro},
    booktitle = {International Conference on Learning Representations},
         year = {2020},
     abstract = {The question of whether deep neural networks are good at generalising beyond their immediate training experience is of critical importance for learning-based approaches to AI. Here, we consider tests of out-of-sample generalisation that require an agent to respond to never-seen-before instructions by manipulating and positioning objects in a 3D Unity simulated room. We first describe a comparatively generic agent architecture that exhibits strong performance on these tests. We then identify three aspects of the training regime and environment that make a significant difference to its performance: (a) the number of object/word experiences in the training set; (b) the visual invariances afforded by the agent's perspective, or frame of reference; and (c) the variety of visual input inherent in the perceptual aspect of the agent's perception. Our findings indicate that the degree of generalisation that networks exhibit can depend critically on particulars of the environment in which a given task is instantiated. They further suggest that the propensity for neural networks to generalise in systematic ways may increase if, like human children, those networks have access to many frames of richly varying, multi-modal observations as they learn.}
}
@misc{HilletalOPENREVIEW-18,
        title = {Understanding Grounded Language Learning Agents},
       author = {Felix Hill and Karl Moritz Hermann and Phil Blunsom and Stephen Clark},
         year = {2018},
     abstract = {Neural network-based systems can now learn to locate the referents of words and phrases in images, answer questions about visual scenes, and even execute symbolic instructions as first-person actors in partially-observable worlds. To achieve this so-called grounded language learning, models must overcome certain well-studied learning challenges that are also fundamental to infants learning their first words. While it is notable that models with no meaningful prior knowledge overcome these learning obstacles, AI researchers and practitioners currently lack a clear understanding of exactly how they do so. Here we address this question as a way of achieving a clearer general understanding of grounded language learning, both to inform future research and to improve confidence in model predictions. For maximum control and generality, we focus on a simple neural network-based language learning agent trained via policy-gradient methods to interpret synthetic linguistic instructions in a simulated 3D world. We apply experimental paradigms from developmental psychology to this agent, exploring the conditions under which established human biases and learning effects emerge. We further propose a novel way to visualise and analyse semantic representation in grounded language learning agents that yields a plausible computational account of the observed effects.},
}
@article{HilletalPNAS-12,
       author = {Hill, S. L. and Wang, Y. and Riachi, I. and Sch\"{u}rmann, F. and Markram, H.},
        title = {{S}tatistical connectivity provides a sufficient foundation for specific functional connectivity in neocortical neural microcircuits},
      journal = {Proceedings of the National Academy of Science U.S.A.},
         year = 2012,
       volume = 109,
        issue = 42,
        pages = {E2885-2894},
     abstract = {It is well-established that synapse formation involves highly selective chemospecific mechanisms, but how neuron arbors are positioned before synapse formation remains unclear. Using {3D} reconstructions of 298 neocortical cells of different types (including nest basket, small basket, large basket, bitufted, pyrami- dal, and Martinotti cells), we constructed a structural model of a cortical microcircuit, in which cells of different types were independently and randomly placed. We compared the positions of physical appositions resulting from the incidental overlap of axonal and dendritic arbors in the model (statistical structural connectivity) with the positions of putative functional synapses (functional synaptic connectivity) in 90 synaptic connections reconstructed from cortical slice preparations. Overall, we found that statistical connectivity predicted an average of 74 plus-or-minus 2.7\% (mean  SEM) synapse location distributions for nine types of cortical connections. This finding suggests that chemospecific attractive and repulsive mechanisms generally do not result in pairwise-specific connectivity. In some cases, however, the predicted distributions do not match precisely, indicating that chemospecific steering and aligning of the arbors may occur for some types of connections. This finding suggests that random alignment of axonal and dendritic arbors provides a sufficient foundation for specific functional connectivity to emerge in local neural microcircuits.}
}
@book{Hiltzik99,
      author = {Michael Hiltzik},
       title = {Dealers of Lightning: {X}erox {PARC} and the Dawn of the Computer Age},
   publisher = {Harper-Collins},
     address = {New York, NY},
        year = 1999,
}
@article{HinoshitaaetalNN-11,
       author = {Wataru Hinoshitaa and Hiroaki Arieb and Jun Tanib and Hiroshi G. Okunoa and Tetsuya Ogataa},
        title = {Emergence of Hierarchical Structure mirroring Linguistic Composition in a Recurrent Neural Network},
      journal = {Neural Networks},
       volume = 24,
         year = 2011,
        pages = {311-320},
abstract = {We show that a Multiple Timescale Recurrent Neural Network (MTRNN) can acquire the capabilities to recognize, generate, and correct sentences by self-organizing in a way that it mirrors the hierarchical structure of sentences: characters grouped into words, and words into sentences. In an experiment, we trained our model over a set of unannotated sentences from an artificial language, represented as sequences of characters. Once trained, the model could recognize and deterministically generate grammatical sentences, even if they were not learned. Moreover, we found that our model could correct a few substitution errors in a sentence, and the correction performance was improved by adding the errors to the training sentences in each training iteration with a certain probability. An analysis of the neural activations in our model revealed that the MTRNN had self-organized, reflecting the hierarchical linguistic structure by taking advantage of the differences in time scale among its neurons: in particular, neurons that change the fastest represented "characters," those that change more slowly, "words," and those that change the slowest, "sentences."}
}
@inproceedings{HinterstoisseretalICCV-11,
     title = {Multimodal Templates for Real-Time Detection of Texture-Less Objects in Heavily Cluttered Scenes},
    author = {Hinterstoisser, S. and Holzer, S. and Cagniart, C. and Ilic, S. and Konolige, K. and Navab, N. and Lepetit, V.},
 booktitle = {{IEEE} International Conference on Computer Vision},
      year = 2011,
     pages = {858-865},
}
@article{HinterstoisseretalPAMI-12,
     title = {Gradient Response Maps for Real-Time Detection of Texture-Less Objects},
    author = {Hinterstoisser, S. and Cagniart, C. and Ilic, S. and Sturm, P. and Navab, N. and Fua, P. and Lepetit, V.},
   journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
 publisher = {IEEE Computer Society},
    volume = 34,
     issue = 5,
      year = 2012,
     pages = {876-888},
}
@article{HintonAIJ-90,
       author = {Hinton, Geoff E.},
        title = {Mapping part-whole hierarchies into connectionist networks},
      journal = {Artificial Intelligence},
       volume = 46,
         year = 1990,
        pages = {47-75},
}
@inproceedings{HintonICANN-99,
     author = {Hinton, G.E.},
      title = {Product of Experts},
  booktitle = {Proceedings of International Conference on Neural Networks},
     volume = {1},
   location = {Edinburgh, Scotland},
      pages = {1-6},
       year = 1999,
}
@inproceedings{HintonIJCAI-05,
    title = {What kind of a graphical model is the brain?},
   author = {Geoffrey E. Hinton},
booktitle = {Proceedings of the 19th International Joint Conference on Artificial Intelligence},
     year = 2005,
}
@inproceedings{HintonIJCAI-81,
    author = {Hinton, Geoffrey F.},
     title = {A parallel computation that assigns canonical object-based frames of reference},
 booktitle = {Proceedings of the 7th international joint conference on Artificial intelligence},
  location = {Vancouver, BC, Canada},
 publisher = {Morgan Kaufmann},
   address = {San Francisco, CA, USA},
      year = 1981,
     pages = {683-685},
}
@article{HintonNC-02,
   author = {Geoffrey E. Hinton},
    title = {Training Products of Experts by Minimizing Contrastive Divergence},
  journal = {Neural Computation},
   volume = 14,
    pages = {1771-1800},
     year = 2002,
}
@incollection{HintonandPlautCSS-87,
       author = {Hinton, G. E. and Plaut, D. C.},
        title = {Using fast weights to deblur old},
    booktitle = {Proceedings of the 9th Annual Conference of the Cognitive Science Society},
    publisher = {Lawrence Erlbaum Associates},
         year = {1987},
        pages = {177-186},
     abstract = {Connectionist models usually have a single weight on each connection. Some interesting new properties emerge if each connection has two weights: A slowly changing, plastic weight which stores long-term knowledge and a fast-changing, elastic weight which stores temporary knowledge and spontaneously decays towards zero. If a network learns a set of associations and then these associations are "blurred" by subsequent learning, all the original associations can be "deblurred" by rehearsing on just a few of them. The rehearsal allows the fast weights to take on values that temporarily cancel out the changes in the slow weights caused by the subsequent learning.}
}
@article{HintonandRoweisNIPS-02,
       author = {Geoffrey Hinton and Sam Roweis},
        title = {Stochastic Neighbor Embedding},
      journal = {Advances in Neural Information Processing Systems},
       volume = 15,
         year = 2002,
        pages = {833-840},
     abstract = {We describe a probabilistic approach to the task of placing objects, described by high-dimensional vectors or by pairwise dissimilarities, in a low-dimensional space in a way that preserves neighbor identities. A Gaussian is centered on each object in the high-dimensional space and the densities under this Gaussian (or the given dissimilarities) are used to define a probability distribution over all the potential neighbors of the object. The aim of the embedding is to approximate this distribution as well as possible when the same operation is performed on the low-dimensional "images" of the objects. A natural cost function is a sum of Kullback-Leibler divergences, one per object, which leads to a simple gradient for adjusting the positions of the low-dimensional images. Unlike other dimensionality reduction methods, this probabilistic framework makes it easy to represent each object by a mixture of widely separated low-dimensional images. This allows ambiguous objects, like the document count vector for the word "bank", to have versions close to the images of both "river" and "finance" without forcing the images of outdoor concepts to be located close to those of corporate concepts.}
}
@article{HintonandSalakhutdinovSCIENCE-06,
        title = {Reducing the Dimensionality of Data with Neural Networks},
       author = {Hinton, Geoff and Salakhutdinov, Ruslan},
      journal = {Science},
        issue = 5786,
       volume = 313,
        pages = {504-507},
         year = 2006
}
@inproceedings{HintonandSejnowskiCVPR-83,
       author = {Hinton, G. E. and Sejnowski, T. J.},
        title = {Optimal Perceptual Inference},
    booktitle = {Proceedings {IEEE} Conference on Computer Vision and Pattern Recognition},
        pages = {448-453},
         year = 1983,
}
@incollection{HintonandSejnowskiPDP-86,
       author = {G. E. Hinton and T. J. Sejnowski},
        title = {Learning and relearning in {B}oltzmann machines},
    booktitle = {Parallel Distributed Processing, Volume 1 - Explorations in the Microstructure of Cognition: Foundations},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
        pages = {282-317},
         year = 1986,
}
@inproceedings{HintonetalAIS-05,
       author = {Geoffrey Hinton and Simon Osindero and Kejie Bao},
        title = {Learning Causally Linked {M}arkov Random Fields},
    booktitle = {Proceedings of the 10th International Workshop on Artificial Intelligence and Statistics, January 6-8, 2005, Savannah Hotel, Barbados},
    publisher = {Society for Artificial Intelligence and Statistics},
       editor = {Robert G. Cowell and Zoubin Ghahramani},
        pages = {128-135},
         year = 2005,
}
@article{HintonetalCoRR-12,
       author = {Geoffrey E. Hinton and Nitish Srivastava and Alex Krizhevsky and Ilya Sutskever and Ruslan Salakhutdinov},
        title = {Improving neural networks by preventing co-adaptation of feature detectors},
      journal = {CoRR},
       volume = {arXiv:1207.0580},
         year = 2012,
     abstract = {When a large feedforward neural network is trained on a small training set, it typically performs poorly on held-out test data. This "overfitting" is greatly reduced by randomly omitting half of the feature detectors on each training case. This prevents complex co-adaptations in which a feature detector is only helpful in the context of several other specific feature detectors. Instead, each neuron learns to detect a feature that is generally helpful for producing the correct answer given the combinatorially large variety of internal contexts in which it must operate. Random "dropout" gives big improvements on many benchmark tasks and sets new records for speech and object recognition},
}
@article{HintonetalCoRR-15,
       author = {Hinton, Geoff Vinyals, Oriol and Dean, Jeff},
        title = {Distilling knowledge in a neural network},
      journal = {CoRR},
       volume = {arXiv:1503.02531},
         year = {2015},
     abstract = {A very simple way to improve the performance of almost any machine learning algorithm is to train many different models on the same data and then to average their predictions. Unfortunately, making predictions using a whole ensemble of models is cumbersome and may be too computationally expensive to allow deployment to a large number of users, especially if the individual models are large neural nets. Caruana and his collaborators have shown that it is possible to compress the knowledge in an ensemble into a single model which is much easier to deploy and we develop this approach further using a different compression technique. We achieve some surprising results on MNIST and we show that we can significantly improve the acoustic model of a heavily used commercial system by distilling the knowledge in an ensemble of models into a single model. We also introduce a new type of ensemble composed of one or more full models and many specialist models which learn to distinguish fine-grained classes that the full models confuse. Unlike a mixture of experts, these specialist models can be trained rapidly and in parallel.}
}
@inproceedings{HintonetalICANN-11,
        title = {Transforming Auto-Encoders},
       author = {Geoffrey E. Hinton and Alex Krizhevsky and Sida D. Wang},
    booktitle = {Proceedings of the International Conference on Artificial Neural Networks},
     location = {Espoo, Finland},
        pages = {44-51},
         year = 2011,
      comment = {Often referred to as the "capsules" paper},
     abstract = {The artificial neural networks that are used to recognize shapes typically use one or more layers of learned feature detectors that produce scalar outputs. By contrast, the computer vision community uses complicated, hand-engineered features, like SIF, that produce a whole vector of outputs including an explicit representation of the pose of the feature. We show how neural networks can be used to learn features that output a whole vector of instantiation parameters and we argue that this is a much more promising way of dealing with variations in position, orientation, scale and lighting than the methods currently employed in the neural networks community. It is also more promising than the hand- engineered features currently used in computer vision because it provides an efficient way of adapting the features to the domain.}
}
@article{HintonetalNC-06,
       author = {Geoffrey Hinton and Simon Osindero and Yee-Whye Teh},
        title = {A fast learning algorithm for deep belief nets},
      journal = {Neural Computation},
       volume = 18,
        issue = 7,
        pages = {1527-1554},
         year = 2006,
}
@incollection{HintonetalPDP-86,
       author = {Hinton, G. E. and McClelland, J. L. and Rumelhart, D. E.},
        title = {Chapter 3: {D}istributed {R}epresentations},
    booktitle = {Parallel Distributed Processing, Volume 1 - Explorations in the Microstructure of Cognition: Foundations},
       editor = {Rumelhart, D. E. and McClelland, J. L.},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1986,
}
@article{HintonetalSCIENCE-02,
        title = {The "wake-sleep" algorithm for unsupervised neural networks},
       author = {G. E. Hinton and P. Dayan and B. J. Frey and R. M. Neal},
      journal = {Science},
       volume = 268,
        issue = 5214,
        pages = {1158-1161},
         year = 1995,
}
@incollection{HiraokaetalDIALOG-SOCIAL-ROBOT-17,
       author = {Hiraoka, Takuya and Neubig, Graham and Yoshino, Koichiro and Toda, Tomoki and Nakamura, Satoshi},
        title = {Active Learning for Example-Based Dialog Systems},
    booktitle = {Dialogues with Social Robots: Enablements, Analyses, and Evaluation},
       editor = {Jokinen, Kristiina and Wilcock, Graham},
    publisher = {Springer Singapore},
      address = {Singapore},
         year = {2017},
        pages = {67-78},
     abstract = {While example-based dialog is a popular option for the construction of dialog systems, creating example bases for a specific task or domain requires significant human effort. To reduce this human effort, in this paper, we propose an active learning framework to construct example-based dialog systems efficiently. Specifically, we propose two uncertainty sampling strategies for selecting inputs to present to human annotators who create system responses for the selected inputs. We compare performance of these proposed strategies with a random selection strategy in simulation-based evaluation on 6 different domains. Evaluation results show that the proposed strategies are good alternatives to random selection in domains where the complexity of system utterances is low.},
}
@book{Hitchcock1995,
        title = {Radio-Frequency and ELF Electromagnetic Energies: A Handbook for Health Professionals},
       author = {Hitchcock, R.T. and Patterson, R.M.},
         year = 1995,
    publisher = {Wiley}
}
@article{HoandErmonCoRR-16,
       author = {Jonathan Ho and Stefano Ermon},
        title = {Generative Adversarial Imitation Learning},
      journal = {CoRR},
       volume = {arXiv:1606.03476},
         year = {2016},
          url = {http://arxiv.org/abs/1606.03476},
     abstract = {Consider learning a policy from example expert behavior, without interaction with the expert or access to reinforcement signal. One approach is to recover the expert's cost function with inverse reinforcement learning, then extract a policy from that cost function with reinforcement learning. This approach is indirect and can be slow. We propose a new general framework for directly extracting a policy from data, as if it were obtained by reinforcement learning following inverse reinforcement learning. We show that a certain instantiation of our framework draws an analogy between imitation learning and generative adversarial networks, from which we derive a model-free imitation learning algorithm that obtains significant performance gains over existing model-free methods in imitating complex behaviors in large, high-dimensional environments.}
}
@inproceedings{HoandErmonNIPS-16,
       author = {Ho, Jonathan and Ermon, Stefano},
        title = {Generative Adversarial Imitation Learning},
    booktitle = {Advances in Neural Information Processing Systems},
       editor = {D. Lee and M. Sugiyama and U. Luxburg and I. Guyon and R. Garnett},
        pages = {4565-4573},
    publisher = {Curran Associates, Inc.},
       volume = {29},
         year = {2016},
     abstract = {Consider learning a policy from example expert behavior, without interaction with the expert or access to a reinforcement signal. One approach is to recover the expert's cost function with inverse reinforcement learning, then extract a policy from that cost function with reinforcement learning. This approach is indirect and can be slow. We propose a new general framework for directly extracting a policy from data as if it were obtained by reinforcement learning following inverse reinforcement learning. We show that a certain instantiation of our framework draws an analogy between imitation learning and generative adversarial networks, from which we derive a model-free imitation learning algorithm that obtains significant performance gains over existing model-free methods in imitating complex behaviors in large, high-dimensional environments.},
}
@book{HobbsandMoore85,
       editor = {Hobbs, Jerry E. and Moore, Robert C.},
        title = {Formal Theories of the Common Sense World},
    booktitle = {Formal Theories of the Common Sense World},
    publisher = {Ablex},
      address = {Norwood, New Jersey},
         year = 1985,
}
@misc{HochreiterPhD-91,
        title = {Untersuchungen zu dynamischen neuronalen Netzen},
       author = {Sepp Hochreiter},
         year = 1991,
 howpublished = {Diploma thesis. Diploma thesis. Institut f\"{u}r Informatik, Lehrstuhl Prof. Brauer, Technische Universit\"{a}t M{\"{u}}nchen}
}
@article{HochreiterandSchmidhuberNC-97,
       author = {Hochreiter, Sepp and Schmidhuber, J\"{u}rgen},
        title = {Long Short-Term Memory},
      journal = {Neural Computing},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
       volume = 9,
        issue = 8,
         year = 1997,
        pages = {1735-1780},
}
@inproceedings{HochreiteretalICANN-01,
       author = {Hochreiter, Sepp and Younger, Arthur and R. Conwell, Peter},
        title = {Learning To Learn Using Gradient Descent},
    booktitle = {{ICANN 2001. Lecture Notes in Computer Science}},
       editor = {Dorffner G., Bischof H., Hornik K.},
    publisher = {Springer, Berlin, Heidelberg},
       volume = {2130},
         year = {2001},
        pages = {87-94},
     abstract = {This paper introduces the application of gradient descent methods to meta-learning. The concept of "meta-learning", i.e. of a system that improves or discovers a learning algorithm, has been of interest in machine learning for decades because of its appealing applications. Previous meta-learning approaches have been based on evolutionary methods and, therefore, have been restricted to small models with few free parameters. We make meta-learning in large systems feasible by using recurrent neural networks with their attendant learning routines as meta-learning systems. Our system derived complex well performing learning algorithms from scratch. In this paper we also show that our approach performs non-stationary time series prediction.}
}
@article{HockettSCIAM-60,
       author = {Hockett, C. F.},
        title = {The origin of speech},
      journal = {Scientific American},
       volume = {203},
        issue = {2},
         year = {1960},
        pages = {5-12},
}
@book{HoddesonandDaitch2002,
        title = {True Genius: The Life and Science of John Bardeen, the Only Winner of Two Nobel Prizes in Physics},
       author = {Hoddeson, Lillian and Daitch, Vicki}, 
         year = 2002,
    publisher = {Joseph Henry Press}
}
@book{Hodges83,
       author = {Andrew Hodges},
        title = {Alan Turing: The Enigma},
    publisher = {Simon and Schuster},
      address = {New York, NY},
         year = 1983,
}
@article{HodgkinJoP-76,
       author = {Hodgkin, Alan L.},
        title = {Chance and design in electrophysiology},
      journal = {The Journal of Physiology},
       volume = 263,
         year = 1976,
        pages = {1-21},
}
@article{HodgkinandHuxleyJoP-52,
       author = {Hodgkin, Alan L. and Huxley, Andrew F.},
        title = {A quantitative description of membrane current and its application to conduction and excitation in nerve},
      journal = {The Journal of Physiology},
       volume = 117,
        issue = 4,
         year = 1952,
        pages = {500-544},
}
@article{Hoeffding63,
       author = {Hoeffding, W.},
        title = {Probability Inequalities for Sums of Bounded Random Variables},
      journal = {Journal of the American Statistical Association},
       volume = 58,
        pages = {13-30},
         year = 1963,
}
@book{HoelPortandStoneI,
       author = {Hoel, Paul G. and Port, Sidney C. and Stone, Charles J.},
        title = {Introduction to Probability Theory},
    publisher = {Houghton Mifflin},
      address = {Boston, Massachusetts},
         year = 1971,
}
@book{HoelPortandStoneIII,
       author = {Hoel, Paul G. and Port, Sidney C. and Stone, Charles J.},
        title = {Introduction to Stochastic Processes},
    publisher = {Houghton Mifflin},
      address = {Boston, Massachusetts},
         year = 1971,
}
@article{HoemannetalCOiBS-17,
        title = {Mixed emotions in the predictive brain},
       author = {Hoemann, Katie and Gendron, Maria and Barrett, Lisa Feldman},
      journal = {Current Opinion in Behavioral Sciences},
    publisher = {Elsevier},
       volume = {15},
        pages = {51-57},
         year = {2017},
     abstract = {Understanding complex or mixed emotions first requires an exploration of the human nervous system underlying emotions, and indeed all experience. We review current research in neuroscience, which describes the brain as a predictive, internal model of the world that flexibly combines features from past experience to construct emotions. We argue that 'mixed emotions' result when these features of past experience correspond to multiple emotion categories. Integrating event perception and cognitive linguistic theories, we propose that 'mixed emotions' are perceived as an episode of distinct, linked emotional events due to attentional shifts which update the predicted model of experience. These proposed mechanisms have profound implications for the study of emotion; we conclude by suggesting methodological improvements for future research.}
}
@inproceedings{HofCaiCia03,
  author = {T. Hofmann and L. Cai and M. Ciaramita},
  title = {Learning with Taxonomies: Classifying Documents and Words},
  booktitle= {Proceedings of the Workshop on Syntax, Semantics, and Statistics (NIPS 16)},
  year = 2003,
}
@article{HoferetalNATURE-NEUROSCIENCE-11,
        title = {Differential connectivity and response dynamics of excitatory and inhibitory neurons in visual cortex},
       author = {Hofer, Sonja B. and Ko, Ho and Pichler, Bruno and Vogelstein, Joshua and Ros, Hana and Zeng, Hongkui and Lein, Ed and Lesica, Nicholas A. and Mrsic-Flogel, Thomas D.},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = 14,
        issue = 8,
         year = 2011,
        pages = {1045-1052},
     abstract = {Neuronal responses during sensory processing are influenced by both the organization of intracortical connections and the statistical features of sensory stimuli. How these intrinsic and extrinsic factors govern the activity of excitatory and inhibitory populations is unclear. Using two-photon calcium imaging in vivo and intracellular recordings in vitro, we investigated the dependencies between synaptic connectivity, feature selectivity and network activity in pyramidal cells and fast-spiking parvalbumin-expressing (PV) interneurons in mouse visual cortex. In pyramidal cell populations, patterns of neuronal correlations were largely stimulus-dependent, indicating that their responses were not strongly dominated by functionally biased recurrent connectivity. By contrast, visual stimulation only weakly modified co-activation patterns of fast-spiking PV cells, consistent with the observation that these broadly tuned interneurons received very dense and strong synaptic input from nearby pyramidal cells with diverse feature selectivities. Therefore, feedforward and recurrent network influences determine the activity of excitatory and inhibitory ensembles in fundamentally different ways.},
}
@book{Hoffman1998,
       author = {Donald Hoffman},
        title = {Visual Intelligence: How We Create What We See},
    publisher = {W. W. Norton},
      address = {New York, NY},
         year = 1998,
}
@inproceedings{Hofmann00,
       author = {Thomas Hofmann},
        title = {Learning Probabilistic Models of the {W}eb},
    booktitle = {Proceedings of the 23rd International Conference on Research and Development in Information Retrieval (ACM SIGIR)},
         year = 2000
}
@inproceedings{Hofmann00b,
       author = {T. Hofmann},
        title = {Learning the similarity of documents: An information-geometric approach to document retrieval and categorization},
    booktitle = {Advances in Neural Information Processing Systems (NIPS 12)},
        pages = {914-920},
    publisher = {MIT Press},
         year = 1999,
}
@article{Hofmann01,
       author = {Thomas Hofmann},
        title = {Unsupervised Learning by Probabilistic Latent Semantic Analysis},
      journal = {Machine Learning},
       volume = 42,
        issue = 1,
        pages = {177-196},
         year = 2001,
}
@inproceedings{Hofmann02b,
       author = {Y. Altun and T. Hofmann and M. Johnson},
        title = {Discriminative Learning for Label Sequences via Boosting},
    booktitle = {Advances in Neural Information Processing Systems (NIPS 15)},
         year = 2002,
}
@inproceedings{HofmannAAAI02,
       author = {S. Andrews  and T. Hofmann and I. Tsochantaridis},
        title = {Multiple Instance Learning With Generalized Support Vector Machines},
    booktitle = {Proceedings of AAAI Conference, Student Paper},
         year = 2002,
}
@inproceedings{HofmannECML01,
       author = {T. Hofmann},
        title = {What People (Don't) Want},
    booktitle = {European Conference on Machine Learning (ECML)},
         year = 2001,
}
@inproceedings{HofmannECML02,
       author = {I. Tsochantaridis and T. Hofmann},
        title = {Support Vector Machines for Polycategorical Classification},
    booktitle = {European Conference on Machine Learning (ECML)},
         year = 2002,
}
@inproceedings{HofmannNIPS02a,
       author = {S. Andrews and I. Tsochantaridis and T. Hofmann},
        title = {Support Vector Machines for Multiple-Instance Learning},
    booktitle = {Advances in Neural Information Processing Systems (NIPS 15)},
         year = 2002,
}
@inproceedings{HofmannNIPS02b,
       author = {Y. Altun and T. Hofmann and M. Johnson},
        title = {Discriminative Learning for Label Sequences via Boosting},
    booktitle = {Advances in Neural Information Processing Systems (NIPS 15)},
         year = 2002,
}
@inproceedings{HofmannSIGIR00,
       author = {Hofmann, T.},
        title = {Learning Probabilistic Models of the {W}eb},
    booktitle = {Proceedings of the 23th International Conference on Research and Development in Information Retrieval (SIGIR)},
         year = 2000,
}
@inproceedings{HofmannUAI99,
       author = {Hofmann, T.},
        title = {Probabilistic Latent Semantic Analysis},
    booktitle = {Proceedings of the 15th Conference on Uncertainty in AI},
        pages = {289-296},
         year = 1999,
}
@book{Hofstadter1979,
      author = {Douglas Hofstadter},
       title = {G\"{o}del, Escher, Bach: An Eternal Golden Braid},
   publisher = {Basic Books},
     address = {New York, NY},
        year = 1979,
}
@book{Hofstadter1996,
        title = {Fluid Concepts and Creative Analogies: Computer Models of the Fundamental Mechanisms of Thought},
       author = {Hofstadter, Douglas},
    publisher = {Basic Books},
         year = {1996},
}
@incollection{Hofstadter2001,
        title = {Epilogue: Analogy as the core of cognition},
       author = {Hofstadter, Douglas},
    booktitle = {The Analogical Mind: Perspectives from Cognitive Science},
       editor = {Gentner, Dedre and Holyoak, Keith J. and Kokinov, Boicho N.},
    publisher = {MIT Press},
          url = {http://prelectur.stanford.edu/lecturers/hofstadter/analogy.html},
         year = {2001},
        pages = {499-538},
}
@book{Hofstadter2007,
        title = {I Am a Strange Loop},
       author = {Hofstadter, Douglas},
    publisher = {Basic Books},
         year = {2007},
}
@book{Hofstadter2008,
        title = {Metamagical Themas: Questing For the Essence of Mind and Pattern},
       author = {Hofstadter, Douglas},
         year = {2008},
    publisher = {Basic Books}
}
@book{Hofstadter2013,
        title = {Surfaces and Essences: Analogy as the Fuel and Fire of Thinking},
       author = {Hofstadter, Douglas and Sander, Emmanuel},
         year = {2013},
    publisher = {Basic Books}
}
@book{Hofstadter79,
      author = {Douglas Hofstadter},
       title = {G\"{o}del, Escher, Bach: An Eternal Golden Braid},
   publisher = {Basic Books},
     address = {New York, NY},
        year = 1979,
}
@misc{HofstadterANALOGY-05,
        title = {Analogy as the Core of Cognition},
       author = {Douglas Hofstadter},
 howpublished = {https://www.media.mit.edu/cogmac/lcc/hofstadter.html},
     abstract = {Analogy is almost always conceived of by psychologists and cognitive scientists as some kind of advanced or sophisticated mental tool used in problem-solving by specialists in a given domain when all else fails. Moreover, the adjective analogical is almost always married to the noun reasoning , which clearly implies that analogical thinking is seen primarily (if not exclusively) as a special variety of reasoning (i.e., a kind of logic). I have a radically different perception of analogy, which is that every concept in our mental repertoire (whether it comes equipped with a ready-made label, such as the word "cotton" or the stock phrase "in the nick of time", or it is unlabeled, such as "that time when my friend Robert and I had planned to meet up in the Danish fishing village and we...") comes from a large set of unconsciously made analogies stretching back to our childhood, and hence that analogy-making is what lies at the core of human thought. This view of analogy has very little to do with reasoning in any standard sense of the term. Moreover, I argue that the sequential process of thinking is driven at all times by unconscious analogy-making carried out simultaneously at various levels of abstraction. Numerous examples supporting this thesis form the crux of this seminar.},
}
@inproceedings{HoiemetalCVPR-06,
      title = {Putting Objects in Perspective},
     author = {Derek Hoiem and Alexei Efros and Martial Hebert},
  booktitle = {Proceedings of the 2006 {IEEE} Conference on Computer Vision and Pattern Recognition},
  publisher = {IEEE Computer Society},
     volume = 2,
      pages = {2137-2144},
       year = 2006,
}
@inproceedings{HoiemetalICCV-07,
     author = {Derek Hoiem and Andrew Stein and Alexei A. Efros and Martial Hebert},
      title = {Recovering Occlusion Boundaries from a Single Image},
 booktitle  = {Proceedings of the 11th {IEEE} International Conference on Computer Vision},
      pages = {1-8},
       year = 2007,
}
@article{HoiemetalIJCV-07,
     author = {Derek Hoiem and Alexei Efros and Martial Hebert},
      title = {Recovering Surface Layout from an Image},
    journal = {International Journal of Computer Vision},
  publisher = {Kluwer Academic Publishers},
     volume = 75,
      issue = 1,
      pages = {151-172},
       year = 2007,
}        
@article{HokeetalBEHAVIORAL-NEUROSCIENCE-07,
       author = {Hoke, K. L. and Ryan, M. J. and Wilczynski, W.},
        title = {Functional coupling between substantia nigra and basal ganglia homologues in amphibians},
      journal = {behavioral Neuroscience},
       volume = {121},
       number = {6},
         year = {2007},
        pages = {1393-1399},
     abstract = {Neuroanatomical and pharmacological experiments support the existence of a homologue of the mammalian substantia nigra-basal ganglia circuit in the amphibian brain. Demarcation of borders between the striatum and pallidum in frogs, however, has been contentious, and direct evidence of functional coupling between the putative nigral and striatal homologues is lacking. To clarify basal ganglia function in anurans, the authors used expression of immediate-early gene egr-1 as a marker of neural activation in the basal ganglia of túngara frogs (Physalaemus pustulosus). Regional variation in egr-1 mRNA levels distinguished striatal and pallidal portions of the basal ganglia and supported the grouping of the striatopallidal transition zone with the dorsal pallidum. As further evidence for a functional coupling between the dopaminergic cells in the posterior tuberculum (the putative substantia nigra homologue) and the basal ganglia, a positive relationship was demonstrated between the size of the dopaminergic cell population and the neural activation levels within the dorsal pallidum.}
}
@article{HokeetalICB-17,
       author = {Hoke, Kim L. and Hebets, Eileen A. and Shizuka, Daizaburo},
        title = {{Neural Circuitry for Target Selection and Action Selection in Animal Behavior}},
      journal = {Integrative and Comparative Biology},
       volume = {57},
       number = {4},
        pages = {808-819},
         year = {2017},
     abstract = {Animal behaviorists have long strived for a comprehensive understanding of the proximate and ultimate causes of complex behavior, and we propose that recent advances in neurobiology can help reshape or clarify this behavior-oriented understanding. We begin with an overview of current views of neural circuit mechanisms that mediate target selection and action selection. In target selection, different stimuli compete for priority in sensory-motor processing. Action selection is the process by which multiple possible motor actions compete for priority in a manner which balances the needs of the animal with opportunities or threats in the environment. We next discuss spatial and temporal aspects of target and action selection, highlighting how neurophysiological responses to complex displays depend on spatial and temporal components of multisensory stimuli. We use two examples-(1) spatial attention as an example of target selection in the vertebrate midbrain and (2) goal-directed locomotion as an example of action selection in the insect central complex-to further clarify neural circuit dynamics as they relate to target and action selection, and their interaction. We suggest that a deeper understanding of neural circuit properties will introduce new hypotheses into behavioral studies, especially those aimed at understanding the evolution of complex displays based on receiver sensory biases. Additionally, knowledge of neural circuit properties can elucidate ways in which current context and previous experience can together modify neural circuit dynamics to produce complex context-dependent behavioral responses that often characterize animal behavior.},
}
@article{HolcmanandYusteNATURE-15,
       author = {Holcman, David and Yuste, Rafael},
        title = {The new nanophysiology: regulation of ionic flow in neuronal subcompartments},
      journal = {Nature Reviews Neuroscience},
         year = {2015},
    publisher = {Nature Publishing Group},
       volume = {16},
        issue = {11},
        pages = {685-692},
     abstract = {Cable theory and the Goldman-Hodgkin-Huxley-Katz models for the propagation of ions and voltage within a neuron have provided a theoretical foundation for electrophysiology and been responsible for many cornerstone advances in neuroscience. However, these theories break down when they are applied to small neuronal compartments, such as dendritic spines, synaptic terminals or small neuronal processes, because they assume spatial and ionic homogeneity. Here we discuss a broader theory that uses the Poisson-Nernst-Planck (PNP) approximation and electrodiffusion to more accurately model the constraints that neuronal nanostructures place on electrical current flow. This extension of traditional cable theory could advance our understanding of the physiology of neuronal nanocompartments},
}
@book{Holland75,
      author = {John Holland},
       title = {Adaptation in Natural and Artificial Systems},
   publisher = {MIT Press},
     address = {Cambridge, MA},
        year = 1975,
}
@book{Hollandetal87,
       author = {Holland, John H. and Holyoak, Keith J. and Nisbett, Richard E. and Thagard, Paul R.},
        title = {Induction: Processes of Inference, Learning, and Discovery},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1987,
}
@techreport{Hollerbach82,
       author = {Hollerbach, John M.},
        title = {Computers, Brains, and the Control of Movement},
         type = {{AI} Memo},
        issue = 686,
  institution = {{MIT} {AI} Laboratory},
         year = 1982,
}
@article{HolmenandFosterIJPP-13,
       author = {Holmen, J.K. and Foster, D.L.},
        title = {Accelerating Single Iteration Performance of {CUDA}-Based {3D} Reaction-Diffusion Simulations},
      journal = {International Journal of Parallel Programming},
    publisher = {Springer},
       volume = {42},
         year = {2013},
        pages = {343-363},
     abstract = {The most commonly used approach for solving reaction-diffusion systems relies upon stencil computations. Although stencil computations feature low compute intensity, they place high demands on memory bandwidth. Fortunately, GPU computing allows for the heavy reliance of stencil computations on neighboring data points to be exploited to significantly increase simulation speeds by reducing these memory bandwidth demands. Upon reviewing previously published works, a wide-variety of efforts have been made to optimize NVIDIA CUDA-based stencil computations. However, a critical aspect contributing to algorithm performance is commonly glossed over: the halo region loading technique utilized in conjunction with a given spatial blocking technique. This paper presents an in-depth examination of this aspect and the associated single iteration performance impacts when using symmetric, nearest neighbor 19-point stencils. This is accomplished by closely examining how the simulated space is partitioned into thread blocks and the balance between memory accesses, divergence, and computing threads. The resulting optimization strategy for accelerating 3-dimensional reaction-diffusion simulations offers up to 2.45 times speedup for single-precision floating point numbers in reference to GPU-based speedups found within the previously published work that this paper directly extends. In reference to our multithreaded CPU-based implementation, the resulting optimization strategy offers up to 8.69 times speedup for single-precision floating point numbers.}
}
@inproceedings{HolyoakOHTR-12,
       author = {Keith J. Holyoak},
        title = {Analogy and Relational Reasoning},
    booktitle = {Oxford Handbook of Thinking and Reasoning},
    publisher = {Oxford University Press},
       editor = {Keith J. Holyoak and Robert G. Morrison},
        pages = {234-259},
         year = {2012},
     abstract = {Analogy is an inductive mechanism based on structured comparisons of mental representations. It is an important special case of role-based relational reasoning, in which inferences are generated on the basis of patterns of relational roles. Analogical reasoning is a complex process involving retrieval of structured knowledge from long-term memory, representing and manipulating role-filler bindings in working memory, identifying elements that play corresponding roles, generating new inferences, and learning abstract schemas. For empirical analogies, analogical inference is guided by causal knowledge about how the source analog operates. Simpler types of relation-based transfer can be produced by relational priming. Human analogical reasoning is heavily dependent on working memory and other executive functions supported by the prefrontal cortex, with the frontopolar subregion being selectively activated when multiple relations must be integrated to solve a problem.}
}
@article{HolzeletalMEDITATION-07,
       author = {Holzel, B. K.  and Ott, U.  and Hempel, H.  and Hackl, A.  and Wolf, K.  and Stark, R.  and Vaitl, D. },
        title = {{{D}ifferential engagement of anterior cingulate and adjacent medial frontal cortex in adept meditators and non-meditators}},
      journal = {Neuroscience Letters},
         year = {2007},
       volume = {421},
       number = {1},
        pages = {16-21},
     abstract = {This study investigated differences in brain activation during meditation between meditators and non-meditators. Fifteen Vipassana meditators (mean practice: 7.9 years, 2h daily) and fifteen non-meditators, matched for sex, age, education, and handedness, participated in a block-design fMRI study that included mindfulness of breathing and mental arithmetic conditions. For the meditation condition (contrasted to arithmetic), meditators showed stronger activations in the rostral anterior cingulate cortex and the dorsal medial prefrontal cortex bilaterally, compared to controls. Greater rostral anterior cingulate cortex activation in meditators may reflect stronger processing of distracting events. The increased activation in the medial prefrontal cortex may reflect that meditators are stronger engaged in emotional processing.}
}
@article{HolzingeretalCoRR-17,
       author = {Andreas Holzinger and Markus Plass and Katharina Holzinger and Gloria Cerasela Crisan and Camelia{-}Mihaela Pintea and Vasile Palade},
        title = {A glass-box interactive machine learning approach for solving NP-hard problems with the human-in-the-loop},
      journal = {CoRR},
       volume = {arXiv:1708.01104},
         year = {2017},
     abstract = {The goal of Machine Learning to automatically learn from data, extract knowledge and to make decisions without any human intervention. Such automatic (aML) approaches show impressive success. Recent results even demonstrate intriguingly that deep learning applied for automatic classification of skin lesions is on par with the performance of dermatologists, yet outperforms the average. As human perception is inherently limited, such approaches can discover patterns, e.g. that two objects are similar, in arbitrarily high-dimensional spaces what no human is able to do. Humans can deal only with limited amounts of data, whilst big data is beneficial for aML; however, in health informatics, we are often confronted with a small number of data sets, where aML suffer of insufficient training samples and many problems are computationally hard. Here, interactive machine learning (iML) may be of help, where a human-in-the-loop contributes to reduce the complexity of NP-hard problems. A further motivation for iML is that standard black-box approaches lack transparency, hence do not foster trust and acceptance of ML among end-users. Rising legal and privacy aspects, e.g. with the new European General Data Protection Regulations, make black-box approaches difficult to use, because they often are not able to explain why a decision has been made. In this paper, we present some experiments to demonstrate the effectiveness of the human-in-the-loop approach, particularly in opening the black-box to a glass-box and thus enabling a human directly to interact with an learning algorithm. We selected the Ant Colony Optimization framework, and applied it on the Traveling Salesman Problem, which is a good example, due to its relevance for health informatics, e.g. for the study of protein folding. From studies of how humans extract so much from so little data, fundamental ML-research also may benefit.},
}
@misc{HoncharUNDERSTANDING-NEURAL-DIFFERENTIAL-EQUATIONS-19,
        title = {Neural ODEs: Breakdown of another deep learning breakthrough},
       author = {Alexandr Honchar},
 howpublished = {https://towardsdatascience.com/neural-odes-breakdown-of-another-deep-learning-breakthrough-3e78c7213795},
         year = {2019},
}
@article{HondaetalPNAS-18,
       author = {Honda, T. and Nagao, S. and Hashimoto, Y. and Ishikawa, K. and Yokota, T. and Mizusawa, H. and Ito, M.},
        title = {Tandem internal models execute motor learning in the cerebellum},
      journal = {Proceedings of the National Academy of Science U.S.A.},
       volume = {115},
       number = {28},
         year = {2018},
        pages = {7428-7433},
     abstract = {In performing skillful movement, humans use predictions from internal models formed by repetition learning. However, the computational organization of internal models in the brain remains unknown. Here, we demonstrate that a computational architecture employing a tandem configuration of forward and inverse internal models enables efficient motor learning in the cerebellum. The model predicted learning adaptations observed in hand-reaching experiments in humans wearing a prism lens and explained the kinetic components of these behavioral adaptations. The tandem system also predicted a form of subliminal motor learning that was experimentally validated after training intentional misses of hand targets. Patients with cerebellar degeneration disease showed behavioral impairments consistent with tandemly arranged internal models. These findings validate computational tandemization of internal models in motor control and its potential uses in more complex forms of learning and cognition.}
}
@article {HongetalSCIENCE-16,
       author = {Hong, Soyon and Beja-Glasser, Victoria F. and Nfonoyim, Bianca M. and Frouin, Arnaud and Li, Shaomin and Ramakrishnan, Saranya and Merry, Katherine M. and Shi, Qiaoqiao and Rosenthal, Arnon and Barres, Ben A. and Lemere,, Cynthia A. and Selkoe, Dennis J. and Stevens, Beth},
        title = {Complement and microglia mediate early synapse loss in Alzheimer mouse models},
      journal = {Science},
         year = 2016,
    publisher = {American Association for the Advancement of Science},
     abstract = {Synapse loss in Alzheimer{\textquoteright}s disease (AD) correlates with cognitive decline. Involvement of microglia and complement in AD has been attributed to neuroinflammation, prominent late in disease. Here we show in mouse models that complement and microglia mediate synaptic loss early in AD. C1q, the initiating protein of the classical complement cascade, is increased and associated with synapses before overt plaque deposition. Inhibition of C1q, C3 or the microglial complement receptor CR3, reduces the number of phagocytic microglia as well as the extent of early synapse loss. C1q is necessary for the toxic effects of soluble Î²-amyloid (\hmbeta{}) oligomers on synapses and hippocampal long-term potentiation (LTP). Finally, microglia in adult brains engulf synaptic material in a CR3-dependent process when exposed to soluble \hmbeta{} oligomers. Together, these findings suggest that the complement-dependent pathway and microglia that prune excess synapses in development are inappropriately activated and mediate synapse loss in AD.},
}
@book{HopcroftandUllman79,
       author = {Hopcroft, John E. and Ullman, Jeffrey D.},
        title = {Introduction to Automata Theory, Languages, and Computation},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1979,
}
@inproceedings{HopeetalKDD-17,
       author = {Hope, Tom and Chan, Joel and Kittur, Aniket and Shahaf, Dafna},
        title = {Accelerating Innovation Through Analogy Mining},
    booktitle = {Proceedings of the 23rd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining},
         year = {2017},
        pages = {235-243},
    publisher = {ACM},
      address = {New York, NY, USA},
     abstract = {The availability of large idea repositories (e.g., the U.S. patent database) could significantly accelerate innovation and discovery by providing people with inspiration from solutions to analogous problems. However, finding useful analogies in these large, messy, real-world repositories remains a persistent challenge for either human or automated methods. Previous approaches include costly hand-created databases that have high relational structure (e.g., predicate calculus representations) but are very sparse. Simpler machine-learning/information-retrieval similarity metrics can scale to large, natural-language datasets, but struggle to account for structural similarity, which is central to analogy. In this paper we explore the viability and value of learning simpler structural representations, specifically, "problem schemas", which specify the purpose of a product and the mechanisms by which it achieves that purpose. Our approach combines crowdsourcing and recurrent neural networks to extract purpose and mechanism vector representations from product descriptions. We demonstrate that these learned vectors allow us to find analogies with higher precision and recall than traditional information-retrieval methods. In an ideation experiment, analogies retrieved by our models significantly increased people's likelihood of generating creative ideas compared to analogies retrieved by traditional methods. Our results suggest a promising approach to enabling computational analogy at scale is to learn and leverage weaker structural representations.}
} 
@article{HopfieldPNAS-82,
       author = {Hopfield, J. J.},
        title = {Neural networks and physical systems with emergent collective computational abilities},
      journal = {Proceedings of the National Academy of Science},
       volume = 79,
        issue = 8,
         year = 1982,
        pages = {2554-2558},
}
@article{HopfieldandTankSCIENCE-86,
       author = {John J. Hopfield and David W. Tank},
        title = {Computing with Neural Circuits: A Model},
      journal = {Science},
       volume = 233,
        issue = 4764,
        pages = {625-633},
         year = 1986,
     abstract = {A new conceptual framework and a minimization principle together provide an understanding of computation in model neural circuits. The circuits consist of nonlinear graded-response model neurons organized into networks with effectively symmetric synaptic connections. The neurons represent an approximation to biological neurons in which a simplified set of important computational properties is retained. Complex circuits solving problems similar to those essential in biology can be analyzed and understood without the need to follow the circuit dynamics in detail. Implementation ofthe model with electronic devices will provide a class of electronic circuits of novel form and fimction.}
}
@article{HorgaandMaiaRiHN-12,
       author = {Horga, G. and Maia, T. V.},
        title = {{C}onscious and unconscious processes in cognitive control: a theoretical perspective and a novel empirical approach},
      journal = {Frontiers in Human Neuroscience},
         year = {2012},
       volume = {6},
        pages = {199},
     abstract = {Controlled processing is often referred to as "voluntary" or "willful" and therefore assumed to depend entirely on conscious processes. Recent studies using subliminal-priming paradigms, however, have started to question this assumption. Specifically, these studies have shown that subliminally presented stimuli can induce adjustments in control. Such findings are not immediately reconcilable with the view that conscious and unconscious processes are separate, with each having its own neural substrates and modus operandi. We propose a different theoretical perspective that suggests that conscious and unconscious processes might be implemented by the same neural substrates and largely perform the same neural computations, with the distinction between the two arising mostly from the quality of representations (although not all brain regions may be capable of supporting conscious representations). Thus, stronger and more durable neuronal firing would give rise to conscious processes; weaker or less durable neuronal firing would remain below the threshold of consciousness but still be causally efficacious in affecting behavior. We show that this perspective naturally explains the findings that subliminally presented primes induce adjustments in cognitive control. We also highlight an important gap in this literature: whereas subliminal-priming paradigms demonstrate that an unconsciously presented prime is sufficient to induce adjustments in cognitive control, they are uninformative about what occurs under standard task conditions. In standard tasks, the stimuli themselves are consciously perceived; however, the extent to which the processes that lead to adjustments in control are conscious or unconscious remains unexplored. We propose a new paradigm suitable to investigate these issues and to test important predictions of our hypothesis that conscious and unconscious processes both engage the same control machinery, differing mostly in the quality of the representations.}
}
@book{HorganTHE_END_OF_SCIENCE-97,
        title = {The End of Science: Facing the Limits of Knowledge in the Twilight of the Scientific Age},
       author = {Horgan, John},
    publisher = {Little, Brown},
         year = {1997},
}
@book{Horn86,
       author = {Horn, Berthold Klaus Paul},
        title = {Robot Vision},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1986,
}        
@article{HorngandSurPBR-06,
       author = {Horng, S. H. and Sur, M.},
        title = {{V}isual activity and cortical rewiring: activity-dependent plasticity of cortical networks},
      journal = {Progress in Brain Research},
         year = {2006},
       volume = {157},
        pages = {3-11},
     abstract = {The mammalian cortex is organized anatomically into discrete areas, which receive, process, and transmit neural signals along functional pathways. These pathways form a system of complex networks that wire up through development and refine their connections into adulthood. Understanding the processes of cortical-pathway formation, maintenance, and experience-dependent plasticity has been among the major goals of contemporary neurobiology. In this chapter, we will discuss an experimental model used to investigate the role of activity in the patterning of cortical networks during development. This model involves the "rewiring" of visual inputs into the auditory thalamus and subsequent remodeling of the auditory cortex to process visual information. We review the molecular, cellular, and physiological mechanisms of visual "rewiring" and activity-dependent shaping of cortical networks.}
}
@article{HornikNEURAL-NETWORKS-91,
       author = {Hornik, Kurt},
        title = {Approximation Capabilities of Multilayer Feedforward Networks},
      journal = {Neural Networks},
       volume = {4},
       number = {2},
        pages = {251-257},
         year = {1991},
    publisher = {Elsevier Science Ltd.},
}
@article{HorniketalNEURAL-NETWORKS-89,
       author = {Hornik, K. and Stinchcombe, M. and White, H.},
        title = {Multilayer Feedforward Networks Are Universal Approximators},
      journal = {Neural Networks},
       volume = {2},
       number = {5},
        pages = {359-366},
         year = {1989},
    publisher = {Elsevier Science Ltd.},
}
@article{HortonandAdamsPTRSB-05,
       author = {Horton, Jonathan C and Adams, Daniel L},
        title = {The cortical column: a structure without a function},
      journal = {Philosophical Transactions of the Royal Society B: Biological Sciences},
       volume = 360,
        issue = 1456,
        pages = {837-862},
         year = 2005,
     abstract = {This year, the field of neuroscience celebrates the 50th anniversary of Mountcastle's discovery of the cortical column. In this review, we summarize half a century of research and come to the disappointing realization that the column may have no function. Originally, it was described as a discrete structure, spanning the layers of the somatosensory cortex, which contains cells responsive to only a single modality, such as deep joint receptors or cutaneous receptors. Subsequently, examples of columns have been uncovered in numerous cortical areas, expanding the original concept to embrace a variety of different structures and principles. A 'column' now refers to cells in any vertical cluster that share the same tuning for any given receptive field attribute. In striate cortex, for example, cells with the same eye preference are grouped into ocular dominance columns. Unaccountably, ocular dominance columns are present in some species, but not others. In principle, it should be possible to determine their function by searching for species differences in visual performance that correlate with their presence or absence. Unfortunately, this approach has been to no avail; no visual faculty has emerged that appears to require ocular dominance columns. Moreover, recent evidence has shown that the expression of ocular dominance columns can be highly variable among members of the same species, or even in different portions of the visual cortex in the same individual. These observations deal a fatal blow to the idea that ocular dominance columns serve a purpose. More broadly, the term 'column' also denotes the periodic termination of anatomical projections within or between cortical areas. In many instances, periodic projections have a consistent relationship with some architectural feature, such as the cytochrome oxidase patches in V1 or the stripes in V2. These tissue compartments appear to divide cells with different receptive field properties into distinct processing streams. However, it is unclear what advantage, if any, is conveyed by this form of columnar segregation. Although the column is an attractive concept, it has failed as a unifying principle for understanding cortical function. Unravelling the organization of the cerebral cortex will require a painstaking description of the circuits, projections and response properties peculiar to cells in each of its various areas.},
}
@article{Horvitz88b,
       author = {E.J. Horvitz and J.S. Breese and M. Henrion},
        title = {Decision Theory in Expert Systems and Artificial Intelligence},
      journal = {International Journal of Approximate Reasoning},
       volume = 2,
         year = 1988,
        pages = {247-302},
}
@inproceedings{HorvitzAAAI-88,
       author = {Horvitz, Eric J.},
        title = {Reasoning Under Varying and Uncertain Resource Constraints},
    booktitle = {Proceedings AAAI-88},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
        pages = {111-116},
         year = 1988,
}
@article{HorvitzDECISION,
       author = {Horvitz, Eric J. and Breese, John S. and Henrion, Max},
        title = {Decision Theory in Expert Systems and Artificial Intelligence},
      journal = {International Journal of Approximate Reasoning},
       volume = 2,
        pages = {247-302},
         year = 1988,
}
@inproceedings{HorvitzIJCAI-89,
       author = {Horvitz, Eric J. and Cooper, Gregory F. and Heckerman, David E.},
        title = {Reflection and Action Under Scarce Resources: Theoretical Principles and Empirical Study},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
         year = 1989,
        pages = {1121-1127},
}
@techreport{HorvitzPARTITION,
       author = {Horvitz, Eric J. and Breese, John S.},
        title = {Ideal Partition of Resources for Metareasoning},
        issue = {KSL-90-26},
  institution = {Stanford Knowledge Systems Laboratory},
         year = 1990,
}
@inproceedings{HorvitzUAI-87,
       author = {Horvitz, Eric J.},
        title = {Reasoning About Beliefs and Actions Under Computational Resource Constraints},
    booktitle = UAI-87,
     location = {Seattle, Washington},
         year = 1987,
}
@inproceedings{HorvitzUAI-89,
       author = {Horvitz, Eric J. and Suermondt, H. Jacques and Cooper, Gregory F.},
        title = {Bounded Conditioning: Flexible Inference for Decisions Under Scarce Resources},
    booktitle = UAI-89,
        pages = {182-193},
     location = {Windsor, Ontario},
         year = 1989,
} 
@book{Hoskinsetal2010,
        title = {Diagnostic Ultrasound: Physics and Equipment},
       author = {Peter R. Hoskins and Kevin Martin and Abigail Thrush},
         year = 2010,
    publisher = {Cambridge University Press},
} 
@book{HoskynetalEXECUTIVE-FUNCTION-17,
        title = {Executive Functions in Children's Everyday Lives: A Handbook for Professionals in Applied Psychology},
       author = {Hoskyn, M.J. and Iarocci, G. and Young, A.R.},
    publisher = {Oxford University Press},
         year = {2017},
     abstract = {Executive Functions in Children's Everyday Lives captures the diversity and complexity of the executive system that underlies children's everyday life experiences. Acquisition of executive functions, such as interpreting communication cues and the perspectives of others, is foundational to and a function of children's early social and communicative competencies. From the soccer field to the classroom, executive functions support children's strategic thinking and control of their environment. Knowing about executive functions and how this system of cognitive resources emerges in young children is important in understanding children's development. Recent research points to the importance of also considering environmental influences on the executive system. This book is unique in its focus on how experiences in children's early lives influence and are influenced by executive functions. Viewing executive functions through this broad lens is critical for professionals who intervene when children's access to executive functions is less than optimal. This book addresses a wide range of topics, including the neurological basis of executive functions in young children, the assessment of children's executive functions, theoretical and historical conceptions of executive functions, the relations between executive functions and theory of mind, multilingualism, early school transitions, and the relationship of executive functions to Autism and ADHD. This volume will be useful to professionals in applied psychology, undergraduate and graduate students, and social science and applied researchers.}
} 
@article{HoukandWiseCEREBRAL-CORTEX-95,
       author = {Houk, J. C. and Wise, S. P.},
        title = {Distributed modular architectures linking basal ganglia, cerebellum, and cerebral cortex: their role in planning and controlling action},
      journal = {Cerebral Cortex},
       volume = {5},
       number = {2},
         year = {1995},
        pages = {95-110},
     abstract = {The motor system includes structures distributed widely through the CNS, and in this feature article we present a scheme for how they might cooperate in the control of action. Distributed modules, which constitute the basic building blocks of our model, include recurrent loops connecting distant brain structures, as well as local circuitry that modulates loop activity. We consider interconnections among the basal ganglia, cerebellum, and cerebral cortex and the specialized properties of certain cell types within each of those structures, namely, striatal spiny neurons, cerebellar Purkinje cells, and neocortical pyramidal cells. In our model, striatal spiny neurons of the basal ganglia function in contextual pattern recognition under the training influence of reinforcement signals transmitted in dopamine fibers. Cerebellar Purkinje cells also function in pattern recognition, in their case to select and execute actions through training supervised by climbing fibers, which signal discoordination. Neocortical pyramidal cells perform collective computations learned through a local training mechanism and also function as information stores for other modular operations. We discuss how distributed modules might function in a parallel, cooperative manner to plan, modulate, and execute action.}
}
@article{HowardIEEE-66,
       author = {Howard, Ronald A.},
        title = {Information Value Theory},
      journal = {{IEEE} Transactions on Systems Science and Cybernetics},
       volume = 2,
        issue = 1,
        pages = {22-26},
         year = 1966,
}
@article{HowardIEEE-67,
       author = {Howard, Ronald A.},
        title = {Value of Information Lotteries},
      journal = {{IEEE} Transactions on Systems Science and Cybernetics},
       volume = 3,
        issue = 1,
        pages = {54-60},
         year = 1967,
}
@book{HowardPhD,
       author = {Howard, Ronald A.},
        title = {Dynamic Programming and {M}arkov Processes},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1960,
}
@article{HowardandKahanaJoMP-02,
       author = {Marc W. Howard and Michael J. Kahana},
        title = {A Distributed Representation of Temporal Context},
      journal = {Journal of Mathematical Psychology},
       volume = {46},
       number = {3},
        pages = {269-299},
         year = {2002},
     abstract = {The principles of recency and contiguity are two cornerstones of the theoretical and empirical analysis of human memory. Recency has been alternatively explained by mechanisms of decay, displacement, and retroactive interference. Another account of recency is based on the idea of variable context (Estes, 1955; Mensink \& Raaijmakers, 1989). Such notions are typically cast in terms of a randomly fluctuating population of elements reflective of subtle changes in the environment or in the subjects' mental state. This random context view has recently been incorporated into distributed and neural network memory models (Murdock, 1997; Murdock, Smith, \& Bai, 2001). Here we propose an alternative model. Rather than being driven by random fluctuations, this formulation, the temporal context model (TCM), uses retrieval of prior contextual states to drive contextual drift. In TCM, retrieved context is an inherently asymmetric retrieval cue. This allows the model to provide a principled explanation of the widespread advantage for forward recalls in free and serial recall. Modeling data from single-trial free recall, we demonstrate that TCM can simultaneously explain recency and contiguity effects across time scales.}
}
@incollection{HowardandMatheson84a,
       author = {Ronald A. Howard and James E. Matheson},
        title = {Influence Diagrams},
       editor = {Ronald A. Howard and James E. Matheson},
    booktitle = {The Principles and Applications of Decision Analysis},
    publisher = {Strategic Decisions Group},
      address = {Menlo Park, CA 94025},
         year = 1984,
}
@book{HowardandMatheson84b,
       editor = {Howard, Ronald A. and Matheson, James E.},
        title = {The Principles and Applications of Decision Analysis},
    booktitle = {The Principles and Applications of Decision Analysis},
         year = 1984,
    publisher = {Strategic Decisions Group},
      address = {Menlo Park, California},
}
@article{HowardetalPR-05,
       author = {Howard, Marc W. and Fotedar, Mrigankka S. and Datey, Aditya V. and Hasselmo, Michael E.},
        title = {The temporal context model in spatial navigation and relational learning: toward a common explanation of medial temporal lobe function across domains},
      journal = {Psychological review},
         year = {2005},
       volume = {112},
        issue = {1},
        pages = {75-116},
     abstract = {The medial temporal lobe (MTL) has been studied extensively at all levels of analysis, yet its function remains unclear. Theory regarding the cognitive function of the MTL has centered along 3 themes. Different authors have emphasized the role of the MTL in episodic recall, spatial navigation, or relational memory. Starting with the temporal context model (M. W. Howard \& M. J. Kahana, 2002a), a distributed memory model that has been applied to benchmark data from episodic recall tasks, the authors propose that the entorhinal cortex supports a gradually changing representation of temporal context and the hippocampus proper enables retrieval of these contextual states. Simulation studies show this hypothesis explains the firing of place cells in the entorhinal cortex and the behavioral effects of hippocampal lesion in relational memory tasks. These results constitute a first step toward a unified computational theory of MTL function that integrates neurophysiological, neuropsychological, and cognitive findings.},
}
@article{HsiehandJasanoffACS-12,
       author = {Hsieh, V. and Jasanoff, A.},
        title = {Bioengineered probes for molecular magnetic resonance imaging in the nervous system},
      journal = {ACS Chemical Neuroscience},
    publisher = {American Chemical Society},
       volume = 3,
        issue = 8,
         year = 2012,
        pages = {593-602},
}
@inproceedings{HsiehetalIEEE-07,
        title = {Localized high-resolution {MR} imaging of rat brain architecture using micro-fabricated receive-only {RF} coil},
       author = {Hsieh, M. C. and Kuo, L. W. and Wu, E. L. and Chen, J. H.},
    booktitle = {{IEEE} Engineering Medical Biology Society},
         year = 2007,
        pages = {6074-6077},
     abstract = {This study introduced a localized approach for magnetic resonance microscopy (MRM) of the rat brain. A single-loop radiofrequency (RF) receiver coil designed for micro-imaging was developed by using electromagnetic simulation software widely used in communication fields. With transmit-only and receive-only (TORO) configuration, receive-only surface coil can achieve higher signal-to-noise ratio (SNR) at localized brain region. Corpus callosum and hippocampus were landmarks to evaluate the capacity of the proposed coil. On a 3T MRI system, high-resolution MRI of the dissected rat brain was acquired with spatial resolution of 117 x 117 x 500 micrometre cubed. The achieved high local SNR and spatial resolution will provide valuable information for resolving the architecture of the rat brain.}
}
@misc{HuPhD-14,
       author = {Hu, Miao},
        title = {Memristor: Modeling, Simulation and Usage in Neuromorphic Computation},
         year = 2014,
 howpublished = {PhD. Thesis. Electrical and Computer Engineering, University of Pittsburgh},
     abstract = {Memristor, the fourth passive circuit element, has attracted increased attention from various areas since the first real device was discovered in 2008. Its distinctive characteristic to record the historic profile of the voltage/current through itself creates great potential in future circuit design. Inspired by its high Scalability, ultra low power consumption and similar functionality to biology synapse, using memristor to build high density, high power efficiency neuromorphic circuits becomes one of most promising and also challenging applications. The challenges can be concluded into three levels: device level, circuit level and application level. At device level, we studied different memristor models and process variations, then we carried out three independent variation models to describe the variation and stochastic behavior of TiO2 memristors. These models can also extend to other memristor models. Meanwhile, these models are also compact enough for large-scale circuit simulation. At circuit level, inspired by the large-scale and unique requirement of memristor-based neuromorphic circuits, we designed a circuit simulator for efficient memristor cross-point array simulations. Our simulator is 4-5 orders of magnitude faster than tradition SPICE simulators. Both linear and nonlinear memristor cross-point arrays are studied for level-based and spike-based neuromorphic circuits, respectively. At application level, we first designed a few compact memristor-based neuromorphic components, including "Macro cell" for efficient and high definition weight storage, memristor-based stochastic neuron and memristor-based spatio temporal synapse. We then studied three typical neural network models and their hardware realization on memristor-based neuromorphic circuits: Brain-State-in-a-Box (BSB) model stands for level-based neural network, and STDP/ReSuMe models stand for spiking neural network for temporal learning. Our result demonstrates the high resilience to variation of memristor-based circuits and ultra-low power consumption. In this thesis, we have proposed a complete and detailed analysis for memristor-based neuromorphic circuit design from the device level to the application level. In each level, both theoretical analysis and experimental data versification are applied to ensure the completeness and accuracy of the work.}
}
@article{HuaetalBC-99,
       author = {Hua, S. E. and Houk, J. C. and Mussa-Ivaldi, F. A.},
        title = {Emergence of symmetric, modular, and reciprocal connections in recurrent networks with {H}ebbian learning},
      journal = {Biological Cybernetics},
         year = {1999},
       volume = {81},
       number = {3},
        pages = {211-225},
     abstract = {While learning and development are well characterized in feedforward networks, these features are more difficult to analyze in recurrent networks due to the increased complexity of dual dynamics - the rapid dynamics arising from activation states and the slow dynamics arising from learning or developmental plasticity. We present analytical and numerical results that consider dual dynamics in a recurrent network undergoing Hebbian learning with either constant weight decay or weight normalization. Starting from initially random connections, the recurrent network develops symmetric or near-symmetric connections through Hebbian learning. Reciprocity and modularity arise naturally through correlations in the activation states. Additionally, weight normalization may be better than constant weight decay for the development of multiple attractor states that allow a diverse representation of the inputs. These results suggest a natural mechanism by which synaptic plasticity in recurrent networks such as cortical and brainstem premotor circuits could enhance neural computation and the generation of motor programs.}
}
@article{HuangandDarwicheIJAR-96,
       author = {C. Huang and A. Darwiche},
        title = {Inference in belief networks: A procedural guide},
      journal = {International Journal of Approximate Reasoning},
       volume = 15,
        issue = 3,
         year = 1996,
        pages = {225-263},
}
@article{HuangandJainCoRR-13,
       author = {Gary B. Huang and Viren Jain},
        title = {Deep and Wide Multiscale Recursive Networks for Robust Image Labeling},
      journal = {CoRR},
       volume = {arXiv:1310.0354},
         year = 2013,
     abstract = {Feedforward multilayer networks trained by supervised learning have recently demonstrated state of the art performance on image labeling problems such as boundary prediction and scene parsing. As even very low error rates can limit practical usage of such systems, methods that perform closer to human accuracy remain desirable. In this work, we propose a new type of network with the following properties that address what we hypothesize to be limiting aspects of existing methods: (1) a `wide' structure with thousands of features, (2) a large field of view, (3) recursive iterations that exploit statistical dependencies in label space, and (4) a parallelizable architecture that can be trained in a fraction of the time compared to benchmark multilayer convolutional networks. For the specific image labeling problem of boundary prediction, we also introduce a novel example weighting algorithm that improves segmentation accuracy. Experiments in the challenging domain of connectomic reconstruction of neural circuity from 3d electron microscopy data show that these "Deep And Wide Multiscale Recursive" (DAWMR) networks lead to new levels of image labeling performance. The highest performing architecture has twelve layers, interwoven supervised and unsupervised stages, and uses an input field of view of 157,464 voxels (54^3) to make a prediction at each image location. We present an associated open source software package that enables the simple and flexible creation of DAWMR networks.}
}
@article{HuangandJiangJMC-13,
       author = {Huang, Zhuo and Jiang, Xingyu},
        title = {Micro/nano-scale materials and structures for constructing neuronal networks and addressing neurons},
      journal = {Journal of Materials Chemistry C},
       volume = 1,
        issue = 46,
         year = 2013,
        pages = {7652-7662},
    publisher = {The Royal Society of Chemistry},
     abstract = {In recent years, a number of new materials and techniques at the micro/nano-scale for neuroscience have been reported, in particular in studies of neuronal development and electronic addressing. They offer new capabilities to fabricate tools in controlling the surface properties (such as topography and chemistry) as well as executing electrical stimulations and measurements. Here we review the basic principles of these micro/nano-scale materials and highlight the important advances in this field, and finally provide some perspectives for the future.}
}
@inproceedings{HuangandMumfordCVPR-99,
   author = {Jinggang Huang and David Mumford},
    title = {Statistics of Natural Images and Models},
booktitle = CVPR,
    pages = {541-547},
     year = 1999,
}
@inproceedings{HuangetalACL-12,
       author = {Huang, Eric H. and Socher, Richard and Manning, Christopher D. and Ng, Andrew Y.},
        title = {Improving word representations via global context and multiple word prototypes},
    booktitle = {Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics: Long Papers - Volume 1},
     location = {Jeju Island, Korea},
    publisher = {Association for Computational Linguistics},
      address = {Stroudsburg, PA, USA},
         year = 2012,
        pages = {873-882},
}
@article{HuangetalARB-09,
        title = {Super-resolution fluorescence microscopy},
       author = {Huang, B. and Bates, M. and Zhuang, X.},
      journal = {Annual Reviews Biochemistry},
       volume = 78,
         year = 2009,
     abstract = {Achieving a spatial resolution that is not limited by the diffraction of light, recent developments of super-resolution fluorescence microscopy techniques allow the observation of many biological structures not resolvable in conventional fluorescence microscopy. New advances in these techniques now give them the ability to image three-dimensional (3D) structures, measure interactions by multicolor colocalization, and record dynamic processes in living cells at the nanometer scale. It is anticipated that super-resolution fluorescence microscopy will become a widely used tool for cell and tissue imaging to provide previously unobserved details of biological structures and processes.}
}
@article {HuangetalBIORXIV-18,
       author = {Huang, Yu-Chi and Wang, Cheng-Te and Su, Ta-Shun and Kao, Kuo-Wei and Lin, Yen-Jen and Chiang, Ann-Shyn and Lo, Chung-Chuan},
        title = {A single-cell level and connectome-derived computational model of the Drosophila brain},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Laboratory},
         year = {2018},
     abstract = {Computer simulations play an important role in testing hypotheses, integrating knowledge, and providing predictions of neural circuit functions. While considerable effort has been dedicated into simulating primate or rodent brains, the fruit fly (Drosophila melanogaster) is becoming a promising model animal in computational neuroscience for its small brain size, complex cognitive behavior, and abundancy of data available from genes to circuits. Moreover, several Drosophila connectome projects have generated a large number of neuronal images that account for a significant portion of the brain, making a systematic investigation of the whole brain circuit possible. Supported by FlyCircuit (http://www.flycircuit.tw), one of the largest Drosophila neuron image databases, we began a long-term project with the goal to construct a whole-brain spiking network model of the Drosophila brain. In this paper, we report the outcome of the first phase of the project. We developed the Flysim platform, which 1) identifies the polarity of each neuron arbor, 2) predicts connections between neurons, 3) translates morphology data from the database into physiology parameters for computational modeling, 4) reconstructs a brain-wide network model, which consists of 20,089 neurons and 1,044,020 synapses, and 5) performs computer simulations of the resting state. We compared the reconstructed brain network with a randomized brain network by shuffling the connections of each neuron. We found that the reconstructed brain can be easily stabilized by implementing synaptic short-term depression, while the randomized one exhibited seizure-like firing activity under the same treatment. Furthermore, the reconstructed Drosophila brain was structurally and dynamically more diverse than the randomized one and exhibited both Poisson-like and patterned firing activities. Despite being at its early stage of development, this single-cell level brain model allows us to study some of the fundamental properties of neural networks including network balance, critical behavior, long-term stability, and plasticity.},
}
@article{HuangetalCoRR-16a,
       author = {Gao Huang and Yu Sun and Zhuang Liu and Daniel Sedra and Kilian Q. Weinberger},
        title = {Deep Networks with Stochastic Depth},
      journal = {CoRR},
       volume = {arXiv:1603.09382},
         year = {2016},
          url = {http://arxiv.org/abs/1603.09382},
     abstract = {Very deep convolutional networks with hundreds or more layers have lead to significant reductions in error on competitive benchmarks like the ImageNet or COCO tasks. Although the unmatched expressiveness of the many deep layers can be highly desirable at test time, training very deep networks comes with its own set of challenges. The gradients can vanish, the forward flow often diminishes and the training time can be painfully slow even on modern computers. In this paper we propose stochastic depth, a training procedure that enables the seemingly contradictory setup to train short networks and obtain deep networks. We start with very deep networks but during training, for each mini-batch, randomly drop a subset of layers and bypass them with the identity function. The resulting networks are short (in expectation) during training and deep during testing. Training Residual Networks with stochastic depth is compellingly simple to implement, yet effective. We show that this approach successfully addresses the training difficulties of deep networks and complements the recent success of Residual and Highway Networks. It reduces training time substantially and improves the test errors on almost all data sets significantly (CIFAR-10, CIFAR-100, SVHN). Intriguingly, with stochastic depth we can increase the depth of residual networks even beyond 1200 layers and still yield meaningful improvements in test error (4.91\%) on CIFAR-10.}
}
@article{HuangetalCoRR-16b,
        title = {Fully-Automatic Synapse Prediction and Validation on a Large Data Set},
       author = {Gary B. Huang, Louis K. Scheffer, Stephen M. Plaza},
      journal = {CoRR},
       volume = {arXiv:1604.03075},
         year = {2016},
     abstract = {Extracting a connectome from an electron microscopy (EM) data set requires identification of neurons and determination of synapses between neurons. As manual extraction of this information is very time-consuming, there has been extensive research effort to automatically segment the neurons to help guide and eventually replace manual tracing. Until recently, there has been comparatively less research on automatically detecting the actual synapses between neurons. This discrepancy can, in part, be attributed to several factors: obtaining neuronal shapes is a prerequisite first step in extracting a connectome, manual tracing is much more time-consuming than annotating synapses, and neuronal contact area can be used as a proxy for synapses in determining connections. However, recent research has demonstrated that contact area alone is not a sufficient predictor of synaptic connection. Moreover, as segmentation has improved, we have observed that synapse annotation is consuming a more significant fraction of overall reconstruction time. This ratio will only get worse as segmentation improves, gating overall possible speed-up. Therefore, we address this problem by developing algorithms that automatically detect pre-synaptic neurons and their post-synaptic partners. In particular, pre-synaptic structures are detected using a Deep and Wide Multiscale Recursive Network, and post-synaptic partners are detected using a MLP with features conditioned on the local segmentation. This work is novel because it requires minimal amount of training, leverages advances in image segmentation directly, and provides a complete solution for polyadic synapse detection. We further introduce novel metrics to evaluate our algorithm on connectomes of meaningful size. These metrics demonstrate that complete automatic prediction can be used to effectively characterize most connectivity correctly.}
}
@article{HuangetalCoRR-17,
       author = {{Huang}, Q. and {Smolensky}, P. and {He}, X. and {Deng}, L. and {Wu}, D.},
        title = {{Tensor Product Generation Networks}},
      journal = {CoRR},
         year = 2017,
     abstract = {We present a new tensor product generation network (TPGN) that generates natural language descriptions for images. The model has a novel architecture that instantiates a general framework for encoding and processing symbolic structure through neural network computation. This framework is built on Tensor Product Representations (TPRs). We evaluated the proposed TPGN on the MS COCO image captioning task. The exeprimental results show that the TPGN outperforms the LSTM based state-of-the-art baseline with a significant margin. Further, we show that our caption generation model can be interpreted as generating sequences of grammatical categories and retrieving words by their categories from a plan encoded as a distributed representation.}
}
@article{HuangetalJCN-13,
       author = {Huang, Tsung-Ren and Hazy, Thomas and A Herd, Seth and O'Reilly, Randall},
        title = {Assembling Old Tricks for New Tasks: A Neural Model of Instructional Learning and Control},
      journal = {Journal of cognitive neuroscience},
       volume = {25},
        issue = {6},
         year = {2013},
        pages = {843-841},
     abstract = {We can learn from the wisdom of others to maximize success. However, it is unclear how humans take advice to flexibly adapt behavior. On the basis of data from neuroanatomy, neurophysiology, and neuroimaging, a biologically plausible model is developed to illustrate the neural mechanisms of learning from instructions. The model consists of two complementary learning pathways. The slow-learning parietal pathway carries out simple or habitual stimulus-response (S-R) mappings, whereas the fast-learning hippocampal pathway implements novel S-R rules. Specifically, the hippocampus can rapidly encode arbitrary S-R associations, and stimulus-cued responses are later recalled into the basal ganglia-gated pFC to bias response selection in the premotor and motor cortices. The interactions between the two model learning pathways explain how instructions can override habits and how automaticity can be achieved through motor consolidation.}
}
@article{HuangetalNATURE-10,
    author = {Huang, Heng and Delikanli, Savas and Zeng, Hao and Ferkey, Denise M. and Pralle, Arnd},
     title = {Remote control of ion channels and neurons through magnetic-field heating of nanoparticles},
   journal = {Nature Nanonotechnology},
 publisher = {Nature Publishing Group},
    volume = 5,
     issue = 8,
      year = 2010,
     pages = {602-606},
}
@book{Hubel95,
        title = {Eye, Brain and Vision (Scientific American Library, Number 22)},
       author = {David H. Hubel},
     location = {New York},
    publisher = {W. H. Freeman and Company},
         year = 1995,
          url = {http://neuro.med.harvard.edu/site/dh/index.html},
}
@article{HubelandWieselJoP-61,
       author = {Hubel, D. H. and Wiesel, T. N},
        title = {Integrative action in the cat's lateral geniculate body},
      journal = {Journal of Physiology},
       volume = {155},
        pages = {385-398},
         year = {1961},
} 
@article{HubelandWieselJoP-62,
       author = {Hubel, D. H. and Wiesel, T. N},
        title = {Receptive fields, binocular interaction and functional architecture in the cat's visual cortex},
      journal = {Journal of Physiology},
       volume = {160},
        pages = {106-154},
         year = {1962},
}
@article{HubelandWieselJoP-68,
       author = {Hubel, D. H. and Wiesel, T. N},
        title = {Receptive fields and functional architecture of monkey striate cortex},
      journal = {Journal of Physiology},
       volume = 195,
        pages = {215-243},
         year = 1968,
}
@article{Huber85,
       author = {Peter J. Huber},
        title = {Projection Pursuit},
      journal = {The Annals of Statistics},
       volume = 13,
        issue = 2,
         year = 1985,
        pages = {435-475},
}
@techreport{HubermanandAdamic99,
       author = {B.A. Huberman and L.A. Adamic},
        title = {Evolutionary Dynamics of the {W}orld {W}ide {W}eb},
  institution = {Xerox PARC},
         year = 1999,
}
@article{HubermanandHoggAIJ-87,
       author = {Huberman, B. A. and Hogg, T.},
        title = {Phase Transitions in Artificial Initelligence Systems},
      journal = AIJ,
       volume = 33,
         year = 1987,
        pages = {155-171},
}
@article{HudsonandManningCoRR-19,
       author = {Drew A. Hudson and Christopher D. Manning},
        title = {Learning by Abstraction: The Neural State Machine},
      journal = {CoRR},
       volume = {arXiv:1907.03950},
         year = {2019},
     abstract = {We introduce the Neural State Machine, seeking to bridge the gap between the neural and symbolic views of AI and integrate their complementary strengths for the task of visual reasoning. Given an image, we first predict a probabilistic graph that represents its underlying semantics and serves as a structured world model. Then, we perform sequential reasoning over the graph, iteratively traversing its nodes to answer a given question or draw a new inference. In contrast to most neural architectures that are designed to closely interact with the raw sensory data, our model operates instead in an abstract latent space, by transforming both the visual and linguistic modalities into semantic concept-based representations, thereby achieving enhanced transparency and modularity. We evaluate our model on VQA-CP and GQA, two recent VQA datasets that involve compositionality, multi-step inference and diverse reasoning skills, achieving state-of-the-art results in both cases. We provide further experiments that illustrate the model's strong generalization capacity across multiple dimensions, including novel compositions of concepts, changes in the answer distribution, and unseen linguistic structures, demonstrating the qualities and efficacy of our approach.}
}
@article{HudsonandManningICLR-18,
       author = {Drew A. Hudson and Christopher D. Manning},
        title = {Compositional Attention Networks for Machine Reasoning},
      journal = {CoRR},
      comment = {Published as a conference paper at ICLR 2018},
       volume = {arXiv:1803.03067},
         year = {2018},
     abstract = {We present the MAC network, a novel fully differentiable neural network architecture, designed to facilitate explicit and expressive reasoning. MAC moves away from monolithic black-box neural architectures towards a design that encourages both transparency and versatility. The model approaches problems by decomposing them into a series of attention-based reasoning steps, each performed by a novel recurrent Memory, Attention, and Composition (MAC) cell that maintains a separation between control and memory. By stringing the cells together and imposing structural constraints that regulate their interaction, MAC effectively learns to perform iterative reasoning processes that are directly inferred from the data in an end-to-end approach. We demonstrate the model's strength, robustness and interpretability on the challenging CLEVR dataset for visual reasoning, achieving a new state-of-the-art 98.9\% accuracy, halving the error rate of the previous best model. More importantly, we show that the model is computationally-efficient and data-efficient, in particular requiring 5x less data than existing models to achieve strong results.}
}
@article{HuetalCoRR-12,
       author = {Yu Hu and James Trousdale and Kre\"{s}imir Jos\'{i}c and Eric Shea-Brown},
        title = {Motif Statistics and Spike Correlations in Neuronal Networks},
      journal = {CoRR},
       volume = {arXiv:1206.3537},
         year = {2015},
     abstract = {Motifs are patterns of subgraphs of complex networks. We studied the impact of such patterns of connectivity on the level of correlated, or synchronized, spiking activity among pairs of cells in a recurrent network model of integrate and fire neurons. For a range of network architectures, we find that the pairwise correlation coefficients, averaged across the network, can be closely approximated using only three statistics of network connectivity. These are the overall network connection probability and the frequencies of two second-order motifs: diverging motifs, in which one cell provides input to two others, and chain motifs, in which two cells are connected via a third intermediary cell. Specifically, the prevalence of diverging and chain motifs tends to increase correlation. Our method is based on linear response theory, which enables us to express spiking statistics using linear algebra, and a resumming technique, which extrapolates from second order motifs to predict the overall effect of coupling on network correlation. Our motif-based results seek to isolate the effect of network architecture perturbatively from a known network state.}
}
@article{HuetalCoRR-17,
       author = {Ronghang Hu and Jacob Andreas and Marcus Rohrbach and Trevor Darrell and Kate Saenko},
        title = {Learning to Reason: End-to-End Module Networks for Visual Question Answering},
      journal = {CoRR},
       volume = {arXiv:1704.05526},
         year = {2017},
     abstract = {Natural language questions are inherently compositional, and many are most easily answered by reasoning about their decomposition into modular sub-problems. For example, to answer "is there an equal number of balls and boxes?" we can look for balls, look for boxes, count them, and compare the results. The recently proposed Neural Module Network (NMN) architecture implements this approach to question answering by parsing questions into linguistic substructures and assembling question-specific deep networks from smaller modules that each solve one subtask. However, existing NMN implementations rely on brittle off-the-shelf parsers, and are restricted to the module configurations proposed by these parsers rather than learning them from data. In this paper, we propose End-to-End Module Networks (N2NMNs), which learn to reason by directly predicting instance-specific network layouts without the aid of a parser. Our model learns to generate network structures (by imitating expert demonstrations) while simultaneously learning network parameters (using the downstream task loss). Experimental results on the new CLEVR dataset targeted at compositional question answering show that N2NMNs achieve an error reduction of nearly 50\% relative to state-of-the-art attentional approaches, while discovering interpretable network architectures specialized for each question.}
}
@article{HuetalCoRR-18,
       author = {Haonan Yu and Haichao Zhang and Wei Xu},
        title = {Interactive Grounded Language Acquisition and Generalization in a {2D} World},
      journal = {CoRR},
       volume = {arXiv:1802.01433},
         year = {2018},
     abstract = {We build a virtual agent for learning language in a 2D maze-like world. The agent sees images of the surrounding environment, listens to a virtual teacher, and takes actions to receive rewards. It interactively learns the teacher's language from scratch based on two language use cases: sentence-directed navigation and question answering. It learns simultaneously the visual representations of the world, the language, and the action control. By disentangling language grounding from other computational routines and sharing a concept detection function between language grounding and prediction, the agent reliably interpolates and extrapolates to interpret sentences that contain new word combinations or new words missing from training sentences. The new words are transferred from the answers of language prediction. Such a language ability is trained and evaluated on a population of over 1.6 million distinct sentences consisting of 119 object words, 8 color words, 9 spatial-relation words, and 50 grammatical words. The proposed model significantly outperforms five comparison methods for interpreting zero-shot sentences. In addition, we demonstrate human-interpretable intermediate outputs of the model in the appendix}
}
@article{HuetalCoRR-19,
       author = {Ronghang Hu and Anna Rohrbach and Trevor Darrell and Kate Saenko},
        title = {Language-Conditioned Graph Networks for Relational Reasoning},
      journal = {CoRR},
       volume = {arXiv:1905.04405},
         year = {2019},
     abstract = {Solving grounded language tasks often requires reasoning about relationships between objects in the context of a given task. For example, to answer the question "What color is the mug on the plate?" we must check the color of the specific mug that satisfies the "on" relationship with respect to the plate. Recent work has proposed various methods capable of complex relational reasoning. However, most of their power is in the inference structure, while the scene is represented with simple local appearance features. In this paper, we take an alternate approach and build contextualized representations for objects in a visual scene to support relational reasoning. We propose a general framework of Language-Conditioned Graph Networks (LCGN), where each node represents an object, and is described by a context-aware representation from related objects through iterative message passing conditioned on the textual input. E.g., conditioning on the "on" relationship to the plate, the object "mug" gathers messages from the object "plate" to update its representation to "mug on the plate", which can be easily consumed by a simple classifier for answer prediction. We experimentally show that our LCGN approach effectively supports relational reasoning and improves performance across several tasks and datasets.},
}
@article {HuetalJEM-18,
       author = {Hu, Xiangyou and Das, Brati and Hou, Hailong and He, Wanxia and Yan, Riqiang},
        title = {{BACE1} deletion in the adult mouse reverses preformed amyloid deposition and improves cognitive functions},
      journal = {Journal of Experimental Medicine},
         year = {2018},
    publisher = {Rockefeller University Press},
     abstract = {BACE1 initiates the generation of the β-amyloid peptide, which likely causes Alzheimer{\textquoteright}s disease (AD) when accumulated abnormally. BACE1 inhibitory drugs are currently being developed to treat AD patients. To mimic BACE1 inhibition in adults, we generated BACE1 conditional knockout (BACE1fl/fl) mice and bred BACE1fl/fl mice with ubiquitin-CreER mice to induce deletion of BACE1 after passing early developmental stages. Strikingly, sequential and increased deletion of BACE1 in an adult AD mouse model (5xFAD) was capable of completely reversing amyloid deposition. This reversal in amyloid deposition also resulted in significant improvement in gliosis and neuritic dystrophy. Moreover, synaptic functions, as determined by long-term potentiation and contextual fear conditioning experiments, were significantly improved, correlating with the reversal of amyloid plaques. Our results demonstrate that sustained and increasing BACE1 inhibition in adults can reverse amyloid deposition in an AD mouse model, and this observation will help to provide guidance for the proper use of BACE1 inhibitors in human patients.},
}
@article{HuetalNATURE-12,
        title = {Evolution of the human-specific {microRNA miR-941}},
       author = {Hu, Hai Yang and He, Liu and Fominykh, Kseniya and Yan, Zheng and Guo, Song and Zhang, Xiaoyu and Taylor, Martin S. and Tang, Lin and Li, Jie and Liu, Jianmei and Wang, Wen and Yu, Haijing and Khaitovich, Philipp},
      journal = {Nature Communications},
    publisher = {Nature Publishing Group},
       volume = 3,
         year = 2012,
        pages = 1145,
     abstract = {MicroRNA-mediated gene regulation is important in many physiological processes. Here we explore the roles of a microRNA, miR-941, in human evolution. We find that miR-941 emerged de novo in the human lineage, between six and one million years ago, from  an evolutionarily volatile tandem repeat sequence. Its copy-number remains polymorphic in humans and shows a trend for decreasing copy-number with migration out of Africa. Emergence of miR-941 was accompanied by accelerated loss of miR-941-binding sites, presumably to escape regulation. We further show that miR-941 is highly expressed in pluripotent cells, repressed upon differentiation and preferentially targets genes in hedgehog- and insulin-signalling pathways, thus suggesting roles in cellular differentiation. Human-specific effects of miR-941 regulation are detectable in the brain and affect genes involved in neurotransmitter signalling. Taken together, these results implicate miR-941 in human evolution, and provide an example of rapid regulatory evolution in the human lineage.},
}
@book{HughesandCresswell68,
       author = {Hughes, G. E. and Cresswell, M. J.},
        title = {An Introduction to Modal Logic},
    publisher = {Methuen},
      address = {London},
         year = 1968,
}
@article{HumpelNEUROSCIENCE-15,
        title = {Organotypic brain slice cultures: A review},
      journal = {Neuroscience},
       volume = {305},
       number = {Supplement C},
        pages = {86-98},
         year = {2015},
       author = {Christian Humpel},
     abstract = {In vitro cell cultures are an important tool for obtaining insights into cellular processes in an isolated system and a supplement to in vivo animal experiments. While primary dissociated cultures permit a single homogeneous cell population to be studied, there is a clear need to explore the function of brain cells in a three-dimensional system where the main architecture of the cells is preserved. Thus, organotypic brain slice cultures have proven to be very useful in investigating cellular and molecular processes of the brain in vitro. This review summarizes (1) the historical development of organotypic brain slices focusing on the membrane technology, (2) methodological aspects regarding culturing procedures, age of donors or media, (3) whether the cholinergic neurons serve as a model of neurodegeneration in Alzheimer's disease, (4) or the nigrostriatal dopaminergic neurons as a model of Parkinson's disease and (5) how the vascular network can be studied, especially with regard to a synthetic blood-brain barrier. This review will also highlight some limits of the model and give an outlook on future applications.}
}
@book{Humphrey02,
     author = {Nicholas Humphrey},
      title = {The Mind Made Flesh: Frontiers of Psychology},
  publisher = {Oxford University Press},
    address = {Oxford},
       year = 2002,
}
@article {HumsetalELIFE-16,
       author = {Hums, Ingrid and Riedl, Julia and Mende, Fanny and Kato, Saul and Kaplan, Harris S and Latham, Richard and Sonntag, Michael and Traunm{\"{u}}ller, Lisa and Zimmer, Manuel},
       editor = {Calabrese, Ronald L},
        title = {Regulation of two motor patterns enables the gradual adjustment of locomotion strategy in \textit{Caenorhabditis elegans}},
      journal = {eLife},
    publisher = {eLife Sciences Publications, Ltd},
       volume = {5},
         year = {2016},
        pages = {e14116},
     abstract = {In animal locomotion a tradeoff exists between stereotypy and flexibility: fast long-distance travelling (LDT) requires coherent regular motions, while local sampling and area-restricted search (ARS) rely on flexible movements. We report here on a posture control system in C. elegans that coordinates these needs. Using quantitative posture analysis we explain worm locomotion as a composite of two modes: regular undulations versus flexible turning. Graded reciprocal regulation of both modes allows animals to flexibly adapt their locomotion strategy under sensory stimulation along a spectrum ranging from LDT to ARS. Using genetics and functional imaging of neural activity we characterize the counteracting interneurons AVK and DVA that utilize FLP-1 and NLP-12 neuropeptides to control both motor modes. Gradual regulation of behaviors via this system is required for spatial navigation during chemotaxis. This work shows how a nervous system controls simple elementary features of posture to generate complex movements for goal-directed locomotion strategies.},
}
@book{Hunter73,
       author = {Hunter, Geoffrey},
        title = {Metalogic: an introdiction to the metatheory of standard first order logic},
    publisher = {University of California Press},
      address = {Berkeley, California},
         year = 1973,
}
@book{HurdandLoeb85,
       author = {Hurd, Albert E. and Loeb, Peter A.},
        title = {An Introduction to Nonstandard Real Analysis},
    publisher = {Harcourt Brace Jovanovich},
      address = {New York},
         year = 1985,
}
@inproceedings{HurtadoetalASR-05,
        title = {A stochastic approach to dialog management based on neural networks},
       author = {Hurtado, L.-F. and Griol, David and Sanchis, E. and Segarra, E.},
    booktitle = {IEEE Workshop on Automatic Speech Recognition and Understanding},
         year = 2005,
        pages = {226-231},
     abstract = {In this article, we present an approach to the development of a stochastic dialog manager. The model used by this dialog manager to generate its turns takes into account both the last turns of the user and system, and the information supplied by the user throughout the dialog. As the space of situations that can be presented in the dialogs is too large, some techniques for reducing this space have been proposed. This system has been developed in the DIHANA project, whose goal is the design and development of a dialog system to access a railway information system using spontaneous speech in Spanish. A training corpus of 900 dialogs, that was acquired through the Wizard of Oz, was used to learn the models. An evaluation of the dialog manager is also presented},
}
@article{HusseinetalACM-17,
       author = {Hussein, Ahmed and Gaber, Mohamed Medhat and Elyan, Eyad and Jayne, Chrisina},
        title = {Imitation Learning: A Survey of Learning Methods},
      journal = {ACM Computing Surveys},
    publisher = {Association for Computing Machinery},
      address = {New York, NY, USA},
       volume = {50},
       number = {2},
         year = {2017},
     abstract = {Imitation learning techniques aim to mimic human behavior in a given task. An agent (a learning machine) is trained to perform a task from demonstrations by learning a mapping between observations and actions. The idea of teaching by imitation has been around for many years; however, the field is gaining attention recently due to advances in computing and sensing as well as rising demand for intelligent applications. The paradigm of learning by imitation is gaining popularity because it facilitates teaching complex tasks with minimal expert knowledge of the tasks. Generic imitation learning methods could potentially reduce the problem of teaching a task to that of providing demonstrations, without the need for explicit programming or designing reward functions specific to the task. Modern sensors are able to collect and transmit high volumes of data rapidly, and processors with high computational power allow fast processing that maps the sensory data to actions in a timely manner. This opens the door for many potential AI applications that require real-time perception and reaction such as humanoid robots, self-driving vehicles, human computer interaction, and computer games, to name a few. However, specialized algorithms are needed to effectively and robustly learn models as learning by imitation poses its own set of challenges. In this article, we survey imitation learning methods and present design options in different steps of the learning process. We introduce a background and motivation for the field as well as highlight challenges specific to the imitation problem. Methods for designing and evaluating imitation learning tasks are categorized and reviewed. Special attention is given to learning methods in robotics and games as these domains are the most popular in the literature and provide a wide array of problems and methodologies. We extensively discuss combining imitation learning approaches using different sources and methods, as well as incorporating other motion learning methods to enhance imitation. We also discuss the potential impact on industry, present major applications, and highlight current and future research directions.},
}
@article{HuthetalNATURE-16,
       author = {Huth, Alexander G. and de Heer, Wendy A. and Griffiths, Thomas L. and Theunissen, Fr\`{e}d\`{e}ric E. and Gallant, Jack L.},
        title = {Natural speech reveals the semantic maps that tile human cerebral cortex},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = {532},
        issue = {7600},
         year = {2016},
        pages = {453-458},
     abstract = {The meaning of language is represented in regions of the cerebral cortex collectively known as the 'semantic system'. However, little of the semantic system has been mapped comprehensively, and the semantic selectivity of most regions is unknown. Here we systematically map semantic selectivity across the cortex using voxel-wise modelling of functional MRI (fMRI) data collected while subjects listened to hours of narrative stories. We show that the semantic system is organized into intricate patterns that seem to be consistent across individuals. We then use a novel generative model to create a detailed semantic atlas. Our results suggest that most areas within the semantic system represent information about specific semantic domains, or groups of related concepts, and our atlas shows which domains are represented in each area. This study demonstrates that data-driven methods--commonplace in studies of human neuroanatomy and functional connectivity--provide a powerful and efficient means for mapping functional representations in the brain.},
}
@book{Huttegger2017,
        title = {The Probabilistic Foundations of Rational Learning},
       author = {Huttegger, S.M.},
       series = {The Probabilistic Foundations of Rational Learning},
         year = {2017},
    publisher = {Cambridge University Press}
}
@article{HynynenetalNEUROIMAGE-05,
        title = {Local and reversible blood-brain barrier disruption by noninvasive focused ultrasound at frequencies suitable for trans-skull sonications},
       author = {Hynynen, K. and McDannold, N. and Sheikov, N.A. and Jolesz, F.A. and Vykhodtseva, N.},
      journal = {Neuroimage},
       volume = 24,
        issue = 1,
         year = 2005,
        pages = {12-20},
     abstract = {The purpose of this study was to test the hypothesis that burst ultrasound in the presence of an ultrasound contrast agent can disrupt the blood-brain barrier (BBB) with acoustic parameters suitable for completely noninvasive exposure through the skull. The 10-ms exposures were targeted in the brains of 22 rabbits with a frequency of 690 kHz, a repetition frequency of 1 Hz, and peak rarefactional pressure amplitudes up to 3.1 MPa. The total exposure (sonication) time was 20 s. Prior to each sonication, a bolus of ultrasound contrast agent was injected intravenously. Contrast-enhanced MR images were obtained after the sonications to detect localized BBB disruption via local enhancement in the brain. Brain sections were stained with H&E, TUNEL, and vanadium acid fuchsin (VAF)-toluidine blue staining. In addition, horseradish peroxidase (HRP) was injected into four rabbits prior to sonications and transmission electron microscopy was performed. The MRI contrast enhancement demonstrated BBB disruption at pressure amplitudes starting at 0.4 MPa with approximately 50\%; at 0.8 MPa, 90\%; and at 1.4 MPa, 100\% of the sonicated locations showed enhancement. The histology findings following 4 h survival indicated that brain tissue necrosis was induced in approximately 70-80\% of the sonicated locations at a pressure amplitude level of 2.3 MPa or higher. At lower pressure amplitudes, however, small areas of erythrocyte extravasation were seen. The electron microscopy findings demonstrated HRP passage through vessel walls via both transendothelial and paraendothelial routes. These results demonstrate that completely noninvasive focal disruption of the BBB is possible.}
}
@article{Hyvarinen-10,
       author = {Aapo Hyv\"{a}rinen},
        title = {Statistical Models of Natural Images and Cortical Visual Representation},
      journal = {Topics in Cognitive Science},
       volume = 2,
        issue = 2,
        pages = {251-264},
         year = 2010,
}
@article{HyvarinenNC-01,
       author = {Hyvarinen, Aapo},
        title = {Complexity Pursuit: Separating Interesting Components from Time Series},
      journal = {Neural Computation},
         year = {2001},
    publisher = {MIT Press},
       volume = {13},
        issue = {4},
        pages = {883-898},
     abstract = {A generalization of projection pursuit for time series, that is, signals with time structure, is introduced. The goal is to find projections of time series that have interesting structure, defined using criteria related to Kolmogoroff complexity or coding length. Interesting signals are those that can be coded with a short code length. We derive a simple approximation of coding length that takes into account both the nongaussianity and the autocorrelations of the time series. Also, we derive a simple algorithm for its approximative optimization. The resulting method is closely related to blind separation of nongaussian, time-dependent source signals.},
}
@article{HyvarinenandHoyerNC-00,
       author = {Aapo Hyv\"{a}rinen and Patrick O. Hoyer},
        title = {Emergence of phase and shift invariant features by decomposition of natural images into independent feature subspaces},
      journal = {Neural Computation},
       volume = 12,
        issue = 7,
        pages = {1705-1720},
         year = 2000,
}
@article{HyvarinenandHoyerVR-01,
       author = {Aapo Hyv\"{a}rinen and Patrik O. Hoyer},
        title = {A Two-Layer Sparse Coding Model Learns Simple and Complex Cell Receptive Fields and Topography From Natural Images},
      journal = {Vision Research},
       volume = 41,
        issue = 18,
        pages = {2413-2423},
         year = 2001
}
@article{HyvarinenandKosterNETWORK-07,
       author = {Aapo Hyv\"{a}rinen and U. Koster},
        title = {Complex cell pooling and the statistics of natural images},
      journal = {Network},
       volume = 18,
        issue = 2,
        pages = {81-100},
         year = 2007,
}
@book{Hyvarinenetal2009,
        title = {Natural Image Statistics: A probabilistic approach to early computational vision},
       author = {Aapo Hyv\"{a}rinen and Jarmo Hurri and Patrik O. Hoyer},
    publisher = {Springer-Verlag},
         year = 2009,
}
@article{HyvarinenetalJOSA-03,
       author = {Aapo Hyv\"{a}rinen and Jarmo Hurri and Jaakko V\"{a}yrynen},
        title = {Bubbles: a unifying framework for low-level statistical properties of natural image sequences},
      journal = {Journal of the Optical Society of America},
    publisher = {Optical Society of America},
       volume = 20,
        issue = 7,
        pages = {1237-1252},
         year = 2003,
}
@article{HyvarinenetalNC-01,
       author = {Aapo Hyv\"{a}rinen and Patrik O. Hoyer},
        title = {Topographic Independent Component Analysis},
      journal = {Neural Computation},
       volume = 17,
        issue = 7,
        pages = {1527-1558},
         year = 2001,
}
@article{IadevaiaetalPLoS-14,
       author = {Iadevaia, Sergio and Nakhleh, Luay K. and Azencott, Robert and Ram, Prahlad T.},
      journal = {{PLoS ONE}},
    publisher = {Public Library of Science},
        title = {Mapping Network Motif Tunability and Robustness in the Design of Synthetic Signaling Circuits},
         year = {2014},
       volume = {9},
       number = {3},
        pages = {1-7},
     abstract = {Cellular networks are highly dynamic in their function, yet evolutionarily conserved in their core network motifs or topologies. Understanding functional tunability and robustness of network motifs to small perturbations in function and structure is vital to our ability to synthesize controllable circuits. In establishing core sets of network motifs, we selected topologies that are overrepresented in mammalian networks, including the linear, feedback, feed-forward, and bifan circuits. Static and dynamic tunability of network motifs were defined as the motif ability to respectively attain steady-state or transient outputs in response to pre-defined input stimuli. Detailed computational analysis suggested that static tunability is insensitive to the circuit topology, since all of the motifs displayed similar ability to attain predefined steady-state outputs in response to constant inputs. Dynamic tunability, in contrast, was tightly dependent on circuit topology, with some motifs performing superiorly in achieving observed time-course outputs. Finally, we mapped dynamic tunability onto motif topologies to determine robustness of motif structures to changes in topology and identify design principles for the rational assembly of robust synthetic networks.},
}
@article{IbarraJACM-75,
       author = {Ibarra, O. H. and Kim, C. E.},
        title = {Fast Approximation Algorithms for the Knapsack and Sum of the Subset Problems},
      journal = {Journal of the {ACM}},
       volume = 22,
         year = 1975,
        pages = {463-468},
}
@article{IfukorCS-05,
       author = {Ifukor, Presley A.},
        title = {Modelling the Mapping Mechanism in Metaphors},
         year = {2005},
      journal = {Cognitive Science},
    publisher = {Institute for Cognitive Science, Seoul National University},
       volume = {6},
        issue = {1},
        pages = {21-44},
     abstract = {This paper examines some current issues in the syntax and semantics of metaphor and analogy. Theories of metaphoric alignment, structure-mapping and an anti-unification algorithm for tracking several kinds of metaphors are explored.},
}
@article{IgelstrometalENEURO-16,
       author = {Igelstr\"{o}m, Kajsa M. and Webb, Taylor W. and Kelly, Yin T. and Graziano, Michael S.},
        title = {Topographical Organization of Attentional, Social, and Memory Processes in the Human Temporoparietal Cortex},
      journal = {eNeuro},
    publisher = {Society for Neuroscience},
       volume = {3},
       number = {2},
         year = {2016},
     abstract = {The temporoparietal junction (TPJ) is activated in association with a large range of functions, including social cognition, episodic memory retrieval, and attentional reorienting. An ongoing debate is whether the TPJ performs an overarching, domain-general computation, or whether functions reside in domain-specific subdivisions. We scanned subjects with fMRI during five tasks known to activate the TPJ, probing social, attentional, and memory functions, and used data-driven parcellation (independent component analysis) to isolate task-related functional processes in the bilateral TPJ. We found that one dorsal component in the right TPJ, which was connected with the frontoparietal control network, was activated in all of the tasks. Other TPJ subregions were specific for attentional reorienting, oddball target detection, or social attribution of belief. The TPJ components that participated in attentional reorienting and oddball target detection appeared spatially separated, but both were connected with the ventral attention network. The TPJ component that participated in the theory-of-mind task was part of the default-mode network. Further, we found that the BOLD response in the domain-general dorsal component had a longer latency than responses in the domain-specific components, suggesting an involvement in distinct, perhaps postperceptual, computations. These findings suggest that the TPJ performs both domain-general and domain-specific computations that reside within spatially distinct functional components.},
}
@article{IkegayaetalPLoS-08,
       author = {Ikegaya, Yuji and Matsumoto, Wataru and Chiou, Huei-Yu and Yuste, Rafael and Aaron, Gloster},
      journal = {{PLoS} ONE},
    publisher = {Public Library of Science},
        title = {Statistical Significance of Precisely Repeated Intracellular Synaptic Patterns},
       volume = 3,
        issue = 12,
         year = 2008,
        pages = {e3983},
     abstract = {Can neuronal networks produce patterns of activity with millisecond accuracy? It may seem unlikely, considering the probabilistic nature of synaptic transmission. However, some theories of brain function predict that such precision is feasible and can emerge from the non-linearity of the action potential generation in circuits of connected neurons. Several studies have presented evidence for and against this hypothesis. Our earlier work supported the precision hypothesis, based on results demonstrating that precise patterns of synaptic inputs could be found in intracellular recordings from neurons in brain slices and {\it{in vivo}}. To test this hypothesis, we devised a method for finding precise repeats of activity and compared repeats found in the data to those found in surrogate datasets made by shuffling the original data. Because more repeats were found in the original data than in the surrogate data sets, we argued that repeats were not due to chance occurrence. Mokeichev et al. (2007) challenged these conclusions, arguing that the generation of surrogate data was insufficiently rigorous. We have now reanalyzed our previous data with the methods introduced from Mokeichev et al. (2007). Our reanalysis reveals that repeats are statistically significant, thus supporting our earlier conclusions, while also supporting many conclusions that Mokeichev et al. (2007) drew from their recent {\it{in vivo}} recordings. Moreover, we also show that the conditions under which the membrane potential is recorded contributes significantly to the ability to detect repeats and may explain conflicting results. In conclusion, our reevaluation resolves the methodological contradictions between Ikegaya et al. (2004) and Mokeichev et al. (2007), but demonstrates the validity of our previous conclusion that spontaneous network activity is non-randomly organized.},
}
@article{IkegayaetalSCIENCE-04,
       author = {Y. Ikegaya and G. Aaron and R. Cossart and D. Aronov and I. Lamp and D. Ferster and R. Yuste},
        title = {Synfire chains and cortical songs: temporal modules of cortical activity},
      journal = {Science},
       volume = 304,
        issue = 5670,
         year = 2004,
        pages = {559-564},
     abstract = {How can neural activity propagate through cortical networks built with weak, stochastic synapses? We find precise repetitions of spontaneous patterns of synaptic inputs in neocortical neurons in vivo and in vitro. These patterns repeat after minutes, maintaining millisecond accuracy. Calcium imaging of slices reveals reactivation of sequences of cells during the occurrence of repeated intracellular synaptic patterns. The spontaneous activity drifts with time, engaging different cells. Sequences of active neurons have distinct spatial structures and are repeated in the same order over tens of seconds, revealing modular temporal dynamics. Higher order sequences are replayed with compressed timing.}
}
@inproceedings{ImpagliazzoandLevinFOCS-90,
       author = {Impagliazzo, Russell and Levin, Leonid A.},
        title = {No Better Ways to Generate Hard {NP} Instances Than Picking Uniformly at Random},
    booktitle = {Proceedings of the Thirty Second Annual Symposium on Foundations of Computer Science},
        pages = {812-821},
         year = 1990,
}
@article{IndiverietalCoRR-13,
       author = {Giacomo Indiveri and Bernab'{e} Linares-Barranco and Robert Legenstein and George Deligeorgis and Themistoklis Prodromakis},
        title = {Integration of nanoscale memristor synapses in neuromorphic computing architectures},
      journal = {Nanotechnology},
       volume = {24},
        issue = {38},
         year = {2013},
        pages = {384010},
     abstract = {Conventional neuro-computing architectures and artificial neural networks have often been developed with no or loose connections to neuroscience. As a consequence, they have largely ignored key features of biological neural processing systems, such as their extremely low-power consumption features or their ability to carry out robust and efficient computation using massively parallel arrays of limited precision, highly variable, and unreliable components. Recent developments in nano-technologies are making available extremely compact and low power, but also variable and unreliable solid-state devices that can potentially extend the offerings of availing CMOS technologies. In particular, memristors are regarded as a promising solution for modeling key features of biological synapses due to their nanoscale dimensions, their capacity to store multiple bits of information per element and the low energy required to write distinct states. In this paper, we first review the neuro- and neuromorphic computing approaches that can best exploit the properties of memristor and scale devices, and then propose a novel hybrid memristor-CMOS neuromorphic circuit which represents a radical departure from conventional neuro-computing approaches, as it uses memristors to directly emulate the biophysics and temporal dynamics of real synapses. We point out the differences between the use of memristors in conventional neuro-computing architectures and the hybrid memristor-CMOS circuit proposed, and argue how this circuit represents an ideal building block for implementing brain-inspired probabilistic computing paradigms that are robust to variability and fault tolerant by design.}
}
@inproceedings{IndykandMotwaniSTOC-98,
       author = {Indyk, Piotr and Motwani, Rajeev},
    booktitle = {Proceedings of the 30th Annual {ACM} Symposium on Theory of Computing},
        title = {Approximate Nearest Neighbors: Towards Removing the Curse of Dimensionality},
         year = 1998,
        pages = {604-613},
}
@book{IngramDHARMA-08,
        title = {Daniel Ingram},
       author = {Mastering the Core Teachings of the Buddha, An Unusually Hardcore Dharma Book},
         year = {Aeon Books},
    publisher = {2008},
}
@article{IngramGENETICS-03,
       author = {Ingram, Vernon M.},
        title = {Sickle-cell anemia hemoglobin: the molecular biology of the first "molecular disease"--the crucial importance of serendipity.},
      journal = {Genetics},
       volume = {167},
        issue = {1},
         year = {2004},
        pages = {1-7},
}
@article{IntratorandEdelmanNETWORK-97,
       author = {Nathan Intrator and Shimon Edelman},
        title = {Learning low-dimensional representations via the usage of multiple-class labels},
      journal = {Network: Computation in Neural Systems},
       volume = 8,
        pages = {259-281},
         year = 1997,
}
@incollection{IntraubFLEETING-99,
        title = {Understanding and Remembering Briefly Glimpsed Pictures: Implications for Visual Scanning and Memory},
       author = {Helene Intraub},
    booktitle = {Fleeting Memories: Cognition of Brief Visual Stimuli},
    publisher = {MIT Press},
      address = {Cambridge, MA},
       editor = {Veronika Coltheart},
         year = 1999,
        pages = {47-70},
     abstract = {The investigation of what people understand and remember from rapidly presented sequences of visual stimuli began in the late 1960s. In this book prominent researchers approach the topic from psychological, neuropsychological, and electrophysiological perspectives. Specific issues include RSVP (rapid serial visual presentation), attentional blink, repetition blindness, and scene perception. The contributors review recent research on our ability to comprehend and remember pictures of objects and scenes, written words, and sentences when the visual stimuli are presented sequentially at rates of up to ten items per second. In short, the book is about our remarkably developed abilities to understand and remember the contents of very briefly presented material.},
}
@article{IoannidisPLoS-05,
       author = {Ioannidis, John P. A.},
      journal = {{PLOS} Medicine},
    publisher = {Public Library of Science},
        title = {Why Most Published Research Findings Are False},
         year = {2005},
       volume = {2},
       number = {8},
     abstract = {Published research findings are sometimes refuted by subsequent evidence, says Ioannidis, with ensuing confusion and disappointment.},
}
@article{IrleandMarkowitschNEUROSCIENCE-82,
       author = {Irle, E. and Markowitsch, H. J.},
        title = {Widespread cortical projections of the hippocampal formation in the cat},
      journal = {Neuroscience},
         year = {1982},
       volume = {7},
        issue = {11},
        pages = {2637-2647},
     abstract = {Efferent projections from the hippocampal formation to the cat's cortex were traced with the retrograde horseradish peroxidase technique. Different areas of the cortex of 31 cats were injected with small amounts of horseradish peroxidase. All subregions of the hippocampal formation were screened for labeled cells. It was found that, with the exception of the entorhinal injections, only subicular areas of the hippocampal formation contain labeled neurons. When HRP was injected into the entorhinal cortex, labeled cells are also found in the hippocampus proper. The most dense projection from the subicular cortex is directed to the medial part of the cortical hemisphere. Here, cingulate, retrosplenial and medial prefrontal fields receive a substantial number of subicular efferents. Furthermore, the entorhinal cortex is reached by a number of axons originating in the subicular area. Scarce projections from the subicular cortex terminate in the dorsal prefrontal, temporal, parietal and prepiriform cortex. It is suggested that the projection from the subicular cortex to the neocortical areas of the frontal pole (medial prefrontal cortex) is of special importance as it may constitute a link between the association areas of the neocortex and those regions of the limbic system thought to play a role in memory (subicular cortex, mamillary bodies, anterior thalamus, cingulate gyrus).},
}
@inproceedings{IsardCVP-03,
       author = {Isard, M.},
        title = {{PAMPAS}: real-valued graphical models for computer vision},
    booktitle = {Proceedings of the 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition: Volume {I}},
        pages = {613-620},
         year = 2003,
          url = {IsardPAMPAS.pdf},
}
@article{IsardandBlakeIJCV-98,
       author = {Isard, M. and Blake, A.},
        title = {{CONDENSATION} -- conditional density propagation for visual tracking},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 29,
        issue = 1,
        pages = {5-28},
         year = 1998,
}
@article{IsiketalJoN-14,
       author = {Leyla Isik and Ethan M. Meyers and Joel Z. Leibo and Tomaso Poggio},
        title = {The dynamics of invariant object recognition in the human visual system},
      journal = {Journal of Neurophysiology},
       volume = 111,
        issue = 1,
        pages = {91-102},
         year = 2014,
     abstract = {The human visual system can rapidly recognize objects despite transformations that alter their appearance. The precise timing of when the brain computes neural representations that are invariant to particular transformations, however, has not been mapped in humans. Here we employ magnetoencephalography decoding analysis to measure the dynamics of size- and position-invariant visual information development in the ventral visual stream. With this method we can read out the identity of objects beginning as early as 60 ms. Size- and position-invariant visual information appear around 125 ms and 150 ms, respectively, and both develop in stages, with invariance to smaller transformations arising before invariance to larger transformations. Additionally, the magnetoencephalography sensor activity localizes to neural sources that are in the most posterior occipital regions at the early decoding times and then move temporally as invariant information develops. These results provide previously unknown latencies for key stages of human-invariant object recognition, as well as new and compelling evidence for a feed-forward hierarchical model of invariant object recognition where invariance increases at each successive visual area along the ventral stream}
}
@article {IssaetalBIORXIV-16,
       author = {Issa, Elias B. and Cadieu, Charles F. and DiCarlo, James J.},
        title = {Evidence that the ventral stream codes the errors used in hierarchical inference and learning},
      journal = {bioRxiv},
         year = {2016},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {Hierarchical feedforward processing makes object identity explicit at the highest stages of the ventral visual stream. We leveraged this computational goal to study the fine-scale temporal dynamics of neural populations in posterior and anterior inferior temporal cortex (pIT, aIT) during face detection. As expected, we found that a neural spiking preference for natural over distorted face images was rapidly produced, first in pIT and then in aIT. Strikingly, in the next 30 milliseconds of processing, this pattern of selectivity in pIT completely reversed, while selectivity in aIT remained unchanged. Although these dynamics were difficult to explain from a pure feedforward perspective, a model class computing errors through feedback closely matched the observed neural dynamics and parsimoniously explained a range of seemingly disparate IT neural response phenomena. This new perspective augments the standard model of online vision by suggesting that neural signals of states (e.g. likelihood of a face being present) are intermixed with the error signals found in deep hierarchical networks.},
}
@book{Ito2012,
        title = {The Cerebellum: Brain for an Implicit Self},
       author = {Ito, Masao},
    publisher = {Financial Times Press},
         year = {2012},
}
@book{Ito2012cerebellum,
        title = {The Cerebellum: Brain for an Implicit Self},
       author = {Ito, Masao},
    publisher = {Financial Times Press},
         year = {2012},
}
@incollection{ItoCEREBELLAR-CORTEX-18,
       author = {Ito, Masao},
        title = {Olivocerebellar System},
    booktitle = {Handbook of Brain Microcircuits},
       editor = {Shepherd, G.M. and Grillner, S.},
    publisher = {Oxford University Press},
         year = {2018},
        pages = {293-299}
}
@article{ItoNRN-08,
       author = {Ito, Masao},
        title = {Control of mental activities by internal models in the cerebellum},
      journal = {Nature Reviews Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {9},
        issue = {4},
         year = 2008,
        pages = {304-313},
     abstract = {The intricate neuronal circuitry of the cerebellum is thought to encode internal models that reproduce the dynamic properties of body parts. These models are essential for controlling the movement of these body parts: they allow the brain to precisely control the movement without the need for sensory feedback. It is thought that the cerebellum might also encode internal models that reproduce the essential properties of mental representations in the cerebral cortex. This hypothesis suggests a possible mechanism by which intuition and implicit thought might function and explains some of the symptoms that are exhibited by psychiatric patients. This article examines the conceptual bases and experimental evidence for this hypothesis.},
}
@article{ItoPiN-06,
       author = {Ito, M.},
        title = {Cerebellar circuitry as a neuronal machine},
      journal = {Progress in Neurobiology},
       volume = {78},
       number = {3-5},
         year = {2006},
        pages = {272-303},
     abstract = {Shortly after John Eccles completed his studies of synaptic inhibition in the spinal cord, for which he was awarded the 1963 Nobel Prize in physiology/medicine, he opened another chapter of neuroscience with his work on the cerebellum. From 1963 to 1967, Eccles and his colleagues in Canberra successfully dissected the complex neuronal circuitry in the cerebellar cortex. In the 1967 monograph, "The Cerebellum as a Neuronal Machine", he, in collaboration with Masao Ito and Janos Szent{\`{a}}gothai, presented blue-print-like wiring diagrams of the cerebellar neuronal circuitry. These stimulated worldwide discussions and experimentation on the potential operational mechanisms of the circuitry and spurred theoreticians to develop relevant network models of the machinelike function of the cerebellum. In following decades, the neuronal machine concept of the cerebellum was strengthened by additional knowledge of the modular organization of its structure and memory mechanism, the latter in the form of synaptic plasticity, in particular, long-term depression. Moreover, several types of motor control were established as model systems representing learning mechanisms of the cerebellum. More recently, both the quantitative preciseness of cerebellar analyses and overall knowledge about the cerebellum have advanced considerably at the cellular and molecular levels of analysis. Cerebellar circuitry now includes Lugaro cells and unipolar brush cells as additional unique elements. Other new revelations include the operation of the complex glomerulus structure, intricate signal transduction for synaptic plasticity, silent synapses, irregularity of spike discharges, temporal fidelity of synaptic activation, rhythm generators, a Golgi cell clock circuit, and sensory or motor representation by mossy fibers and climbing fibers. Furthermore, it has become evident that the cerebellum has cognitive functions, and probably also emotion, as well as better-known motor and autonomic functions. Further cerebellar research is required for full understanding of the cerebellum as a broad learning machine for neural control of these functions.}
}
@article{ItoetalNEURON-14,
       author = {Ito, K. and Shinomiya, K. and Ito, M. and Armstrong, J. D. and Boyan, G. and Hartenstein, V. and Harzsch, S. and Heisenberg, M. and Homberg, U. and Jenett, A. and Keshishian, H. and Restifo, L. L. and Rossler, W. and Simpson, J. H. and Strausfeld, N. J. and Strauss, R. and Vosshall, L. B.},
        title = {{{A} systematic nomenclature for the insect brain}},
      journal = {Neuron},
         year = {2014},
       volume = {81},
       number = {4},
        pages = {755-765},
     abstract = {Despite the importance of the insect nervous system for functional and developmental neuroscience, descriptions of insect brains have suffered from a lack of uniform nomenclature. Ambiguous definitions of brain regions and fiber bundles have contributed to the variation of names used to describe the same structure. The lack of clearly determined neuropil boundaries has made it difficult to document precise locations of neuronal projections for connectomics study. To address such issues, a consortium of neurobiologists studying arthropod brains, the Insect Brain Name Working Group, has established the present hierarchical nomenclature system, using the brain of Drosophila melanogaster as the reference framework, while taking the brains of other taxa into careful consideration for maximum consistency and expandability. The following summarizes the consortium's nomenclature system and highlights examples of existing ambiguities and remedies for them. This nomenclature is intended to serve as a standard of reference for the study of the brain of Drosophila and other insects.}
}
@article{ItohetalNATURE-04,
        title = {Mechanically driven {ATP} synthesis by {F}$_1$-{ATP}ase},
       author = {Hiroyasu Itoh and Akira Takahashi and Kengo Adachi and Hiroyuki Noji and Ryohei Yasuda and Masasuke Yoshida and Kazuhiko Kinosita Jr.},
      journal = {Nature},
       volume = 427,
         year = 2004,
        pages = {465-468}
}
@inproceedings{IttiSPIE-02,
       author = {L. Itti},
        title = {Toward highly capable neuromorphic autonomous robots: beobots},
    booktitle = {Proceedings of {SPIE 47}: Annual International Symposium on Optical Science and Technology},
       volume = 4787,
        pages = {37-45},
         year = 2002,
}
@inproceedings{IttiandBaldiCVPR-05,
       author = {L. Itti and P. Baldi},
        title = {A Principled Approach to Detecting Surprising Events in Video},
    booktitle = {Proceedings {IEEE} Conference on Computer Vision and Pattern Recognition ({CVPR})},
      address = {San Siego, CA},
         year = 2005,
        pages = {631-637},
     abstract = {Primates demonstrate unparalleled ability at rapidly orienting towards important events in complex dynamic environments. During rapid guidance of attention and gaze towards potential objects of interest or potential threats, however, often there is no time for detailed visual analysis. Thus, heuristic computations are necessary to locate the most interesting events in quasi real-time. We present a new theory of sensory surprise, which provides a principled and computable shortcut to important information. We develop a model that computes instantaneous low-level surprise at every location in video streams. The algorithm significantly correlates with eye movements of two human observers watching complex video clips, including television programs (17,936 frames, 2,152 saccadic gaze shifts). The resulting system allows more sophisticated and time-consuming image analysis to be efficiently focused onto the most surprising subsets of the incoming data.},
}
@article{IttietalPAMI-98,
       author = {L. Itti and C. Koch and E. Niebur},
        title = {A Model of Saliency-Based Visual Attention for Rapid Scene Analysis},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
       volume = 20,
        issue = 11,
        pages = {1254-1259},
         year = 1998,
}
@article{Iturria-MedinaetalNEUROIMAGE-08,
       author = {Iturria-Medina, Y. and Sotero, R. C. and Canales-Rodriguez, E. J. and Aleman-Gomez, Y. and Melie-Garcia, L.},
        title = {{S}tudying the human brain anatomical network via diffusion-weighted {M}{R}{I} and {G}raph {T}heory},
      journal = {Neuroimage},
         year = {2008},
       volume = {40},
       number = {3},
        pages = {1064-1076},
     abstract = {Our goal is to study the human brain anatomical network. For this, the anatomical connection probabilities (ACP) between 90 cortical and subcortical brain gray matter areas are estimated from diffusion-weighted Magnetic Resonance Imaging (DW-MRI) techniques. The ACP between any two areas gives the probability that those areas are connected at least by a single nervous fiber. Then, the brain is modeled as a non-directed weighted graph with continuous arc weights given by the ACP matrix. Based on this approach, complex networks properties such as small-world attributes, efficiency, degree distribution, vulnerability, betweenness centrality and motifs composition are studied. The analysis was carried out for 20 right-handed healthy subjects (mean age: 31.10, S.D.: 7.43). According to the results, all networks have small-world and broad-scale characteristics. Additionally, human brain anatomical networks present bigger local efficiency and smaller global efficiency than the corresponding random networks. In a vulnerability and betweenness centrality analysis, the most indispensable and critical anatomical areas were identified: putamens, precuneus, insulas, superior parietals and superior frontals. Interestingly, some areas have a negative vulnerability (e.g. superior temporal poles, pallidums, supramarginals and hechls), which suggest that even at the cost of losing in global anatomical efficiency, these structures were maintained through the evolutionary processes due to their important functions. Finally, symmetrical characteristic building blocks (motifs) of size 3 and 4 were calculated, obtaining that motifs of size 4 are the expanded version of motif of size 3. These results are in agreement with previous anatomical studies in the cat and macaque cerebral cortex.}
}
@incollection{IuculanoetalHFNC-18,
       author = {Teresa Iuculano and Aarthi Padmanabhan and Vinod Menon},
        title = {Systems Neuroscience of Mathematical Cognition and Learning: Basic Organization and Neural Sources of Heterogeneity in Typical and Atypical Development},
    booktitle = {Heterogeneity of Function in Numerical Cognition},
    publisher = {Academic Press},
         year = {2018},
        pages = {287-336}
}
@article{IversenandMishkinEBR-70,
       author = {Iversen, S. D. and Mishkin, M.},
        title = {Perseverative interference in monkeys following selective lesions of the inferior prefrontal convexity},
      journal = {Experimental Brain Research},
         year = {1970},
       volume = {11},
       number = {4},
        pages = {376-386},
     abstract = {Monkeys with lesions of the inferior frontal convexity were impaired relative to controls in retaining an auditory frequency differentiation (although subsequent thresholds were normal) and in learning object and spatial reversals. Performance was characterized by perseverative interference, a frontal symptom which now seems attributable to damage to the inferior convexity.}
}
@article{IversonandZuckerPAMI-95,
       author = {Lee A. Iverson and Steven W. Zucker},
        title = {Logical / Linear Operators for Image Curves},
      journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
       volume = 17,
        issue = 10,
         year = 1995,
        pages = {982-996},
}
@article{IwamuroetalEJN-17,
       author = {Iwamuro, H. and Tachibana, Y. and Ugawa, Y. and Saito, N. and Nambu, A.},
        title = {{I}nformation processing from the motor cortices to the subthalamic nucleus and globus pallidus and their somatotopic organizations revealed electrophysiologically in monkeys},
      journal = {European Journal Neuroscience},
       volume = {46},
       number = {11},
        pages = {2684-2701},
         year = {2017},
     abstract = {To understand how the information derived from different motor cortical areas representing different body parts is organized in the basal ganglia, we examined the neuronal responses in the subthalamic nucleus (STN), and the external (GPe) and internal (GPi) segments of the globus pallidus (input, relay and output nuclei, respectively) to stimulation of the orofacial, forelimb and hindlimb regions of the primary motor cortex (MI) and supplementary motor area (SMA) in macaque monkeys under the awake state. Most STN and GPe/GPi neurons responded exclusively to stimulation of either the MI or SMA, and one-fourth to one-third of neurons responded to both. STN neurons responding to the hindlimb, forelimb and orofacial regions of the MI were located along the medial-lateral axis in the posterolateral STN, while neurons responding to the orofacial region of the SMA were located more medially than the others in the anteromedial STN. GPe/GPi neurons responding to the hindlimb, forelimb and orofacial regions of the MI were found along the dorsal-ventral axis in the posterolateral GPe/GPi, and neurons responding to the corresponding regions of the SMA were similarly but less clearly distributed in more anteromedial regions. Moreover, neurons responding to the distal and proximal forelimb MI regions were found along the lateral-medial axis in the STN and the ventral-dorsal axis in the GPe/GPi. Most STN and GPe/GPi neurons showed kinaesthetic responses with similar somatotopic maps. These observations suggest that the somatotopically organized inputs from the MI and SMA are well preserved in the STN and GPe/GPi with partial convergence.}
}
@article{IwasakiAIJ-86,
        title = {Causality in Device Behavior},
       author = {Iwasaki, Yumi and Simon, Herbert A.},
      journal = AIJ,
       volume = 29,
         year = 1986,
        pages = {3-32},
}
@article{IyeretalPLoS-13,
       author = {Iyer, Ramakrishnan and Menon, Vilas and Buice, Michael and Koch, Christof and Mihalas, Stefan},
      journal = {PLoS Computational Biology},
    publisher = {Public Library of Science},
        title = {The Influence of Synaptic Weight Distribution on Neuronal Population Dynamics},
         year = {2013},
       volume = {9},
        pages = {1-16},
     abstract = {Neurons communicate via action potentials. Typically, depolarizations caused by presynaptic firing are small, such that many synaptic inputs are necessary to exceed the firing threshold. This is the assumption made by standard mathematical approaches such as the Fokker-Planck formalism. However, in some cases the synaptic weight can be large. On occasion, a single input is capable of exceeding threshold. Although this phenomenon can be studied with computational simulations, these can be impractical for large scale brain simulations or suffer from the problem of insufficient knowledge of the relevant parameters. Improving upon the standard Fokker-Planck approach, we develop a hybrid approach combining semi-analytical with computational methods into an efficient technique for analyzing the effect that rare and large synaptic weights can have on neural network activity. Our method has both neurobiological as well as methodological implications. Sparse but powerful synapses provide networks with response celerity, enhanced bandwidth and stability, even when the networks are matched for average input. We introduce a measure characterizing this response. Furthermore, our method can characterize the sub-threshold membrane potential distribution and spiking statistics of very large networks of distinct but homogeneous populations of 10s to 100s of distinct neuronal cell types throughout the brain.},
}
@book{Izhikevich2007,
       author = {Eugene M. Izhikevich},
        title = {Dynamical Systems in Neuroscience: The Geometry of Excitability and Bursting},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2007
}
@article{IzhikevichCC-07,
       author = {Izhikevich, Eugene. M.},
        title = {Solving the distal reward problem through linkage of {STDP} and dopamine signaling},
      journal = {Cerebral Cortex},
         year = {2007},
       volume = {17},
       number = {10},
        pages = {2443-2452},
     abstract = {In Pavlovian and instrumental conditioning, reward typically comes seconds after reward-triggering actions, creating an explanatory conundrum known as "distal reward problem": How does the brain know what firing patterns of what neurons are responsible for the reward if 1) the patterns are no longer there when the reward arrives and 2) all neurons and synapses are active during the waiting period to the reward? Here, we show how the conundrum is resolved by a model network of cortical spiking neurons with spike-timing-dependent plasticity (STDP) modulated by dopamine (DA). Although STDP is triggered by nearly coincident firing patterns on a millisecond timescale, slow kinetics of subsequent synaptic plasticity is sensitive to changes in the extracellular DA concentration during the critical period of a few seconds. Random firings during the waiting period to the reward do not affect STDP and hence make the network insensitive to the ongoing activity-the key feature that distinguishes our approach from previous theoretical studies, which implicitly assume that the network be quiet during the waiting period or that the patterns be preserved until the reward arrives. This study emphasizes the importance of precise firing patterns in brain dynamics and suggests how a global diffusive reinforcement signal in the form of extracellular DA can selectively influence the right synapses at the right time.}
}
@article{IzhikevichNN-00,
       author = {Eugene M Izhikevich},
        title = {Computing with oscillators},
      journal = {Unpublished},
         year = 2000,
     abstract = {We study neuro-computational properties of non-linear oscillatory systems. Since we use the canonical phase-models approach, our results do not depend on the physical nature of each oscillator or the detailed form of mathematical equations that are used to describe its dynamics. In particular, we show that anything that can oscillate can also compute, the only problem is how to couple the oscillators. We apply this theory to detailed models of electrical, optical, and mechanical oscillators and we investigate the possibility to use such oscillators to build an oscillatory neurocomputer having autocorrelative associative memory. It stores and retrieves complex oscillatory patterns as synchronized states having appropriate phase relations between neurons.}
}
@article{IzhikevichandEdelmanPNAS-08,
       author = {Eugene M. Izhikevich and Gerald M. Edelman},
        title = {Large-scale model of mammalian thalamo-cortical systems},
      journal = {Proceedings of the National Academy of Science},
       volume = 105,
        issue = 9,
        pages = {3593-3598},
         year = 2008,
}
@article{IzhikevichandFitzHughSCHOLARPEDIA-06,
       author = {Izhikevich E.M. and FitzHugh R.},
        title = {Fitz{H}ugh-{N}agumo Model},
      journal = {Scholarpedia},
       volume = {1},
        issue = {9},
         year = {2006},
        pages = {1349},
}        
@article{IzquierdoandBeerCOIN-16,
        title = {The whole worm: brain-body-environment models of {\it{Caenorhabditis elegans}}},
       author = {Eduardo J Izquierdo and Randall D Beer},
      journal = {Current Opinion in Neurobiology},
       volume = {40},
         year = {2016},
        pages = {14-22},
     abstract = {Brain, body and environment are in continuous dynamical interaction, and it is becoming increasingly clear that an animal's behavior must be understood as a product not only of its nervous system, but also of the ongoing feedback of this neural activity through the biomechanics of its body and the ecology of its environment. Modeling has an essential integrative role to play in such an understanding. But successful whole-animal modeling requires an animal for which detailed behavioral, biomechanical and neural information is available and a modeling methodology which can gracefully cope with the constantly changing balance of known and unknown biological constraints. Here we review recent progress on both optogenetic techniques for imaging and manipulating neural activity and neuromechanical modeling in the nematode worm Caenorhabditis elegans. This work demonstrates both the feasibility and challenges of whole-animal modeling.}
}        
@inproceedings{JaakkolaandHausslerNIPS-98,
       author = {Jaakkola, T. and Haussler, D.},
        title = {Exploiting generative models in discriminative classifiers},
    booktitle = {Advances in Neural Information Processing Systems 11},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1998,
}
@inproceedings{JaakkolaandJordanAIS-96,
       author = {Tommi Jaakkola and Michael Jordan},
        title = {A variational approach to Bayesian logistic regression problems and their extensions},
    booktitle = {Proceedings of the sixth international workshop on artificial intelligence and statistics},
         year = 1996,
}
@article{JabesetalJoCN-11,
       author = {Jab\`{e}s, Adeline and Lavenex, Pamela Banta and Amaral, David G. and Lavenex, Pierre},
        title = {Postnatal development of the hippocampal formation: a stereological study in macaque monkeys},
      journal = {The Journal of comparative neurology},
         year = {2011},
       volume = {519},
        issue = {6},
        pages = {1051-1070},
     abstract = {We performed a stereological analysis of neuron number, neuronal soma size, and volume of individual regions and layers of the macaque monkey hippocampal formation during early postnatal development. We found a protracted period of neuron addition in the dentate gyrus throughout the first postnatal year and a concomitant late maturation of the granule cell population and individual dentate gyrus layers that extended beyond the first year of life. Although the development of CA3 generally paralleled that of the dentate gyrus, the distal portion of CA3, which receives direct entorhinal cortex projections, matured earlier than the proximal portion of CA3. CA1 matured earlier than the dentate gyrus and CA3. Interestingly, CA1 stratum lacunosum-moleculare, in which direct entorhinal cortex projections terminate, matured earlier than CA1 strata oriens, pyramidale, and radiatum, in which the CA3 projections terminate. The subiculum developed earlier than the dentate gyrus, CA3, and CA1, but not CA2. However, similarly to CA1, the molecular layer of the subiculum, in which the entorhinal cortex projections terminate, was overall more mature in the first postnatal year compared with the stratum pyramidale in which most of the CA1 projections terminate. Unlike other hippocampal fields, volumetric measurements suggested regressive events in the structural maturation of presubicular neurons and circuits. Finally, areal and neuron soma size measurements revealed an early maturation of the parasubiculum. We discuss the functional implications of the differential development of distinct hippocampal circuits for the emergence and maturation of different types of ``hippocampus-dependent'' memory processes, including spatial and episodic memories.},
}
@book{Jackendoff02,
       author = {Ray Jackendoff},
        title = {Foundations of Language: Brain, Meaning, Grammar, Evolution},
    publisher = {Oxford University Press},
      address = {Oxford, UK},
         year = 2002,
}
@article{JackendoffBBS-03,
        title = {Pr\'{e}cis of Foundations of Language: Brain, Meaning, Grammar, Evolution},
       author = {Jackendoff, Ray},
      journal = {Behavioral and Brain Sciences},
       volume = {26},
       number = {6},
    publisher = {Cambridge University Press},
         year = {2003},
        pages = {651-665},
     abstract = {The goal of this study is to reintegrate the theory of generative grammar into the cognitive sciences. Generative grammar was right to focus on the child's acquisition of language as its central problem, leading to the hypothesis of an innate Universal Grammar. However, generative grammar was mistaken in assuming that the syntactic component is the sole course of combinatoriality, and that everything else is "interpretive." The proper approach is a parallel architecture, in which phonology, syntax, and semantics are autonomous generative systems linked by interface components. The parallel architecture leads to an integration within linguistics, and to a far better integration with the rest of cognitive neuroscience. It fits naturally into the larger architecture of the mind/brain and permits a properly mentalistic theory of semantics. It results in a view of linguistic performance in which the rules of grammar are directly involved in processing. Finally, it leads to a natural account of the incremental evolution of the language capacity. Comments and the author's reply are included.}
}
@book{Jackson1958selected,
        title = {Selected Writings of John Hughlings Jackson: Evolution and dissolution of the nervous system},
       author = {Jackson, John Hughlings},
       series = {Selected Writings of John Hughlings Jackson},
         year = {1958},
    publisher = {Basic Books}
}
@book{Jackson90,
       author = {Jackson, Peter},
        title = {Introduction to Expert Systems},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1990,
}
@article{JacksonetalNEUROPSYCHOLOGIA-06,
       author = {Philip L. Jackson and Eric Brunet and Andrew N. Meltzoff and Jean Decety},
        title = {Empathy examined through the neural mechanisms involved in imagining how I feel versus how you feel pain},
      journal = {Neuropsychologia},
       volume = {44},
       number = {5},
        pages = {752-761},
         year = {2006},
     abstract = {Perspective-taking is a stepping stone to human empathy. When empathizing with another individual, one can imagine how the other perceives the situation and feels as a result. To what extent does imagining the other differs from imagining oneself in similar painful situations? In this functional magnetic resonance imaging experiment, participants were shown pictures of people with their hands or feet in painful or non-painful situations and instructed to imagine and rate the level of pain perceived from different perspectives. Both the Self's and the Other's perspectives were associated with activation in the neural network involved in pain processing, including the parietal operculum, anterior cingulate cortex (ACC; BA32) and anterior insula. However, the Self-perspective yielded higher pain ratings and involved the pain matrix more extensively in the secondary somatosensory cortex, the ACC (BA 24a′/24b′), and the insula proper. Adopting the perspective of the Other was associated with specific increase in the posterior cingulate/precuneus and the right temporo-parietal junction. These results show the similarities between Self- and Other-pain representation, but most interestingly they also highlight some distinctiveness between these two representations, which is a crucial aspect of human empathy. It may be what allows us to distinguish empathic responses to others versus our own personal distress. These findings are consistent with the view that empathy does not involve a complete Self-Other merging.}
}
@inproceedings{JacobsetalSIGGRAPH-95,
       author = {Charles E. Jacobs and Adam Finkelstein and David H. Salesin},
        title = {Fast Multiresolution Image Querying},
    booktitle = {Proceedings of {SIGGRAPH}},
        pages = {277-286},
         year = 1995,
}
@article{JaderbergetalCoRR-16,
       author = {Max Jaderberg and Volodymyr Mnih and Wojciech Marian Czarnecki and Tom Schaul and Joel Z. Leibo and David Silver and Koray Kavukcuoglu},
        title = {Reinforcement Learning with Unsupervised Auxiliary Tasks},
      journal = {CoRR},
       volume = {arXiv:1611.05397},
         year = {2016},
     abstract = {Deep reinforcement learning agents have achieved state-of-the-art results by directly maximising cumulative reward. However, environments contain a much wider variety of possible training signals. In this paper, we introduce an agent that also maximises many other pseudo-reward functions simultaneously by reinforcement learning. All of these tasks share a common representation that, like unsupervised learning, continues to develop in the absence of extrinsic rewards. We also introduce a novel mechanism for focusing this representation upon extrinsic rewards, so that learning can rapidly adapt to the most relevant aspects of the actual task. Our agent significantly outperforms the previous state-of-the-art on Atari, averaging 880\% expert human performance, and a challenging suite of first-person, three-dimensional \emph{Labyrinth} tasks leading to a mean speedup in learning of 10\% and averaging 87\% expert human performance on Labyrinth.}
}
@article{JaderbergetalCoRR-16b,
       author = {Max Jaderberg and Wojciech Marian Czarnecki and Simon Osindero and Oriol Vinyals and Alex Graves and Koray Kavukcuoglu},
        title = {Decoupled Neural Interfaces using Synthetic Gradients},
      journal = {CoRR},
       volume = {arXiv:1608.05343},
         year = {2016},
     abstract = {Training directed neural networks typically requires forward-propagating data through a computation graph, followed by backpropagating error signal, to produce weight updates. All layers, or more generally, modules, of the network are therefore locked, in the sense that they must wait for the remainder of the network to execute forwards and propagate error backwards before they can be updated. In this work we break this constraint by decoupling modules by introducing a model of the future computation of the network graph. These models predict what the result of the modelled subgraph will produce using only local information. In particular we focus on modelling error gradients: by using the modelled synthetic gradient in place of true backpropagated error gradients we decouple subgraphs, and can update them independently and asynchronously i.e. we realise decoupled neural interfaces. We show results for feed-forward models, where every layer is trained asynchronously, recurrent neural networks (RNNs) where predicting one's future gradient extends the time over which the RNN can effectively model, and also a hierarchical RNN system with ticking at different timescales. Finally, we demonstrate that in addition to predicting gradients, the same framework can be used to predict inputs, resulting in models which are decoupled in both the forward and backwards pass -- amounting to independent networks which co-learn such that they can be composed into a single functioning corporation.}
}
@article{JaderbergetalICLR-17,
       author = {Max Jaderberg and Volodymyr Mnih and Wojciech Marian Czarnecki and Tom Schaul and Joel Z. Leibo and David Silver and Koray Kavukcuoglu},
        title = {Reinforcement Learning with Unsupervised Auxiliary Tasks},
      journal = {International Conference on Learning Representations},
         year = {2017},
     abstract = {Deep reinforcement learning agents have achieved state-of-the-art results by directly maximising cumulative reward. However, environments contain a much wider variety of possible training signals. In this paper, we introduce an agent that also learns separate policies for maximising many other pseudo-reward functions simultaneously by reinforcement learning. All of these tasks share a common representation that, like unsupervised learning, continues to develop in the absence of extrinsic rewards. We also introduce a novel mechanism for focusing this representation upon extrinsic rewards, so that learning can rapidly adapt to the most relevant aspects of the actual task. Our agent significantly outperforms the previous state-of-the-art on Atari, averaging 880\% expert human performance, and a challenging suite of first-person, three-dimensional Labyrinth tasks leading to a mean speedup in learning of 10\% and averaging 87\% expert human performance on Labyrinth.},
}
@article{JaegerCoRR-14a,
       author = {Herbert Jaeger},
        title = {Controlling Recurrent Neural Networks by Conceptors},
      journal = {CoRR},
       volume = {arXiv:1403.3369},
         year = {2014},
      comment = {Technical Report No. 31, Jacobs University, Bremem, School of Engineering and Science, March 2014},
     abstract = {The human brain is a dynamical system whose extremely complex sensor driven neural processes give rise to conceptual, logical cognition. Understanding the interplay between nonlinear neural dynamics and concept-level cognition remains a major scientific challenge. Here I propose a mechanism of neurodynamical organization, called conceptors, which unites nonlinear dynamics with basic principles of conceptual abstraction and logic. It becomes possible to learn, store, abstract, focus, morph, generalize, de-noise and recognize a large number of dynamical patterns within a single neural system; novel patterns can be added without interfering with previously acquired ones; neural noise is automatically filtered. Conceptors help explaining how conceptual-level information processing emerges naturally and robustly in neural systems, and remove a number of roadblocks in the theory and applications of recurrent neural networks.},
}
@article{JaegerCoRR-14b,
       author = {Herbert Jaeger},
        title = {Conceptors: an easy introduction},
      journal = {CoRR},
       volume = {arXiv:1406.2671},
         year = {2014},
     abstract = {Conceptors provide an elementary neuro-computational mechanism which sheds a fresh and unifying light on a diversity of cognitive phenomena. A number of demanding learning and processing tasks can be solved with unprecedented ease, robustness and accuracy. Some of these tasks were impossible to solve before. This entirely informalpaper introduces the basic principles of conceptors and highlights some of their usages.},
}
@incollection{JaegerNIPS-02,
        title = {Adaptive Nonlinear System Identification with Echo State Networks},
       author = {Herbert Jaeger},
    booktitle = {Advances in Neural Information Processing Systems 15},
       editor = {S. Becker and S. Thrun and K. Obermayer},
    publisher = {MIT Press},
         year = {2003},
        pages = {609-616},
     abstract = {Echo state networks (ESN) are a novel approach to recurrent neural network training. An ESN consists of a large, fixed, recurrent "reservoir" network, from which the desired output is obtained by training suitable output connection weights. Determination of optimal output weights becomes a linear, uniquely solvable task of MSE minimization. This article reviews the basic ideas and describes an online adaptation scheme based on the RLS algorithm known from adaptive linear systems. As an example, a 10-th order NARMA system is adaptively identified. The known benefits of the RLS algorithms carryover from linear systems to nonlinear ones; specifically, the convergence rate and misadjustment can be determined at design time.},
}
@article{JaegerandHaasSCIENCE-04,
       author = {Jaeger, H. and Haas, H.},
        title = {Harnessing nonlinearity: predicting chaotic systems and saving energy in wireless communication},
      journal = {Science},
         year = {2004},
       volume = {304},
       number = {5667},
        pages = {78-80},
     abstract = {We present a method for learning nonlinear systems, echo state networks (ESNs). ESNs employ artificial recurrent neural networks in a way that has recently been proposed independently as a learning mechanism in biological brains. The learning method is computationally efficient and easy to use. On a benchmark task of predicting a chaotic time series, accuracy is improved by a factor of 2400 over previous techniques. The potential for engineering applications is illustrated by equalizing a communication channel, where the signal error rate is improved by two orders of magnitude.}
}
@techreport{JailletTR-85,
       author = {Jaillet, P.},
        title = {Probabilistic Traveling Salesman Problems},
        issue = {TR-185},
  institution = {Operations Research Center, MIT},
         year = 1985,
}
@article{JainCoRR-17,
       author = {Viren Jain},
        title = {Adversarial Image Alignment and Interpolation},
      journal = {CoRR},
       volume = {arXiv:1707.00067},
         year = {2017},
}
@article{JainHHMI_MPI_CONNECTOMICS-19,
        title = {{HHMI/MPI} Connectomics Meeting},
       author = {Viren Jain},
     location = {Berlin},
         year = {2019},
     abstract = {Remarkable demonstration of Google (i.e., Neuromancer's) influence on the field. I counted something like seventeen talks that mentioned results / tools / collaborations related to Neuromancer. Jeff Lichtman said to me privately that this level of saturation "shows how Google's involvement has completely revolutionized the field". Many labs are using FAFB to trace their (drosophila) neurons of interest, and there is tremendous excitement about Peter's results and our recent preprint (which, by the way, has been downloaded about 800 times in just a few days). There's curiosity about FlyEM's hemibrain project but since it's not going to be public for roughly another year, people are not sitting around waiting and instead are aggressively pursuing what they can in FAFB. Sebastian also showed off a nice segmentation of a realigned version of FAFB that they are planning to release (sometime in the summer, if I understand correctly, but I'm not sure). Taken together, these observations suggest that for some time to come FAFB will have huge impact on drosophila connectomics and therefore we should triple-check we've done all we can on the current segmentation before finalizing/releasing it, and perhaps even consider follow-up efforts beyond that. A general lesson of FAFB is that getting interesting data into peoples hands sooner rather than later is a hugely powerful thing, and Jeff and I agreed that this should inform the approach to the human dataset (i.e., trying to release something relatively early, rather than waiting for paper acceptance etc). Sebastian has a new type of alignment pipeline (one part of which is described here) that avoids any kind of spring mesh optimization; given that they have successfully applied it to their petascale MICRONS data we should probably try to understand it further. The battle between TEM / SEM / blockface / sectioning rages on. Since people (including us) are now generating good TEM segmentations, there is some additional wind behind that sail, but overall, it remains a remarkably unsettled question. Neuroglancer is ubiquitous, and Jeremy's professionalism and pragmatism in designing and maintaining the tool has earned Google a huge amount of goodwill in the connectomics community. China is creating a connectomics "science park" with a \$350 million investment. They already have some kind of facility with ~10 EM's to create an "assembly line" connectomics approach. Interesting and kind of scary. Was hard to figure out what they're really going to do (unsurprisingly). Jeff used his presentation to advocate for the whole mouse brain project, and there was no real push back as far as I could tell. There is some political friction as to who might lead such a project and what technology would be used, but I didn't see any serious negativity about the idea of putting together a project like this. Analysis of connectomes remains a hugely unsolved issue of course, and Gerry used his talk to argue that even for simple drosophila behaviors there are very complex circuits involved and figuring these out even with an EM connectome is going to be hard. Expansion microscopy (and super-resolution microscopy in general) was less buzzy than I expected, perhaps because EM reconstruction is finally starting to work at scale. But even Joergen (now with Michale Fee at MIT) has started using expansion microscopy for sparse reconstruction in songbird tissue. Very little innovative work on automated segmentation was presented. Seems like people have instead been focused on scaling up existing approaches to larger datasets, since those methods at least sort of work.}
}
@article{JainetalCON-10,
        title = {Machines that learn to segment images: a crucial technology for connectomics},
       author = {Viren Jain and H. Sebastian Seung and Srinivas C. Turag},
      journal = {Current Opinion in Neurobiology},
       volume = 20,
        issue = 5,
         year = 2010,
        pages = {1-14},
}
@inproceedings{JainetalCVPR-08,
        title = {Fast image search for learned metrics},
       author = {Jain, Prateek and Kulis, Brian and Grauman, Kristen},
    booktitle = {Computer Vision and Pattern Recognition, 2008. CVPR 2008. IEEE Conference on},
 organization = {IEEE},
         year = {2008},
        pages = {1-8},
}
@book{James62,
       author = {William James},
        title = {Essays on Faith and Morals},
    publisher = {Selected from {\emph{The Will to Believe and Other Essays}}, 1898; {\emph{Talks to Teachers on Psychology}}, 1899; {\emph{Memories and Studies}}, 1911. Meridian Books},
      address = {New York, NY},
         year = 1962,
}
@book{James63,
      author = {William James},
       title = {Pragmatism and Other Essays},
   publisher = {Selected from {\emph{The Will to Believe and Other Essays}}, 1896; {\emph{Pragmatism}}, 1907; {\emph{The Meaning of Truth}}, 1909. Washington Square Press},
     address = {New York, NY},
        year = 1963,
}
@article{JansenPAC-14,
       author = {Martin, Jansen},
        title = {The energy landscape concept and its implications for synthesis planning},
      journal = {Pure and Applied Chemistry},
         year = {2014},
       volume = {86},
        issue = {6},
     abstract = {Synthesis of novel solids, in a chemical sense, is one of the spearheads of innovation in materials research. However, such an undertaking is substantially impaired by lack of control and predictability. We present a concept that points the way towards rational planning of syntheses in solid state and materials chemistry. The foundation of our approach is the representation of the whole material world, i.e., the known and not-yet-known chemical compounds, on an energy landscape, which implies information about the free energies of these configurations. From this it follows at once that all chemical compounds capable of existence (both thermodynamically stable and metastable ones) are already present in virtuo in this landscape. For the first step of synthesis planning, i.e., the identification of candidates that are capable of existence, we computationally search the respective potential energy landscapes for (meta)stable structure candidates. Recently we have extended our techniques to finite temperatures and pressures and calculated phase diagrams, including metastable manifestations of matter, without resorting to any experimental pre-information. The conception developed is physically consistent, and its feasibility has been proven. Applying appropriate experimental tools has enabled us to realize, e.g., elusive Na3 N, including almost all of its predicted polymorphs, many years after the predictions were published.}
}
@article{JanuszewskietalCoRR-16,
       author = {Michal Januszewski and Jeremy Maitin-Shepard and Peter Li and J\"{o}rgen Kornfeld and Winfried Denk and Viren Jain},
        title = {Flood-Filling Networks},
      journal = {CoRR},
       volume = {arXiv:1611.00421},
         year = {2016},
     abstract = {State-of-the-art image segmentation algorithms generally consist of at least two successive and distinct computations: a boundary detection process that uses local image information to classify image locations as boundaries between objects, followed by a pixel grouping step such as watershed or connected components that clusters pixels into segments. Prior work has varied the complexity and approach employed in these two steps, including the incorporation of multi-layer neural networks to perform boundary prediction, and the use of global optimizations during pixel clustering. We propose a unified and end-to-end trainable machine learning approach, flood-filling networks, in which a recurrent 3d convolutional network directly produces individual segments from a raw image. The proposed approach robustly segments images with an unknown and variable number of objects as well as highly variable object sizes. We demonstrate the approach on a challenging 3d image segmentation task, connectomic reconstruction from volume electron microscopy data, on which flood-filling neural networks substantially improve accuracy over other state-of-the-art methods. The proposed approach can replace complex multi-step segmentation pipelines with a single neural network that is learned end-to-end.}
} 
@article{JanuszewskietalNATURE-METHODS-18,
       author = {Januszewski, Michal and Kornfeld, J\"{o}rgen and Li, Peter H and Pope, Art and Blakely, Tim and Lindsey, Larry and Maitin-Shepard, Jeremy B and Tyka, Mike and Denk, Winfried and Jain, Viren},
        title = {High-Precision Automated Reconstruction of Neurons with Flood-filling Networks},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = {15},
         year = {2017},
        pages = {605-610},
     abstract = {Reconstruction of neural circuits from volume electron microscopy data requires the tracing of complete cells including all their neurites. Automated approaches have been developed to perform the tracing, but without costly human proofreading their error rates are too high to obtain reliable circuit diagrams. We present a method for automated segmentation that, like the majority of previous efforts, employs convolutional neural networks, but contains in addition a recurrent pathway that allows the iterative optimization and extension of the reconstructed shape of individual neural processes. We used this technique, which we call flood-filling networks, to trace neurons in a data set obtained by serial block-face electron microscopy from a male zebra finch brain. Our method achieved a mean error-free neurite path length of 1.1 mm, an order of magnitude better than previously published approaches applied to the same dataset. Only 4 mergers were observed in a neurite test set of 97 mm path length.},
}
@article{JarosiewiczetalJoN-02,
       author = {Jarosiewicz, Beata and McNaughton, Bruce L. and Skaggs, William E.},
        title = {Hippocampal Population Activity during the Small-Amplitude Irregular Activity State in the Rat},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {22},
       number = {4},
         year = {2002},
        pages = {1373-1384},
     abstract = {The sleeping rat cycles between two well-characterized physiological states, slow-wave sleep (SWS) and rapid-eye-movement sleep (REM), often identified by the presence of large-amplitude irregular activity (LIA) and theta activity, respectively, in the hippocampal EEG. Inspection of the activity of ensembles of hippocampal CA1 complex-spike cells along with the EEG reveals the presence of a third physiological state within SWS. We characterize the hippocampal EEG and population activity of this third state relative to theta activity and LIA, its incidence relative to REM and LIA, and the functional correlates of its population activity. This state occurs repeatedly within stretches of SWS, occupying \~{}33\% of SWS and \~{}20\% of total sleep, and it follows nearly every REM episode; however, it never occurs just before a REM episode. The EEG during this state becomes low in amplitude for a few seconds, probably corresponding to {\textquotedblleft}small-amplitude irregular activity' (SIA) described in the literature; we will call its manifestation during sleep {\textquotedblleft}S-SIA.' During S-SIA, a small subset of cells becomes active, whereas the rest remain nearly silent, with the same subset of cells active across long sequences of S-SIA episodes. These cells are physiologically indistinguishable from ordinary complex-spike cells; thus, the question arises as to whether they have any special functional correlates. Indeed, many of these cells are found to have place fields encompassing the location where the rat sleeps, raising the possibility that S-SIA is a state of increased alertness in which the animal{\textquoteright}s location in the environment is represented in the brain.},
}
@inproceedings{JarrettetalICCV-09,
        title = {What is the Best Multi-Stage Architecture for Object Recognition?},
       author = {Jarrett, Kevin and Kavukcuoglu, Koray and Ranzato, Marc'Aurelio and LeCun, Yann},
    booktitle = {Proceedings of the International Conference on Computer Vision},
    publisher = {IEEE Computer Society},
         year = 2009,
}
@book{Jasanoff2018biological,
       author = {Jasanoff, Alan},
        title = {The Biological Mind: How Brain, Body, and Environment Collaborate to Make Us Who We Are},
         year = {2018},
    publisher = {Basic Books},
     abstract = {To many, the brain is the seat of personal identity and autonomy. But the way we talk about the brain is often rooted more in mystical conceptions of the soul than in scientific fact. This blinds us to the physical realities of mental function. We ignore bodily influences on our psychology, from chemicals in the blood to bacteria in the gut, and overlook the ways that the environment affects our behavior, via factors varying from subconscious sights and sounds to the weather. As a result, we alternately overestimate our capacity for free will or equate brains to inorganic machines like computers. But a brain is neither a soul nor an electrical network: it is a bodily organ, and it cannot be separated from its surroundings. Our selves aren't just inside our heads--they're spread throughout our bodies and beyond. Only once we come to terms with this can we grasp the true nature of our humanity.}
}
@misc{Jay_Alammar_The_Illustrated_BERT_and_CO-18,
        title = {The Illustrated Bert and Company},
       author = {Jay Alammar},
         year = 2018,
        month = {December},
 howpublished = {http://jalammar.github.io},
}
@misc{Jay_Alammar_The_Illustrated_Transformer-18,
        title = {The Illustrated Transformer},
       author = {Jay Alammar},
         year = 2018,
        month = {June},
 howpublished = {http://jalammar.github.io},
}
@incollection{Jaynes79,
       author = {Edwin T. Jaynes},
        pages = {15-118},
        title = {Where do we stand on maximum entropy},
    booktitle = {The Maximum Entropy Formalism},
       editor = {Levin and Tribus},
    publisher = {{MIT} Press},
         year = 1979,
}
@incollection{JenattonetalNIPS-12,
     title = {A latent factor model for highly multi-relational data},
    author = {Jenatton, Rodolphe and Nicolas L. Roux and Bordes, Antoine and Guillaume R. Obozinski},
 booktitle = {Advances in Neural Information Processing Systems 25},
    editor = {F. Pereira and C.J.C. Burges and L. Bottou and K.Q. Weinberger},
      year = 2012,
     pages = {3176-3184},
}
@article{JenettetalCELL-REPORTS-12,
       author = {Jenett, Arnim and Rubin, Gerald M. and Ngo, Teri-T.B. and Shepherd, David and Murphy, Christine and Dionne, Heather and Pfeiffer, Barret D. and Cavallaro, Amanda and Hall, Donald and Jeter, Jennifer and Iyer, Nirmala and Fetter, Dona and Hausenfluck, Joanna H. and Peng, Hanchuan and Trautman, Eric T. and Svirskas, Robert R. and Myers, Eugene W. and Iwinski, Zbigniew R. and Aso, Yoshinori and DePasquale, Gina M. and Enos, Adrianne and Hulamm, Phuson and Lam, Shing Chun Benny and Li, Hsing-Hsi and Laverty, Todd R. and Long, Fuhui and Qu, Lei and Murphy, Sean D. and Rokicki, Konrad and Safford, Todd and Shaw, Kshiti and Simpson, Julie H. and Sowell, Allison and Tae, Susana and Yu, Yang and Zugates, Christopher T.},
        title = {A {GAL4}-Driver Line Resource for {\it{Drosophila}} Neurobiology},
      journal = {Cell Reports},
    publisher = {Elsevier},
       volume = {2},
        issue = {4},
         year = {2012},
        pages = {991-1001},
     abstract = {We established a collection of 7,000 transgenic lines of Drosophila melanogaster. Expression of GAL4 in each line is controlled by a different, defined fragment of genomic DNA that serves as a transcriptional enhancer. We used confocal microscopy of dissected nervous systems to determine the expression patterns driven by each fragment in the adult brain and ventral nerve cord. We present image data on 6,650 lines. Using both manual and machine-assisted annotation, we describe the expression patterns in the most useful lines. We illustrate the utility of these data for identifying novel neuronal cell types, revealing brain asymmetry, and describing the nature and extent of neuronal shape stereotypy. The GAL4 lines allow expression of exogenous genes in distinct, small subsets of the adult nervous system. The set of DNA fragments, each driving a documented expression pattern, will facilitate the generation of additional constructs for manipulating neuronal function.},
}
@article{JenkinsMataricIJHR-04,
       author = {Odest Chadwicke Jenkins and Maja J. Matari\'c},
        title = {Performance-Derived Behavior Vocabularies: Data-driven Acqusition of Skills from Motion},
      journal = {International Journal of Humanoid Robotics},
         year = 2004,
       volume = 1,
        issue = 2,
        pages = {237-288},
          url = {http://www.cs.brown.edu/~cjenkins/papers/cjenkins_ijhr2004.pdf},
}
@inproceedings{JenkinsetalAAAI-04,
       author = {Odest Chadwicke Jenkins and Monica Nicolescu and Maja J. Matari\'{c}},
        title = {Autonomy and Supervision for Robot Skills and Tasks Learned from Demonstration},
    booktitle = {AAAI-04 Workshop on Supervisory Control of Learning and Adaptive Systems},
         year = 2004,
          url = {http://www.cs.brown.edu/~cjenkins/papers/aaai-sc-camera.pdf},
}
@book{Jensen00,
       author = {Finn Jensen},
        title = {Bayesian graphical models},
    publisher = {John Wiley and Sons},
      address = {New York, NY},
         year = 2000,
}
@book{Jensen01,
       author = {Finn Jensen},
        title = {Bayesian Networks and Decision Graphs},
    publisher = {Springer-Verlag},
      address = {New York, NY},
         year = 2001,
}
@article{JensenetalCSQ-90,
       author = {F.V. Jensen and S.L. Lauritzen and K.G. Olesen},
        title = {Bayesian updating in recursive graphical models by local computations},
      journal = {Computational Statistical Quarterly},
         year = 1990,
        pages = {269-282},
       volume = 4,
}
@article{JensenetalNANO-07,
        title = {Nanotube radio},
       author = {Jensen, K. and Weldon, J. and Garcia, H. and Zettl, A.},
      journal = {Nano Letters},
       volume = 7,
        issue = 11,
         year = 2007,
        pages = {3508-3511},
     abstract = {We have constructed a fully functional, fully integrated radio receiver from a single carbon nanotube. The nanotube serves simultaneously as all essential components of a radio: antenna, tunable band-pass filter, amplifier, and demodulator. A direct current voltage source, as supplied by a battery, powers the radio. Using carrier waves in the commercially relevant 40-400 MHz range and both frequency and amplitude modulation techniques, we demonstrate successful music and voice reception.}
} 
@techreport{JensenetalTR-88,
       author = {Finn V. Jensen and Kristian G. Oleson and Stig Kj{\ae}r Anderson},
        title = {An Algebra of {B}ayesian Belief Universes for Knowledge-based Systems},
  institution = {Institute for Electronic Systems, Department of Mathematics and Computer Science, University of Aalborg},
        issue = {R 88-25},
         year = 1988,
}
@techreport{JensenetalTR-89,
       author = {Finn V. Jensen and Lauritzen, Steffen L. and Olesen, Kristian G.},
        title = {Bayesian Updating in Recursive Graphical Models by Local Computations},
  institution = {Institute for Electronic Systems, Department of Mathematics and Computer Science, University of Aalborg},
        issue = {R 89-15},
         year = 1989,
}
@article{JercogetalCSHPB-16,
       author = {Jercog, P. and Rogerson, T. and Schnitzer, M. J.},
        title = {Large-Scale Fluorescence Calcium-Imaging Methods for Studies of Long-Term Memory in Behaving Mammals},
      journal = {Cold Spring Harbor Perspectives in Biology},
         year = {2016},
       volume = {8},
       number = {5},
     abstract = {During long-term memory formation, cellular and molecular processes reshape how individual neurons respond to specific patterns of synaptic input. It remains poorly understood how such changes impact information processing across networks of mammalian neurons. To observe how networks encode, store, and retrieve information, neuroscientists must track the dynamics of large ensembles of individual cells in behaving animals, over timescales commensurate with long-term memory. Fluorescence Ca(2+)-imaging techniques can monitor hundreds of neurons in behaving mice, opening exciting avenues for studies of learning and memory at the network level. Genetically encoded Ca(2+) indicators allow neurons to be targeted by genetic type or connectivity. Chronic animal preparations permit repeated imaging of neural Ca(2+) dynamics over multiple weeks. Together, these capabilities should enable unprecedented analyses of how ensemble neural codes evolve throughout memory processing and provide new insights into how memories are organized in the brain.}
}
@article{JerniganetalPBR-11,
       author = {Jernigan, Terry L. and Baar, William F. C. and Stiles, Joan and Madsen, Kathrine Skak},
        title = {Postnatal brain development: structural imaging of dynamic neurodevelopmental processes},
      journal = {Progress in brain research},
         year = {2011},
       volume = {189},
        pages = {77-92},
     abstract = {After birth, there is striking biological and functional development of the brain's fiber tracts as well as remodeling of cortical and subcortical structures. Behavioral development in children involves a complex and dynamic set of genetically guided processes by which neural structures interact constantly with the environment. This is a protracted process, beginning in the third week of gestation and continuing into early adulthood. Reviewed here are studies using structural imaging techniques, with a special focus on diffusion weighted imaging, describing age-related brain maturational changes in children and adolescents, as well as studies that link these changes to behavioral differences. Finally, we discuss evidence for effects on the brain of several factors that may play a role in mediating these brain-behavior associations in children, including genetic variation, behavioral interventions, and hormonal variation associated with puberty. At present longitudinal studies are few, and we do not yet know how variability in individual trajectories of biological development in specific neural systems map onto similar variability in behavioral trajectories.},
}
@inproceedings{JerrumandSinclairSTOC-88,
       author = {Mark Jerrum and Alistair Sinclair},
        title = {Conductance and the Rapid Mixing Property for {M}arkov Chains: the Approximation of the Permanent Resolved (Preliminary Version)},
    booktitle = STOC,
        pages = {235-244},
         year = 1988,
}
@inproceedings{JhuangetalICCV-07,
       author = {Hueihan Jhuang and Thomas Serre and Lior Wolf and Tomaso Poggio},
        title = {A Biologically Inspired System for Action Recognition},
    booktitle = {IEEE International Conference on Computer Vision},
         year = 2007
}
@article{JiangetalSCIENCE-15,
       author = {Jiang, X. and Shen, S. and Cadwell, C. R. and Berens, P. and Sinz, F. and Ecker, A. S. and Patel, S. and Tolias, A. S.},
        title = {Principles of connectivity among morphologically defined cell types in adult neocortex},
      journal = {Science},
       volume = {350},
       number = {6264},
         year = {2015},
        pages = {9462},
     abstract = {Since the work of Ramón y Cajal in the late 19th and early 20th centuries, neuroscientists have speculated that a complete understanding of neuronal cell types and their connections is key to explaining complex brain functions. However, a complete census of the constituent cell types and their wiring diagram in mature neocortex remains elusive. By combining octuple whole-cell recordings with an optimized avidin-biotin-peroxidase staining technique, we carried out a morphological and electrophysiological census of neuronal types in layers 1, 2/3, and 5 of mature neocortex and mapped the connectivity between more than 11,000 pairs of identified neurons. We categorized 15 types of interneurons, and each exhibited a characteristic pattern of connectivity with other interneuron types and pyramidal cells. The essential connectivity structure of the neocortical microcircuit could be captured by only a few connectivity motifs.}
}
@article{JilketalJETAI-08,
       author = {Jilk, David and Lebiere, Christian and O'Reilly, Randall and R. Anderson, John},
        title = {{SAL}: An explicitly pluralistic cognitive architecture},
      journal = {Journal Experimental and Theoretical Artificial Intelligence},
        pages = {197-218},
         year = {2008},
       volume = {20},
     abstract = {The SAL cognitive architecture is a synthesis of two well-established constituents: ACT-R, a hybrid symbolic-subsymbolic cognitive architecture, and Leabra, a neural architecture. These component architectures have vastly different origins yet suggest a surprisingly convergent view of the brain, the mind and behaviour. Furthermore, both of these architectures are internally pluralistic, recognising that models at a single level of abstraction cannot capture the required richness of behaviour. In this article, we offer a brief principled defence of epistemological pluralism in cognitive science and artificial intelligence, and elaborate on the SAL architecture as an example of how pluralism can be highly effective as an approach to research in cognitive science.}
}
@inproceedings{JimenezandPerezICPR-06,
       author = {Manuel J. Mar\'{i}n-Jim\'{e}nez and Nicol\'{a}s P\'{e}rez de la Blanca},
        title = {Empirical Study of Multi-scale Filter Banks for Object Categorization},
    booktitle = {Proceedings of the 18th International Conference on Pattern Recognition},
    publisher = {IEEE Computer Society},
        pages = {578-581},
         year = 2006,
}        
@inproceedings{JinandGemanCVPR-06, 
      title = {Context and Hierarchy in a Probabilistic Image Model}, 
     author = {Ya Jin and Stuart Geman},
  booktitle = {Proceedings of the 2006 {IEEE} Conference on Computer Vision and Pattern Recognition}, 
  publisher = {IEEE Computer Society},
     volume = 2, 
      pages = {2145-2152}, 
       year = 2006, 
}
@article{JinetalCoRR-18,
       author = {Haifeng Jin, Qingquan Song, Xia Hu},
        title = {Auto-Keras: Efficient Neural Architecture Search with Network Morphism},
      journal = {CoRR},
       volume = {arXiv:1806.10282},
         year = {2018},
     abstract = {Neural architecture search (NAS) has been proposed to automatically tune deep neural networks, but existing search algorithms usually suffer from expensive computational cost. Network morphism, which keeps the functionality of a neural network while changing its neural architecture, could be helpful for NAS by enabling a more efficient training during the search. In this paper, we propose a novel framework enabling Bayesian optimization to guide the network morphism for efficient neural architecture search by introducing a neural network kernel and a tree-structured acquisition function optimization algorithm, which more efficiently explores the search space. Intensive experiments have been done to demonstrate the superior performance of the developed framework over the state-of-the-art methods. Moreover, we build an open-source AutoML system on our method, namely Auto-Keras. The system runs in parallel on CPU and GPU, with an adaptive search strategy for different GPU memory limits.},
}
@inproceedings{JinetalICML-20,
        title = {Multi-Objective Molecule Generation using Interpretable Substructures},
       author = {Wengong Jin and Regina Barzilay and Tommi Jaakkola},
    booktitle = {Proceedings of the 37th International Conference on Machine Learning},
       volume = {arXiv:2002.03244},
         year = {2020},
     abstract = {Drug discovery aims to find novel compounds with specified chemical property profiles. In terms of generative modeling, the goal is to learn to sample molecules in the intersection of multiple property constraints. This task becomes increasingly challenging when there are many property constraints. We propose to offset this complexity by composing molecules from a vocabulary of substructures that we call molecular rationales. These rationales are identified from molecules as substructures that are likely responsible for each property of interest. We then learn to expand rationales into a full molecule using graph generative models. Our final generative model composes molecules as mixtures of multiple rationale completions, and this mixture is fine-tuned to preserve the properties of interest. We evaluate our model on various drug design tasks and demonstrate significant improvements over state-of-the-art baselines in terms of accuracy, diversity, and novelty of generated compounds.},
}
@article{JinetalNEURON-12,
       author = {Jin, L. and Han, Z. and Platisa, J. and Wooltorton, J. R. and Cohen, L. B. and Pieribone, V. A.},
        title = {Single action potentials and subthreshold electrical events imaged in neurons with a fluorescent protein voltage probe},
      journal = {Neuron},
         year = {2012},
       volume = {75},
       number = {5},
        pages = {779-785},
     abstract = {Monitoring neuronal electrical activity using fluorescent protein-based voltage sensors has been limited by small response magnitudes and slow kinetics of existing probes. Here we report the development of a fluorescent protein voltage sensor, named ArcLight, and derivative probes that exhibit large changes in fluorescence intensity in response to voltage changes. ArcLight consists of the voltage-sensing domain of Ciona intestinalis voltage-sensitive phosphatase and super ecliptic pHluorin that carries the point mutation A227D. The fluorescence intensity of ArcLight A242 decreases by 35\% in response to a 100 mV depolarization when measured in HEK293 cells, which is more than five times larger than the signals from previously reported fluorescent protein voltage sensors. We show that the combination of signal size and response speed of these new probes allows the reliable detection of single action potentials and excitatory potentials in individual neurons and dendrites.}
}
@inproceedings{JingandBalujaWWW-08,
       author = {Jing, Yushi and Baluja, Shumeet},
        title = {PageRank for Product Image Search},
    booktitle = {Proceedings of the 17th World Wide Web Conference},
     location = {Beijing, China},
         year = 2008,
}
@article{JissendietalJoR-08,
        title = {Diffusion tensor imaging ({DTI}) and tractography of the cerebellar projections to prefrontal and posterior parietal cortices: A study at {3T}},
       author = {P. Jissendi and S. Baudry and D. Bal\'{e}riaux},
      journal = {Journal of Neuroradiology},
       volume = {35},
        issue = {1},
         year = {2009},
        pages = {42-50},
     abstract = {Objectives: Previous diffusion tensor imaging (DTI) studies have identified cerebellar pathways and supratentorial connections, but none of them have isolated cerebellar projections to prefrontal and posterior parietal cortices using tractography. The aim of our study was to identify and visualize on 3D projections, as well as on 2D cross-sectional images, the cerebellar projections to prefrontal (PF) and posterior parietal (PP) cortices.

Material and methods: The study included 10 healthy volunteers, four males and six females aged 25 to 45 years (mean age 31 years). A DTI sequence was applied at 3 Tesla using diffusion sensitizing gradients in 32 directions. White matter tracts were reconstructed by applying a multiple ROI (region of interest) tractography technique.

Results: PF projections were obtained in all subjects. PP projections were obtained in six over 10 subjects. On 2D cross-sectional images, the tracts showed the same anatomical location in each ROI in all subjects.

Conclusion: This DTI study at 3T resulted in a selective and full visualization of cerebellar projections to PF and PP cortices for the first time and is introductive for further optimized and quantitative DTI study of these tracts.},
}
@inproceedings{Joachims98,
       author = {Thorsten Joachims},
        title = {Text categorization with support vector machines: learning with many relevant features},
    booktitle = {Proceedings of {ECML}-98, 10th European Conference on Machine Learning},
        issue = 1398,
    publisher = {Springer Verlag, Heidelberg, DE},
      address = {Chemnitz, DE},
       editor = {Claire N{\'{e}}dellec and C{\'{e}}line Rouveirol},
        pages = {137-142},
         year = 1998,
}
@article{JoeletalNN-02,
       author = {Joel, D. and Niv, Y. and Ruppin, E.},
        title = {Actor-critic models of the basal ganglia: new anatomical and computational perspectives},
      journal = {Neural Networks},
         year = {2002},
       volume = {15},
       number = {4-6},
        pages = {535-547},
     abstract = {A large number of computational models of information processing in the basal ganglia have been developed in recent years. Prominent in these are actor-critic models of basal ganglia functioning, which build on the strong resemblance between dopamine neuron activity and the temporal difference prediction error signal in the critic, and between dopamine-dependent long-term synaptic plasticity in the striatum and learning guided by a prediction error signal in the actor. We selectively review several actor-critic models of the basal ganglia with an emphasis on two important aspects: the way in which models of the critic reproduce the temporal dynamics of dopamine firing, and the extent to which models of the actor take into account known basal ganglia anatomy and physiology. To complement the efforts to relate basal ganglia mechanisms to reinforcement learning (RL), we introduce an alternative approach to modeling a critic network, which uses Evolutionary Computation techniques to 'evolve' an optimal RL mechanism, and relate the evolved mechanism to the basic model of the critic. We conclude our discussion of models of the critic by a critical discussion of the anatomical plausibility of implementations of a critic in basal ganglia circuitry, and conclude that such implementations build on assumptions that are inconsistent with the known anatomy of the basal ganglia. We return to the actor component of the actor-critic model, which is usually modeled at the striatal level with very little detail. We describe an alternative model of the basal ganglia which takes into account several important, and previously neglected, anatomical and physiological characteristics of basal ganglia-thalamocortical connectivity and suggests that the basal ganglia performs reinforcement-biased dimensionality reduction of cortical inputs. We further suggest that since such selective encoding may bias the representation at the level of the frontal cortex towards the selection of rewarded plans and actions, the reinforcement-driven dimensionality reduction framework may serve as a basis for basal ganglia actor models. We conclude with a short discussion of the dual role of the dopamine signal in RL and in behavioral switching.}
}
@article{JoeschetalNATURE-10,
       author = {Joesch, M. and Schnell, B. and Raghu, S. V. and Reiff, D. F. and Borst, A.},
        title = {{{O}{N} and {O}{F}{F} pathways in {D}rosophila motion vision}},
      journal = {Nature},
       volume = {468},
       number = {7321},
         year = {2010},
        pages = {300-304},
     abstract = {Motion vision is a major function of all visual systems, yet the underlying neural mechanisms and circuits are still elusive. In the lamina, the first optic neuropile of Drosophila melanogaster, photoreceptor signals split into five parallel pathways, L1-L5. Here we examine how these pathways contribute to visual motion detection by combining genetic block and reconstitution of neural activity in different lamina cell types with whole-cell recordings from downstream motion-sensitive neurons. We find reduced responses to moving gratings if L1 or L2 is blocked; however, reconstitution of photoreceptor input to only L1 or L2 results in wild-type responses. Thus, the first experiment indicates the necessity of both pathways, whereas the second indicates sufficiency of each single pathway. This contradiction can be explained by electrical coupling between L1 and L2, allowing for activation of both pathways even when only one of them receives photoreceptor input. A fundamental difference between the L1 pathway and the L2 pathway is uncovered when blocking L1 or L2 output while presenting moving edges of positive (ON) or negative (OFF) contrast polarity: blocking L1 eliminates the response to moving ON edges, whereas blocking L2 eliminates the response to moving OFF edges. Thus, similar to the segregation of photoreceptor signals in ON and OFF bipolar cell pathways in the vertebrate retina, photoreceptor signals segregate into ON-L1 and OFF-L2 channels in the lamina of Drosophila.}
}
@article{JoetalNANO-LETTERS-10,
       author = {Jo, S. H. and Chang, T. and Ebong, I. and Bhadviya, B. B. and Mazumder, P. and Lu, W.},
        title = {Nanoscale memristor device as synapse in neuromorphic systems},
      journal = {Nano Letters},
         year = 2010,
       volume = 10,
        issue = 4,
        pages = {1297-1301},
     abstract = {A memristor is a two-terminal electronic device whose conductance can be precisely modulated by charge or flux through it. Here we experimentally demonstrate a nanoscale silicon-based memristor device and show that a hybrid system composed of complementary metal-oxide semiconductor neurons and memristor synapses can support important synaptic functions such as spike timing dependent plasticity. Using memristors as synapses in neuromorphic circuits can potentially offer both high connectivity and high density required for efficient computing.}
}
@article{JoetalNANO-LETTERS-12,
       author = {Jo, Sung Hyun and Chang, Ting and Ebong, Idongesit and Bhadviya, Bhavitavya B. and Mazumder, Pinaki and Lu, Wei},
        title = {Nanoscale Memristor Device as Synapse in Neuromorphic Systems},
      journal = {Nano Letters},
       volume = 10,
        issue = 4,
        pages = {1297-1301},
         year = 2010,
     abstract = {A memristor is a two-terminal electronic device whose conductance can be precisely modulated by charge or flux through it. Here we experimentally demonstrate a nanoscale silicon-based memristor device and show that a hybrid system composed of complementary metal−oxide semiconductor neurons and memristor synapses can support important synaptic functions such as spike timing dependent plasticity. Using memristors as synapses in neuromorphic circuits can potentially offer both high connectivity and high density required for efficient computing.}
}
@article{JoetalNSR-15,
       author = {Jo, Taeho and Hou, Jie and Eickholt, Jesse and Cheng, Jianlin},
        title = {Improving Protein Fold Recognition by Deep Learning Networks},
      journal = {Scientific Reports},
    publisher = {Nature Publishing Group},
       volume = {5},
       number = {17573},
         year = 2015,
     abstract = {For accurate recognition of protein folds, a deep learning network method (DN-Fold) was developed to predict if a given query-template protein pair belongs to the same structural fold. The input used stemmed from the protein sequence and structural features extracted from the protein pair. We evaluated the performance of DN-Fold along with 18 different methods on Lindahl's benchmark dataset and on a large benchmark set extracted from SCOP 1.75 consisting of about one million protein pairs, at three different levels of fold recognition (i.e., protein family, superfamily, and fold) depending on the evolutionary distance between protein sequences. The correct recognition rate of ensembled DN-Fold for Top 1 predictions is 84.5\%, 61.5\%, and 33.6\% and for Top 5 is 91.2\%, 76.5\%, and 60.7\% at family, superfamily, and fold levels, respectively. We also evaluated the performance of single DN-Fold (DN-FoldS), which showed the comparable results at the level of family and superfamily, compared to ensemble DN-Fold. Finally, we extended the binary classification problem of fold recognition to real-value regression task, which also show a promising performance. DN-Fold is freely available through a web server at http://iris.rnet.missouri.edu/dnfold.},
}
@misc{JohanssonandLanser,
       author = {Christopher Johansson and Anders Lansner},
        title = {Towards Cortex Sized Artificial Nervous Systems},
}
@techreport{JohanssonandLansnerTR-03,
       author = {Johansson, C. and Lansner, A.},
        title = {Mapping of the {BCPNN} onto cluster computers},
  institution = {Department of Numerical Analysis and Computing Science Royal Institute of Technology, Stockholm, Sweden},
        issue = {TRITA-NA-P0305},
         year = 2003,
}
@article{John_Hughlings_JacksonMEDICAL-HISTORY-07,
        title = {{An Introduction to the Life and Work of John Hughlings Jackson: Introduction}},
       author = {John Hughlings Jackson [Biography]},
      journal = {Medical History. Supplement},
         year = {2007},
    publisher = {Medical History},
        issue = {26},
        pages = {3-34},
     abstract = {John Hughlings Jackson created the conceptual framework for clinical neurophysiology, the discipline that underlies diagnostic neurology. He began by establishing a consistent scientific method based on the systematic analysis of anatomy, pathology and physiology. This method revolved around his concept of the focal lesion, a concept that he refined and which became a cornerstone of bedside neurology. He recognized that focal epilepsy and focal necrosis can be viewed as reciprocal physiological processes, and concluded that somatotopic representation is found in the entire nervous system including the cortex.}
}
@article{Johnson67,
       author = {Johnson, Stephan C.},
        title = {Hierarchical Clustering Schemes},
      journal = {Psychometrika},
       volume = 32,
         year = 1967,
        pages = {241-254},
}
@article {JohnsonCD-01,
        title = {Functional Brain Development in Infants: Elements of an Interactive Specialization Framework},
       author = {Johnson, Mark H.},
      journal = {Child Development},
       volume = 71,
        issue = 1,
    publisher = {Blackwell Publishers Inc.},
         year = 2001,
        pages = {75-81},
}
@misc{JohnsonGOOG-18,
       author = {Justin Johnson},
        title = {Compositional Visual Intelligence through Language},
     abstract = {The use of deep neural networks has led to fantastic recent progress on fundamental problems in computer vision. However the most well-studied tasks, such as image classification and object detection, encompass only a small fraction of the rich visual intelligence that people display. In order to make progress toward the goal of visual intelligence, we must define new tasks and construct new datasets that push the boundaries of artificial visual intelligence. An important facet of visual intelligence is composition - our understanding of the whole derives from an understanding of the parts. One avenue for studying the compositional nature of visual intelligence is through connections with natural language, which is also inherently compositional. To this end I will present three research directions toward compositional visual intelligence through language. In each we will see how incorporating compositionality forces us to rethink our tasks and datasets, but results in systems with richer visual intelligence. I will first discuss image captioning, where moving from sentence captions to dense captions and descriptive paragraphs results in richer image descriptions. I will next discuss visual question answering, where an emphasis on compositional reasoning gives rise to new datasets and models. I will then discuss text-to-image synthesis, where replacing freeform natural language with explicitly compositional scene graphs of objects and relationships allows us to generate more complex images. I will conclude by discussing future directions for improving visual intelligence through composition.}
}
@inproceedings{JohnsonICLR-17,
        title = {Learning Graphical State Transitions},
       author = {Daniel D. Johnson},
    booktitle = {International Conference on Learning Representations},
         year = {2017},
     abstract = {Graph-structured data is important in modeling relationships between multiple entities, and can be used to represent states of the world as well as many data structures. Li et al. (2016) describe a model known as a Gated Graph Sequence Neural Network (GGS-NN) that produces sequences from graph-structured input. In this work I introduce the Gated Graph Transformer Neural Network (GGT-NN), an extension of GGS-NNs that uses graph-structured data as an intermediate representation. The model can learn to construct and modify graphs in sophisticated ways based on textual input, and also to use the graphs to produce a variety of outputs. For example, the model successfully learns to solve almost all of the bAbI tasks (Weston et al., 2016), and also discovers the rules governing graphical formulations of a simple cellular automaton and a family of Turing machines. TL;DR: I introduce a set of differentiable graph transformations, and use them to build a model with a graphical internal state that can extract structured data from text and use it to answer queries.}
}
@incollection{JohnsonPLM-70,
       author = {Neal F. Johnson},
        title = {The Role of Chunking and Organization in The Process of Recall},
       editor = {Gordon H. Bower},
    booktitle = {Psychology of Learning and Motivation},
    publisher = {Academic Press},
       volume = {4},
        pages = {171-247},
         year = {1970},
     abstract = {This chapter discusses the role of chunking and organization in the process of recall. There are four concepts associated with the organization of memory: chunk, memory code, decode, and recode. Chunks have been operationally defined as behavior sequences, which tend to occur either adjacently or in an all-or-none manner. Theoretically, they can be defined as item or information sets, which are stored within the same memory code, with the code and the chunk being distinct. Recoding refers to the process of learning the code for a chunk, and decoding is the process of translating the code into the information it represents. Organization plays a major role in the associative relations between an item in a sequence and all the other items in the sequence. When the items are from the same chunk, there appears to be a transferable association, but if they are from different chunks the transfer is small, if it exists at all. Chunks may also represent decision units in recall.}
}
@incollection{JohnsonUAI-88,
       author = {Johnson, Rodney W.},
        title = {Independence and {B}ayesian Updating Methods},
     crossref = UAI-88,
}
@article{JohnsonandOlshausenVR-05,
       author = {Jeffrey S. Johnson and Bruno A. Olshausen},
        title = {The recognition of partially visible natural objects in the presence and absence of their occluders},
      journal = {Vision Research},
       volume = 45,
        pages = {3262-3276},
         year = 2005,
     abstract = {The visual system is adept at compensating for the missing information in scenes that results from occlusion, but how this is done is not fully understood. In particular, the role of the occluding object in visual processing and its effect on the subsequent recognition of the occluded object is unclear. We report three human behavioral experiments suggesting that the recognition of partially visible objects is facilitated when the missing object information is replaced by an occluder rather than simply removed. Furthermore, we provide EEG evidence suggesting that the processes responsible for facilitated recognition occur relatively early in the visual stream.},
}
@article{JohnsonandStevensNATURE-18,
       author = {Matthew B. Johnson and Beth Stevens},
        title = {Pruning hypothesis comes of age},
      journal = {Nature},
         year = {2018},
    publisher = {Macmillan Publishers Limited},
       volume = {554},
        pages = {38-39},
     abstract = {The idea that disrupted pruning of neuronal connections in the brain during adolescence is a cause of schizophrenia was proposed in 1983. This proved prescient, as subsequent imaging, genetic and molecular research has shown.}
}
@article{JohnsonetalCoRR-16,
       author = {Justin Johnson and Bharath Hariharan and Laurens van der Maaten and Li Fei{-}Fei and C. Lawrence Zitnick and Ross B. Girshick},
        title = {{CLEVR:} {A} Diagnostic Dataset for Compositional Language and Elementary Visual Reasoning},
      journal = {CoRR},
       volume = {arXiv:1612.06890},
         year = {2016},
     abstract = {When building artificial intelligence systems that can reason and answer questions about visual data, we need diagnostic tests to analyze our progress and discover shortcomings. Existing benchmarks for visual question answering can help, but have strong biases that models can exploit to correctly answer questions without reasoning. They also conflate multiple sources of error, making it hard to pinpoint model weaknesses. We present a diagnostic dataset that tests a range of visual reasoning abilities. It contains minimal biases and has detailed annotations describing the kind of reasoning each question requires. We use this dataset to analyze a variety of modern visual reasoning systems, providing novel insights into their abilities and limitations.}
}
@article{JohnsonetalFiSN-19,
       author = {Johnson, Elizabeth L. and King-Stephens, David and Weber, Peter B. and Laxer, Kenneth D. and Lin, Jack J. and Knight, Robert T.},
        title = {Spectral Imprints of Working Memory for Everyday Associations in the Frontoparietal Network},
      journal = {Frontiers in Systems Neuroscience},
       volume = {12},
        pages = {65},
         year = {2019},
     abstract = {How does the human brain rapidly process incoming information in working memory? In growing divergence from a single-region focus on the prefrontal cortex (PFC), recent work argues for emphasis on how distributed neural networks are rapidly coordinated in support of this central neurocognitive function. Previously, we showed that working memory for everyday what,' `where,' and 'when' associations depends on multiplexed oscillatory systems, in which signals of different frequencies simultaneously link the PFC to parieto-occipital and medial temporal regions, pointing to a complex web of sub-second, bidirectional interactions. Here, we used direct brain recordings to delineate the frontoparietal oscillatory correlates of working memory with high spatiotemporal precision. Seven intracranial patients with electrodes simultaneously localized to prefrontal and parietal cortices performed a visuospatial working memory task that operationalizes the types of identity and spatiotemporal information we encounter every day. First, task-induced oscillations in the same delta-theta (2-7 Hz) and alpha-beta (9-24 Hz) frequency ranges previously identified using scalp electroencephalography (EEG) carried information about the contents of working memory. Second, maintenance was linked to directional connectivity from the parietal cortex to the PFC. However, presentation of the test prompt to cue identity, spatial, or temporal information changed delta-theta coordination from a unidirectional, parietal-led system to a bidirectional, frontoparietal system. Third, the processing of spatiotemporal information was more bidirectional in the delta-theta range than was the processing of identity information, where alpha-beta connectivity did not exhibit sensitivity to the contents of working memory. These findings implicate a bidirectional delta-theta mechanism for frontoparietal control over the contents of working memory.}
}
@article{JohnsonetalJAH-09,
       author = {Johnson, Sara B. and Blum, Robert W. and Giedd, Jay N.},
        title = {Adolescent Maturity and the Brain: The Promise and Pitfalls of Neuroscience Research in Adolescent Health Policy},
      journal = {Journal of Adolescent Health},
         year = {2009},
       volume = {45},
        issue = {3},
        pages = {216-221},
     abstract = {Longitudinal neuroimaging studies demonstrate that the adolescent brain continues to mature well into the 20s. This has prompted intense interest in linking neuromaturation to maturity of judgment. Public policy is struggling to keep up with burgeoning interest in cognitive neuroscience and neuroimaging. However, empirical evidence linking neurodevelopmental processes and adolescent real-world behavior remains sparse. Nonetheless, adolescent brain development research is already shaping public policy debates about when individuals should be considered mature for policy purposes. With this in mind, in this article we summarize what is known about adolescent brain development and what remains unknown, as well as what neuroscience can and cannot tell us about the adolescent brain and behavior. We suggest that a conceptual framework that situates brain science in the broader context of adolescent developmental research would help to facilitate research-to-policy translation. Furthermore, although contemporary discussions of adolescent maturity and the brain often use a deficit-based approach, there is enormous opportunity for brain science to illuminate the great strengths and potentialities of the adolescent brain. So, too, can this information inform policies that promote adolescent health and well-being.},
}
@article{JohnsonetalNATURE-16,
       author = {Johnson, Carolyn M. and Peckler, Hannah and Tai, Lung-Hao and Wilbrecht, Linda},
        title = {Rule learning enhances structural plasticity of long-range axons in frontal cortex},
      journal = {Nature Communications},
    publisher = {Nature Publishing Group},
       volume = {7},
         year = {2016},
     abstract = {Rules encompass cue-action-outcome associations used to guide decisions and strategies in a specific context. Subregions of the frontal cortex including the orbitofrontal cortex (OFC) and dorsomedial prefrontal cortex (dmPFC) are implicated in rule learning, although changes in structural connectivity underlying rule learning are poorly understood. We imaged OFC axonal projections to dmPFC during training in a multiple choice foraging task and used a reinforcement learning model to quantify explore-exploit strategy use and prediction error magnitude. Here we show that rule training, but not experience of reward alone, enhances OFC bouton plasticity. Baseline bouton density and gains during training correlate with rule exploitation, while bouton loss correlates with exploration and scales with the magnitude of experienced prediction errors. We conclude that rule learning sculpts frontal cortex interconnectivity and adjusts a thermostat for the explore-exploit balance.},
}
@inproceedings{JojicandFreyCVPR-01,
       author = {N. Jojic and B. Frey},
        title = {Learning Flexible Sprites in Video Layers},
    booktitle = {Proceedings {IEEE} Conference on Computer Vision and Pattern Recognition},
        pages = {199-206},
       volume = 1,
         year = 2001,
}
@book{Jolliffe86,
       author = {Jolliffe, I. T.},
        title = {Principal Component Analysis},
    publisher = {Springer-Verlag},
      address = {New York},
         year = 1986,
}
@article{JonasandKordingBIORXIV-16,
       author = {Jonas, Eric and Kording, Konrad},
        title = {Could a neuroscientist understand a microprocessor?},
      journal = {bioRxiv},
         year = {2016},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {There is a popular belief in neuroscience that we are primarily data limited, that producing large, multimodal, and complex datasets will, enabled by data analysis algorithms, lead to fundamental insights into the way the brain processes information. Microprocessors are among those artificial information processing systems that are both complex and that we understand at all levels, from the overall logical flow, via logical gates, to the dynamics of transistors. Here we take a simulated classical microprocessor as a model organism, and use our ability to perform arbitrary experiments on it to see if popular data analysis methods from neuroscience can elucidate the way it processes information. We show that the approaches reveal interesting structure in the data but do not meaningfully describe the hierarchy of information processing in the processor. This suggests that current approaches in neuroscience may fall short of producing meaningful models of the brain.},
}
@article{JonasandKordingCoRR-14,
       author = {Eric Jonas and Konrad Kording},
        title = {Automatic discovery of cell types and microcircuitry from neural connectomics},
      journal = {CoRR},
       volume = {arXiv:1407.4137},
         year = 2014,
     abstract = {Neural connectomics has begun producing massive amounts of data, necessitating new analysis methods to discover the biological and computational structure. It has long been assumed that discovering neuron types and their relation to microcircuitry is crucial to understanding neural function. Here we developed a nonparametric Bayesian technique that identifies neuron types and microcircuitry patterns in connectomics data. It combines the information traditionally used by biologists, including connectivity, cell body location and the spatial distribution of synapses, in a principled and probabilistically-coherent manner. We show that the approach recovers known neuron types in the retina and enables predictions of connectivity, better than simpler algorithms. It also can reveal interesting structure in the nervous system of C. elegans, and automatically discovers the structure of a  microprocessor. Our approach extracts structural meaning from connectomics, enabling new approaches of automatically deriving anatomical insights from these emerging datasets.}
}
@article{JonasandKordingELIFE-15,
       author = {Eric Jonas and Konrad Kording},
        title = {Automatic discovery of cell types and microcircuitry from neural connectomics},
      journal = {eLIFE},
       volume = 4,
         year = 2015,
        pages = {e04250},
     abstract = {Neural connectomics has begun producing massive amounts of data, necessitating new analysis methods to discover the biological and computational structure. It has long been assumed that discovering neuron types and their relation to microcircuitry is crucial to understanding neural function. Here we developed a non-parametric Bayesian technique that identifies neuron types and microcircuitry patterns in connectomics data. It combines the information traditionally used by biologists in a principled and probabilistically coherent manner, including connectivity, cell body location, and the spatial distribution of synapses. We show that the approach recovers known neuron types in the retina and enables predictions of connectivity, better than simpler algorithms. It also can reveal interesting structure in the nervous system of Caenorhabditis elegans and an old man-made microprocessor. Our approach extracts structural meaning from connectomics, enabling new approaches of automatically deriving anatomical insights from these emerging datasets.},
}
@article{JonasandKordingPLoS-17,
       author = {Jonas, Eric and Kording, Konrad Paul},
      journal = {{PLoS} Computational Biology},
    publisher = {Public Library of Science},
        title = {Could a Neuroscientist Understand a Microprocessor?},
         year = {2017},
       volume = {13},
       number = {1},
          url = {https://doi.org/10.1371/journal.pcbi.       series = 268},
        pages = {1-24},
     abstract = {Neuroscience is held back by the fact that it is hard to evaluate if a conclusion is correct; the complexity of the systems under study and their experimental inaccessability make the assessment of algorithmic and data analytic technqiues challenging at best. We thus argue for testing approaches using known artifacts, where the correct interpretation is known. Here we present a microprocessor platform as one such test case. We find that many approaches in neuroscience, when used naïvely, fall short of producing a meaningful understanding.},
}
@misc{Jonathan_Hui_DQN_Deep_Q-Network,
       author = {Jonathan Hui},
        title = {{RL With DQN Deep Q-Networks}},
 howpublished = {https://medium.com/@jonathan_hui/rl-dqn-deep-q-network-e207751f7ae4},
         year = {2018}
}
@book{Jones2008,
        title = {Soft machines: Nanotechnology And life},
       author = {Richard Anthony Lewis Jones},
    publisher = {Oxford University Press},
      address = {Oxford, UK},
         year = 2008,
}
@article{Jones87,
       author = {J.P. Jones and L.A. Palmer},
        title = {The two-dimensional Spatial Structure of Simple Receptive Fields in the Cat Striate Cortex},
      journal = {Journal of Neurophysiology},
       volume = 58,
        pages = {1187-11211},
         year = 1987,
}
@article{JonesCOIN-16,
        title = {When brain rhythms aren't 'rhythmic': implication for their mechanisms and meaning},
       author = {Stephanie R. Jones},
      journal = {Current Opinion in Neurobiology},
       volume = {40},
         year = {2016},
        pages = {72-80},
     abstract = {Rhythms are a prominent signature of brain activity. Their expression is correlated with numerous examples of healthy information processing and their fluctuations are a marker of disease states. Yet, their causal or epiphenomenal role in brain function is still highly debated. We review recent studies showing brain rhythms are not always 'rhythmic', by which we mean representative of repeated cycles of activity. Rather, high power and continuous rhythms in averaged signals can represent brief transient events on single trials whose density accumulates in the average. We also review evidence showing time-domain signals with vastly different waveforms can exhibit identical spectral-domain frequency and power. Further, non-oscillatory waveform feature can create spurious high spectral power. Knowledge of these possibilities is essential when interpreting rhythms and is easily missed without considering pre-processed data. Lastly, we discuss how these findings suggest new directions to pursue in our quest to discover the mechanism and meaning of brain rhythms.}
}
@article{JonesPNAS-00,
       author = {Jones, E. G.},
        title = {Microcolumns in the cerebral cortex},
      journal = {Proceedings of the National Academy of Science},
         year = 2000,
       volume = 97,
        issue = 10,
        pages = {5019-5021},
}
@article{JonesandFernyhoughCPR-07,
        title = {Neural correlates of inner speech and auditory verbal hallucinations: A critical review and theoretical integration},
      journal = {Clinical Psychology Review},
       author = {Simon R. Jones and Charles Fernyhough},
       volume = {27},
       number = {2},
        pages = {140-154},
         year = {2007},
     abstract = {The neuroimaging and neurophysiological literature on inner speech in healthy participants and those who experience auditory verbal hallucinations (AVHs) is reviewed. AVH-hearers in remission and controls do not differ neurologically on tasks involving low levels of verbal self-monitoring (VSM), such as reciting sentences in inner speech. In contrast, on tasks involving high levels of VSM, such as auditory verbal imagery, AVH-hearers in remission show less activation in areas including the middle and superior temporal gyri. This pattern of findings leads to a conundrum, given that mentation involving low levels of VSM is typically held to form the raw material for AVHs. We address this by noting that existing neuroimaging and neurophysiological studies have been based on unexamined assumptions about the form and developmental significance of inner speech. We set out a Vygotskian approach to AVHs which can account for why they are generally experienced as the voice of another person, with specific acoustic properties, and a tendency to take the form of commands. On this approach, which we argue is consistent with the neural correlates evidence, AVHs result from abnormalities in the transition between condensed and expanded dialogic inner speech. Further potential empirical tests of this model are discussed.}
}
@article{JonesandMewhortPR-07,
       author = {Jones, M. N. and Mewhort, D. J. K.},
        title = {Representing word meaning and order information in a composite holographic lexicon},
      journal = {Psychological Review},
       volume = 114,
         year = 2007,
        pages = {1-37},
     abstract = {The authors present a computational model that builds a holographic lexicon representing both word meaning and word order from unsupervised experience with natural language. The model uses simple convolution and superposition mechanisms (cf. B. B. Murdock, 1982) to learn distributed holographic representations for words. The structure of the resulting lexicon can account for empirical data from classic experiments studying semantic typicality, categorization, priming, and semantic constraint in sentence completions. Furthermore, order information can be retrieved from the holographic representations, allowing the model to account for limited word transitions without the need for built-in transition rules. The model demonstrates that a broad range of psychological data can be accounted for directly from the structure of lexical representations learned in this way, without the need for complexity to be built into either the processing mechanisms or the representations. The holographic representations are an appropriate knowledge representation to be used by higher order models of language comprehension, relieving the complexity required at the higher level.},
}
@article{JonesetalCB-97,
        title = {Top-down Learning of Low-level Vision Tasks},
       author = {Michael Jones and Pawan Sinha and Thomas Vetter and Tomaso Poggio},
      journal = {Current Biology},
       volume = 7,
        pages = {991-994},
         year = 1997,
}
@article{JooandFrankNATURE-REVIEWS-NEUROSCIENCE-18,
       author = {Joo, Hannah R. and Frank, Loren M.},
        title = {The hippocampal sharp wave-ripple in memory retrieval for immediate use and consolidation},
      journal = {Nature Reviews Neuroscience},
         year = {2018},
       volume = {19},
        issue = {12},
        pages = {744-757},
     abstract = {Various cognitive functions have long been known to require the hippocampus. Recently, progress has been made in identifying the hippocampal neural activity patterns that implement these functions. One such pattern is the sharp wave-ripple (SWR), an event associated with highly synchronous neural firing in the hippocampus and modulation of neural activity in distributed brain regions. Hippocampal spiking during SWRs can represent past or potential future experience, and SWR-related interventions can alter subsequent memory performance. These findings and others suggest that SWRs support both memory consolidation and memory retrieval for processes such as decision-making. In addition, studies have identified distinct types of SWR based on representational content, behavioural state and physiological features. These various findings regarding SWRs suggest that different SWR types correspond to different cognitive functions, such as retrieval and consolidation. Here, we introduce another possibility - that a single SWR may support more than one cognitive function. Taking into account classic psychological theories and recent molecular results that suggest that retrieval and consolidation share mechanisms, we propose that the SWR mediates the retrieval of stored representations that can be utilized immediately by downstream circuits in decision-making, planning, recollection and/or imagination while simultaneously initiating memory consolidation processes.},
}
@book{Jordan98,
       editor = {Jordan, M.I.},
        title = {Learning in Graphical Models},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1998,
}
@article{JordanandJacobsNC-94,
       author = {Jordan, Michael I. and Jacobs, Robert A.},
        title = {Hierarchical mixtures of experts and the {EM} algorithm},
      journal = {Neural Computation},
       volume = 6,
         year = 1994,
        pages = {181-214},
}
@article{JordanandRumelhartCS-92,
       author = {Jordan, Michael I. and Rumelhart, David E.},
        title = {Forward models: Supervised learning with a distal teacher},
      journal = {Cognitive Science},
       volume = 16,
         year = 1992,
        pages = {307-354},
}
@article{JordanetalML-99,
       author = {Jordan, Michael I. and Ghahramani, Zoubin and Jaakkola, Tommi S. and Saul, Lawrence K.},
        title = {An Introduction to Variational Methods for Graphical Models},
      journal = {Machine Learning},
    publisher = {Kluwer Academic Publishers},
       volume = {37},
       number = {2},
         year = {1999},
        pages = {183-233},
}
@inproceedings{JoshietalLISA-09,
       author = {Joshi, J. and Chih-Chieh Hsu and Parker, A.C. and Deshmukh, P.},
    booktitle = {Life Science Systems and Applications Workshop, 2009. {LiSSA 2009. IEEE/NIH}},
        title = {A carbon nanotube cortical neuron with excitatory and inhibitory dendritic computations},
         year = {2009},
        pages = {133-136},
     abstract = {A cortical neuron with carbon nanotube circuit elements that performs nonlinear dendritic computations with excitatory and inhibitory post-synaptic potentials is presented. An inhibitory synapse with controllable parameters that implement plasticity is described. The circuit design was simulated using carbon nanotube spice models, showing that the neuron fires as long as the inhibitory post-synaptic potential is weak or absent. Strong inhibitory potentials prevent the neuron from firing.},
}
@article{JoslinandRoachAIJ-89,
        title = {A theoretical analysis of conjunctive-goal problems},
       author = {David Joslin and John Roach},
      journal = {Artificial Intelligence},
       volume = {41},
       number = {1},
         year = {1989},
        pages = {97-106},
     abstract = {Region analysis is a new technique for analyzing search problems by applying graph theory to problem state spaces. The analysis here is of search problems, not search algorithms; analyzing problems and classes of problems lets us understand the underlying structure and inherent complexity of those problems. The analysis technique is demonstrated in the domain of robot planning problems. Region analysis of conjunctive-goal planning problems gives us a characterization of subgoal interactions that is independent of the problem representation. We give a formal characterization of nonlinear planning problems, and show that nonlinearity is a weak characterization of the difficulty in planning problems.}
}
@article{JuangandRabinerATT-85,
       author = {Juang, B. H. and Rabiner, L. R.},
        title = {A Probabilistic Distance Measure for Hidden {M}arkov Models},
      journal = {AT\&T Technical Journal},
       volume = 64,
        issue = 2,
        pages = {391-408},
         year = 1985,
}
@book{Judson1979,
        title = {The Eighth Day of Creation: Makers of the Revolution in Biology},
       author = {Judson, Horace Freeland},
       series = {Touchstone Books},
      address = {New York, NY},
         year = {1979},
    publisher = {Simon and Schuster}
}
@book{Judson96,
       author = {Horace Freeland Judson},
        title = {The Eighth Day of Creation: Makers of the Revolution in Biology},
    publisher = {Cold Spring Harbor Laboratory Press},
      address = {Long Island, New York},
         year = 1996,
}
@article{JulianietalCoRR-18,
  author    = {Arthur Juliani and Vincent{-}Pierre Berges and Esh Vckay and Yuan Gao and Hunter Henry and Marwan Mattar and Danny Lange},
  title     = {Unity: {A} General Platform for Intelligent Agents},
  journal   = {CoRR},
  volume    = {arXiv:1809.02627},
  year      = {2018},
  abstract = {Recent advances in Deep Reinforcement Learning and Robotics have been driven by the presence of increasingly realistic and complex simulation environments. Many of the existing platforms, however, provide either unrealistic visuals, inaccurate physics, low task complexity, or a limited capacity for interaction among artificial agents. Furthermore, many platforms lack the ability to flexibly configure the simulation, hence turning the simulation environment into a black-box from the perspective of the learning system. Here we describe a new open source toolkit for creating and interacting with simulation environments using the Unity platform: Unity ML-Agents Toolkit. By taking advantage of Unity as a simulation platform, the toolkit enables the development of learning environments which are rich in sensory and physical complexity, provide compelling cognitive challenges, and support dynamic multi-agent interaction. We detail the platform design, communication protocol, set of example environments, and variety of training scenarios made possible via the toolkit.}
}
@article{JunandDezhePLoS-07,
       author = {Jun, Joseph K. and Jin, Dezhe Z.},
        title = {Development of Neural Circuitry for Precise Temporal Sequences through Spontaneous Activity, Axon Remodeling, and Synaptic Plasticity},
      journal = {{PLoS} ONE},
    publisher = {Public Library of Science},
       volume = 2,
        issue = 8,
         year = 2007,
        pages = {e723},
     abstract = {Temporally precise sequences of neuronal spikes that span hundreds of milliseconds are observed in many brain areas, including songbird premotor nucleus, cat visual cortex, and primary motor cortex. Synfire chains--networks in which groups of neurons are connected via excitatory synapses into a unidirectional chain--are thought to underlie the generation of such sequences. It is unknown, however, how synfire chains can form in local neural circuits, especially for long chains. Here, we show through computer simulation that long synfire chains can develop through spike-time dependent synaptic plasticity and axon remodeling--the pruning of prolific weak connections that follows the emergence of a finite number of strong connections. The formation process begins with a random network. A subset of neurons, called training neurons, intermittently receive superthreshold external input. Gradually, a synfire chain emerges through a recruiting process, in which neurons within the network connect to the tail of the chain started by the training neurons. The model is robust to varying parameters, as well as natural events like neuronal turnover and massive lesions. Our model suggests that long synfire chain can form during the development through self-organization, and axon remodeling, ubiquitous in developing neural circuits, is essential in the process.},
}
@article{JungandSchnitzerOPTICS-03,
        title = {Multiphoton endoscopy},
       author = {Jung, J.C. and Schnitzer, M.J.},
      journal = {Optics Letters},
       volume = 28,
        issue = 11,
        pages = {902-904},
         year = 2003,
     abstract = {Despite widespread use of multiphoton fluorescence microscopy, development of endoscopes for nonlinear optical imaging has been stymied by the degradation of ultrashort excitation pulses that occurs within optical fiber as a result of the combined effects of group-velocity dispersion and self-phase modulation. We introduce microendoscopes (350-         year =  microm in diameter) based on gradient-index microlenses that effectively eliminate self-phase modulation within the endoscope. Laser-scanning multiphoton fluorescence endoscopy exhibits micrometer-scale resolution. We used multiphoton endoscopes to image fluorescently labeled neurons and dendrites.}
}
@article{JungetalCoRR-19,
       author = {Minju Jung and Takazumi Matsumoto and Jun Tani},
        title = {Goal-Directed Behavior under Variational Predictive Coding: Dynamic Organization of Visual Attention and Working Memory},
      journal = {CoRR},
       volume = {arXiv:1903.04932},
         year = {2019},
     abstract = {Mental simulation is a critical cognitive function for goal-directed behavior because it is essential for assessing actions and their consequences. When a self-generated or externally specified goal is given, a sequence of actions that is most likely to attain that goal is selected among other candidates via mental simulation. Therefore, better mental simulation leads to better goal-directed action planning. However, developing a mental simulation model is challenging because it requires knowledge of self and the environment. The current paper studies how adequate goal-directed action plans of robots can be mentally generated by dynamically organizing top-down visual attention and visual working memory. For this purpose, we propose a neural network model based on variational Bayes predictive coding, where goal-directed action planning is formulated by Bayesian inference of latent intentional space. Our experimental results showed that cognitively meaningful competencies, such as autonomous top-down attention to the robot end effector (its hand) as well as dynamic organization of occlusion-free visual working memory, emerged. Furthermore, our analysis of comparative experiments indicated that introduction of visual working memory and the inference mechanism using variational Bayes predictive coding significantly improve the performance in planning adequate goal-directed actions.}
}
@article{JungmannetalNATURE-METHODS-14,
       author = {Jungmann, R. and Avendano, M. S. and Woehrstein, J. B. and Dai, M. and Shih, W. M. and Yin, P.},
        title = {Multiplexed {3D} Cellular Super-Resolution Imaging with {DNA-PAINT and Exchange-PAINT}},
      journal = {Nature Methods},
         year = 2014,
       volume = 11,
        issue = 3,
        pages = {313-318},
     abstract = {While super-resolution fluorescence microscopy is a powerful tool for biological research, obtaining multiplexed images for a large number of distinct target species remains challenging. Here we use the transient binding of short fluorescently labeled oligonucleotides (DNA-PAINT, point accumulation for imaging in nanoscale topography) for simple and easy-to-implement multiplexed {3D} super-resolution imaging inside fixed cells and achieve sub-10 nm spatial resolution in vitro using synthetic DNA structures. We also report a novel approach for multiplexing (Exchange-PAINT) that allows sequential imaging of multiple targets using only a single dye and a single laser source. We experimentally demonstrate ten-color super-resolution imaging in vitro on synthetic DNA structures and four-color imaging of proteins in a fixed cell.},
}
@book{JurafskyandMartin2018natural,
  title = {Speech and Language Processing: An Introduction to Natural Language Processing, Computational Linguistics, and Speech Recognition},
       author = {Daniel Jurafsky and James H. Martin},
      comment = {Third Edition},
          url = {https://web.stanford.edu/~jurafsky/slp3/},
         year = {2018},
    publisher = {Pearson}
}
@article{JustandBauerHBM-15,
        title = {Monitoring the growth of the neural representations of new animal concepts},
       author = {Marcel Just and Andrew Bauer},
      journal = {Human Brain Mapping},
    publisher = {Wiley Periodicals},
         year = 2015,
     abstract = {Although enormous progress has recently been made in identifying the neural representations of individual object concepts, relatively little is known about the growth of a neural knowledge representation as a novel object concept is being learned. In this fMRI study, the growth of the neural representations of eight individual extinct animal concepts was monitored as participants learned two features of each animal, namely its habitat (i.e., a natural dwelling or scene) and its diet or eating habits. Dwelling/scene information and diet/eating-related information have each been shown to activate their own characteristic brain regions. Several converging methods were used here to capture the emergence of the neural representation of a new animal feature within these characteristic, a priori-specified brain regions. These methods include statistically reliable identification (classification) of the eight newly acquired multivoxel patterns, analysis of the neural representational similarity among the newly learned animal concepts, and conventional GLM assessments of the activation in the critical regions. Moreover, the representation of a recently learned feature showed some durability, remaining intact after another feature had been learned. This study provides a foundation for brain research to trace how a new concept makes its way from the words and graphics used to teach it, to a neural representation of that concept in a learner's brain.}
}
@book{KR-89,
       editor = {Brachman, Ronald J. and Levesque, Hector J. and Reiter, Raymond},
        title = {Proceedings of the First International Conference on Principles of Knowledge Representation and Reasoning},
    booktitle = {Proceedings of the First International Conference on Principles of Knowledge Representation and Reasoning},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1989,
}
@article{KabanandGirolamiJIIS-03,
        title = {A Dynamic Probabilistic Model to Visualise Topic Evolution in Text Streams},
       author = {Kab\'{a}n, A. and Girolami, M.},
      journal = {Journal of Intelligent Information Systems},
      comment = {Modeling homogeneous temporal segments in data stream},
       volume = 18,
        issue = {2-3},
        pages = {107-125},
         year = 2003,
}
@inproceedings{KabanzaAAAI-90,
       author = {Kabanza, F.},
        title = {Synthesis of Reactive Plans for Multi-Path Environments},
    booktitle = {Proceedings AAAI-90},
     location = {Boston, Massachusetts},
 organization = {AAAI},
        pages = {164-169},
         year = 1990,
}        
@book{Kabat2013,
        title = {Full Catastrophe Living (Revised Edition): Using the Wisdom of Your Body and Mind to Face Stress, Pain, and Illness},
       author = {Kabat-Zinn, J. and Hanh, T.N.},
         year = {2013},
    publisher = {Random House Publishing Group}
}
@book{Kaelbling92,
       author = {Kaelbling, Leslie Pack},
        title = {Learning in Embedded Systems},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1992,
}
@inproceedings{KaelblingAAAI-88,
       author = {Kaelbling, Leslie Pack},
        title = {Goals as Parallel Program Specifications},
    booktitle = {Proceedings AAAI-88},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
        pages = {60-65},
         year = 1988,
}
@inproceedings{KaelblingICML-93,
       author = {Kaelbling, Leslie Pack},
        title = {Hierarchical Reinforcement Learning: A Preliminary Report},
    booktitle = {Proceedings Tenth International Conference on Machine Learning},
        pages = {167-173},
         year = 1993,
     abstract = {This paper presents the HDG learning algorithm, which uses a hierarchical decomposition of the state space to make learning to achieve goals more efficient with a small penalty in path quality. Special care must be taken when performing hierarchical planning and learning in stochastic domains, because macro-operators cannot be executed ballistically. The HDG algorithm, which is a descendent of Watkins' Q-learning algorithm, is described here and preliminary empirical results are presented.}
}
@inproceedings{KaelblingML-93,
       author = {Kaelbling, Leslie Pack},
        title = {Hierarchical Learning in Stochastic Domains: A Preliminary Report},
    booktitle = {Proceedings Tenth International Conference on Machine Learning},
        pages = {167-173},
         year = 1993,
     abstract = {This paper presents the HDG learning algorithm, which uses a hierarchical decomposition of the state space to make learning to achieve goals more efficient with a small penalty in path quality. Special care must be taken when performing hierarchical planning and learning in stochastic domains, because macro-operators cannot be executed ballistically. The HDG algorithm, which is a descendent of Watkins' Q-learning algorithm, is described here and preliminary empirical results are presented.}
}
@incollection{KaelblingPW-86,
       author = {Kaelbling, Leslie Pack},
        title = {An Architecture for Intelligent Reactive Systems},
     crossref = {GeorgeffandLanskyPW-86},
        pages = {395-410},
}
@techreport{KaelblingPhD,
       author = {Kaelbling, Leslie Pack},
        title = {Learning in Embedded Systems},
         type = {{Ph.D.} Thesis},
  institution = {Stanford University},
         year = 1990,
}
@article{Kaelblingetal96,
       author = {Leslie Pack Kaelbling and Michael L. Littman and Andrew P. Moore},
        title = {Reinforcement Learning: A Survey},
      journal = {Journal of Artificial Intelligence Research},
       volume = 4,
        pages = {237-285},
          url = {http://citeseer.nj.nec.com/kaelbling96reinforcement.html},
         year = 1996,
}
@article{KaelblingetalAIJ-98,
       author = {Leslie Pack Kaelbling and Michael L. Littman and Anthony R. Cassandra},
        title = {Planning and acting in partially observable stochastic domains},
      journal = AIJ,
       volume = 101,
         year = 1998,
        pages = {99-134},
}
@article{KaelblingetalJAIR-96,
       author = {Kaelbling, Leslie P. and Littman, Michael L. and Moore, Andrew W.},
        title = {Reinforcement Learning: A Survey},
      journal = jair,
       volume = 4,
        pages = {237-285},
         year = 1996,
}
@inproceedings{KaelblingetalYale-92,
       author = {Kaelbling, Leslie and Basye, Kenneth and Dean, Thomas},
        title = {Learning Labelled Graphs from Noisy Data},
    booktitle = {Proceedings of the 7th Yale Workshop on Adaptive and Learning Systems},
     location = {Yale University, New Haven, Connecticut},
 organization = {Center for Systems Science},
        pages = {149-154},
         year = 1992,
}
@article{KaewTrakulPongandBowdenICV-03,
       author = {Pakorn KaewTrakulPong and Richard Bowden},
        title = {A real time adaptive visual surveillance system for tracking low-resolution colour targets in dynamically changing scenes},
      journal = {Image Vision Computing},
       volume = 21,
        issue = 10,
         year = 2003,
        pages = {913-929},
}
@article{KahnandGorryAIJ-77,
       author = {Kahn, Kenneth and Gorry, G. Anthony},
        title = {Mechanizing Temporal Knowledge},
      journal = AIJ,
       volume = 9,
         year = 1977,
        pages = {87-108},
}	
@book{Kahneman2011,
       author = {Daniel Kahneman},
        title = {Thinking Fast and Slow},
    publisher = {Farrar, Straus and Giroux},
      address = {New York},
         year = 2011,
}
@article{KahnemanandLovalloMS-93,
        title = {Timid choices and bold forecasts: A cognitive perspective on risk-taking},
       author = {Kahneman, D. and Lovallo, D.},
      journal = {Management Science},
       volume = 39,
        issue = 1,
         year = 1993,
        pages = {17-31},
}
@article{KahnemanandTverskyAP-84,
       author = {Kahneman, D. and Tversky, A.},
        title = {Choices, Values, and Frames},
      journal = {American Psychologist},
       volume = {39},
       number = {4},
         year = {1984},
        pages = {341-350},
     abstract = {Discusses the cognitive and the psychophysical determinants of choice in risky and riskless contexts. The psychophysics of value induce risk aversion in the domain of gains and risk seeking in the domain of losses. The psychophysics of chance induce overweighting of sure things and of improbable events, relative to events of moderate probability. Decision problems can be described or framed in multiple ways that give rise to different preferences, contrary to the invariance criterion of rational choice. The process of mental accounting, in which people organize the outcomes of transactions, explains some anomalies of consumer behavior. In particular, the acceptability of an option can depend on whether a negative outcome is evaluated as a cost or as an uncompensated loss. The relationships between decision values and experience values and between hedonic experience and objective states are discussed}
}
@article{KahnemanandTverskyCP-73,
        title = {Availability: A heuristic for judging frequency and probability},
       author = {Kahneman, Daniel and Tversky, Amos},
      journal = {Cognitive Psychology},
       volume = 5,
        issue = 2,
         year = 1973,
        pages = {207-232},
}
@article{KahnemanandTverskyECON-79,
        title = {Prospect theory: An analysis of decisions under risk},
       author = {Kahneman, Daniel and Tversky, Amos},
      journal = {Econometrica},
       volume = 47,
        issue = 2,
         year = 1979,
        pages = {263-291},
}
@article{KaiserCoRR-11,
       author = {Marcus Kaiser},
        title = {A Tutorial in Connectome Analysis: Topological and Spatial Features of Brain Networks},
      journal = {CoRR},
       volume = {arXiv:1105.4705},
         year = {2011},
     abstract = {High-throughput methods for yielding the set of connections in a neural system, the connectome, are now being developed. This tutorial describes ways to analyze the topological and spatial organization of the connectome at the macroscopic level of connectivity between brain regions as well as the microscopic level of connectivity between neurons. We will describe topological features at three different levels: the local scale of individual nodes, the regional scale of sets of nodes, and the global scale of the complete set of nodes in a network. Such features can be used to characterize components of a network and to compare different networks, e.g. the connectome of patients and control subjects for clinical studies. At the global scale, different types of networks can be distinguished and we will describe Erd\"{o}s-R\'{e}nyi random, scale-free, small-world, modular, and hierarchical archetypes of networks. Finally, the connectome also has a spatial organization and we describe methods for analyzing wiring lengths of neural systems. As an introduction for new researchers in the field of connectome analysis, we discuss the benefits and limitations of each analysis approach.}
}
@article{KaiseretalCoRR-19,
       author = {Lukasz Kaiser and Mohammad Babaeizadeh and Piotr Milos and Blazej Osinski and Roy H. Campbell and Konrad Czechowski and Dumitru Erhan and Chelsea Finn and Piotr Kozakowski and Sergey Levine and Ryan Sepassi and George Tucker and Henryk Michalewski},
        title = {Model-Based Reinforcement Learning for Atari},
      journal = {CoRR},
       volume = {arXiv:1903.00374},
         year = {2019},
     abstract = {Model-free reinforcement learning (RL) can be used to learn effective policies for complex tasks, such as Atari games, even from image observations. However, this typically requires very large amounts of interaction -- substantially more, in fact, than a human would need to learn the same games. How can people learn so quickly? Part of the answer may be that people can learn how the game works and predict which actions will lead to desirable outcomes. In this paper, we explore how video prediction models can similarly enable agents to solve Atari games with fewer interactions than model-free methods. We describe Simulated Policy Learning (SimPLe), a complete model-based deep RL algorithm based on video prediction models and present a comparison of several model architectures, including a novel architecture that yields the best results in our setting. Our experiments evaluate SimPLe on a range of Atari games in low data regime of 100K interactions between the agent and the environment, which corresponds to two hours of real-time play. In most games SimPLe outperforms state-of-the-art model-free algorithms, in some games by over an order of magnitude.}
}
@misc{KakadePERFORMANCE-DILEMMA-03,
        title = {On the Sample Complexity of Reinforcement Learning},
       author = {Sham Machandranath Kakade},
 howpublished = {PhD Thesis - Gatsby Computational Neuroscience Unit, University College London},
         year = {2003}
}
@book{Kaku2015future,
        title = {The Future of the Mind: The Scientific Quest to Understand, Enhance, and Empower the Mind},
       author = {Kaku, Michio},
         year = {2015},
    publisher = {Anchor Books, a division of Random House LLC}
}
@inproceedings{KalaiSTOC-92,
       author = {G. Kalai},
        title = {A subexponential randomized simplex algorithm},
    booktitle = {Proceedings of 24th Annual {ACM} Symposium on the Theory of Computing},
     location = {Victoria, Canada},
        pages = {475-482},
         year = 1992,
}
@inproceedings{KalaletalCVPR-10,
       author = {Kalal, Z. and Matas, J. and Mikolajczyk, K.},
    booktitle = {Computer Vision and Pattern Recognition (CVPR), 2010 IEEE Conference on},
        title = {P-N learning: Bootstrapping binary classifiers by structural constraints},
         year = {2010},
        pages = {49-56},
     abstract = {This paper shows that the performance of a binary classifier can be significantly improved by the processing of structured unlabeled data, i.e. data are structured if knowing the label of one example restricts the labeling of the others. We propose a novel paradigm for training a binary classifier from labeled and unlabeled examples that we call P-N learning. The learning process is guided by positive (P) and negative (N) constraints which restrict the labeling of the unlabeled set. P-N learning evaluates the classifier on the unlabeled data, identifies examples that have been classified in contradiction with structural constraints and augments the training set with the corrected samples in an iterative process. We propose a theory that formulates the conditions under which P-N learning guarantees improvement of the initial classifier and validate it on synthetic and real data. P-N learning is applied to the problem of on-line learning of object detector during tracking. We show that an accurate object detector can be learned from a single example and an unlabeled video sequence where the object may occur. The algorithm is compared with related approaches and state-of-the-art is achieved on a variety of objects (faces, pedestrians, cars, motorbikes and animals).},
}
@article{KalaletalPAMI-12,
       author = {Kalal, Z. and Mikolajczyk, K. and Matas, J.},
      journal = {Pattern Analysis and Machine Intelligence, IEEE Transactions on},
        title = {Tracking-Learning-Detection},
         year = {2012},
       volume = {34},
       number = {7},
        pages = {1409-1422},
     abstract = {This paper investigates long-term tracking of unknown objects in a video stream. The object is defined by its location and extent in a single frame. In every frame that follows, the task is to determine the object's location and extent or indicate that the object is not present. We propose a novel tracking framework (TLD) that explicitly decomposes the long-term tracking task into tracking, learning, and detection. The tracker follows the object from frame to frame. The detector localizes all appearances that have been observed so far and corrects the tracker if necessary. The learning estimates the detector's errors and updates it to avoid these errors in the future. We study how to identify the detector's errors and learn from them. We develop a novel learning method (P-N learning) which estimates the errors by a pair of "experts": (1) P-expert estimates missed detections, and (2) N-expert estimates false alarms. The learning process is modeled as a discrete dynamical system and the conditions under which the learning guarantees improvement are found. We describe our real-time implementation of the TLD framework and the P-N learning. We carry out an extensive quantitative evaluation which shows a significant improvement over state-of-the-art approaches.},
}
@article{KalchbrennerandBlunsomACL-13,
        title = {Recurrent Convolutional Neural Networks for Discourse Compositionality},
       author = {Kalchbrenner, Nal and Blunsom, Phil},
         year = 2013,
      journal = {Proceedings of the 2013 Workshop on Continuous Vector Space Models and their Compositionality},
    publisher = {Association for Computational Linguistics},
     abstract = {The compositionality of meaning extends beyond the single sentence. Just as words combine to form the meaning of sentences, so do sentences combine to form the meaning of paragraphs, dialogues and general discourse. We introduce both a sentence model and a discourse model corresponding to the two levels of compositionality. The sentence model adopts convolution as the central operation for composing semantic vectors and is based on a novel hierarchical convolutional neural network. The discourse model extends the sentence model and is based on a recurrent neural network that is conditioned in a novel way both on the current sentence and on the current speaker. The discourse model is able to capture both the sequentiality of sentences and the interaction between different speakers. Without feature engineering or pretraining and with simple greedy decoding, the discourse model coupled to the sentence model obtains state of the art performance on a dialogue act classification experiment.},
}
@inproceedings{KalchbrennerandBlunsomEMNLP-13,
        title = {Recurrent Continuous Translation Models},
       author = {Kalchbrenner, Nal and Blunsom, Phil},
    booktitle = {Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing},
    publisher = {Association for Computational Linguistics},
         year = 2013,
     abstract = {We introduce a class of probabilistic continuous translation models called Recurrent Continuous Translation Models that are purely based on continuous representations for words, phrases and sentences and do not rely on alignments or phrasal translation units. The models have a generation and a conditioning aspect. The generation of the translation is modelled with a target Recurrent Language Model, whereas the conditioning on the source sentence is modelled with a Convolutional Sentence Model. Through various experiments, we show first that our models obtain a perplexity with respect to gold translations that is > 43\% lower than that of state-of-the-art alignment-based translation models. Secondly, we show that they are remarkably sensitive to the word order, syntax, and meaning of the source sentence despite lacking alignments. Finally we show that they match a state-of-the-art system when rescoring n-best lists of translations.},
}
@article{KalchbrenneretalACL-14,
        title = {A Convolutional Neural Network for Modelling Sentences},
       author = {Nal Kalchbrenner and Edward Grefenstette and Phil Blunsom},
         year = 2014,
      journal = {Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics},
     location = {Baltimore, USA},
    publisher = {Association for Computational Linguistics},
     abstract = {The ability to accurately represent sentences is central to language understanding. We describe a convolutional architecture dubbed the Dynamic Convolutional Neural Network (DCNN) that we adopt for the semantic modelling of sentences. The network uses Dynamic k-Max Pooling, a global pooling operation over linear sequences. The network handles input sentences of varying length and induces a feature graph over the sentence that is capable of explicitly capturing short and long-range relations. The network does not rely on a parse tree and is easily applicable to any language. We test the DCNN in four experiments: small scale binary and multi-class sentiment prediction, six-way question classification and Twitter sentiment prediction by distant supervision. The network achieves excellent performance in the first three tasks and a greater than 25\% error reduction in the last task with respect to the strongest baseline.},
}
@article{KaliandDayanNATURE-NEUROSCIENCE-04,
       author = {K\'{a}li, Szabolcs and Dayan, Peter},
        title = {Off-line replay maintains declarative memories in a model of hippocampal-neocortical interactions},
      journal = {Nature Neuroscience},
         year = {2004},
       volume = {7},
        issue = {3},
        pages = {286-294},
     abstract = {During sleep, neural activity in the hippocampus and neocortex seems to recapitulate aspects of its earlier, awake form. This replay may be a substrate for the consolidation of long-term declarative memories, whereby they become independent of the hippocampus and are stored in neocortex. In contrast to storage, other crucial facets of competent long-term memory, such as maintenance of access to stored traces and preservation of their correct interpretation, have received little attention. We investigate long-term episodic and semantic memory in a theoretical model of neocortical-hippocampal interaction. We find that, in the absence of regular hippocampal reactivation, even supposedly consolidated episodic memories are fragile in the face of cortical semantic plasticity. Replay allows access to episodes stored in the hippocampus to be maintained, by keeping them in appropriate register with changing neocortical representations. Hippocampal storage and replay also has a constructive role in the recall of structured, semantic information.},
}
@book{Kalmanetal69,
       author = {Kalman, R. E. and Falb, P. L. and Arbib, M. A.},
        title = {Topics in Mathematical System Theory},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1969,
}
@article{KambhampatiSIGART-95,
       author = {Subbarao Kambhampati},
        title = {A comparative analysis of partial-order planning and task-reduction planning},
      journal = {{ACM} {SIGART} Bulletin},
         year = 1995,
       volume = 6,
        issue = 1,
}
@article{KambhampatiandHendlerAIJ-92,
       author = {Kambhampati, Subbarao and Hendler, James},
        title = {A Validation Structure Based Theory of Plan Modification and Reuse},
      journal = AIJ,
       volume = 55,
        issue = {2-3},
        pages = {193-258},
         year = 1992,
}
@article{KambhampatietalAIJ-95,
       author = {Subbarao Kambhampati and Craig Knoblock and Qiang Yang},
        title = {Refinement Search as a unifying framework for evaluating design tradeoffs in partial order planning},
      journal = AIJ,
       volume = 76,
        issue = {1-2},
        pages = {167-238},
         year = 1995,
}
@techreport{KammererPhD-15,
       author = {Axel Kammerer},
        title = {Memory Capacity in the Hippocampus: Influence of Inhibition and Remapping},
        issue = {{PhD Dissertation}},
  institution = {Graduate School of Systemic Neurosciences der Ludwig-Maximilians-Universit\"{a}t M\"{U}unchen},
         year = 2015,
     abstract = {Neural assemblies in hippocampus encode positions. During rest, the hippocampus replays sequences of neural activity seen during awake behavior. This replay is linked to memory consolidation and mental exploration of the environment. Recurrent networks can be used to model the replay of sequential activity. Multiple sequences can be stored in the synaptic connections. To achieve a high memory capacity, recurrent networks require a pattern separation mechanism. Such a mechanism is global remapping, observed in place cell populations. A place cell fires at a particular position of an environment and is silent elsewhere. Multiple place cells usually cover an environment with their firing fields. Small changes in the environment or context of a behavioral task can cause global remapping, i.e. profound changes in place cell firing fields. Global remapping causes some cells to cease firing, other silent cells to gain a place field, and other place cells to move their firing field and change their peak firing rate. The effect is strong enough to make global remapping a viable pattern separation mechanism.

We model two mechanisms that improve the memory capacity of recurrent networks. The effect of inhibition on replay in a recurrent network is modeled using binary neurons and binary synapses. A mean field approximation is used to determine the optimal parameters for the inhibitory neuron population. Numerical simulations of the full model were carried out to verify the predictions of the mean field model. A second model analyzes a hypothesized global remapping mechanism, in which grid cell firing is used as feed forward input to place cells. Grid cells have multiple firing fields in the same environment, arranged in a hexagonal grid. Grid cells can be used in a model as feed forward inputs to place cells to produce place fields. In these grid-to-place cell models, shifts in the grid cell firing patterns cause remapping in the place cell population. We analyze the capacity of such a system to create sets of separated patterns, i.e. how many different spatial codes can be generated. The limiting factor are the synapses connecting grid cells to place cells. To assess their capacity, we produce different place codes in place and grid cell populations, by shuffling place field positions and shifting grid fields of grid cells. Then we use Hebbian learning to increase the synaptic weights between grid and place cells for each set of grid and place code. The capacity limit is reached when synaptic interference makes it impossible to produce a place code with sufficient spatial acuity from grid cell firing. Additionally, it is desired to also maintain the place fields compact, or sparse if seen from a coding standpoint. Of course, as more environments are stored, the sparseness is lost. Interestingly, place cells lose the sparseness of their firing fields much earlier than their spatial acuity.

For the sequence replay model we are able to increase capacity in a simulated recurrent network by including an inhibitory population. We show that even in this more complicated case, capacity is improved. We observe oscillations in the average activity of both excitatory and inhibitory neuron populations. The oscillations get stronger at the capacity limit. In addition, at the capacity limit, rather than observing a sudden failure of replay, we find sequences are replayed transiently for a couple of time steps before failing. Analyzing the remapping model, we find that, as we store more spatial codes in the synapses, first the sparseness of place fields is lost. Only later do we observe a decay in spatial acuity of the code. We found two ways to maintain sparse place fields while achieving a high capacity: inhibition between place cells, and partitioning the place cell population so that learning affects only a small fraction of them in each environment. We present scaling predictions that suggest that hundreds of thousands of spatial codes can be produced by this pattern separation mechanism. 

The effect inhibition has on the replay model is two-fold. Capacity is increased, and the graceful transition from full replay to failure allows for higher capacities when using short sequences. Additional mechanisms not explored in this model could be at work to concatenate these short sequences, or could perform more complex operations on them. The interplay of excitatory and inhibitory populations gives rise to oscillations, which are strongest at the capacity limit. The oscillation draws a picture of how a memory mechanism can cause hippocampal oscillations as observed in experiments. In the remapping model we showed that sparseness of place cell firing is constraining the capacity of this pattern separation mechanism. Grid codes outperform place codes regarding spatial acuity, as shown in Mathis et al. (2012). Our model shows that the grid-to-place transformation is not harnessing the full spatial information from the grid code in order to maintain sparse place fields. This suggests that the two codes are independent, and communication between the areas might be mostly for synchronization. High spatial acuity seems to be a specialization of the grid code, while the place code is more suitable for memory tasks. 

In a detailed model of hippocampal replay we show that feedback inhibition can increase the number of sequences that can be replayed. The effect of inhibition on capacity is determined using a meanfield model, and the results are verified with numerical simulations of the full network. Transient replay is found at the capacity limit, accompanied by oscillations that resemble sharp wave ripples in hippocampus. In a second model Hippocampal replay of neuronal activity is linked to memory consolidation and mental exploration. Furthermore, replay is a potential neural correlate of episodic memory. To model hippocampal sequence replay, recurrent neural networks are used. Memory capacity of such networks is of great interest to determine their biological feasibility. And additionally, any mechanism that improves capacity has explanatory power. We investigate two such mechanisms. The first mechanism to improve capacity is global, unspecific feedback inhibition for the recurrent network. In a simplified meanfield model we show that capacity is indeed improved.

The second mechanism that increases memory capacity is pattern separation. In the spatial context of hippocampal place cell firing, global remapping is one way to achieve pattern separation. Changes in the environment or context of a task cause global remapping. During global remapping, place cell firing changes in unpredictable ways: cells shift their place fields, or fully cease firing, and formerly silent cells acquire place fields. Global remapping can be triggered by subtle changes in grid cells that give feed-forward inputs to hippocampal place cells. We investigate the capacity of the underlying synaptic connections, defined as the number of different environments that can be represented at a given spatial acuity. We find two essential conditions to achieve a high capacity and sparse place fields: inhibition between place cells, and partitioning the place cell population so that learning affects only a small fraction of them in each environments. We also find that sparsity of place fields is the constraining factor of the model rather than spatial acuity. Since the hippocampal place code is sparse, we conclude that the hippocampus does not fully harness the spatial information available in the grid code. The two codes of space might thus serve different purposes.},
}
@article{KampsetalCOGNITION-18,
       author = {Kamps, Frederik S. and Julian, Joshua B. and Battaglia, Peter and Landau, Barbara and Kanwisher, Nancy and Dilks, Daniel D.},
        title = {Dissociating intuitive physics from intuitive psychology: Evidence from Williams syndrome},
      journal = {Cognition},
       volume = {168},
         year = {2017},
        pages = {146-153},
     abstract = {Prior work suggests that our understanding of how things work (``intuitive physics'') and how people work (``intuitive psychology'') are distinct domains of human cognition. Here we directly test the dissociability of these two domains by investigating knowledge of intuitive physics and intuitive psychology in adults with Williams syndrome (WS) - a genetic developmental disorder characterized by severely impaired spatial cognition, but relatively spared social cognition. WS adults and mental-age matched (MA) controls completed an intuitive physics task and an intuitive psychology task. If intuitive physics is a distinct domain (from intuitive psychology), then we should observe differential impairment on the physics task for individuals with WS compared to MA controls. Indeed, adults with WS performed significantly worse on the intuitive physics than the intuitive psychology task, relative to controls. These results support the hypothesis that knowledge of the physical world can be disrupted independently from knowledge of the social world.},
}
@inproceedings{KanazaandDeanIJCAI-89,
       author = {Kanazawa, Keiji and Dean, Thomas},
        title = {A Model for Projection and Action},
    booktitle = {Proceedings {IJCAI}-89},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
     location = {Detroit, Michigan},
 organization = {IJCAI},
         year = 1989,
        pages = {985-990},
}
@inproceedings{KanazawaAAAI-91,
       author = {Kanazawa, Keiji},
        title = {A logic and time nets for probabilistic inference},
    booktitle = {Proceedings AAAI-91},
     location = {Anaheim, California},
 organization = {AAAI},
         year = 1991,
        pages = {360-365},
}
@phdthesis{KanazawaPhD,
       author = {Kanazawa, Keiji},
        title = {Reasoning about Time and Probability},
       school = {Brown University},
         year = 1991,
      address = {Providence, RI},
} 
@inproceedings{KanazawaetalAAAI-92,
       author = {Kanazawa, Keiji and Dean, Thomas and Kirman, Jak},
        title = {Compact Probabilistic Network Models for Semi-{M}arkov Processes},
    booktitle = {Submitted to {AAAI}-92},
     location = {San Jose, California},
 organization = {AAAI},
         year = 1992,
}
@book{Kandel2006,
       author = {Eric R. Kandel},
        title = {In Search of Memory: The Emergence of a New Scince of Mind},
    publisher = {W. W. Norton},
      address = {New York, NY},
         year = 2006,
}
@book{Kandel2018disordered,
        title = {The Disordered Mind: What Unusual Brains Tell Us About Ourselves},
       author = {Kandel, Eric R.},
         year = {2018},
    publisher = {Farrar, Straus and Giroux}
}
@misc{KandelNOBEL-PRIZE-2000,
        title = {The Nobel Prize in Physiology or Medicine 2000 Awarded to Eric Kandel},
       author = {Eric R. Kandel},
 howpublished = {\url{https://www.nobelprize.org/prizes/medicine/2000/kandel/facts/}},
         year = {2020},
}
@book{KandelSchwartzandJessell2000,
        title = {Principles of neural science (Fourth Edition)},
       author = {Kandel, E.R. and Schwartz, J.H. and Jessell, T.M.},
    publisher = {McGraw-Hill, Health Professions Division},
         year = 2000,
}
@book{KandelSchwartzandJessell2012,
        title = {Principles of neural science (Fifth Edition)},
       author = {Kandel, E.R. and Schwartz, J.H. and Jessell, T.M. and Siegelbaum, S.A. and Hudspeth, A. J.},
    publisher = {McGraw-Hill, Health Professions Division},
         year = 2012,
}
@article{KanervaCC-09,
        title = {Hyperdimensional Computing: An Introduction to Computing in Distributed Representation with High-Dimensional Random Vectors},
       author = {Kanerva, Pentti},
      journal = {Cognitive Computation},
       volume = 1,
        issue = 2,
         year = 2009,
        pages = {139-159},
}
@inproceedings{KanervaICANN-94,
       author = {Kanerva, P.},
        title = {The binary spatter code for encoding concepts at many levels},
    booktitle = {Proceedings of International Conference on Artificial Neural Networks},
    publisher = {Springer-Verlag},
       editor = {Marinaro, M. and Morasso, P.},
         year = 1994,
        pages = {226-9},
     abstract = {The Spatter Code is a high-dimensional (e.g., N = 10,000), random code that encodes "high-level concepts" in terms of their "low-level attributes" so that concepts at different levels can be mixed freely. The binary spatter code is the simplest. It has two N-bit codewords for each concept or item, a "high-level," or dense, word with many randomly placed 1s and a "low-level," or sparse, word with a few (that are contained in the many). The dense codewords can be used as inputs to an associative memory. The sparse codewords are used in encoding new concepts. When several items (attributes, concepts, chunks) are combined to form a new item, the two codewords for the new item are made from the sparse codewords of its constituents as follows: the new dense word is the logical OR of the constituents (i.e., their sum thresholded at 0.5), and the new sparse word has 1s where the constituent words overlap (i.e., their sum thresholded at 1.5). When the parameters for the code are chosen properly, the number of 1s in the codewords is maintained as new items are encoded from combinations of old ones.},
}
@inproceedings{KanevaetalIEEE-10,
       author = {Biliana Kaneva and Josef Sivic and Antonio Torralba and Shai Avidan and William T. Freeman},
        title = {Infinite Images: Creating and Exploring a Large Photorealistic Virtual Space},
    booktitle = {Proceedings of the {IEEE}},
         year = 2010
}
@article{KangetalNATURE-11,
        title = {Spatio-temporal transcriptome of the human brain},
       author = {Kang, Hyo Jung and Kawasawa, Yuka Imamura and Cheng, Feng and Zhu Ying and Xu, Xuming and Li, Mingfeng and Sousa, Andre M. M. and Pletikos, Mihovil and Meyer, Kyle A. and Sedmak, Goran and Guennel Tobias and Shin, Yurae and Johnson, Matthew B. and Krsnik, Zeljka and Mayer, Simone and Fertuzinhos, Sofia and Umlauf, Sheila and Lisgo, Steven N. and Vortmeyer, Alexander and Weinberger, Daniel R. and Mane, Shrikant and Hyde, Thomas M. and Huttner, Anita and Reimers, Mark and Kleinman, Joel E. and Sestan, Nenad},
      journal = {Nature},
       volume = 478,
        issue = 7373,
         year = 2011,
        pages = {483-489},
     abstract = {Brain development and function depend on the precise regulation of gene expression. However, our understanding of the complexity and dynamics of the transcriptome of the human brain is incomplete. Here we report the generation and analysis of exon-level transcriptome and associated genotyping data, representing males and females of different ethnicities, from multiple brain regions and neocortical areas of developing and adult post-mortem human brains. We found that 86 per cent of the genes analysed were expressed, and that 90 per cent of these were differentially regulated at the whole-transcript or exon level across brain regions and/or time. The majority of these spatio-temporal differences were detected before birth, with subsequent increases in the similarity among regional transcriptomes. The transcriptome is organized into distinct co-expression networks, and shows sex-biased gene expression and exon usage. We also profiled trajectories of genes associated with neurobiological categories and diseases, and identified associations between single nucleotide polymorphisms and gene expression. This study provides a comprehensive data set on the human brain transcriptome and insights into the transcriptional foundations of human neurodevelopment.},
}
@article{KaniandElsheikhetalCoRR-17,
       author = {J. Nagoor Kani and Ahmed H. Elsheikh},
        title = {{DR-RNN:} {A} deep residual recurrent neural network for model reduction},
      journal = {CoRR},
       volume = {arXiv:1709.00939},
          url = {http://arxiv.org/abs/1709.00939},
     abstract = {We introduce a deep residual recurrent neural network (DR-RNN) as an efficient model reduction technique for nonlinear dynamical systems. The developed DR-RNN is inspired by the iterative steps of line search methods in finding the residual minimiser of numerically discretized differential equations. We formulate this iterative scheme as stacked recurrent neural network (RNN) embedded with the dynamical structure of the emulated differential equations. Numerical examples demonstrate that DR-RNN can effectively emulate the full order models of nonlinear physical systems with a significantly lower number of parameters in comparison to standard RNN architectures. Further, we combined DR-RNN with Proper Orthogonal Decomposition (POD) for model reduction of time dependent partial differential equations. The presented numerical results show the stability of proposed DR-RNN as an explicit reduced order technique. We also show significant gains in accuracy by increasing the depth of proposed DR-RNN similar to other applications of deep learning.}
}
@article{Kanizsa55,
       author = {Gaetano Kanizsa},
        title = {Margini quasi-percettivi in campi con stimolazione omogenea},
      journal = {Rivista di Psicologia},
       volume = 49,
        issue = 1,
        pages = {7-30},
         year = 1955,
}
@inproceedings{KanskyetalICMLR-17,
       author = {Ken Kansky and Tom Silver and David A. M{\'{e}}ly and Mohamed Eldawy and Miguel L{\'{a}}zaro-Gredilla and Xinghua Lou and Nimrod Dorfman and Szymon Sidor and D. Scott Phoenix and Dileep George},
        title = {Schema Networks: Zero-shot Transfer with a Generative Causal Model of Intuitive Physics},
      journal = {Internation Conference on Machine Learning and Representation},
         year = {2017},
     abstract = {The recent adaptation of deep neural network-based methods to reinforcement learning and planning domains has yielded remarkable progress on individual tasks. Nonetheless, progress on task-to-task transfer remains limited. In pursuit of efficient and robust generalization, we introduce the Schema Network, an object-oriented generative physics simulator capable of disentangling multiple causes of events and reasoning backward through causes to achieve goals. The richly structured architecture of the Schema Network can learn the dynamics of an environment directly from data. We compare Schema Networks with Asynchronous Advantage Actor-Critic and Progressive Networks on a suite of Breakout variations, reporting results on training efficiency and zero-shot generalization, consistently demonstrating faster, more robust learning and better transfer. We argue that generalizing from limited data and learning causal relationships are essential abilities on the path toward generally intelligent systems.},
} 
@book{Kant1781critiqueofpurereason,
       author = {Immanuel Kant},
        title = {Critique of Pure Reason},
    publisher = {Reprinted 1965, St. Martin's Press},
      address = {New York, NY},
         year = 1781,
}
@book{Kant65,
       author = {Immanuel Kant},
        title = {Critique of Pure Reason},
    publisher = {Reprinted 1965, St. Martin's Press},
      address = {New York, NY},
         year = 1781,
}
@article{KantCoRR-18,
       author = {Neel Kant},
        title = {Recent Advances in Neural Program Synthesis},
      journal = {CoRR},
       volume = {arXiv:1802.02353},
         year = {2018},
     abstract = {In recent years, deep learning has made tremendous progress in a number of fields that were previously out of reach for artificial intelligence. The successes in these problems has led researchers to consider the possibilities for intelligent systems to tackle a problem that humans have only recently themselves considered: program synthesis. This challenge is unlike others such as object recognition and speech translation, since its abstract nature and demand for rigor make it difficult even for human minds to attempt. While it is still far from being solved or even competitive with most existing methods, neural program synthesis is a rapidly growing discipline which holds great promise if completely realized. In this paper, we start with exploring the problem statement and challenges of program synthesis. Then, we examine the fascinating evolution of program induction models, along with how they have succeeded, failed and been reimagined since. Finally, we conclude with a contrastive look at program synthesis and future research recommendations for the field.},
}
@article{KaplanetalFiCN-13,
       author = {Kaplan, Barry B  and  Kar, Amar N  and  Gioio, Anthony  and  Aschrafi, Armaz},
        title = {MicroRNAs in the axon and presynaptic nerve terminal},
      journal = {Frontiers in Cellular Neuroscience},
       volume = {7},
       number = {126},
         year = {2013},
        pages = {1-5},
     abstract = {The distal structural/functional domains of the neuron, to include the axon and presynaptic nerve terminal, contain a large, heterogeneous population of mRNAs and an active protein synthetic system. These local components of the genetic expression machinery play a critical role in the development, function, and long-term viability of the neuron. In addition to the local mRNA populations these presynaptic domains contain a significant number of non-coding RNAs that regulate gene expression post-transcriptionally. Here, we review a small, but rapidly evolving literature on the composition and function of microRNAs that regulate gene expression locally in the axon and nerve terminal. In this capacity, these small regulatory RNAs have a profound effect on axonal protein synthesis, local energy metabolism, and the modulation of axonal outgrowth and branching.}
}
@article{KaravasilisetalJoR-19,
        title = {Ipsilateral and contralateral cerebro-cerebellar white matter connections: {A} diffusion tensor imaging study in healthy adults},
       author = {Efstratios Karavasilis and Foteini Christidi and Georgios Velonakis and Zoi Giavri and Nikolaos L. Kelekis and Efstathios P. Efstathopoulos and Ioannis Evdokimidis and Georges Dellatolas},
      journal = {Journal of Neuroradiology},
       volume = {46},
        issue = {1},
         year = {2019},
        pages = {52-60},
     abstract = {Objectives: The cerebellum has a pivotal role in regulating human behavior; yet whether this function is mediated only through contralateral cerebro-cerebellar pathways is under-investigated. Thus, we examined feed-backward and feed-forward ipsilateral and contralateral cerebro-cerebellar connections using a detereministic diffusion tensor imaging (DTI) algorithm, the robustness of which was also estimated using phantom DTI data.

Materials and methods: Fifty-one healthy controls (22-60 years old; 15 males/36 females) were scanned in a 3T MRI scanner with a 30-direction DTI sequence. Multiple region-of-interest (ROI) method was applied for the reconstruction of the ipsilateral and contralateral (based on cerebellar seed ROI) fronto-ponto-cerebellar (FPC), parieto-ponto-cerebellar (PPC), temporo-ponto-cerebellar (TPC), occipito-ponto-cerebellar (OPC) and dentate-rubro-thalamo-cortical (DRTC) tract bilaterally using the Brainance DTI Suite. A realistic diffusion MR phantom was used to evaluate the fiber tracking methodology for 16 fibers containing crossing, kissing, splitting and bending configurations.

Results: Both contralateral and ipsilateral FPC, PPC, OPC and ipsilateral DRTC tracts were successfully reconstructed; the contralateral DRTC tract was not reconstructed in all subjects. Also, the TPC tract was not reproduced in several subjects mostly regarding the contralateral connection. Descriptive DTI measures (number of fibers, fractional anisotropy, radial and axial diffusivity) are presented for each tract. Regarding phantom data, Brainance DTI Suite returned a dataset of 16 fibers that almost perfectly matched the 16 ground truth fibers.

Conclusions: We identified ipsilateral and contralateral connections using a clinically applicable DTI sequence, a robust deterministic algorithm and an unbiased methodology, which can be applied in daily practice in different brain pathologies.},
}
@inproceedings{Kargeretal97,
       author = {D. Karger and E. Lehman and F. T. Leighton and M. Levine and D. Lewin and R. Panigrahy},
        title = {Consistent hashing and random trees:  Distributed caching protocols for relieving hot spots on the {W}orld {W}ide {W}eb},
    booktitle = {Proceedings of the 29th Annual ACM Symposium on Theory of Computing},
        pages = {654-663},
         year = 1997,
}
@article{KarklinandLewickiNATURE-09,
        title = {Emergence of complex cell properties by learning to generalize in natural scenes},
       author = {Karklin, Yan and Lewicki, Michael S.},
      journal = {Nature},
       volume = 457,
        issue = 7225,
         year = 2009,
        pages = {83-86},
}
@article{KarlheinzetalCURRENT-BIOLOGY-02,
       author = {Rein, Karlheinz and Z\"{o}ckler, Malte and Mader, Michael T. and Gr\"{u}bel, Cornelia and Heisenberg, Martin},
        title = {The Drosophila Standard Brain},
      journal = {Current Biology},
    publisher = {Elsevier},
       volume = {12},
        issue = {3},
         year = {2002},
        pages = {227-231},
     abstract = {Organisms and organs come in sizes and shapes. With size, science has no problems, but how to quantify shape? How similar are two birds or two brains? This problem is particularly pressing in cases like brains where structure reflects function. The problem is not new, but satisfying solutions have yet to be worked out. For brain anatomy, no general methodology for a statistically secured quantitative description is available. Using the small brain of the fly Drosophila melanogaster, we have explored a new approach combining immunohistochemistry, high-resolution 3D confocal microscopy, and advanced graphics computing. For a genetic model organism such as Drosophila, a quantitative assessment of brain structure is particularly rewarding, since it allows for the identification of genetic variants with subtle brain structure phenotypes and, even more importantly, the organization of the wealth of gene expression patterns in the brain into a genetic atlas linking molecular and organismic gene function. We now provide a representative standard for the brain of D. melanogaster wild-type with means and variances for several aspects of its shape. Its application to volumetry, mutants, and gene expression patterns is demonstrated.},
}
@book{KarlinandTaylor75,
       author = {Karlin, Samuel and Taylor, Howard M.},
        title = {A First Course in Stochastic Processes, Second Edition},
    publisher = {Academic Press},
      address = {New York},
         year = 1975,
}
@article{KarlssonandFrankNATURE-NEUROSCIENCE-09,
        title = {Awake replay of remote experiences in the hippocampus},
       author = {Karlsson, Mattias P and Frank, Loren M},
      journal = {Nature neuroscience},
       volume = {12},
       number = {7},
        pages = {913},
         year = {2009},
    publisher = {Nature Publishing Group}
}
@article {KarlssonetalSCIENCE-12,
       author = {Karlsson, Mattias P. and Tervo, Dougal G. R. and Karpova, Alla Y.},
        title = {Network Resets in Medial Prefrontal Cortex Mark the Onset of Behavioral Uncertainty},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {338},
       number = {6103},
         year = {2012},
        pages = {135-139},
     abstract = {The ability to display behavioral flexibility depends on an internal representation of the environment-a framework of beliefs that can be adjusted on the basis of experience. Recording from multiple electrodes in the rat medial prefrontal cortex, Karlsson et al. (p. 135) investigated how ensembles of neurons changed their activity during the performance of a task in which the animal had to update its knowledge of reward contingencies. The results suggest that changes in perceived action-outcome contingencies were associated with abrupt switches in neuronal representations in the rat medial prefrontal cortex.Regions within the prefrontal cortex are thought to process beliefs about the world, but little is known about the circuit dynamics underlying the formation and modification of these beliefs. Using a task that permits dissociation between the activity encoding an animal{\textquoteright}s internal state and that encoding aspects of behavior, we found that transient increases in the volatility of activity in the rat medial prefrontal cortex accompany periods when an animal{\textquoteright}s belief is modified after an environmental change. Activity across the majority of sampled neurons underwent marked, abrupt, and coordinated changes when prior belief was abandoned in favor of exploration of alternative strategies. These dynamics reflect network switches to a state of instability, which diminishes over the period of exploration as new stable representations are formed.},
}
@article{Karmarkar84,
       author = {N. Karmarkar},
        title = {A New polynomial time Algorithm for Linear Programming},
      journal = {Combinatorica},
       volume = 4,
        issue = 4,
         year = 1984,
        pages = {373-395},
}
@article{KarnikandGerlaiBBS-12,
       author = {Karnik, I. and Gerlai, R.},
        title = {{C}an zebrafish learn spatial tasks? {A}n empirical analysis of place and single {C}{S}-{U}{S} associative learning},
      journal = {Behavioral Brain Research},
         year = {2012},
       volume = {233},
       number = {2},
        pages = {415-421},
     abstract = {The zebrafish may be an ideal tool with which genes underlying learning and memory can be identified and functionally investigated. From a translational viewpoint, relational learning and episodic memory are particularly important as their impairment is the hallmark of prevalent human neurodegenerative diseases. Recent reports suggest that zebrafish are capable of solving complex relational-type associative learning tasks, namely spatial learning tasks. However, it is not known whether good performance in these tasks was truly based upon relational learning or upon a single CS-US association. Here we study whether zebrafish can find a rewarding stimulus (sight of conspecifics) based upon a single associative cue or/and upon the location of the reward using a method conceptually similar to 'context and cue dependent fear conditioning' employed with rodents. Our results confirm that zebrafish can form an association between a salient visual cue and the rewarding stimulus and at the same time they can also learn where the reward is presented. Although our results do not prove that zebrafish form a dynamic spatial map of their surroundings and use this map to locate their reward, they do show that these fish perform similarly to rodents whose hippocampal function is unimpaired. These results further strengthen the notion that complex cognitive abilities exist in the zebrafish and thus they may be analyzed using the excellent genetic tool set developed for this simple vertebrate.}
}
@incollection{Karp72,
       author = {Karp, Richard M.},
        title = {Reducibility Among Combinatorial Problems},
       editor = {Miller, R. E. and Thatcher, J. W.},
    booktitle = {Complexity of Computer Computations},
    publisher = {Plenum Press},
     location = {New York},
         year = 1972,
        pages = {85-103},
}
@article{Karp77,
       author = {Karp, Richard},
        title = {Probabilistic Analysis of Partitioning Algorithms for the Traveling Salesman Problem in the Plane},
      journal = {Math. Oper. Res.},
       volume = 2,
         year = 1977,
        pages = {209-224},
}
@article{KarpSIAMJC-88,
       author = {Karp, Richard and Motwani, Rajeev and Prabhakar, Raghavan},
        title = {Deferred Data Structuring},
      journal = {SIAM Journal of Computing},
       volume = 17,
         year = 1988,
        pages = {883-902},
}
@misc{KarparthyCONVOLUTIONAL-NEURAL-NETWORKS-16,
        title = {Convolutional Neural Networks for Visual Recognition},
       author = {Andrej Karparthy},
 howpublished = {http://cs231n.github.io/convolutional-networks/},
         year = {2016},
}
@misc{KarparthyUNREASONABLY-EFFECTIVE-RNN-15,
        title = {The Unreasonable Effectivenss of Recursive Neutal Networks},
       author = {Andrej Karparthy},
 howpublished = {http://karpathy.github.io/2015/05/21/rnn-effectiveness/},
         year = {2015},
}
@inproceedings{KarpathayandFei-FeiCVPR-15,
        title = {Deep Visual-Semantic Alignments for Generating Image Descriptions},
       author = {Andrej Karpathy and Li Fei-Fei},
    booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
         year = 2015,
     abstract = {We present a model that generates natural language descriptions of images and their regions. Our approach leverages datasets of images and their sentence descriptions to learn about the inter-modal correspondences between language and visual data. Our alignment model is based on a novel combination of Convolutional Neural Networks over image regions, bidirectional Recurrent Neural Networks over sentences, and a structured objective that aligns the two modalities through a multimodal embedding. We then describe a Multimodal Recurrent Neural Network architecture that uses the inferred alignments to learn to generate novel descriptions of image regions. We demonstrate that our alignment model produces state of the art results in retrieval experiments on Flickr8K, Flickr30K and MSCOCO datasets. We then show that the generated descriptions significantly outperform retrieval baselines on both full images and on a new dataset of region-level annotations}
}
@inproceedings{KarpathyandFei-FeiCVPR-15,
       author = {Andrej Karpathy and Fei-Fei Li},
        title = {Deep visual-semantic alignments for generating image descriptions},
    booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
         year = {2015},
        pages = {3128-3137},
     abstract = {We present a model that generates natural language descriptions of images and their regions. Our approach leverages datasets of images and their sentence descriptions to learn about the inter-modal correspondences between language and visual data. Our alignment model is based on a novel combination of Convolutional Neural Networks over image regions, bidirectional Recurrent Neural Networks over sentences, and a structured objective that aligns the two modalities through a multimodal embedding. We then describe a Multimodal Recurrent Neural Network architecture that uses the inferred alignments to learn to generate novel descriptions of image regions. We demonstrate that our alignment model produces state of the art results in retrieval experiments on Flickr8K, Flickr30K and MSCOCO datasets. We then show that the generated descriptions significantly outperform retrieval baselines on both full images and on a new dataset of region-level annotations.}
}
@article{KarpathyandFei-FeiPAMI-16,
       author = {A. Karpathy and L. Fei-Fei},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
        title = {Deep Visual-Semantic Alignments for Generating Image Descriptions},
         year = {2016},
       volume = {PP},
       number = {99},
        pages = {1-1},
     abstract = {We present a model that generates natural language descriptions of images and their regions. Our approach leverages datasets of images and their sentence descriptions to learn about the inter-modal correspondences between language and visual data. Our alignment model is based on a novel combination of Convolutional Neural Networks over image regions, bidirectional Recurrent Neural Networks (RNN) over sentences, and a structured objective that aligns the two modalities through a multimodal embedding. We then describe a Multimodal Recurrent Neural Network architecture that uses the inferred alignments to learn to generate novel descriptions of image regions. We demonstrate that our alignment model produces state of the art results in retrieval experiments on Flickr8K, Flickr30K and MSCOCO datasets. We then show that the generated descriptions outperform retrieval baselines on both full images and on a new dataset of region-level annotations. Finally, we conduct large-scale analysis of our RNN language model on the Visual Genome dataset of 4.1 million captions and highlight the differences between image and region-level caption statistics.},
}
@inproceedings{KarpathyetalCVPR-14,
       author = {Karpathy, Andrej and Toderici, George and Shetty, Sanketh and Leung, Thomas and Sukthankar, Rahul and Fei-Fei, Li},
        title = {Large-Scale Video Classification with Convolutional Neural Networks},
    booktitle = {Proceedings of the 2014 {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 2014,
        pages = {1725-1732},
     abstract = {Convolutional Neural Networks (CNNs) have been established as a powerful class of models for image recognition problems. Encouraged by these results, we provide an extensive empirical evaluation of CNNs on largescale video classification using a new dataset of 1 million YouTube videos belonging to 487 classes. We study multiple approaches for extending the connectivity of the a CNN in time domain to take advantage of local spatio-temporal information and suggest a multiresolution, foveated architecture as a promising way of speeding up the training. Our best spatio-temporal networks display significant performance improvements compared to strong feature-based baselines (55.3\% to 63.9\%), but only a surprisingly modest improvement compared to single-frame models (59.3\% to 60.9\%). We further study the generalization performance of our best model by retraining the top layers on the UCF-101 Action Recognition dataset and observe significant performance improvements compared to the UCF-101 baseline model (63.3\% up from 43.9\%).},
} 
@article{KarretalCELL-12,
        title = {A whole-cell computational model predicts phenotype from genotype},
       author = {Karr, J.R. and Sanghvi, J.C. and Macklin, D.N. and Gutschow, M.V. and Jacobs, J.M. and Bolival Jr, B. and Assad-Garcia, N. and Glass, J.I. and Covert, M.W.},
      journal = {Cell},
       volume = 150,
        issue = 2,
        pages = {389-401},
         year = 2012,
     abstract = {Understanding how complex phenotypes arise from individual molecules and their interactions is a primary challenge in biology that computational approaches are poised to tackle. We report a whole-cell computational model of the life cycle of the human pathogen Mycoplasma genitalium that includes all of its molecular components and their interactions. An integrative approach to modeling that combines diverse mathematics enabled the simultaneous inclusion of fundamentally different cellular processes and experimental measurements. Our whole-cell model accounts for all annotated gene functions and was validated against a broad range of data. The model provides insights into many previously unobserved cellular behaviors, including in vivo rates of protein-DNA association and an inverse relationship between the durations of DNA replication initiation and replication. In addition, experimental analysis directed by model predictions identified previously undetected kinetic parameters and biological functions. We conclude that comprehensive whole-cell models can be used to facilitate biological discovery.}
}
@inproceedings{KartsaklisetalACL-14,
        title = {Resolving Lexical Ambiguity in Tensor Regression Models of Meaning},
       author = {Dimitri Kartsaklis and Nal Kalchbrenner and Mehrnoosh Sadrzadeh},
         year = 2014,
      address = {Baltimore, USA},
    booktitle = {Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics},
        pages = {212-217},
    publisher = {Association for Computational Linguistics},
     abstract = {This paper provides a method for improving tensor-based compositional distributional models of meaning by the addition of an explicit disambiguation step prior to composition. In contrast with previous research where this hypothesis has been successfully tested against relatively simple compositional models, in our work we use a robust model trained with linear regression. The results we get in two experiments show the superiority of the prior disambiguation method and suggest that the effectiveness of this approach is model independent.},
}
@inproceedings{KarunaratneandBostromCI-06,
       author = {Thashmee Karunaratne and Henrik Bostr{\"{o}}m},
        title = {Learning to classify structured data by graph propositionalization},
    booktitle = {Proceedings of the Second {IASTED} International Conference on Computational Intelligence},
         year = 2006,
        pages = {283-288},
}
@inproceedings{KarunaratneetalICMLA-10,
       author = {Thashmee Karunaratne and Henrik Bostr{\"{o}}m and Ulf Norinder},
        title = {Pre-Processing Structured Data for Standard Machine Learning Algorithms by Supervised Graph Propositionalization - {A} Case Study with Medicina Chemistry Datasets},
    booktitle = {The Ninth International Conference on Machine Learning and Applications},
         year = 2010,
        pages = {828-833},
}
@article{KassetalIJCV-88,
     author = {M. Kass and A. Witkin and D. Terzopoulos},
      title = {Snakes: {Active} Contour Models},
    journal = {International Journal of Computer Vision},
  publisher = {Kluwer Academic Publishers},
     volume = 1,
      issue = 4,
      pages = {321-331},
       year = 1988,
}
@article{KasthurietalCELL-15,
       author = {Kasthuri, Narayanan and Hayworth, Kenneth Jeffrey and Berger, Daniel Raimund and Schalek, Richard Lee and Conchello, Jose Angel and Knowles-Barley, Seymour and Lee, Dongil and Vazquez-Reina, Amelio and Kaynig, Verena and Jones, Thouis Raymond and Roberts, Mike and Morgan, Josh Lyskowski and Tapia, Juan Carlos and Seung, H. Sebastian and Roncal, William Gray and Vogelstein, Joshua Tzvi and Burns, Randal and Sussman, Daniel Lewis and Priebe, Carey Eldin and Pfister, Hanspeter and Lichtman, Jeff William},
        title = {Saturated Reconstruction of a Volume of Neocortex},
      journal = {Cell},
         year = 2015,
    publisher = {Elsevier},
       volume = 162,
        issue = 3,
        pages = {648-661},
     abstract = {We describe automated technologies to probe the structure of neural tissue at nanometer resolution and use them to generate a saturated reconstruction of a sub-volume of mouse neocortex in which all cellular objects (axons, dendrites, and glia) and many sub-cellular components (synapses, synaptic vesicles, spines, spine apparati, postsynaptic densities, and mitochondria) are rendered and itemized in a database. We explore these data to study physical properties of brain tissue. For example, by tracing the trajectories of all excitatory axons and noting their juxtapositions, both synaptic and non-synaptic, with every dendritic spine we refute the idea that physical proximity is sufficient to predict synaptic connectivity (the so-called Peter's rule). This online minable database provides general access to the intrinsic complexity of the neocortex and enables further data-driven inquiries.},
}
@article{KatoetalCELL-15,
       author = {Kato, Saul and Kaplan, Harris S. and Schr\"{o}del, Tina and Skora, Susanne and Lindsay, Theodore H. and Yemini, Eviatar and Lockery, Shawn and Zimmer, Manuel},
        title = {Global Brain Dynamics Embed the Motor Command Sequence of {\it{Caenorhabditis elegans}}},
      journal = {Cell},
    publisher = {Elsevier},
       volume = {163},
        issue = {3},
         year = {2015},
        pages = {656-669},
     abstract = {While isolated motor actions can be correlated with activities of neuronal networks, an unresolved problem is how the brain assembles these activities into organized behaviors like action sequences. Using brain-wide calcium imaging in Caenorhabditis elegans, we show that a large proportion of neurons across the brain share information by engaging in coordinated, dynamical network activity. This brain state evolves on a cycle, each segment of which recruits the activities of different neuronal sub-populations and can be explicitly mapped, on a single trial basis, to the animals' major motor commands. This organization defines the assembly of motor commands into a string of run-and-turn action sequence cycles, including decisions between alternative behaviors. These dynamics serve as a robust scaffold for action selection in response to sensory input. This study shows that the coordination of neuronal activity patterns into global brain dynamics underlies the high-level organization of behavior.},
}
@article{KatoetalCELL-15.supplement,
       author = {Kato, Saul and Kaplan, Harris S. and Schr\"{o}del, Tina and Skora, Susanne and Lindsay, Theodore H. and Yemini, Eviatar and Lockery, Shawn and Zimmer, Manuel},
        title = {Global Brain Dynamics Embed the Motor Command Sequence of {\it{Caenorhabditis elegans}}: Supplementary Materials},
      journal = {Cell},
       volume = {163},
        issue = {3},
         year = {2015},
}
@article{KatoetalNEURON-14,
       author = {Saul Kato and Yifan Xu and Christine E. Cho and L.F. Abbott and Cornelia I. Bargmann},
        title = {Temporal Responses of {\it{Caenorhabditis elegans}} Chemosensory Neurons Are Preserved in Behavioral Dynamics},
      journal = {Neuron},
       volume = {81},
       number = {3},
         year = {2014},
        pages = {616-628},
     abstract = {Animals track fluctuating stimuli over multiple timescales during natural olfactory behaviors. Here, we define mechanisms underlying these computations in Caenorhabditis elegans. By characterizing neuronal calcium responses to rapidly fluctuating odor sequences, we show that sensory neurons reliably track stimulus fluctuations relevant to behavior. \{AWC\} olfactory neurons respond to multiple odors with subsecond precision required for chemotaxis, whereas \{ASH\} nociceptive neurons integrate noxious cues over several seconds to reach a threshold for avoidance behavior. Each neuron's response to fluctuating stimuli is largely linear and can be described by a biphasic temporal filter and dynamical model. A calcium channel mutation alters temporal filtering and avoidance behaviors initiated by \{ASH\} on similar timescales. A sensory G-alpha protein mutation affects temporal filtering in \{AWC\} and alters steering behavior in a way that supports an active sensing model for chemotaxis. Thus, temporal features of sensory neurons can be propagated across circuits to specify behavioral dynamics.}
}
@article{KatsukiandConstantinidisFiIN-12,
       author = {Katsuki, Fumi and Constantinidis, Christos},
        title = {Unique and shared roles of the posterior parietal and dorsolateral prefrontal cortex in cognitive functions},
      journal = {Frontiers in integrative neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = {6},
         year = {2012},
        pages = {17-17},
     abstract = {The dorsolateral prefrontal cortex (PFC) and posterior parietal cortex (PPC) are two parts of a broader brain network involved in the control of cognitive functions such as working-memory, spatial attention, and decision-making. The two areas share many functional properties and exhibit similar patterns of activation during the execution of mental operations. However, neurophysiological experiments in non-human primates have also documented subtle differences, revealing functional specialization within the fronto-parietal network. These differences include the ability of the PFC to influence memory performance, attention allocation, and motor responses to a greater extent, and to resist interference by distracting stimuli. In recent years, distinct cellular and anatomical differences have been identified, offering insights into how functional specialization is achieved. This article reviews the common functions and functional differences between the PFC and PPC, and their underlying mechanisms.},
}
@article{KatsukietalJOVE-17,
        title = {Preparation of a chronic brain-imaging window in Drosophila melanogaster},
       author = {Takeo Katsuki and Dhruv Grover and Ralph J. Greenspan},
      journal = {Journal of Visualized Experiments},
         year = {2017},
        pages = {submitted},
     abstract = {The fruit fly Drosophila melanogaster offers an array of powerful genetic tools useful for recording and manipulating neurons in a relatively small brain that is capable of displaying highly complex behaviors. One outstanding problem in Drosophila neuroscience is the ability to gain optical access to the brain through a non-transparent cuticle layer. Here we present a method to prepare a transparent glass window on the head of a fly for chronic in vivo imaging experiments. The window is created by removing the cuticle from the dorsal side of the fly head using fine forceps and covering the opening with biologically inert, transparent silicon. A piece of diced coverslip is then placed on top to ensure an optically flat surface for imaging. The window allows imaging the protocerebrum including antennal lobes, central complex, and mushroom bodies. After a successful surgery, which typically takes 10 min, the operated fly can survive a few days to a few weeks. The brain of the operated flies can be imaged repeatedly under an upright fluorescence microscope. It also allows for real-time monitoring of brain activity in freely walking flies when used with Flyception, a galvanometer mirror-based tracking system recently developed for imaging the brain in an unrestrained naturally behaving fly. This simple, economical technique enables a variety of new imaging approaches for studying neuronal anatomy, activity, and function.},
}
@article{KatzandFrostTiN-96,
        title = {Intrinsic neuromodulation: altering neuronal circuits from within},
       author = {Paul S. Katz and William N. Frost},
      journal = {Trends in Neurosciences},
       volume = 19,
        issue = 2,
        pages = {54-61},
         year = 1996,
     abstract = {There are two sources of neuromodulation for neuronal circuits: extrinsic inputs and intrinsic components of the circuits themselves. Extrinsic neuromodulation is known to be pervasive in nervous systems, but intrinsic neuromodulation is less recognized, despite the fact that it has now been demonstrated in sensory and neuromuscular circuits and in central pattern generators. By its nature, intrinsic neuromodulation produces local changes in neuronal computation, whereas extrinsic neuromodulation can cause global changes, often affecting many circuits simultaneously. Studies in a number of systems are defining the different properties of these two forms of neuromodulation.}
}
@article{KatzandWillnerACIE-04,
       author = {Katz, E. and Willner, I.},
        title = {Integrated nanoparticle-biomolecule hybrid systems: synthesis, properties, and applications},
      journal = {Angewandte Chemie International Edition},
       volume = 43,
        issue = 45,
         year = 2004,
        pages = {6042-6108},
     abstract = {Nanomaterials, such as metal or semiconductor nanoparticles and nanorods, exhibit similar dimensions to those of biomolecules, such as proteins (enzymes, antigens, antibodies) or DNA. The integration of nanoparticles, which exhibit unique electronic, photonic, and catalytic properties, with biomaterials, which display unique recognition, catalytic, and inhibition properties, yields novel hybrid nanobiomaterials of synergetic properties and functions. This review describes recent advances in the synthesis of biomolecule-nanoparticle/nanorod hybrid systems and the application of such assemblies in the generation of 2D and {3D} ordered structures in solutions and on surfaces. Particular emphasis is directed to the use of biomolecule-nanoparticle (metallic or semiconductive) assemblies for bioanalytical applications and for the fabrication of bioelectronic devices.}
}
@article{KaufmanetalELIFE-15,
       author = {Kaufman, M. T. and Churchland, M. M. and Ryu, S. I. and Shenoy, K. V.},
        title = {Vacillation, indecision and hesitation in moment-by-moment decoding of monkey motor cortex},
      journal = {Elife},
       volume = {4},
         year = {2015},
        pages = {e04677},
     abstract = {When choosing actions, we can act decisively, vacillate, or suffer momentary indecision. Studying how individual decisions unfold requires moment-by-moment readouts of brain state. Here we provide such a view from dorsal premotor and primary motor cortex. Two monkeys performed a novel decision task while we recorded from many neurons simultaneously. We found that a decoder trained using 'forced choices' (one target viable) was highly reliable when applied to 'free choices'. However, during free choices internal events formed three categories. Typically, neural activity was consistent with rapid, unwavering choices. Sometimes, though, we observed presumed 'changes of mind': the neural state initially reflected one choice before changing to reflect the final choice. Finally, we observed momentary 'indecision': delay forming any clear motor plan. Further, moments of neural indecision accompanied moments of behavioral indecision. Together, these results reveal the rich and diverse set of internal events long suspected to occur during free choice.}
}
@article{KaufmanetalNATURE-NEUROSCIENCE-14,
       author = {Kaufman, Matthew T. and Churchland, Mark M. and Ryu, Stephen I. and Shenoy, Krishna V.},
        title = {Cortical activity in the null space: permitting preparation without movement},
      journal = {Nature Neuroscience},
         year = {2014},
    publisher = {Nature Publishing Group},
       volume = {17},
        issue = {3},
        pages = {440-448},
     abstract = {Neural circuits must perform computations and then selectively output the results to other circuits. Yet synapses do not change radically at millisecond timescales. A key question then is: how is communication between neural circuits controlled? In motor control, brain areas directly involved in driving movement are active well before movement begins. Muscle activity is some readout of neural activity, yet it remains largely unchanged during preparation. Here we find that during preparation, while the monkey holds still, changes in motor cortical activity cancel out at the level of these population readouts. Motor cortex can thereby prepare the movement without prematurely causing it. Further, we found evidence that this mechanism also operates in dorsal premotor cortex, largely accounting for how preparatory activity is attenuated in primary motor cortex. Selective use of 'output-null' vs. 'output-potent' patterns of activity may thus help control communication to the muscles and between these brain areas.},
}
@book{Kaufmann75,
       author = {Kaufmann, Arnold},
        title = {Introduction to the Theory of Fuzzy Subsets},
    publisher = {Academic Press},
      address = {New York},
         year = 1975,
}
@inproceedings{KautzAAAI-86,
       author = {Kautz, Henry},
        title = {The Logic of Persistence},
    booktitle = {Proceedings AAAI-86},
     location = {Philadelphia, Pennsylvania},
 organization = {AAAI},
         year = 1986,
        pages = {401-405},
}
@inproceedings{KautzandAllenAAAI-86,
       author = {Kautz, Henry and Allen, James},
        title = {Generalized Plan Recognition},
    booktitle = {Proceedings AAAI-86},
     location = {Philadelphia, Pennsylvania},
 organization = {AAAI},
         year = 1986,
        pages = {32-37},
}
@misc{Kavli_FoundationFUTURES-SYMPOSIUM-13,
        title = {10th Kavli Futures Symposium: Data Deluge from the Brain Activity Map},
       author = {Terry Sejnowski and Miyoung Chun},
         year = {2013},
 howpublished = {https://www.kavlifoundation.org/sites/default/files/docs/BAMHistory-Jan2013b.pdf},
}
@inproceedings{KavrakiandLatombeIEEE-94,
       author = {Kavraki, Lydia and Latombe, Jean-Claude},
        title = {Randomized preprocessing of configuration space for fast path planning},
    booktitle = {{IEEE} International Conference on Robotics \& Automation},
      address = {San Diego},
        pages = {2138-2145},
         year = 1994,
}
@inproceedings{KavukcuogluetalCVPR-09,
        title = {Learning Invariant Features through Topographic Filter Maps},
       author = {Kavukcuoglu, Koray and Ranzato, Marc'Aurelio and Fergus, Rob and LeCun, Yann},
    booktitle = {Proceedings International Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 2009,
}
@article{KavukcuogluetalCoRR-16,
       author = {Max Jaderberg and Wojciech Marian Czarnecki and Simon Osindero and Oriol Vinyals and Alex Graves and Koray Kavukcuoglu},
        title = {Decoupled Neural Interfaces using Synthetic Gradients},
      journal = {CoRR},
       volume = {arXiv:1608.05343},
         year = {2016},
}
@inproceedings{KavukcuogluetalNIPS-10,
       author = {Kavukcuoglu, Koray and Sermanet, Pierre and Boureau, Y-Lan and Gregor, Karol and Mathieu, Micha\"{e}l and LeCun, Yann},
        title = {Learning Convolutional Feature Hierarchies for Visual Recognition},
    booktitle = {Proceedings of the 23rd International Conference on Neural Information Processing Systems - Volume 1},
    publisher = {Curran Associates Inc.},
         year = {2010},
     location = {Vancouver, British Columbia, Canada},
        pages = {1090-1098},
}
@techreport{KavukcuogluetalTR-08,
        title = {Fast Inference in Sparse Coding Algorithms with Applications to Object Recognition},
       author = {Kavukcuoglu, Koray and Ranzato, Marc'Aurelio and LeCun, Yann},
  institution = {Computational and Biological Learning Lab, Courant Institute, NYU},
        issue = {CBLL-TR-2008-12-01},
         year = 2008,
          url = {http://cs.nyu.edu/~koray/publis/koray-psd-08.pdf},
}
@inproceedings{KawanoetalNEMS-07,
       author = {Kawano, T. and Chung Yeung Cho and Liwei Lin},
    booktitle = {IEEE International Conference on Nano/Micro Engineered and Molecular Systems (NEMS)},
        title = {Carbon Nanotube-based Nanoprobe Electrode},
         year = 2007,
        pages = {895-898},
     abstract = {This paper reports a 10 nm in diameter nanoprobe electrode made of carbon nanotube (CNT) for possible electrophysiological measurements for biomedical applications. The nanoprobe is based on an individual carbon nanotube fabricated by a controlled local growth process and subsequently encapsulated with an insulating layer of Parylene-C. It is integrated with a silicon microstructure with a total length of 5mum and its tip at the distal end is locally heated to expose about 100nm-long CNT as the sensing port. We believe this nano scale CNT probe, with its high strength and Young's modulus, may act as a low-invasive intracellular electrode for measurements inside cells or neurons.},
}
@article{KawashimaetalCELL-16,
       author = {Kawashima, Takashi and Zwart, Maarten F. and Yang, Chao-Tsung and Mensh, Brett D. and Ahrens, Misha B.},
        title = {The Serotonergic System Tracks the Outcomes of Actions to Mediate Short-Term Motor Learning},
      journal = {Cell},
         year = {2016},
    publisher = {Elsevier},
     abstract = {To execute accurate movements, animals must continuously adapt their behavior to changes in their bodies and environments. Animals can learn changes in the relationship between their locomotor commands and the resulting distance moved, then adjust command strength to achieve a desired travel distance. It is largely unknown which circuits implement this form of motor learning, or how. Using whole-brain neuronal imaging and circuit manipulations in larval zebrafish, we discovered that the serotonergic dorsal raphe nucleus (DRN) mediates short-term locomotor learning. Serotonergic DRN neurons respond phasically to swim-induced visual motion, but little to motion that is not self-generated. During prolonged exposure to a given motosensory gain, persistent DRN activity emerges that stores the learned efficacy of motor commands and adapts future locomotor drive for tens of seconds. The DRN's ability to track the effectiveness of motor intent may constitute a computational building block for the broader functions of the serotonergic system.},
}
@article{KayandFrankHIPPOCAMPUS-19,
       author = {Kay, Kenneth and Frank, Loren M.},
        title = {Three brain states in the hippocampus and cortex},
      journal = {Hippocampus},
       volume = {29},
       number = {3},
         year = {2019},
        pages = {184-238},
     abstract = {Contemporary brain research seeks to understand how cognition is reducible to neural activity. Crucially, much of this effort is guided by a scientific paradigm that views neural activity as essentially driven by external stimuli. In contrast, recent perspectives argue that this paradigm is by itself inadequate and that understanding patterns of activity intrinsic to the brain is needed to explain cognition. Yet, despite this critique, the stimulus-driven paradigm still dominates-possibly because a convincing alternative has not been clear. Here, we review a series of findings suggesting such an alternative. These findings indicate that neural activity in the hippocampus occurs in one of three brain states that have radically different anatomical, physiological, representational, and behavioral correlates, together implying different functional roles in cognition. This three-state framework also indicates that neural representations in the hippocampus follow a surprising pattern of organization at the timescale of \~1 s or longer. Lastly, beyond the hippocampus, recent breakthroughs indicate three parallel states in the cortex, suggesting shared principles and brain-wide organization of intrinsic neural activity.},
}
@article{KayetalNATURE-08,
       author = {Kay, K.N. and Naselaris, T. and Prenger, R.J. and Gallant, J.L.},
        title = {Identifying natural images from human brain activity},
      journal = {Nature},
       volume = 452,
        pages = {352-355},
         year = 2008,
}
@article{KayetalNATURE-13,
       author = {Tye, Kay M. and Mirzabekov, Julie J. and Warden, Melissa R. and Ferenczi, Emily A. and Tsai, Hsing-Chen and Finkelstein, Joel and Kim, Sung-Yon and Adhikari, Avishek and Thompson, Kimberly R. and Andalman, Aaron S. and Gunaydin, Lisa A. and Witten, Ilana B. and Deisseroth, Karl},
      journal = {Nature},
        issue = 7433,
        pages = {537-541},
    publisher = {Nature Publishing Group},
        title = {{Dopamine neurons modulate neural encoding and expression of depression-related behaviour}},
       volume = 493,
         year = 2013,
     abstract = {Major depression is characterized by diverse debilitating symptoms that include hopelessness and anhedonia. Dopamine neurons involved in reward and motivation are among many neural populations that have been hypothesized to be relevant, and certain antidepressant treatments, including medications and brain stimulation therapies, can influence the complex dopamine system. Until now it has not been possible to test this hypothesis directly, even in animal models, as existing therapeutic interventions are unable to specifically target dopamine neurons. Here we investigated directly the causal contributions of defined dopamine neurons to multidimensional depression-like phenotypes induced by chronic mild stress, by integrating behavioural, pharmacological, optogenetic and electrophysiological methods in freely moving rodents. We found that bidirectional control (inhibition or excitation) of specified midbrain dopamine neurons immediately and bidirectionally modulates (induces or relieves) multiple independent depression symptoms caused by chronic stress. By probing the circuit implementation of these effects, we observed that optogenetic recruitment of these dopamine neurons potently alters the neural encoding of depression-related behaviours in the downstream nucleus accumbens of freely moving rodents, suggesting that processes affecting depression symptoms may involve alterations in the neural encoding of action in limbic circuitry.},
}
@article{KayetalNEUROSCIENCE-11,
        title = {Retinal ganglion cells with distinct directional preferences differ in molecular identity, structure, and central projections},
       author = {Kay, J.N. and De la Huerta, I. and Kim, I.J. and Zhang, Y. and Yamagata, M. and Chu, M.W. and Meister, M. and Sanes, J.R.},
      journal = {Journal of Neuroscience},
       volume = 25,
        issue = 31,
         year = 2011,
        pages = {7753-7762},
}
@misc{KazemnejadPOSITIONAL-ENCODING-19,
        title = {Transformer Architecture: The Role of Positional Encoding},
       author = {Amirhossein Kazemnejad},
 howpublished = {\url{https://kazemnejad.com/blog/transformer_architecture_positional_encoding/}},
         year = {2019},
}
@book{Kean2014dueling,
        title = {The Tale of the Dueling Neurosurgeons: The History of the Human Brain as Revealed by True Stories of Trauma, Madness, and Recovery},
       author = {Kean, Sam},
         year = {2014},
    publisher = {Little, Brown}
}
@article{KeandTongPRE-08,
        title = {Easily adaptable complexity measure for finite time series},
       author = {Ke, Da-Guan and Tong, Qin-Ye},
      journal = {Physical Review E},
    publisher = {American Physical Society},
       volume = {77},
        issue = {6},
        pages = {066215},
         year = {2008},
}
@article{KearnsPNAS-10,
        title = {Behavioral dynamics and influence in networked coloring and consensus},
       author = {Stephen Judd and Michael Kearns Yevgeniy Vorobeychik},
      journal = {Proceedings of the National Academy of Science},
       volume = 107,
        issue = 34,
         year = 2010,
}
@inproceedings{KearnsandSingh98,
       author = {M. Kearns and S. Singh},
        title = {Near-optimal reinforcement learning in polynomial time},
    booktitle = {Proceedings of the International Conference on Machine Learning},
         year = 1998,
}
@inproceedings{KearnsandSinghICML-98,
       author = {Kearns, M. and Singh, S.},
        title = {Near-Optimal Peformance for Reinforcement Learning in Polynomial Time},
    booktitle = {Proceedings Fifteenth International Conference on Machine Learning},
         year = 1998,
}
@inproceedings{KearnsandValiantSTOC-89,
       author = {Kearns, M. and Valiant, L. G.},
        title = {Cryptographic Limitations on Learning Boolean Functions and Finite Automata},
    booktitle = {Proceedings of the Twenty First Annual {ACM} Symposium on Theoretical Computing},
        pages = {433-444},
         year = 1989,
}
@inproceedings{KearnsetalCOLT-95,
       author = {Michael Kearns and Yishai Mansour and Andrew Ng and Dana Ron},
        title = {An experimental and theoretical comparison of model selection methods},
    booktitle = {Proceedings of the Eighth Annual {ACM} Conference onComputational Learning Theory},
        pages = {21-30},
         year = 1995,
}
@inproceedings{KearnsetalIJCAI-99,
       author = {M. Kearns and Y. Mansour and A. Ng},
        title = {A Sparse Sampling Algorithm for Near-Optimal Planning in Large {M}arkov Decision Processes},
    booktitle = {Proceedings {IJCAI} 16},
     location = {Stockholm, Sweden},
 organization = {IJCAII},
         year = 1999,
}
@inproceedings{KechadiHPCN-99,
       author = {M-Tahar Kechadi},
        title = {Recurrent neural network approach for partitioning irregular graphs},
    booktitle = {International Conference on High-Performance Computing and Networking},
    publisher = {Springer},
      address = {Berlin Heidelberg},
       editor = {Sloot, Peter and Bubak, Marian and Hoekstra, Alfons and Hertzberger, Bob},
         year = {1999},
        pages = {450-459},
     abstract = {This paper is concerned with utilizing a neural network approach to solve the k-way partitioning problem. The k-way partitioning is modeled as a constraint satisfaction problem with linear inequalities and binary variables. A new recurrent neural network architecture is proposed for k-way partitioning. This network is based on an energy function that controls the competition between the partition's external cost and the penalty function. This method is implemented and compared to other global search techniques such as simulated annealing and genetic algorithms. It is shown that it converges better than these techniques.}
}
@inproceedings{KeeetalVS-07,
       author = {Yan Ke and Sukthankar Rahul and Martial Hebert},
        title = {Behavior Recognition via Sparse Spatio-Temporal Features},
    booktitle = {Proceedings of the CVPR Workshop on Visual Surveillance},
         year = 2007,
}
@article{KeeneyOR-81,
       author = {Keeney, Ralph L.},
        title = {Analysis of preference dependencies among objectives},
      journal = {Operations Research},
       volume = 29,
         year = 1981,
        pages = {1105-1120},
}
@book{KeeneyandRaiffa76,
       author = {Keeney, Ralph L. and Raiffa, Howard},
        title = {Decisions with Multiple Objectives: Preferences and Value Tradeoffs},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1976,
}
@article{KeetalCELL-16,
       author = {Ke, Meng-Tsen and Nakai, Yasuhiro and Fujimoto, Satoshi and Takayama, Rie and Yoshida, Shuhei and Kitajima, Tomoya S. and Sato, Makoto and Imai, Takeshi},
        title = {Super-Resolution Mapping of Neuronal Circuitry With an Index-Optimized Clearing Agent},
      journal = {Cell Reports},
         year = {2016},
    publisher = {Elsevier},
     abstract = {Super-resolution imaging deep inside tissues has been challenging, as it is extremely sensitive to light scattering and spherical aberrations. Here, we report an optimized optical clearing agent for high-resolution fluorescence imaging (SeeDB2). SeeDB2 matches the refractive indices of fixed tissues to that of immersion oil (1.518), thus minimizing both light scattering and spherical aberrations. During the clearing process, fine morphology and fluorescent proteins were highly preserved. SeeDB2 enabled super-resolution microscopy of various tissue samples up to a depth of > 100\hmmu{}m, an order of magnitude deeper than previously possible under standard mounting conditions. Using this approach, we demonstrate accumulation of inhibitory synapses on spine heads in NMDA-receptor-deficient neurons. In the fly medulla, we found unexpected heterogeneity in axon bouton orientations among Mi1 neurons, a part of the motion detection circuitry. Thus, volumetric super-resolution microscopy of cleared tissues is a powerful strategy in connectomic studies at synaptic levels.},
}
@inproceedings{KeetalCVPR-05,
       author = {Yan Ke and Sukthankar Rahul and Martial Hebert},
        title = {Efficient Visual Event Detection using Volumetric Features},
    booktitle = {Proceedings of the 2005 {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 2005,
}
@article{KeetalNATURE-13,
       author = {Ke, Meng-Tsen and Fujimoto, Satoshi and Imai, Takeshi},
        title = {{SeeDB}: a simple and morphology-preserving optical clearing agent for neuronal circuit reconstruction},
      journal = {Nature Neuroscience},
       volume = 0,
        issue = 0,
         year = 2013,
        pages = {000-000},
     abstract = {We report a water-based optical clearing agent, SeeDB, which clears fixed brain samples in a few days without quenching many types of fluorescent dyes, including fluorescent proteins and lipophilic neuronal tracers. Our method maintained a constant sample volume during the clearing procedure, an important factor for keeping cellular morphology intact, and facilitated the quantitative reconstruction of neuronal circuits. Combined with two-photon microscopy and an optimized objective lens, we were able to image the mouse brain from the dorsal to the ventral side. We used SeeDB to describe the near-complete wiring diagram of sister mitral cells associated with a common glomerulus in the mouse olfactory bulb. We found the diversity of dendrite wiring patterns among sister mitral cells, and our results provide an anatomical basis for non-redundant odor coding by these neurons. Our simple and efficient method is useful for imaging intact morphological architecture at large scales in both the adult and developing brains.}
}
@inproceedings{KeetalNIPS-05,
       author = {Yan Ke and Sukthankar Rahul and Martial Hebert},
        title = {Efficient Temporal Mean Shift for Activity Recognition in Video},
    booktitle = {{NIPS} Workshop on Activity Recognition and Discovery},
         year = 2005
}
@article{Kelly94,
       author = {Kelly, Kevin},
        title = {Cracking Wall Street},
      journal = {Wired},
         year = 1994,
}
@inproceedings{KelmetalICPR-06,
       author = {B. Michael Kelm and Chris Pal and Andrew McCallum},
        title = {Combining Generative and Discriminative Methods for Pixel Classification with Multi-Conditional Learning},
    booktitle = {Proceedings of the 18th International Conference on Pattern Recognition},
    publisher = {IEEE Computer Society},
        pages = {828-832},
         year = 2006,
}
@book{KemenyandSnell60,
       author = {Kemeny, J. G. and Snell, J. L.},
        title = {Finite {M}arkov Chains},
    publisher = {D. Van Nostrand},
      address = {New York},
         year = 1960,
}
@book{Kemmerer2014cognitive,
        title = {Cognitive Neuroscience of Language},
       author = {Kemmerer, D.L.},
    publisher = {Psychology Press},
         year = {2014},
}
@article{KempaetalAM-07,
       author = {Kempa, K. and Rybczynski, J. and Huang, Z. and Gregorczyk, K. and Vidan, A. and Kimball, B. and Carlson, J. and Benham, G. and Wang, Y. and Herczynski, A. and Ren, Z. F.},
        title = {Carbon Nanotubes as Optical Antennae},
      journal = {Advanced Materials},
       volume = 19,
        issue = 3,
         year = 2007,
        pages = {421-426},
}
@inproceedings{KempeetalAAAI-06,
       author = {Kemp, Charles and Tenenbaum, Joshua B. and Griffiths, Thomas L. and Yamada, Takeshi and Ueda, Naonori},
        title = {Learning Systems of Concepts with an Infinite Relational Model},
    booktitle = {Proceedings of the 21st National Conference on Artificial Intelligence},
    publisher = {AAAI Press},
     location = {Boston, Massachusetts},
         year = {2006},
        pages = {381-388},
     abstract = {Relationships between concepts account for a large proportion of semantic knowledge. We present a nonparametric Bayesian model that discovers systems of related concepts. Given data involving several sets of entities, our model discovers the kinds of entities in each set and the relations between kinds that are possible or likely. We apply our approach to four problems: clustering objects and features, learning ontologies, discovering kinship systems, and discovering structure in political data.}
}
@inproceedings{KempeetalICALP-05,
       author = {David Kempe and Jon Kleinberg and \'{E}va Tardos},
        title = {Influential Nodes in a Diffusion Model for Social Networks},
    booktitle = {Proceedings of the 32nd International Colloquium on Automata, Languages and Programming},
 localtion    = {Lisboa, Portugal},
    publisher = {Springer Verlag},
        pages = {1127-1138},
         year = 2005,
}
@inproceedings{KempeetalKDD-03,
      author = {Kempe, David and Kleinberg, Jon and Tardos, \'{E}va},
       title = {Maximizing the spread of influence through a social network},
   booktitle = {Proceedings of the 9th {ACM SIGKDD} International Conference on Knowledge Discovery and Data Mining},
    location = {Washington, DC, USA},
   publisher = {ACM},
     address = {New York, NY, USA},
       pages = {137-146},
        year = 2003,
}
@article{KempesetalECAP-05,
       author = {Kempes, M. and Matthys, W. and de Vries, H. and van Engeland, H.},
        title = {Reactive and proactive aggression in children--a review of theory, findings and the relevance for child and adolescent psychiatry},
      journal = {European Child and Adolescent Psychiatry},
       volume = {14},
       number = {1},
         year = {2005},
        pages = {11-19},
     abstract = {The clinical population of aggressive children diagnosed as having an oppositional defiant disorder (ODD) or a conduct disorder (CD) is heterogeneous, both with respect to behaviour and aetiology. Recently, the following distinction has been proposed that might further clarify this heterogeneity: reactive aggression is an aggressive response to a perceived threat or provocation, whereas proactive aggression is defined as behaviour that anticipates a reward. In this article we examine various aspects of this distinction. We will [1] examine the evidence that reactive and proactive aggression are distinct phenomena by discussing the theories underlying the distinction between the subtypes in humans and we briefly review evidence for a similar distinction in animals; [2] we critically review the literature on the measurement in children via questionnaires and behavioural observations; we then point out that the correlation observed between the subtypes is due to the fact that many children show both types of aggression; [3] we review the literature on specific characteristics of the subtypes giving attention to social information processing, peer status, biological correlates and developmental history, and demonstrate that there is some evidence to suggest that reactive and proactive aggression are distinct dimensions; [4] we discuss the relevance of the distinction between reactive and proactive aggression for child and adolescent psychiatry.}
}
@article{KenetetalNATURE-03,
       author = {Kenet, T. and Bibitchkov, D. and Tsodyks, M. and Grinvald, A. and Arieli, A.},
        title = {Spontaneously emerging cortical representations of visual attributes},
      journal = {Nature},
       volume = {425},
       number = {6961},
         year = {2003},
        pages = {954-956},
     abstract = {Spontaneous cortical activity--ongoing activity in the absence of intentional sensory input--has been studied extensively, using methods ranging from EEG (electroencephalography), through voltage sensitive dye imaging, down to recordings from single neurons. Ongoing cortical activity has been shown to play a critical role in development, and must also be essential for processing sensory perception, because it modulates stimulus-evoked activity, and is correlated with behaviour. Yet its role in the processing of external information and its relationship to internal representations of sensory attributes remains unknown. Using voltage sensitive dye imaging, we previously established a close link between ongoing activity in the visual cortex of anaesthetized cats and the spontaneous firing of a single neuron. Here we report that such activity encompasses a set of dynamically switching cortical states, many of which correspond closely to orientation maps. When such an orientation state emerged spontaneously, it spanned several hypercolumns and was often followed by a state corresponding to a proximal orientation. We suggest that dynamically switching cortical states could represent the brain's internal context, and therefore reflect or influence memory, perception and behaviour.}
}
@article{KengetalCPR-11,
       author = {Keng, Shian-Ling and Smoski, Moria J. and Robins, Clive J.},
        title = {Effects of Mindfulness on Psychological Health: A Review of Empirical Studies},
      journal = {Clinical Psychology Review},
         year = {2011},
       volume = {31},
        issue = {6},
        pages = {1041-1056},
     abstract = {Within the past few decades, there has been a surge of interest in the investigation of mindfulness as a psychological construct and as a form of clinical intervention. This article reviews the empirical literature on the effects of mindfulness on psychological health. We begin with a discussion of the construct of mindfulness, differences between Buddhist and Western psychological conceptualizations of mindfulness, and how mindfulness has been integrated into Western medicine and psychology, before reviewing three areas of empirical research: cross-sectional, correlational research on the associations between mindfulness and various indicators of psychological health; intervention research on the effects of mindfulness-oriented interventions on psychological health; and laboratory-based, experimental research on the immediate effects of mindfulness inductions on emotional and behavioral functioning. We conclude that mindfulness brings about various positive psychological effects, including increased subjective well-being, reduced psychological symptoms and emotional reactivity, and improved behavioral regulation. The review ends with a discussion on mechanisms of change of mindfulness interventions and suggested directions for future research.}
}
@book{Kenneally2007first,
        title = {The First Word: The Search for the Origins of Language},
       author = {Kenneally, Christine},
         year = {2007},
    publisher = {Viking}
}
@incollection{KenworthyECB-07,
        title = {Cell Membrane Dynamics},
       author = {Kenworthy, Anne K.},
       editor = {Begley, Tadhg P.},
    publisher = {John Wiley \& Sons, Inc.},
    booktitle = {Wiley Encyclopedia of Chemical Biology},
         year = 2007,
     abstract = {Cell membranes are two-dimensional fluids that exhibit a wide range of dynamic behaviors. Recent technical advances have enabled unprecedented views of membrane dynamics in living cells. In this technical review, we provide a brief overview of three well-studied examples of membrane dynamics: lateral diffusion of proteins and lipids in the plane of the membrane, vesicular trafficking between intracellular compartments, and exchange of proteins on and off membranes. We then discuss experimental approaches to monitor membrane protein and lipid dynamics, and we place a special emphasis on the use of genetically encoded fluorescent probes and live cell-imaging techniques.},
}
@inproceedings{KeoghetalICDM-01,
       author = {Eamonn J. Keogh and Selina Chu and David Hart and Michael J. Pazzani},
        title = {An Online Algorithm for Segmenting Time Series},
    booktitle = {Proceedings of the IEEE International Conference on Data Mining},
        pages = {289-296},
         year = 2001,
}
@inproceedings{KeoghetalKDD-02,
       author = {Eamonn Keogh and Stefano Lonardi and Bill 'Yuan-chi' Chiu},
        title = {Finding surprising patterns in a time series database in linear time and space},
    booktitle = {Proceedings of the eighth ACM SIGKDD International Conference on Knowledge Discovery and Data mining},
         year = 2002,
        pages = {550-556},
     location = {Edmonton, Alberta, Canada},
    publisher = {ACM Press},
}
@inproceedings{KepnerHPEC-01,
       author = {Kepner, Jeremy},
        title = {Parallel Programming with {MatlabMPI}},
    booktitle = {Proceedings of the High Performance Embedded Computing ({HPEC} 2001) Workshop},
         year = 2001,
}
@article{KerchnerandNicollNATURE-REVIEWS-NEUROSCIENCE-08,
       author = {Kerchner, G. A. and Nicoll, R. A.},
        title = {Silent synapses and the emergence of a postsynaptic mechanism for {LTP}},
      journal = {Nature Reviews Neuroscience},
         year = {2008},
       volume = {9},
       number = {11},
        pages = {813-825},
     abstract = {Silent synapses abound in the young brain, representing an early step in the pathway of experience-dependent synaptic development. Discovered amidst the debate over whether long-term potentiation reflects a presynaptic or a postsynaptic modification, silent synapses--which in the hippocampal CA1 subfield are characterized by the presence of NMDA receptors but not AMPA receptors--have stirred some mechanistic controversy of their own. Out of this literature has emerged a model for synapse unsilencing that highlights the central role for postsynaptic AMPA-receptor trafficking in the expression of excitatory synaptic plasticity.}
}
@article{KerretalSCIAM-08,
       author = {Kerr, R. A. and Bartol, T. M. and Kaminsky, B. and Dittrich, M. and Chang, J. C. and Baden, S. B. and Sejnowski, T. J. and Stiles, J. R.},
        title = {Fast {Monte Carlo} Simulation Methods for Biological Reaction-Diffusion Systems in Solution and on Surfaces},
      journal = {{SIAM} Journal Scientific Computing},
         year = 2008,
       volume = 30,
        issue = 6,
        pages = 3126,
}
@article{KerretalSJSC-08,
       author = {Rex A. Kerr and Thomas M. Bartol and Boris Kaminsky and Markus Dittrich and Jen-Chien Jack Chang and Scott B. Baden and Terrence J. Sejnowski and Joel R. Stiles},
        title = {Fast Monte Carlo Simulation Methods for Biological Reaction-Diffusion Systems in Solution and on Surfaces},
      journal = {SIAM Journal on Scientific Computing},
       volume = 30,
        issue = 6,
         year = 2008,
        pages = {3126-3149},
     abstract = {Many important physiological processes operate at time and space scales far beyond those accessible to atom-realistic simulations, and yet discrete stochastic rather than continuum methods may best represent finite numbers of molecules interacting in complex cellular spaces. We describe and validate new tools and algorithms developed for a new version of the MCell simulation program (MCell3), which supports generalized Monte Carlo modeling of diffusion and chemical reaction in solution, on surfaces representing membranes, and combinations thereof. A new syntax for describing the spatial directionality of surface reactions is introduced, along with optimizations and algorithms that can substantially reduce computational costs (e.g., event scheduling, variable time and space steps). Examples for simple reactions in simple spaces are validated by comparison to analytic solutions. Thus we show how spatially realistic Monte Carlo simulations of biological systems can be far more cost-effective than often is assumed, and provide a level of accuracy and insight beyond that of continuum methods.}
}
@article{KeseretalFiN-15,
       author = {Keser, Zafer and Hasan, Khader M. and Mwangi, Benson I. and Kamali, Arash and Ucisik-Keser, Fehime Eymen and Riascos, Roy F. and Yozbatiran, Nuray and Francisco, Gerard E. and Narayana, Ponnada A.},
        title = {Diffusion tensor imaging of the human cerebellar pathways and their interplay with cerebral macrostructure},
      journal = {Frontiers in neuroanatomy},
    publisher = {Frontiers Media S.A.},
       volume = {9},
         year = {2015},
        pages = {41-41},
     abstract = {Cerebellar white matter (WM) connections to the central nervous system are classified functionally into the Spinocerebellar (SC), vestibulocerebellar (VC), and cerebrocerebellar subdivisions. The SC pathways project from spinal cord to cerebellum, whereas the VC pathways project from vestibular organs of the inner ear. Cerebrocerebellar connections are composed of feed forward and feedback connections between cerebrum and cerebellum including the cortico-ponto-cerebellar (CPC) pathways being of cortical origin and the dentate-rubro-thalamo-cortical (DRTC) pathway being of cerebellar origin. In this study we systematically quantified the whole cerebellar system connections using diffusion tensor magnetic resonance imaging (DT-MRI). Ten right-handed healthy subjects (7 males and 3 females, age range 20-51 years) were studied. DT-MRI data were acquired with a voxel size = 2 mm x 2 mm x 2 mm at a 3.0 Tesla clinical MRI scanner. The DT-MRI data were prepared and analyzed using anatomically-guided deterministic tractography methods to reconstruct the SC, DRTC, fronto-ponto-cerebellar (FPC), parieto-ponto-cerebellar (PPC), temporo-ponto-cerebellar (TPC) and occipito-ponto-cerebellar (OPC). The DTI-attributes or the cerebellar tracts along with their cortical representation (Brodmann areas) were presented in standard Montreal Neurological Institute space. All cerebellar tract volumes were quantified and correlated with volumes of cerebral cortical, subcortical gray matter (GM), cerebral WM and cerebellar GM, and cerebellar WM. On our healthy cohort, the ratio of total cerebellar GM-to-WM was ~3.29 -  0.24, whereas the ratio of cerebral GM-to-WM was approximately 1.10 - 0.11. The sum of all cerebellar tract volumes is ~25.8 7.3 mL, or a percentage of 1.6 - 0.45 of the total intracranial volume (ICV).},
}
@article{KesnerandRollsNBR-15,
       author = {Kesner, R. P. and Rolls, E. T.},
        title = {A computational theory of hippocampal function, and tests of the theory: {N}ew developments}}
@article{KetzetalCoRR-19,
       author = {Nicholas Ketz and Soheil Kolouri and Praveen K. Pilly},
        title = {Using World Models for Pseudo-Rehearsal in Continual Learning},
      journal = {CoRR},
       volume = {arXiv:1903.02647},
         year = {2019},
     abstract = {The utility of learning a dynamics/world model of the environment in reinforcement learning has been shown in a many ways. When using neural networks, however, these models suffer catastrophic forgetting when learned in a lifelong or continual fashion. Current solutions to the continual learning problem require experience to be segmented and labeled as discrete tasks, however, in continuous experience it is generally unclear what a sufficient segmentation of tasks would be. Here we propose a method to continually learn these internal world models through the interleaving of internally generated episodes of past experiences (i.e., pseudo-rehearsal). We show this method can sequentially learn unsupervised temporal prediction, without task labels, in a disparate set of Atari games. Empirically, this interleaving of the internally generated rollouts with the external environment's observations leads to a consistent reduction in temporal prediction loss compared to non-interleaved learning and is preserved over repeated random exposures to various tasks. Similarly, using a network distillation approach, we show that modern policy gradient based reinforcement learning algorithms can use this internal model to continually learn to optimize reward based on the world model's representation of the environment.},
}
@article{Khachian79,
       author = {L. G. Khachian},
        title = {A Polynomial algorithm for linear programming},
      journal = {Soviet Math Dokl.},
       volume = 20,
        pages = {191-194},
         year = 1979,
}
@article{KhalidetalNEUROIMAGE-14,
        title = {Tracing the evolution of multi-scale functional networks in a mouse model of depression using persistent brain network homology},
       author = {Arshi Khalid and Byung Sun Kim and Moo K. Chung and Jong Chul Ye and Daejong Jeon},
      journal = {NeuroImage},
       volume = {101},
         year = {2014},
        pages = {351-363},
     abstract = {Many brain diseases or disorders, such as depression, are known to be associated with abnormal functional connectivity in neural networks in the brain. Some bivariate measures of electroencephalography (EEG) for coupling analysis have been used widely in attempts to explain abnormalities related with depression. However, brain network evolution based on persistent functional connections in \{EEG\} signals could not be easily unveiled. For a geometrical exploration of brain network evolution, here, we used persistent brain network homology analysis with \{EEG\} signals from a corticosterone (CORT)-induced mouse model of depression. \{EEG\} signals were obtained from eight cortical regions (frontal, somatosensory, parietal, and visual cortices in each hemisphere). The persistent homology revealed a significantly different functional connectivity between the control and \{CORT\} model, but no differences in common coupling measures, such as cross correlation and coherence, were apparent. The \{CORT\} model showed a more localized connectivity and decreased global connectivity than the control. In particular, the somatosensory and parietal cortices were loosely connected in the \{CORT\} model. Additionally, the \{CORT\} model displayed altered connections among the cortical regions, especially between the frontal and somatosensory cortices, versus the control. This study demonstrates that persistent homology is useful for brain network analysis, and our results indicate that the CORT-induced depression mouse model shows more localized and decreased global connectivity with altered connections, which may facilitate characterization of the abnormal brain network underlying depression.},
}
@article{Khaligh-RazaviandKriegeskortePLoS-14,
       author = {Khaligh-Razavi, Seyed-Mahdi and Kriegeskorte, Nikolaus},
      journal = {PLoS Computational Biology},
    publisher = {Public Library of Science},
        title = {Deep Supervised, but Not Unsupervised, Models May Explain IT Cortical Representation},
         year = {2014},
       volume = {10},
       number = {11},
        pages = {e1003915},
     abstract = {Computers cannot yet recognize objects as well as humans can. Computer vision might learn from biological vision. However, neuroscience has yet to explain how brains recognize objects and must draw from computer vision for initial computational models. To make progress with this chicken-and-egg problem, we compared 37 computational model representations to representations in biological brains. The more similar a model representation was to the high-level visual brain representation, the better the model performed at object categorization. Most models did not come close to explaining the brain representation, because they missed categorical distinctions between animates and inanimates and between faces and other objects, which are prominent in primate brains. A deep neural network model that was trained by supervision with over a million category-labeled images and represents the state of the art in computer vision came closest to explaining the brain representation. Our brains appear to impose upon the visual input certain categorical divisions that are important for successful behavior. Brains might learn these divisions through evolution and individual experience. Computer vision similarly requires learning with many labeled images so as to emphasize the right categorical divisions.},
}
@article{KhambhatietalNEUROIMAGE-17,
        title = {Modeling and interpreting mesoscale network dynamics},
      journal = {NeuroImage},
         year = {2017},
       author = {Ankit N. Khambhati and Ann E. Sizemore and Richard F. Betzel and Danielle S. Bassett},
     abstract = {Recent advances in brain imaging techniques, measurement approaches, and storage capacities have provided an unprecedented supply of high temporal resolution neural data. These data present a remarkable opportunity to gain a mechanistic understanding not just of circuit structure, but also of circuit dynamics, and its role in cognition and disease. Such understanding necessitates a description of the raw observations, and a delineation of computational models and mathematical theories that accurately capture fundamental principles behind the observations. Here we review recent advances in a range of modeling approaches that embrace the temporally-evolving interconnected structure of the brain and summarize that structure in a dynamic graph. We describe recent efforts to model dynamic patterns of connectivity, dynamic patterns of activity, and patterns of activity atop connectivity. In the context of these models, we review important considerations in statistical testing, including parametric and non-parametric approaches. Finally, we offer thoughts on careful and accurate interpretation of dynamic graph architecture, and outline important future directions for method development.}
}
@misc{KhandelwalBEAM-SEARCH-18,
        title = {An intuitive explanation of Beam Search - Towards Data Science, February 1, 2018},
       author = {Renu Khandelwal},
 howpublished = {\url{https://towardsdatascience.com/an-intuitive-explanation-of-beam-search-9b1d744e7a0f}},
         year = {2018},
}
@article{KhatibIJRR-86,
       author = {Khatib, Oussama},
        title = {Real-time obstacle avoidance for manipulators and mobile robots},
      journal = {International Journal of Robotics Research},
       volume = 5,
         year = 1986,
        pages = {90-99},
}
@book{KhatibetalRR-89,
       editor = {Khatib, Oussama and Craig, John H. and Lozano-P\'{e}rez, Tom\'{a}s},
        title = {Robotics Review 1},
    booktitle = {Robotics Review 1},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1989,
}
@article{KhazenetalPLoS-12,
       author = {Khazen, Georges and Hill, Sean L. and Sch\"{u}rmann, Felix and Markram, Henry},
        title = {Combinatorial Expression Rules of Ion Channel Genes in Juvenile Rat ({\it{Rattus norvegicus}}) Neocortical Neurons},
      journal = {{PLoS} ONE},
    publisher = {Public Library of Science},
       volume = 7,
        issue = 4,
         year = 2012,
        pages = {e34786},
     abstract = {The electrical diversity of neurons arises from the
   expression of different combinations of ion channels. The gene
   expression rules governing these combinations are not known. We
   examined the expression of twenty-six ion channel genes in a
   broad range of single neocortical neuron cell types. Using
   expression data from a subset of twenty-six ion channel genes in
   ten different neocortical neuronal types, classified according
   to their electrophysiological properties, morphologies and
   anatomical positions, we first developed an incremental Support
   Vector Machine (iSVM) model that prioritizes the predictive
   value of single and combinations of genes for the rest of the
   expression pattern. With this approach we could predict the
   expression patterns for the ten neuronal types with an average
   10-fold cross validation accuracy of 87\% and for a further
   fourteen neuronal types not used in building the model, with an
   average accuracy of 75\%. The expression of the genes for
   {\it{HCN4}}, {\it{Kv2.2}}, {\it{Kv3.2}} and {\it{Ca$\Beta$3}} were
   found to be particularly strong predictors of ion channel gene
   combinations, while expression of the {\it{Kv1.4}} and
   {\it{Kv3.3}} genes has no predictive value. Using a logic gate
   analysis, we then extracted a spectrum of observed combinatorial
   gene expression rules of twenty ion channels in different
   neocortical neurons. We also show that when applied to a
   completely random and independent data, the model could not
   extract any rules and that it is only possible to extract them
   if the data has consistent expression patterns. This novel
   strategy can be used for predictive reverse engineering
   combinatorial expression rules from single-cell data and could
   help identify candidate transcription regulatory processes.},
}
@article{KiebeletalPLoS-08,
       author = {Kiebel, Stefan J. and Daunizeau, Jean and Friston, Karl J.},
      journal = {PLOS Computational Biology},
    publisher = {Public Library of Science},
        title = {A Hierarchy of Time-Scales and the Brain},
         year = {2008},
       volume = {4},
       number = {11},
        pages = {1-12},
     abstract = {Currently, there is no theory that explains how the large-scale organization of the human brain can be related to our environment. This is astonishing because neuroscientists generally assume that the brain represents events in our environment by decoding sensory input. Here, we propose that the brain models the entire environment as a collection of hierarchical, dynamical systems, where slower environmental changes provide the context for faster changes. We suggest that there is a simple mapping between this temporal hierarchy and the anatomical hierarchy of the brain. Our theory provides a framework for explaining a wide range of neuroscientific findings by a single principle.},
}
@article{KienandAltmanCOMP-BIO-PHYS-92,
       author = {Jenny Kien and Jennifer S Altman},
        title = {Preparation and execution of movement: Parallels between insect and mammalian motor systems},
      journal = {Comparative Biochemistry and Physiology Part A: Physiology},
       volume = {103},
       number = {1},
         year = {1992},
        pages = {15-24},
     abstract = {1. The organization of the motor systems underlying locomotion in insects and mammals is surprisingly similar. There are also parallels between the insect motor system and the system underlying reaching and the occulomotor system in primates. 2. The movements generated by all these systems are planned or prepared before their execution and there is a partial separation of circuits for preparation and execution. 3. These circuits consist of multiple descending pathways interconnected to form overlapping loops which work co-operatively to determine the motor output. Thus, both insect and mammalian motor systems can be treated as parallel distributed (PDP) systems. 4. This enables a comparison of functional levels of processing in the different systems and also provides a basis for modelling motor systems with attractor neural networks. }
}
@article{KillackeyandShermanJoN-03,
       author = {Killackey, Herbert P. and Sherman, S. Murray},
        title = {Corticothalamic Projections from the Rat Primary Somatosensory Cortex},
      journal = {Journal of Neuroscience},
       volume = {23},
       number = {19},
         year = {2003},
        pages = {7381--7384},
     abstract = {To study the cells of origin of corticothalamic inputs to the ventral posterior and posterior medial nuclei of the somatosensory thalamus in rats, we injected small aliquots of tracer into each nucleus and analyzed the pattern of retrograde labeling in the posteromedial barrel subfield of primary somatosensory cortex, which can be divided into barrel and nonbarrel zones. The ventral posterior nucleus is innervated by neurons in layer VIa of both zones, whereas the posterior medial nucleus is innervated by neurons in layers Vb and VIb of both zones with additional innervation from layer VIa of nonbarrel cortex. Thus, only the posterior medial nucleus receives a layer Vb input. Because the layer Vb input is interpreted as the initiation of a feedforward cortico-thalamocortical pathway, this implies that the target of the posterior medial nucleus, which includes the nonbarrel cortex, is a higher-order cortical area. We thus suggest that this cortical zone, which is classically considered part of the primary somatosensory cortex, should be reclassified as higher-order cortex.},
}
@inproceedings{KimEMNLP-14,
        title = {Convolutional Neural Networks for Sentence Classification},
       author = {Kim, Yoon},
    booktitle = {Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing},
    publisher = {Association for Computational Linguistics},
         year = {2014},
        pages = {1746-1751},
     abstract = {We report on a series of experiments with convolutional neural networks (CNN) trained on top of pre-trained word vectors for sentence-level classification tasks. We show that a simple CNN with little hyperparameter tuning and static vec- tors achieves excellent results on multi- ple benchmarks. Learning task-specific vectors through fine-tuning offers further gains in performance. We additionally propose a simple modification to the architecture to allow for the use of both task-specific and static vectors. The CNN models discussed herein improve upon the state of the art on 4 out of 7 tasks, which include sentiment analysis and question classification.},
}
@article{KimandDeanAIJ-03,
        title = {Solving Factored {M}arkov Decision Processes Using Non-homogeneous Partitions},
       author = {Kim, Kee-Eung and Dean, Thomas},
      journal = AIJ,
       volume = 147,
        issue = {1-2},
        pages = {225-251},
         year = 2003,
}
@inproceedings{KimandDeanIJCAI-01,
       author = {Kim, Kee-Eung and Dean, Thomas},
        title = {Solving Factored {MDP}s using Non-Homogeneous Partitions},
    booktitle = {Proceedings {IJCAI}-01},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
     location = {Seattle, Washington},
 organization = {IJCAI},
        pages = {683-689},
         year = 2001,
}
@inproceedings{KimandDeanPRICAI-02,
       author = {Kim, Kee-Eung and Dean, Thomas},
        title = {Solving Factored {MDP}s with Large Action Spaces Using Algebraic Decision Diagrams},
    booktitle = {Proceedings of the 7th Pacific Rim International Conference on Artificial Intelligence ({PRICAI}-02)}, 
    publisher = {Springer},
       editor = {Ishizuka, Mitsuru and Satter, Abdul},
     location = {Tokyo, Japan},
 organization = {PRIJCAI},
        pages = {80-89},
         year = 2002,
}
@inproceedings{KimandPearlIJCAI-83,
       author = {Jin H. Kim and Judea Pearl},
        title = {A Computational Model for Causal and Diagnostic Reasoning in Inference Systems},
    booktitle = {Proceedings {IJCAI} 8},
     location = {Karlsruhe, West Germany},
 organization = {IJCAII},
         year = 1983,
        pages = {190-193},
}
@article{KimandTullman-ErcekCOiB-12,
       author = {Edward Y Kim, Danielle Tullman-Ercek},
        title = {Engineering nanoscale protein compartments for synthetic organelles},
      journal = {Current Opinion in Biotechnology},
       volume = {December},
         year = 2012,
     abstract = {Advances in metabolic engineering have given rise to the biological production of novel fuels and chemicals, but yields are often low without significant optimization. One generalizable solution is to create a specialized organelle for the sequestration of engineered metabolic pathways. Bacterial microcompartments are an excellent scaffold for such an organelle. These compartments consist of a porous protein shell that encapsulates enzymes. To repurpose these structures, researchers have begun to determine how the protein shell is assembled, how pores may be used to control small molecule transport across the protein shell, and how to target heterologous enzymes to the compartmentinterior. With these advances, it will soon be possible to use engineered forms of these protein shells to create designer organelles.}
}
@inproceedings{KimetalAIPS-00,
       author = {Kim, Kee-Eung and Meuleau, Nicolas and Dean, Thomas},
        title = {Approximate Solutions to Factored {M}arkov Decision Processes via Greedy Search in the Space of Finite State Controllers},
    booktitle = {Proceedings of the 5th International Conference on Artificial Intelligence Planning Systems ({ICAPS}-2000},
    publisher = {AAAI Press},
      address = {Menlo Park, California},
        pages = {323-330},
         year = 2000,
}
@article{KimetalBBR-12,
       author = {Kim, Woon Ryoung and Christian, Kimberly and Ming, Guo-Li and Song, Hongjun},
        title = {Time-dependent involvement of adult-born dentate granule cells in behavior},
      journal = {Behavioural Brain Research},
         year = {2012},
       volume = {227},
        issue = {2},
        pages = {470-479},
     abstract = {Adult-born neurons are continuously generated and incorporated into the circuitry of the hippocampus throughout life in mammals. Cumulative evidence supports a physiological role for adult-born neurons, yet it not clear whether this subset of dentate granule cells makes a unique contribution to hippocampal function. Perturbation or ablation of adult hippocampal neurogenesis leads to deficits in the acquisition of learned associations or memory recall, whereas an increase in adult hippocampal neurogenesis enhances some forms of learning and memory. The observed effects thus far appear to be task-dependent, species-specific, and sensitive to the timing of manipulations. Here, we review the recent evidence correlating adult-born dentate granule cells (DGCs) with hippocampal-dependent behavior and focus on the dynamic properties of this neuronal population that may underlie its function. We further discuss a framework for future investigations of how newly integrated neurons may contribute to hippocampal processing using advanced genetic techniques with enhanced temporal resolution.},
}
@article{KimetalCoRR-15,
       author = {Hyongsuk Kim and Maheshwar Pd. Sah and Shyam Prasad Adhikari},
        title = {Pinched Hysteresis Loop is the Fingerprint of Memristive Devices},
      journal = {CoRR},
       volume = {arXiv:1202.2437},
         year = 2015,
     abstract = {This short note clarifies that the "pinched hysteresis loop" fingerprint of a memristor, or a memristive device, must hold for all amplitudes, for all frequencies, and for all initial conditions, of any periodic testing waveform, such as sinusoidal or triangular signals, which assumes both positive and negative values over each period of the waveform. We proved that the systems presented in [1] are not memristive devices because their hysteresis loops are not pinched at the origin for all amplitudes, and for all initial conditions.}
}
@inproceedings{KimetalECV-09,
       author = {Junsik Kim and Myung Hwangbo and Takeo Kanade},
        title = {Realtime Affine-photometric {KLT} Feature Tracker on {GPU} in {CUDA} Framework},
    booktitle = {Workshop on Embedded Computer Vision (ECV), 2009 (held in conjunction with ICCV)},
         year = 2009,
}
@article{KimetalFiCN-17,
       author = {Kim, Taegyo and Hamade, Khaldoun C. and Todorov, Dmitry and Barnett, William H. and Capps, Robert A. and Latash, Elizaveta M. and Markin, Sergey N. and Rybak, Ilya A. and Molkov, Yaroslav I.},
        title = {Reward Based Motor Adaptation Mediated by Basal Ganglia},
      journal = {Frontiers in Computational Neuroscience},
       volume = {11},
         year = {2017},
        pages = {19},
     abstract = {It is widely accepted that the basal ganglia (BG) play a key role in action selection and reinforcement learning. However, despite considerable number of studies, the BG architecture and function are not completely understood. Action selection and reinforcement learning are facilitated by the activity of dopaminergic neurons, which encode reward prediction errors when reward outcomes are higher or lower than expected. The BG are thought to select proper motor responses by gating appropriate actions, and suppressing inappropriate ones. The direct striato-nigral (GO) and the indirect striato-pallidal (NOGO) pathways have been suggested to provide the functions of BG in the two-pathway concept. Previous models confirmed the idea that these two pathways can mediate the behavioral choice, but only for a relatively small number of potential behaviors. Recent studies have provided new evidence of BG involvement in motor adaptation tasks, in which adaptation occurs in a non-error-based manner. In such tasks, there is a continuum of possible actions, each represented by a complex neuronal activity pattern. We extended the classical concept of the two-pathway BG by creating a model of BG interacting with a movement execution system, which allows for an arbitrary number of possible actions. The model includes sensory and premotor cortices, BG, a spinal cord network, and a virtual mechanical arm performing 2D reaching movements. The arm is composed of 2 joints (shoulder and elbow) controlled by 6 muscles (4 mono-articular and 2 bi-articular). The spinal cord network contains motoneurons, controlling the muscles, and sensory interneurons that receive afferent feedback and mediate basic reflexes. Given a specific goal-oriented motor task, the BG network through reinforcement learning constructs a behavior from an arbitrary number of basic actions represented by cortical activity patterns. Our study confirms that, with slight modifications, the classical two-pathway BG concept is consistent with results of previous studies, including non-error based motor adaptation experiments, pharmacological manipulations with BG nuclei, and functional deficits observed in BG-related motor disorders.}
}
@article{KimetalJBO-08,
        title = {In vivo confocal and multiphoton microendoscopy},
       author = {Kim, P. and Puoris'haag, M. and Co Te, D. and Lin, C.P. and Yun, S.H.},
      journal = {Journal of Biomedical Optics},
       volume = 13,
        issue = 1,
         year = 2008,
        pages = 010501,
     abstract = {The ability to conduct high-resolution fluorescence imaging in internal organs of small animal models in situ and over time can make a significant impact in biomedical research. Toward this goal, we developed a real-time confocal and multiphoton endoscopic imaging system. Using 1-mm-diameter endoscopes based on gradient index lenses, we demonstrate video-rate multicolor multimodal imaging with cellular resolution in live mice.}
}
@article{KimetalNATURE-13,
       author = {Kim, Yoonseob and Zhu, Jian and Yeom, Bongjun and Di Prima, Matthew and Su, Xianli and Kim, Jin-Gyu and Yoo, Seung Jo and Uher, Ctirad and Kotov, Nicholas A.},
        title = {Stretchable nanoparticle conductors with self-organized conductive pathways},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = 500,
        issue = 7460,
         year = 2013,
        pages = {59-63},
     abstract = {Research in stretchable conductors is fuelled by diverse technological needs. Flexible electronics, neuroprosthetic and cardiostimulating implants, soft robotics and other curvilinear systems require materials with high conductivity over a tensile strain of 100 per cent (refs 1-3). Furthermore, implantable devices or stretchable displays need materials with conductivities a thousand times higher while retaining a strain of 100 per cent. However, the molecular mechanisms that operate during material deformation and stiffening make stretchability and conductivity fundamentally difficult properties to combine. The macroscale stretching of solids elongates chemical bonds, leading to the reduced overlap and delocalization of electronic orbitals. This conductivity-stretchability dilemma can be exemplified by liquid metals, in which conduction pathways are retained on large deformation but weak interatomic bonds lead to compromised strength. The best-known stretchable conductors use polymer matrices containing percolated networks of high-aspectratio nanometre-scale tubes or nanowires to address this dilemma to some extent6-11. Further improvements have been achieved by using fillers (the conductive component) with increased aspect ratio, of all-metallic composition12, or with specific alignment (the way the fillers are arranged in the matrix). However, the synthesis and separation of high-aspect-ratio fillers is challenging, stiffness increases with the volume content of metallic filler, and anisotropy increases with alignment15. Pre-strained substrates, buckled microwires and three-dimensional microfluidic polymer networks have also been explored. Here we demonstrate stretchable conductors of polyurethane containing spherical nanoparticles deposited by either layer-by-layer assembly or vacuum-assisted flocculation. High conductivity and stretchability were observed in both composites despite the minimal aspect ratio of the nanoparticles. These materials also demonstrate the electronic tunability of mechanical properties, which arise from the dynamic self-organization of the nanoparticles under stress. A modified percolation theory incorporating the self-assembly behaviour of nanoparticles gave an excellent match with the experimental data.}
}
@article{KimetalNATURE-14,
       author = {Kim, Jinseop S. and Greene, Matthew J. and Zlateski, Aleksandar and Lee, Kisuk and Richardson, Mark and Turaga, Srinivas C. and Purcaro, Michael and Balkam, Matthew and Robinson, Amy and Behabadi, Bardia F. and Campos, Michael and Denk, Winfried and Seung, H. Sebastian and the EyeWirers},
        title = {Space-time wiring specificity supports direction selectivity in the retina},
      journal = {Nature},
         year = 2014,
    publisher = {Nature Publishing Group},
       volume = 509,
        issue = 7500,
         year = 2014,
        pages = {331-336},
     abstract = {How does the mammalian retina detect motion? This classic problem in visual neuroscience has remained unsolved for 50 years. In search of clues, here we reconstruct Off-type starburst amacrine cells (SACs) and bipolar cells (BCs) in serial electron microscopic images with help from EyeWire, an online community of "citizen neuroscientists". On the basis of quantitative analyses of contact area and branch depth in the retina, we find evidence that one BC type prefers to wire with a SAC dendrite near the SAC soma, whereas another BC type prefers to wire far from the soma. The near type is known to lag the far type in time of visual response. A mathematical model shows how such "space-time wiring specificity" could endow SAC dendrites with receptive fields that are oriented in space-time and therefore respond selectively to stimuli that move in the outward direction from the soma.},
}
@article{KimetalNATURE-METHODS-17,
       author = {Dalhyung Kim and Jungsoo Kim and Joao C. Marques and Abhinav Grama and David G. C. Hildebrand and Wenchao Gu and Jennifer M. Li and Drew N. Robson},
        title = {Pan-neuronal cellular resolution calcium imaging in freely swimming zebrafish by tracking microscopy},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = {SUBMITTED},
         year = {2017},
}
@article{KimetalNSR-16,
       author = {Kim, Wonjoo and Chattopadhyay, Anupam and Siemon, Anne and Linn, Eike and Waser, Rainer and Rana, Vikas},
        title = {Multistate Memristive Tantalum Oxide Devices for Ternary Arithmetic},
      journal = {Scientific Reports},
         year = {2016},
       volume = {6},
        pages = {36652},
     abstract = {Redox-based resistive switching random access memory (ReRAM) offers excellent properties to implement future non-volatile memory arrays. Recently, the capability of two-state ReRAMs to implement Boolean logic functionality gained wide interest. Here, we report on seven-states Tantalum Oxide Devices, which enable the realization of an intrinsic modular arithmetic using a ternary number system. Modular arithmetic, a fundamental system for operating on numbers within the limit of a modulus, is known to mathematicians since the days of Euclid and finds applications in diverse areas ranging from e-commerce to musical notations. We demonstrate that multistate devices not only reduce the storage area consumption drastically, but also enable novel in-memory operations, such as computing using high-radix number systems, which could not be implemented using two-state devices. The use of high radix number system reduces the computational complexity by reducing the number of needed digits. Thus the number of calculation operations in an addition and the number of logic devices can be reduced.}
}
@misc{KimetalROLAND-INSTITUTE-17,
       author = {Kim, D. and Gu, W. and Li, J.M. and Robson, D.N.},
        title = {Pan-neuronal calcium imaging in freely swimming larval zebrafish by high-speed tracking microscopy},
         year = {2017},
 howpublished = {Roland Institute},
}
@article{KimetalSCIENCE-17,
       author = {Kim, Sung Soo and Rouault, Herv\'{e} and Druckmann, Shaul and Jayaraman, Vivek},
        title = {Ring attractor dynamics in the Drosophila central brain},
         year = {2017},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
     abstract = {Ring attractors are a class of recurrent networks hypothesized to underlie the representation of heading direction. Such network structures, schematized as a ring of neurons whose connectivity depends on their heading preferences, can sustain a bump-like activity pattern whose location can be updated by continuous shifts along either turn direction. We recently reported that a population of fly neurons represents the animal's heading via bump-like activity dynamics. We combined two-photon calcium imaging in head-fixed flying flies with optogenetics to overwrite the existing population representation with an artificial one, which was then maintained by the circuit with naturalistic dynamics. A network with local excitation and global inhibition enforces this unique and persistent heading representation. Ring attractor networks have long been invoked in theoretical work; our study provides physiological evidence of their existence and functional architecture.},
}
@article{KimetalSCIENCE-17.Supplement,
       author = {Kim, Sung Soo and Rouault, Herv{\'e} and Druckmann, Shaul and Jayaraman, Vivek},
        title = {Ring attractor dynamics in the Drosophila central brain},
         year = {2017},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
     abstract = {Supplementary Materials},
}
@inproceedings{KimetalSPIE-11,
        title = {Combined ultrasonic and photoacoustic system for deep tissue imaging},
       author = {Kim, Chulhong and Erpelding, Todd N. and Jankovic, Ladislav and Wang, Lihong V.},
    booktitle = {Proceedings SPIE 7899, Photons Plus Ultrasound: Imaging and Sensing},
         year = 2011,
}
@article{KimmeletalPAMI-11,
       author = {R. Kimmel and C. Zhang and A. M. Bronstein and M. M. Bronstein},
        title = {Are {MSER} features really interesting?},
      journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
        pages = {(to appear)},
         year = 2011,
}
@article{KingmaandWellingCoRR-13,
       author = {Diederik P. Kingma and Max Welling},
        title = {Auto-Encoding Variational {B}ayes},
      journal = {CoRR},
       volume = {arXiv:1312.6114},
         year = 2013,
     abstract = {How can we perform efficient inference and learning in directed probabilistic models, in the presence of continuous latent variables with intractable posterior distributions, and large datasets? We introduce a stochastic variational inference and learning algorithm that scales to large datasets and, under some mild differentiability conditions, even works in the intractable case. Our contributions is two-fold. First, we show that a reparameterization of the variational lower bound yields a lower bound estimator that can be straightforwardly optimized using standard stochastic gradient methods. Second, we show that for i.i.d. datasets with continuous latent variables per datapoint, posterior inference can be made especially efficient by fitting an approximate inference model (also called a recognition model) to the intractable posterior using the proposed lower bound estimator. Theoretical advantages are reflected in experimental results.}
}
@inproceedings{KingmaetalNIPS-14,
       author = {Kingma, Diederik P. and Rezende, Danilo J. and Mohamed, Shakir and Welling, Max},
        title = {Semi-supervised Learning with Deep Generative Models},
    booktitle = {Proceedings of the 27th International Conference on Neural Information Processing Systems},
         year = {2014},
     location = {Montreal, Canada},
        pages = {3581-3589},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
     abstract = {The ever-increasing size of modern data sets combined with the difficulty of obtaining label information has made semi-supervised learning one of the problems of significant practical importance in modern data analysis. We revisit the approach to semi-supervised learning with generative models and develop new models that allow for effective generalisation from small labelled data sets to large unlabelled ones. Generative approaches have thus far been either inflexible, inefficient or non-scalable. We show that deep generative models and approximate Bayesian inference exploiting recent advances in variational methods can be used to provide significant improvements, making generative approaches highly competitive for semi-supervised learning.}
}
@misc{KinneyPhD-09,
        title = {Investigation of neurotransmitter diffusion in three-dimensional reconstructions of hippocampal neuropil},
       author = {Justin Kinney},
         year = 2008,
 howpublished = {Ph.D. Thesis. University of California, San Diego},
}
@article{KipfandWellingCoRR-17,
       author = {Thomas N. Kipf and Max Welling},
        title = {Semi-Supervised Classification with Graph Convolutional Networks},
      journal = {CoRR},
       volume = {arXiv:1609.02907},
         year = {2017},
     abstract = {We present a scalable approach for semi-supervised learning on graph-structured data that is based on an efficient variant of convolutional neural networks which operate directly on graphs. We motivate the choice of our convolutional architecture via a localized first-order approximation of spectral graph convolutions. Our model scales linearly in the number of graph edges and learns hidden layer representations that encode both local graph structure and features of nodes. In a number of experiments on citation networks and on a knowledge graph dataset we demonstrate that our approach outperforms related methods by a significant margin.}
}
@inproceedings{KipfandWellingICLR-17,
       author = {Thomas N. Kipf and Max Welling},
        title = {Semi-Supervised Classification with Graph Convolutional Networks},
    booktitle = {International Conference on Learning Representations (ICLR)},
         year = {2017},
}
@misc{KirbyHELMHOLTZ-06,
        title = {A Tutorial on Helmholtz Machines},
       author = {Kevin G. Kirby},
 howpublished = {Department of Computer Science, Northern Kentucky University},
         year = {2019},
          url = {https://www.nku.edu/~kirby/docs/HelmholtzTutorialKoeln.pdf},
}
@article{KirchnerandThorpeVR-06,
        title = {Ultra-rapid object detection with saccadic eye movements: Visual processing speed revisited},
      journal = {Vision Research},
       volume = 46,
        issue = 11,
        pages = {1762 - 1776},
         year = 2006,
       author = {Holle Kirchner and Simon J. Thorpe},
     abstract = {Previous ultra-rapid go/no-go categorization studies with manual responses have demonstrated the remarkable speed and efficiency with which humans process natural scenes. Using a forced-choice saccade task we show here that when two scenes are simultaneously flashed in the left and right hemifields, human participants can reliably make saccades to the side containing an animal in as little as 120ms. Low level differences between target and distractor images were unable to account for these exceptionally fast responses. The results suggest a very fast and unexpected route linking visual processing in the ventral stream with the programming of saccadic eye movements.}
}
@article{KirchoffetalPTRS_B-18,
       author = {Kirchhoff, Michael and Parr, Thomas and Palacios, Ensor and Friston, Karl and Kiverstein, Julian},
        title = {The {M}arkov blankets of life: Autonomy, active inference and the free energy principle},
      journal = {Journal of the Royal Society, Interface},
         year = {2018},
    publisher = {The Royal Society},
       volume = {15},
        issue = {138},
        pages = {20170792},
     abstract = {This work addresses the autonomous organization of biological systems. It does so by considering the boundaries of biological systems, from individual cells to Home sapiens, in terms of the presence of Markov blankets under the active inference scheme-a corollary of the free energy principle. A Markov blanket defines the boundaries of a system in a statistical sense. Here we consider how a collective of Markov blankets can self-assemble into a global system that itself has a Markov blanket; thereby providing an illustration of how autonomous systems can be understood as having layers of nested and self-sustaining boundaries. This allows us to show that: (i) any living system is a Markov blanketed system and (ii) the boundaries of such systems need not be co-extensive with the biophysical boundaries of a living organism. In other words, autonomous systems are hierarchically composed of Markov blankets of Markov blankets-all the way down to individual cells, all the way up to you and me, and all the way out to include elements of the local environment.},
}
@article{KiritanietalJN-16,
        title = {Hierarchical Connectivity and Connection-Specific Dynamics in the Corticospinal-Corticostriatal Microcircuit in Mouse Motor Cortex (vol 32, pg 4992, 2012)},
       author = {Kiritani, Taro and Wickersham, Ian R. and Seung, H. Sebastian and Shepherd, Gordon M.G.},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {36},
       number = {37},
        pages = {9755-9755},
         year = {2016},
     abstract = {The generation of purposive movement by mammals involves coordinated activity in the corticospinal and corticostriatal systems, which are involved in different aspects of motor control. In the motor cortex, corticospinal and corticostriatal neurons are closely intermingled, raising the question of whether and how information flows intracortically within and across these two channels. To explore this, we developed an optogenetic technique based on retrograde transfection of neurons with deletion-mutant rabies virus encoding channelrhodopsin-2, and used this in conjunction with retrograde anatomical labeling to stimulate and record from identified projection neurons in mouse motor cortex. We also used paired recordings to measure unitary connections. Both corticospinal and callosally projecting corticostriatal neurons in layer 5B formed within-class (recurrent) connections, with higher connection probability among corticostriatal than among corticospinal neurons. In contrast, across-class connectivity was extraordinarily asymmetric, essentially unidirectional from corticostriatal to corticospinal. Corticostriatal neurons in layer 5A and corticocortical neurons (callosal projection neurons similar to corticostriatal neurons) similarly received a paucity of corticospinal input. Connections involving presynaptic corticostriatal neurons had greater synaptic depression, and those involving postsynaptic corticospinal neurons had faster decaying EPSPs. Consequently, the three connections displayed a diversity of dynamic properties reflecting the different combinations of presynaptic and postsynaptic projection neurons. Collectively, these findings delineate a four-way specialized excitatory microcircuit formed by corticospinal and corticostriatal neurons. The "rectifying" corticostriatal-to-corticospinal connectivity implies a hierarchical organization and functional compartmentalization of corticospinal activity via unidirectional signaling from higher-order (corticostriatal) to lower-order (corticospinal) output neurons.},
}
@book{KirkandHwu2010,
      title = {Programming Massively Parallel Processors: A Hands-on Approach},
     author = {David B. Kirk and Wen-mei David W. Hwu}, 
  publisher = {Morgan Kaufmann},
    address = {San Francisco, CA, USA},
       year = 2010,
}
@article{KirkpatrickandSelmanSCIENCE-94,
        title = {Critical Behavior in the Satisfiability of Random Boolean Formuae},
       author = {Kirkpatrick, S. and Selman, B.},
      journal = {Science},
       volume = 264,
         year = 1994,
        pages = {1297-1301},
}
@article{KirkpatrickelalSCIENCE-83,
        title = {Optimization by Simulated Annealing},
       author = {Kirkpatrick, S. and Gelatt, C. D. and Vecchi, M. P.},
      journal = {Science},
       volume = 220,
         year = 1983,
        pages = {671-680},
}
@article{KirkpatricketalCoRR-16,
       author = {James Kirkpatrick and Razvan Pascanu and Neil C. Rabinowitz and Joel Veness and Guillaume Desjardins and Andrei A. Rusu and Kieran Milan and John Quan and Tiago Ramalho and Agnieszka Grabska{-}Barwinska and Demis Hassabis and Claudia Clopath and Dharshan Kumaran and Raia Hadsell},
        title = {Overcoming catastrophic forgetting in neural networks},
      journal = {CoRR},
       volume = {arXiv:1612.00796},
         year = {2016},
      comment = {http://arxiv.org/abs/1612.00796},
     abstract = {The ability to learn tasks in a sequential fashion is crucial to the development of artificial intelligence. Neural networks are not, in general, capable of this and it has been widely thought that catastrophic forgetting is an inevitable feature of connectionist models. We show that it is possible to overcome this limitation and train networks that can maintain expertise on tasks which they have not experienced for a long time. Our approach remembers old tasks by selectively slowing down learning on the weights important for those tasks. We demonstrate our approach is scalable and effective by solving a set of classification tasks based on the MNIST hand written digit dataset and by learning several Atari 2600 games sequentially.},
}
@phdthesis{KirmanPhD,
       author = {Kirman Jak},
        title = {Predicting Real-Time Planner Performance by Domain Characterization},
       school = {Department of Computer Science, Brown University},
         year = 1994,
}
@inproceedings{KirmanetalEWSP-94,
       author = {Kirman, Jak and Nicholson, Ann and Lejter, Moises and Dean, Thomas and Santos Jr., Eugene},
        title = {Using Goals to Find Plans with High Expected Utility},
       editor = {Sandewall, E. and Backstrom, C.},
    booktitle = {Current Trends in {AI} Planning},
    publisher = {IOS Press},
      address = {Amsterdam},
         year = 1994,
}
@inproceedings{KirmanetalRA-91,
       author = {Kirman, Jak and Basye, Kenneth and Dean, Thomas},
        title = {Sensor Abstractions for Control of Navigation},
    booktitle = {Proceedings of the {IEEE} International Conference on Robotics and Automation},
     location = {Sacramento, California},
        pages = {2812-2817},
         year = 1991,
}
@article{KirosetalCoRR-15,
        title = {Skip-Thought Vectors},
       author = {Ryan Kiros and Yukun Zhu and Ruslan Salakhutdinov and Richard S. Zemel and Antonio Torralba and Raquel Urtasun and Sanja Fidler},
      journal = {CoRR},
       volume = {arXiv:1506.06726},
         year = 2015,
     abstract = {We describe an approach for unsupervised learning of a generic, distributed sentence encoder. Using the continuity of text from books, we train an encoder-decoder model that tries to reconstruct the surrounding sentences of an encoded passage. Sentences that share semantic and syntactic properties are thus mapped to similar vector representations. We next introduce a simple vocabulary expansion method to encode words that were not seen as part of training, allowing us to expand our vocabulary to a million words. After training our model, we extract and evaluate our vectors with linear models on 8 tasks: semantic relatedness, paraphrase detection, image-sentence ranking, question-type classification and 4 benchmark sentiment and subjectivity datasets. The end result is an off-the-shelf encoder that can produce highly generic sentence representations that are robust and perform well in practice. We will make our encoder publicly available.}
}
@techreport{KissmeyerandTallantTR-86,
       author = {Kissmeyer, Kirsten Y. and Tallant, Anne M.},
        title = {MACPLAN, a Mixed-Initiative Approach to Airlift Planning},
  institution = {MITRE Corporation},
         year = 1986,
}
@book{Kita2003,
       editor = {Sotaro Kita},
        title = {Pointing: Where language, culture, and cognition meet},
    publisher = {Lawrence Erlbaum Associates, Inc.},
      address = {Mahwah, NJ, USA},
         year = 2003
}
@article{KitabatakeetalNCNM-07,
       author = {Kitabatake, Y. and Sailor, K. A. and Ming, G. L. and Song, H. },
        title = {Adult neurogenesis and hippocampal memory function: new cells, more plasticity, new memories?},
      journal = {Neurosurgery Clinics of North America},
         year = {2007},
       volume = {18},
       number = {1},
        pages = {105-113},
     abstract = {The discovery of active adult neurogenesis in mammals, a process of generating functional neurons from neural stem cells, suggests that the adult brain is more dynamic than once imagined. The coincidence of this phenomenon occurring in the hippocampus, a region critical to the learning process, begs the question of whether adult neurogenesis is involved in memory formation. Here, the authors review rapidly accumulating evidence showing a strong correlation between certain types of memory functions and adult neurogenesis in the hippocampus. Establishment of the potential link between memory formation and adult neurogenesis is instrumental, at a basic science level, to understand the function of neural networks and is essential, at a clinical level, to develop effective therapies for various cognitive dysfunctions.}
}
@article {KitamuraetalSCIENCE-17,
       author = {Kitamura, Takashi and Ogawa, Sachie K. and Roy, Dheeraj S. and Okuyama, Teruhiro and Morrissey, Mark D. and Smith, Lillian M. and Redondo, Roger L. and Tonegawa, Susumu},
        title = {Engrams and circuits crucial for systems consolidation of a memory},
      journal = {Science},
       volume = {356},
       number = {6333},
        pages = {73-78},
         year = {2017},
    publisher = {American Association for the Advancement of Science},
     abstract = {Memories are thought to be formed in the hippocampus and later moved to the neocortex for long-term storage. However, little is known about the mechanisms that underlie the formation and maturation of neocortical memories and their interaction with the hippocampal network. Kitamura et al. discovered that at the onset of learning, neurons for contextual fear memory are quickly produced in the prefrontal cortex. This process depends on the activity of afferents from both the hippocampus and the amygdala. Over time, the prefrontal neurons consolidate their role in memory expression. In contrast, the hippocampal neurons slowly lose this function.Science, this issue p. 73Episodic memories initially require rapid synaptic plasticity within the hippocampus for their formation and are gradually consolidated in neocortical networks for permanent storage. However, the engrams and circuits that support neocortical memory consolidation have thus far been unknown. We found that neocortical prefrontal memory engram cells, which are critical for remote contextual fear memory, were rapidly generated during initial learning through inputs from both the hippocampal{\textendash}entorhinal cortex network and the basolateral amygdala. After their generation, the prefrontal engram cells, with support from hippocampal memory engram cells, became functionally mature with time. Whereas hippocampal engram cells gradually became silent with time, engram cells in the basolateral amygdala, which were necessary for fear memory, were maintained. Our data provide new insights into the functional reorganization of engrams and circuits underlying systems consolidation of memory.},
}
@inproceedings{KivinenetalICCV-07,
        title = {Learning Multiscale Representations of Natural Scenes Using Dirichlet Processes},
       author = {Jyri J. Kivinen and Erik B. Sudderth and Michael I. Jordan},
    booktitle = {Proceedings of the 11th {IEEE} International Conference on Computer Vision},
         year = 2007,
}
@inproceedings{KlaeseretalBMVC-08,
       author = {Kl\"{a}ser, Alexander and Marszalek, Marcin and Schmid, Cordelia},
        title = {A spatio-temporal descriptor based on 3{D}-gradients},
    booktitle = {Proceedings of the British Machine Vision Conference},
         year = 2008,
        pages = {995-1004},
}
@article{KleckneretalNATURE-17,
        title = {Evidence for a large-scale brain system supporting allostasis and interoception in humans},
       author = {Kleckner, Ian and Zhang, Jiahe and Touroutoglou, Alexandra and Chanes, Lorena and Xia, Chenjie and Simmons, W Kyle and Quigley, Karen and Dickerson, Bradford and Barrett, Lisa},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Labs Journals},
         year = {2017},
        pages = {098970},
     abstract = {Large-scale intrinsic brain systems have been identified for exteroceptive senses (such as sight, hearing and touch). We introduce an analogous system for representing sensations from within the body, called interoception, and demonstrate its relation to regulating peripheral systems in the body, called allostasis. Employing the recently introduced Embodied Predictive Interoception Coding (EPIC) model, we used tract-tracing studies of macaque monkeys, followed by two intrinsic functional magnetic resonance imaging samples (N = 280 and N = 270) to evaluate the existence of an intrinsic allostatic-interoceptive system in the human brain. Another sample (N = 41) allowed us to evaluate the convergent validity of the hypothesized allostatic-interoceptive system by showing that individuals with stronger connectivity between system hubs performed better on an implicit index of interoceptive ability related to autonomic fluctuations. Implications include insights for the brain's functional architecture, dissolving the artificial boundary between mind and body, and unifying mental and physical illness.}
}
@incollection{KleeandMinty72,
       author = {Victor Klee and G. J. Minty},
        title = {How good is the simplex algorithm?},
    booktitle = {Inequalities, {III}},
         year = 1972,
       editor = {O. Shisha},
    publisher = {Academic Press},
      address = {New York},
        pages = {159-175},
}
@book{Klein1998sources,
        title = {Sources of Power: {H}ow People Make Decisions},
       author = {Klein, G.A.},
         year = {1998},
    publisher = {MIT Press}
}
@article{KleinQJEO-16,
       author = {Klein, S.B.},
        title = {Autonoetic consciousness: Reconsidering the role of episodic memory in future-oriented self-projection},
      journal = {Quarterly Journal of Experimental Psychology},
         year = {2016},
       volume = {69},
       number = {2},
        pages = {381-401},
     abstract = {Following the seminal work of Ingvar (1985. "Memory for the future": An essay on the temporal organization of conscious awareness. Human Neurobiology, 4, 127-136), Suddendorf (1994. The discovery of the fourth dimension: Mental time travel and human evolution. Master's thesis. University of Waikato, Hamilton, New Zealand), and Tulving (1985. Memory and consciousness. Canadian Psychology, 26, 1-12), exploration of the ability to anticipate and prepare for future contingencies that cannot be known with certainty has grown into a thriving research enterprise. A fundamental tenet of this line of inquiry is that future-oriented mental time travel, in most of its presentations, is underwritten by a property or an extension of episodic recollection. However, a careful conceptual analysis of exactly how episodic memory functions in this capacity has yet to be undertaken. In this paper I conduct such an analysis. Based on conceptual, phenomenological, and empirical considerations, I conclude that the autonoetic component of episodic memory, not episodic memory per se, is the causally determinative factor enabling an individual to project him or herself into a personal future.}
}
@article{KleinbergHITS,
       author = {Jon M. Kleinberg},
        title = {Authoritative Sources in a Hyperlinked Environment},
      journal = {Journal of the ACM},
       volume = 46,
        issue = 5,
        pages = {604-632},
         year = 1999,
}
@inproceedings{KleinbergSODA-98,
       author = {J. Kleinberg},
        title = {Authoritative sources in a hyperlinked environment},
    booktitle = {Proceedings of 9th Annual ACM-SIAM Symposium on Discrete Algorithms},
        pages = {668-677},
         year = 1998,
}
@inproceedings{KleinbergSTOC-00,
       author = {Jon M. Kleinberg},
        title = {The {S}mall-{W}orld {P}henomenon: An Algorithmic Perspective},
    booktitle = {Proceedings of the 32nd ACM Symposium on Theory of Computing},
          url = {http://citeseer.nj.nec.com/kleinberg00smallworld.html},
         year = 2000,
}
@inproceedings{Kleinbergetal99,
       author = {J. Kleinberg and R. Kumar and P. Raghavan and S. Rajagopalan and A. Tomkins},
        title = {The {W}eb as a graph: Measurements, Models and Methods},
    booktitle = {Proceedings of the International Conference on Combinatorics and Computing},
        pages = {},
         year = 1999,
}
@inproceedings{KleinbergetalCC-99,
       author = {J. Kleinberg and S.R. Kumar and P. Raghavan and S. Rajagopalan and A. Tomkins},
        title = {The {W}eb as a graph: Measurements, models and methods},
    booktitle = {Proceedings of the International Conference on Combinatorics and Computing},
         year = 1999,
}
@article{KleinetalAJNR-16,
       author = {Klein, A.P. and Ulmer, J.L. and Quinet, S.A. and Mathews, V. and Mark, L.P.},
        title = {Nonmotor Functions of the Cerebellum: An Introduction},
      journal = {American Journal of Neuroradiology},
       volume = {37},
       number = {6},
         year = {2016},
	pages = {1005-1009},
}
@article{KlinzingetalNBM-16,
       author = {Klinzing, J. G. and Rasch, B.  and Born, J. and Diekelmann, S.},
        title = {{{S}leep's role in the reconsolidation of declarative memories}},
      journal = {Neurobiological Learning and Memory},
       volume = {136},
         year = {2016},
        pages = {166-173},
     abstract = {Sleep is known to support the consolidation of newly encoded and initially labile memories. Once consolidated, remote memories can return to a labile state upon reactivation and need to become reconsolidated in order to persist. Here we asked whether sleep also benefits the reconsolidation of remote memories after their reactivation and how reconsolidation during sleep compares to sleep-dependent consolidation processes. In three groups, participants were trained on a visuo-spatial learning task in the presence of a contextual odor. Participants in the 'reconsolidation' group learned the task on day 1. On day 2, they were subjected to a reactivation procedure by presenting the odor cue and a mock recall test in the learning context before a 40-min sleep or wake period. Participants in the 'remote consolidation' group followed the same procedure but did not receive reactivation on day 2. Participants in the 'recent consolidation' group skipped the procedure on day 1 and learned the task immediately before the sleep or wake period. After the sleep or wake interval, memory stability was tested in all subjects. The results show that this short 40-min sleep period significantly facilitated the reconsolidation of reactivated memories, whereas the consolidation of non-reactivated remote memories was less affected and recently encoded memories did not benefit at all. These findings tentatively suggest that sleep has a beneficial effect on the reconsolidation of remote memories, acting at a faster rate than sleep-associated consolidation.}
}
@article{KlocknerandPintoPARALLEL-COMPUTING-11,
        title = {PyCUDA and PyOpenCL: A Scripting-Based Approach to GPU Run-Time Code Generation},
       author = {Kl{\"o}ckner, Andreas and Pinto, Nicolas and Lee, Yunsup and Catanzaro, Bryan and Ivanov, Paul and Fasih, Ahmed},
      journal = {Parallel Computing},
    publisher = {Elsevier},
       volume = {38},
        issue = {3},
         year = {2012},
        pages = {157-174},
}
@techreport{KloeckneretalTR-09,
        title = {{PyCUDA: GPU} Run-Time Code Generation for High-Performance Computing},
       author = {A. Kl\"{o}eckner and N. Pinto and Y. Lee and B. Catanzaro and P. Ivanov and A. Fasih},
  institution = {Scientific Computing Group, Brown University},
        issue = {2009-40},
      address = {Providence, RI, USA},
         year = 2009,
}
@incollection{KnappandMortonEXECUTIVE-FUNCTION-17,
        title = {Executive Functioning: A Developmental Cognitive Neuroscience Perspective},
       author = {Katie Knapp and Bruce Morton},
    booktitle = {Executive Functions in Children's Everyday Lives: A Handbook for Professionals in Applied Psychology},
       editor = {Hoskyn, M.J. and Iarocci, G. and Young, A.R.},
    publisher = {Oxford University Press},
         year = {2017},
        pages = {9-17}
}
@article{KneusselandWagnerNATURE-13,
       author = {Kneussel, Matthias and Wagner, Wolfgang},
        title = {Myosin motors at neuronal synapses: drivers of membrane transport and actin dynamics},
      journal = {Nature Reviews Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {14},
        issue = {4},
         year = {2013},
        pages = {233-247},
}
@inproceedings{KnoblockAAAI-91,
       author = {Knoblock, Craig A.},
        title = {Search Reduction in Hierarchical Problem Solving},
    booktitle = {Proceedings AAAI-91},
     location = {Anaheim, California},
 organization = {AAAI},
        pages = {686-691},
         year = 1991,
}
@book{Knudsen99,
      author = {Knudsen, J. B.},
       title = {The Unofficial Guide to Lego Mindstorms Robots},
   publisher = {O'Reilly},
     address = {Sebastopol, CA},
        year = 1999,
}
@book{Knuth01,
      author = {Donald Knuth},
       title = {Things a Computer Scientist Rarely Talks About},
   publisher = {University of Chicago Press},
     comment = {Originally by CSLI Publications},
     address = {Chicago, IL},
        year = 2001,
}
@book{KnuthTAOCP1,
      author = {Donald Knuth},
       title = {Fundamental Algorithms},
   publisher = {Third edition, Addison-Wesley},
     address = {Reading, MA},
        year = 1997,
}
@book{KnuthTAOCP2,
      author = {Donald Knuth},
       title = {Seminumerical Algorithms},
   publisher = {Third edition, Addison-Wesley},
     address = {Reading, MA},
        year = 1997,
}
@book{KnuthTAOCP3,
      author = {Donald Knuth},
       title = {Sorting and Searching},
   publisher = {Third edition, Addison-Wesley},
     address = {Reading, MA},
        year = 1998,
}
@inproceedings{KoberetalSPIE-79,
       author = {Kober, W. and Grosch, C. and Rosenthal, T.},
        title = {Three Dimensional Surface Shell Correlator},
    booktitle = {Proceedings of the {SPIE} Conference on Advances in Intelligent Robotic Systems},
 organization = {SPIE},
       volume = 186,
         year = 1979,
}
@article{KocabasPNAS-08,
        title = {Radio frequency analog electronics based on carbon nanotube transistors},
       author = {Kocabas, C. and Kim, H. and Banks, T. and Rogers, J.A. and Pesetski, A.A. and Baumgardner, J.E. and Krishnaswamy, S.V. and Zhang, H.},
      journal = {Proceedings of the National Academy of Science},
       volume = 105,
        issue = 5,
         year = 2008,
        pages = {1405-1409},
     abstract = {The potential to exploit single-walled carbon nanotubes (SWNTs) in advanced electronics represents a continuing, major source of interest in these materials. However, scalable integration of SWNTs into circuits is challenging because of difficulties in controlling the geometries, spatial positions, and electronic properties of individual tubes. We have implemented solutions to some of these challenges to yield radio frequency (RF) SWNT analog electronic devices, such as narrow band amplifiers operating in the VHF frequency band with power gains as high as 14 dB. As a demonstration, we fabricated nanotube transistor radios, in which SWNT devices provide all of the key functions, including resonant antennas, fixed RF amplifiers, RF mixers, and audio amplifiers. These results represent important first steps to practical implementation of SWNTs in high-speed analog circuits. Comparison studies indicate certain performance advantages over silicon and capabilities that complement those in existing compound semiconductor technologies.}
}
@book{Koch2004,
        title = {The Quest for Consciousness: A Neurobiological Approach},
       author = {Christof Koch},
         year = {2004},
    publisher = {MIT Press}
}
@book{Koch2012,
        title = {Consciousness: Confessions of a Romantic Reductionist},
       author = {Christof Koch},
         year = {2012},
    publisher = {MIT Press}
}
@incollection{KochFICN-12,
       author = {Koch, Christof},
        title = {Project {MindScope}},
    booktitle = {Frontiers in Computational Neuroscience},
    publisher = {Bernstein Conference Proceedings},
     location = {Munich, Germany},
         year = {2012},
       series = {33},
     abstract = {The Allen Institute for Brain Science is engaged in a ten year project to study the principles by which information is encoded, transformed and represented in the mammalian cerebral cortex and related structures. The goals of MindScope are to provide a quantitative taxonomy of cell types and their interconnections in visual cortex and associated brain regions, to observe their dynamics under physiological conditions in behaving mice, to construct cellular models of how their dynamics and function arise from the structural description, and to understand how this  function relates to visual perception. This is a large-scale, in-house team effort to synthesize anatomical, physiological and theoretical knowledge into a description of the wiring scheme of the cortex, at both the structural and the functional levels. The fruits of this cerebroscope will be freely available to the public.}
}
@book{KochanandWood89,
      author = {Stephen G. Kochan and Patrick H. Wood},
       title = {Unix Shell Programming},
   publisher = {Second edition, Sams},
     address = {Indianapolis, IN},
        year = 1989,
}
@article{KochandBuiceCELL-15,
       author = {Christof Koch and Michael A. Buice},
        title = {A Biological Imitation Game},
      journal = {Cell},
       volume = {163},
       number = {2},
        pages = {277-280},
         year = {2015},
     abstract = {The digital reconstruction of a slice of rat somatosensory cortex from the Blue Brain Project provides the most complete simulation of a piece of excitable brain matter to date. To place these efforts in context and highlight their strengths and limitations, we introduce a Biological Imitation Game, based on Alan Turing's Imitation Game, that operationalizes the difference between real and simulated brains.}
}
@article{KochandClayNATURE-12,
        title = {Neuroscience: Observatories of the mind},
       author = {Koch, Christof and Reid, R. Clay},
      journal = {Nature},
       volume = 483,
        issue = 7390,
         year = 397,
        pages = 398,
}
@article{KochandHeppNATURE-06,
       author = {Koch, C. and Hepp, K.},
        title = {Quantum mechanics in the brain},
      journal = {Nature},
       volume = {440},
       number = {7084},
         year = {2006},
        pages = {611-612},
     abstract = {Does the enormous computing power of neurons mean consciousness can be explained within a purely neurobiological framework, or is there scope for quantum computation in the brain?},
}
@book{KochandSegev1998,
       editor = {Koch, Christof and Segev, Idan},
        title = {Methods in Neuronal Modeling: From Ions to Networks},
         year = 1998,
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
} 
@article{KochandUllmanHN-85,
       author = {Christof Koch and Shimon Ullman},
        title = {Shifts in selective visual attention: towards the underlying neural circuitry},
      journal = {Human Neurobiology},
       volume = 4,
        issue = 44,
        pages = {219-227},
         year = 1985,
}
@article{KodandaramaiahetalNATURE-12,
        title = {{Automated whole-cell patch-clamp electrophysiology of neurons in vivo}},
       author = {Kodandaramaiah, Suhasa B. and Franzesi, Giovanni T. and Chow, Brian Y. and Boyden, Edward S. and Forest, Craig R.},
      journal = {Nature Methods},
       volume = 9,
        issue = 6,
         year = 2012,
        pages = {585-587},
}
@incollection{KoditschekEAI-87,
       author = {Koditschek, D.},
        title = {Robot Control Systems},
     crossref = {EAI-87},
        pages = {902-923},
}
@inproceedings{KoditschekIEEE-87,
       author = {Koditschek, D.},
        title = {Exact Robot Navigation by Means of Potential Functions: Some Topological Considerations},
    booktitle = {{IEEE} International Conference on Robotics and Automation},
     location = {Raleigh, NC},
         year = 1987,
        pages = {1-6},
}
@incollection{KoditschekRR-89,
       author = {Koditschek, D.},
        title = {Robot Planning and Control via Potential Functions},
     crossref = {KhatibetalRR-89},
        pages = {349-367},
}
@article{KoechlinandJubaultNEURON-06,
        title = {Broca's Area and the Hierarchical Organization of Human Behavior},
       author = {Koechlin, Etienne and Jubault, Thomas},
      journal = {Neuron},
       volume = 50,
        issue = 6,
         year = 2006,
        pages = {963-974},
     abstract = {The prefrontal cortex subserves executive control, i.e., the organization of action or thought in relation to internal goals. This brain region hosts a system of executive processes extending from premotor to the most anterior prefrontal regions that governs the temporal organization of behavior. Little is known, however, about the prefrontal executive system involved in the hierarchical organization of behavior. Here, we show using magnetic resonance imaging in humans that the posterior portion of the prefrontal cortex, including Broca's area and its homolog in the right hemisphere, contains a system of executive processes that control start and end states and the nesting of functional segments that combine in hierarchically organized action plans. Our results indicate that Broca's area and its right homolog process hierarchically structured behaviors regardless of their temporal organization, suggesting a fundamental segregation between prefrontal executive systems involved in the hierarchical and temporal organization of goal-directed behaviors.},
}
@article{KoechlinandSummerfieldTiCS-07,
       author = {Koechlin, E. and Summerfield, C.},
        title = {An information theoretical approach to prefrontal executive function},
      journal = {Trends in Cognitive Science},
         year = {2007},
       volume = {11},
       number = {6},
        pages = {229-235},
     abstract = {The prefrontal cortex subserves executive control--that is, the ability to select actions or thoughts in relation to internal goals. Here, we propose a theory that draws upon concepts from information theory to describe the architecture of executive control in the lateral prefrontal cortex. Supported by evidence from brain imaging in human subjects, the model proposes that action selection is guided by hierarchically ordered control signals, processed in a network of brain regions organized along the anterior-posterior axis of the lateral prefrontal cortex. The theory clarifies how executive control can operate as a unitary function, despite the requirement that information be integrated across multiple distinct, functionally specialized prefrontal regions.}
} 
@article{KoechlinetalPNAS-00,
      journal = {Proceedings of the National Academy of Sciences},
       author = {Koechlin, Etienne and Corrado, Gregory and Pietrini, Pietro and Grafman, Jordan},
        title = {Dissociating the role of the medial and lateral anterior prefrontal cortex in human planning},
       volume = 97,
        issue = 13,
         year = 2000,
        pages = {7651-7656},
     abstract = {The anterior prefrontal cortex is known to subserve higher cognitive functions such as task management and planning. Less is known, however, about the functional specialization of this cortical region in humans. Using functional MRI, we report a double dissociation: the medial anterior prefrontal cortex, in association with the ventral striatum, was engaged preferentially when subjects executed tasks in sequences that were expected, whereas the polar prefrontal cortex, in association with the dorsolateral striatum, was  involved preferentially when subjects performed tasks in sequences that  were contingent on unpredictable events. These results parallel the  functional segregation previously described between the medial and  lateral premotor cortex underlying planned and contingent motor control  and extend this division to the anterior prefrontal cortex, when task  management and planning are required. Thus, our findings support the  assumption that common frontal organizational principles underlie motor  and higher executive functions in humans.},
}
@article{KoechlinetalSCIENCE-03,
       author = {Etienne Koechlin and Chryst\`{e}le Ody and Fr\'{e}d\'{e}rique Kouneiher},
        title = {The architecture of cognitive control in the human prefrontal cortex},
      journal = {Science},
       volume = 302,
         year = 2003,
        pages = {1181-1185},
     abstract = {The prefrontal cortex (PFC) subserves cognitive control: the ability to coordinate thoughts or actions in relation with internal goals. Its functional architecture, however, remains poorly understood. Using brain imaging in humans, we showed that the lateral PFC is organized as a cascade of executive processes from premotor to anterior PFC regions that control behavior according to stimuli, the present perceptual context, and the temporal episode in which stimuli occur, respectively. The results support an unified modular model of cognitive control that describes the overall functional organization of the human lateral PFC and has basic methodological and theoretical implications.},
}
@article{KoepplandHaeuslerWCSB-09,
       author = {H. Koeppl and S. Haeusler},
        title = {Motifs, algebraic connectivity and computational performance of two data-based cortical circuit templates},
      journal = {Proceedings of the sixth International Workshop on Computational Systems Biology},
         year = {2009},
        pages = {83-86},
     abstract = {Based on extensive experimental work two templates for the laminae-specific connectivity pattern of cortical microcircuits have recently been published. We examine their structural and computational properties and attempt to correlate those properties. To characterize structure or circuit topology we consider the distribution of network motifs as well the algebraic connectivity of the network. In order to understand the computational properties of these two templates, we have generated computer models of them, consisting of Hodgkin-Huxley point neurons with conductance based synapses that have a biologically realistic short-term plasticity. The performance is assessed using nine generic computational tasks that require accumulation and merging of information contained in two afferent spike inputs.},
}
@article{KoetalBIOMOLECULES-09,
        title = {DNA nanotubes as combinatorial vehicles for cellular delivery},
       author = {Ko, S. and Liu, H. and Chen, Y. and Mao, C.},
      journal = {Biomacromolecules},
       volume = 9,
        issue = 11,
        pages = {3039-43},
         year = 2008,
     abstract = {This work explores using self-assembled DNA nanostructures as carriers for drug delivery. We have recently developed an organic nanotube system that is assembled from a single component: a 52-base-long DNA single strand. In this work, functional agents (folate as a cancer cell target agent and Cy3 as a fluorescence imaging agent) are conjugated with the DNA strands; the conjugates self-assemble into micrometers-long nanotubes (NTs). The conjugated DNA-NTs can be effectively taken by cancer cells as demonstrated by fluorescence imaging and fluorescence-activated cell sorting. No obvious toxicity has been observed under current experimental conditions.}
} 
@article{KoetalNATURE-11,
       author = {Ko, Ho and Hofer, Sonja B. and Pichler, Bruno and Buchanan, Katherine A. and Sjostrom, P. Jesper and Mrsic-Flogel, Thomas D.},
        title = {Functional specificity of local synaptic connections in neocortical networks},
      journal = {Nature},
         year = 2011,
    publisher = {Nature Publishing Group},
       volume = 473,
        issue = 7345,
        pages = {87-91},
     abstract = {Neuronal connectivity is fundamental to information processing in the brain. Therefore, understanding the mechanisms of sensory processing requires uncovering how connection patterns between neurons relate to their function. On a coarse scale, long-range projections can preferentially link cortical regions with similar responses to sensory stimuli1-4. But on the local scale, where den- drites and axons overlap substantially, the functional specificity of connections remains unknown. Here we determine synaptic connectivity between nearby layer 2/3 pyramidal neurons in vitro, the response properties of which were first characterized in mouse visual cortex in vivo. We found that connection probability was related to the similarity of visually driven neuronal activity. Neurons with the same preference for oriented stimuli connected at twice the rate of neurons with orthogonal orientation preferences. Neurons responding similarly to naturalistic stimuli formed connections at much higher rates than those with uncorrelated responses. Bidirectional synaptic connections were found more frequently between neuronal pairs with strongly correlated visual responses. Our results reveal the degree of functional specificity of local synaptic connections in the visual cortex, and point to the existence of fine-scale subnetworks dedicated to processing related sensory information.},
}
@article{KoetalNATURE-13,
        title = {The emergence of functional microcircuits in visual cortex},
       author = {Ko, Ho and Cossell, Lee and Baragli, Chiara and Antolik, Jan and Clopath, Claudia and Hofer, Sonja B. and Mrsic-Flogel, Thomas D.},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = 496,
        issue = 7443,
         year = 2013,
        pages = {96-100},
     abstract = {n primary visual cortex (V1), connectivity between layer 2/3 (L2/3) excitatory neurons undergoes extensive reorganization after the onset of visual experience whereby neurons with similar feature selectivity form functional microcircuits (Ko et al., 2011, 2013). It remains unknown whether visual experience is required for the developmental refinement of intracortical circuitry or whether this maturation is guided intrinsically. Here, we correlated the connectivity between V1 L2/3 neurons assayed by simultaneous whole-cell recordings in vitro to their response properties measured by two-photon calcium imaging in vivo in dark-reared mice. We found that neurons with similar responses to oriented gratings or natural movies became preferentially connected in the absence of visual experience. However, the relationship between connectivity and similarity of visual responses to natural movies was not as strong in dark-reared as in normally reared mice. Moreover, dark rearing prevented the normally occurring loss of connections between visually nonresponsive neurons after eye opening (Ko et al., 2013). Therefore, our data suggest that the absence of visual input does not prevent the emergence of functionally specific recurrent connectivity in cortical circuits; however, visual experience is required for complete microcircuit maturation.}
}
@article{KohanetalCoRR-18,
       author = {Adam A. Kohan and Edward A. Rietman and Hava T. Siegelmann},
        title = {Error Forward-Propagation: Reusing Feedforward Connections to Propagate Errors in Deep Learning},
      journal = {CoRR},
       volume = {arxiv:1808.03357},
         year = {2018},
     abstract = {We introduce Error Forward-Propagation, a biologically plausible mechanism to propagate error feedback forward through the network. Architectural constraints on connectivity are virtually eliminated for error feedback in the brain; systematic backward connectivity is not used or needed to deliver error feedback. Feedback as a means of assigning credit to neurons earlier in the forward pathway for their contribution to the final output is thought to be used in learning in the brain. How the brain solves the credit assignment problem is unclear. In machine learning, error backpropagation is a highly successful mechanism for credit assignment in deep multilayered networks. Backpropagation requires symmetric reciprocal connectivity for every neuron. From a biological perspective, there is no evidence of such an architectural constraint, which makes backpropagation implausible for learning in the brain. This architectural constraint is reduced with the use of random feedback weights. Models using random feedback weights require backward connectivity patterns for every neuron, but avoid symmetric weights and reciprocal connections. In this paper, we practically remove this architectural constraint, requiring only a backward loop connection for effective error feedback. We propose reusing the forward connections to deliver the error feedback by feeding the outputs into the input receiving layer. This mechanism, Error Forward-Propagation, is a plausible basis for how error feedback occurs deep in the brain independent of and yet in support of the functionality underlying intricate network architectures. We show experimentally that recurrent neural networks with two and three hidden layers can be trained using Error Forward-Propagation on the MNIST and Fashion MNIST datasets, achieving 1.90\% and 11\% generalization errors respectively.},
}
@article{KohatsuandYamamotoNATURE-COMMUNICATIONS-15,
       author = {Kohatsu, Soh and Yamamoto, Daisuke},
        title = {Visually induced initiation of Drosophila innate courtship-like following pursuit is mediated by central excitatory state},
      journal = {Nature Communications},
    publisher = {Nature Publishing Group},
       volume = {6},
         year = {2015},
        pages = {6457},
     abstract = {The courtship ritual of male Drosophila represents an innate behaviour that is initiated by female-derived sensory stimuli. Here we report that moving light spots can induce courtshiplike following pursuit in tethered wild-type male flies provided the fly is primed by optogenetic stimulation of specific dsx-expressing neuronal clusters in the lateral protocerebrum (LPR). Namely, stimulation of the pC1 neuronal cluster initiates unilateral wing extension and vibration of both sides, whereas stimulation of the pC2l cluster initiates only contralateral wing displays. In addition, stimulation of pC2l but not pC1 neurons induced abdominal bending and proboscis extension. Ca2+ imaging of the pC1 cluster revealed periodic Ca2+ rises, each corresponding to a turn of the male fly during courtship. In contrast, group-reared fruit mutant males exhibit light spot-induced courtship pursuit without optogenetic priming.
Ca2+ imaging revealed enhanced responses of LPR neurons to visual stimuli in the mutants,
suggesting a neural correlate of the light spot-induced courtship behaviour.}
}
@book{Kohler1925mentality,
        title = {The Mentality of Apes},
       author = {K\"{o}hler, W. and Winter, E.},
         year = {1925},
    publisher = {Harcourt, Brace \& World}
}
@article{KohonenBC-82,
       author = {Kohonen, Teuvo},
        title = {Self-Organized Formation of Topologically Correct Feature Maps},
      journal = {Biological Cybernetics},
       volume = 43,
         year = 1982,
        pages = {59-69},
}
@article{KolaitisandVardiIC-90,
       author = {Kolaitis, P. and Vardi, M.},
        title = {0-1 Laws and Decision Problems for Fragments of Second-Order Logic},
      journal = {Information and Computation},
       volume = 87,
        pages = {302-338},
         year = 1990,
}
@article{KolaitisandVardiIC-92,
       author = {Kolaitis, P. and Vardi, M.},
        title = {Infinitary Logics and 0-1 Laws},
      journal = {Information and Computation},
       volume = 98,
        pages = {258-294},
         year = 1992,
}
@inproceedings{KolaitisandVardiLICS-88,
       author = {Kolaitis, P. and Vardi, M.},
        title = {0-1 Laws and Decision Problems for Fragments of Second-Order Logic},
    booktitle = {Proceedings of the Third {IEEE} Symposium on Logic in Computer Science},
         year = 1986,
        pages = {2-11},
}
@article{KolbetalCLICK-01,
       author = {Kolb, Hartmuth C. and Finn, M. G. and Sharpless, K. Barry},
        title = {Click Chemistry: Diverse Chemical Function from a Few Good Reactions},
      journal = {Angewandte Chemie International Edition},
       volume = 40,
        issue = 11,
         year = 2001,
        pages = {2004-2021},
     abstract = {Examination of nature's favorite molecules reveals a striking preference for making carbon-heteroatom bonds over carbon-carbon bonds--surely no surprise given that carbon dioxide is nature's starting material and that most reactions are performed in water. Nucleic acids, proteins, and polysaccharides are condensation polymers of small subunits stitched together by carbon-heteroatom bonds. Even the 35 or so building blocks from which these crucial molecules are made each contain, at most, six contiguous C-C bonds, except for the three aromatic amino acids. Taking our cue from nature's approach, we address here the development of a set of powerful, highly reliable, and selective reactions for the rapid synthesis of useful new compounds and combinatorial libraries through heteroatom links (C-X-C), an approach we call "click chemistry". Click chemistry is at once defined, enabled, and constrained by a handful of nearly perfect "spring-loaded" reactions. The stringent criteria for a process to earn click chemistry status are described along with examples of the molecular frameworks that are easily made using this spartan, but powerful, synthetic strategy.},
}
@article{KolderandBaderSIAM-08,
       author = {Tamara G. Kolda and Brett W. Bader},
        title = {Tensor Decompositions and Applications},
      journal = {{SIAM} Review},
         year = 2009,
       volume = 51,
        issue = 3,
         year = 2009,
        pages = {455-500},
     abstract = {This survey provides an overview of higher-order tensor decompositions, their applications, and available software. A tensor is a multidimensional or N-way array. Decompositions of higher-order tensors (i.e., N-way arrays with N $\geq$ 3) have applications in psychometrics, chemometrics, signal processing, numerical linear algebra, computer vision, numerical analysis, data mining, neuroscience, graph analysis, and elsewhere. Two particular tensor decompositions can be considered to be higher-order extensions of the matrix singular value decomposition: CANDECOMP/PARAFAC (CP) decomposes a tensor as a sum of rank-one tensors, and the Tucker decomposition is a higher-order form of principal component analysis.},
} 
@article{KoleandStuartNEURON-12,
       author = {Kole, Maarten H. P. and Stuart, Greg J.},
        title = {Signal Processing in the Axon Initial Segment},
      journal = {Neuron},
    publisher = {Elsevier},
         year = {2012},
       volume = {73},
        issue = {2},
        pages = {235-247},
     abstract = {The axon initial segment (AIS) is a specialized membrane region in the axon of neurons where action potentials are initiated. Crucial to the function of the AIS is the presence of specific voltage-gated channels clustered at high densities, giving the AIS unique electrical properties. Here we review recent data on the physiology of the AIS. These data indicate that the role of the AIS is far richer than originally thought, leading to the idea that it represents a dynamic signal processing unit within neurons, regulating the integration of synaptic inputs, intrinsic excitability, and transmitter release. Furthermore, these observations point to a critical role of the AIS in disease.},
} 
@article{KoleskeNATURE-REVIEWS-NEUROSCIENCE-13,
       author = {Koleske, Anthony J.},
        title = {Molecular mechanisms of dendrite stability},
      journal = {Nature Reviews Neuroscience},
         year = {2013},
    publisher = {Nature Publishing Group},
       volume = {14},
        issue = {8},
        pages = {536-550},
     abstract = {In the developing brain, dendrite branches and dendritic spines form and turn over dynamically. By contrast, most dendrite arbors and dendritic spines in the adult brain are stable for months, years and possibly even decades. Emerging evidence reveals that dendritic spine and dendrite arbor stability have crucial roles in the correct functioning of the adult brain and that loss of stability is associated with psychiatric disorders and neurodegenerative diseases. Recent findings have provided insights into the molecular mechanisms that underlie long-term dendrite stabilization, how these mechanisms differ from those used to mediate structural plasticity and how they are disrupted in disease.},
}
@inproceedings{KollerandFratinaICML-98,
       author = {D. Koller and R. Fratkina},
        title = {Using learning for approximation in stochastic processes},
    booktitle = {Proceedings Fifteenth International Conference on Machine Learning},
     location = {Madison, Wisconsin},
         year = 1998,
}
@book{KollerandFriedman09,
       author = {Daphne Koller and Nir Friedman},
        title = {Probabilistic Graphical Models: {P}rinciples and Techniques},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2009,
}
@inproceedings{KolleretlaUAI-99,
       author = {Daphne Koller and Uri Lerner and Dragomir Anguelov},
        title = {A general algorithm for approximate inference and its application to hybrid bayes nets},
    booktitle = {Proceedings of the the Fifteenth Annual Conference on Uncertainty in Artificial Intelligence},
        pages = {324-333},
         year = 1999,
}
@article{KolsteretalJoN-09,
       author = {Kolster, H. and Mandeville, J. B. and Arsenault, J. T. and Ekstrom, L. B. and Wald, L. L. and Vanduffel, W.},
        title = {Visual field map clusters in macaque extrastriate visual cortex},
      journal = {Journal Neuroscience},
         year = {2009},
       volume = {29},
       number = {21},
        pages = {7031-7039},
     abstract = {The macaque visual cortex contains >30 different functional visual areas, yet surprisingly little is known about the underlying organizational principles that structure its components into a complete "visual" unit. A recent model of visual cortical organization in humans suggests that visual field maps are organized as clusters. Clusters minimize axonal connections between individual field maps that represent common visual percepts, with different clusters thought to carry out different functions. Experimental support for this hypothesis, however, is lacking in macaques, leaving open the question of whether it is unique to humans or a more general model for primate vision. Here we show, using high-resolution blood oxygen level-dependent functional magnetic resonance imaging data in the awake monkey at 7 T, that the middle temporal area (area MT/V5) and its neighbors are organized as a cluster with a common foveal representation and a circular eccentricity map. This novel view on the functional topography of area MT/V5 and satellites indicates that field map clusters are evolutionarily preserved and may be a fundamental organizational principle of the Old World primate visual cortex.}
}
@inproceedings{KombrinketalICASSP-12,
       author = {Stefan Kombrink and Tom\`{a}s Mikolov and Martin Karafi{\'a}t and Luk{\'a}s Burget},
        title = {Improving language models for ASR using translated in-domain data},
    booktitle = {2012 IEEE International Conference on Acoustics, Speech and Signal Processing,},
    publisher = {IEEE},
         year = 2012,
        pages = {4405-4408},
}
@article{KonagurthuandLeskBMC-08,
       author = {Konagurthu, Arun S. and Lesk, Arthur M.},
        title = {On the origin of distribution patterns of motifs in biological networks},
      journal = {BMC Systems Biology},
         year = {2008},
       volume = {2},
        issue = {1},
        pages = {1-8},
     abstract = {Inventories of small subgraphs in biological networks have identified commonly-recurring patterns, called motifs. The inference that these motifs have been selected for function rests on the idea that their occurrences are significantly more frequent than random.},
}
@article{KonenandKastnerNN-08,
       author = {Konen, Christina S. and Kastner, Sabine},
        title = {Two hierarchically organized neural systems for object information in human visual cortex},
      journal = {Nature Neuroscience},
       volume = 11,
        issue = 2,
        pages = {224-231},
         year = 2008,
      comment = {Contrary to the two-pathways hypothesis, our findings indicate that basic object information related to shape, size and viewpoint may be represented similarly in two parallel and hierarchically organized neural systems in the ventral and dorsal visual pathways.},
}
@inproceedings{KonishietalCVPR-99,
       author = {S. M. Konishi and A.L. Yuille and J. M. Coughlan and S. C. Zhu},
        title = {Fundamental Bounds on Edge Detection: An Information Theoretic Evaluation of Different Edge Cues},
    booktitle = CVPR,
        pages = {573-579},
         year = 1999,
}
@article{KonkelandCohenFiN-09,
       author = {Konkel, Alex and Cohen, Neal J.},
        title = {Relational memory and the hippocampus: representations and methods},
      journal = {Frontiers in Neuroscience},
    publisher = {Frontiers Research Foundation},
       volume = {3},
        issue = {2},
         year = {2009},
        pages = {166-174},
     abstract = {Since the discovery of the importance of the hippocampus for normal memory, considerable research has endeavored to characterize the precise role played by the hippocampus. Previously we have offered the relational memory theory, which posits that the hippocampus forms representations of arbitrary or accidentally occurring relations among the constituent elements of experience. In a recent report we emphasized the role of the hippocampus in all manner of relations, supporting this claim with the finding that amnesic patients with hippocampal damage were similarly impaired on probes of memory for spatial, sequential, and associative relations. In this review we place these results in the context of the broader literature, including how different kinds of relational or source information are tested, and consider the importance of specifying hippocampal function in terms of the representations it supports.},
}
@article{KonkleandCaramazzaJoN-13,
       author = {Konkle, Talia and Caramazza, Alfonso},
        title = {Tripartite Organization of the Ventral Stream by Animacy and Object Size},
      journal = {Journal of Neuroscience},
       volume = {33},
       number = {25},
        pages = {10235-10242},
         year = {2013},
    publisher = {Society for Neuroscience},
     abstract = {Occipito-temporal cortex is known to house visual object representations, but the organization of the neural activation patterns along this cortex is still being discovered. Here we found a systematic, large-scale structure in the neural responses related to the interaction between two major cognitive dimensions of object representation: animacy and real-world size. Neural responses were measured with functional magnetic resonance imaging while human observers viewed images of big and small animals and big and small objects. We found that real-world size drives differential responses only in the object domain, not the animate domain, yielding a tripartite distinction in the space of object representation. Specifically, cortical zones with distinct response preferences for big objects, all animals, and small objects, are arranged in a spoked organization around the occipital pole, along a single ventromedial, to lateral, to dorsomedial axis. The preference zones are duplicated on the ventral and lateral surface of the brain. Such a duplication indicates that a yet unknown higher-order division of labor separates object processing into two substreams of the ventral visual pathway. Broadly, we suggest that these large-scale neural divisions reflect the major joints in the representational structure of objects and thus place informative constraints on the nature of the underlying cognitive architecture.},
}
@article{KonkleandOlivaCS-07,
        title = {Normative representation of objects: {E}vidence for an ecological bias in object perception and memory},
       author = {Konkle, Talia and Oliva,Aude},
    booktitle = {29th Annual Conference of the Cognitive Science Society},
         year = {2007},
        pages = {407-412},
     abstract = {An apple can be held in your hand if you are eating it, or be perceived in a fruit bowl from a few meters away. While the apple's real-world size is constant, the visual angle it subtends in your visual field varies based on your distance to the apple. Given the range of visual angles that an object can subtend in the visual field, is there a visual angle that is preferred? The current experiments show evidence supporting the idea of a privileged visual size for object representation, termed the normative size. In Experiment 1, observers adjusted the visual size of objects on a monitor, selecting the subjectively 'right' size to see the object. Inter-observer selections were strikingly consistent, and showed a correlation between the size selected on the screen and the actual physical size of the objects in the real world. These size selections were taken as estimates of the normative size for each object. In Experiments 2 and 3, using a size memory task and a change detection task, we found evidence that both long-term and short-term memory for the visual size of objects is biased towards the normative size. Altogether the results support the claims that perception of objects is sensitive to a normative size and that object memory is biased toward this perceptual norm.}
}
@book{Konolige86,
       author = {Konolige, Kurt},
        title = {A Deduction Model of Belief},
    publisher = {Pitman Publishing},
      address = {London},
         year = 1986,
}
@techreport{KonoligeTR-85,
       author = {Konolige, Kurt},
        title = {Research on Distributed Artificial Intelligence},
        issue = {Final Report for Project 1350},
  institution = {SRI International},
         year = 1985,
}        
@techreport{KoomenTR-88,
       author = {Koomen, Johannes A.},
        title = {The TIMELOGIC Temporal Reasoning System},
  institution = {University of Rochester Department of Computer Science},
        issue = 231,
         year = 1988,
}
@article{KoomeyetalIEEE-11,
       author = {Koomey, J.G. and Berard, S. and Sanchez, M. and Wong, H.},
      journal = {Annals of the History of Computing, IEEE},
        title = {Implications of Historical Trends in the Electrical Efficiency of Computing},
       volume = 33,
        issue = 3,
         year = 2011,
        pages = {46-54},
     abstract = {The electrical efficiency of computation has doubled roughly every year and a half for more than six decades, a pace of change comparable to that for computer performance and electrical efficiency in the microprocessor era. These efficiency improvements enabled the creation of laptops, smart phones, wireless sensors, and other mobile computing devices, with many more such innovations yet to come. The Web Extra appendix outlines the data and methods used in this study.},
}
@article{KoonandChanFiCN-17,
       author = {Koon, Alex. C and Chan, Ho Yin Edwin},
        title = {Drosophila melanogaster As a Model Organism to Study {RNA} Toxicity of Repeat Expansion-Associated Neurodegenerative and Neuromuscular Diseases},
      journal = {Frontiers in Cellular Neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = {11},
         year = {2017},
        pages = {70},
     abstract = {For nearly a century, the fruit fly, Drosophila melanogaster, has proven to be a valuable tool in our understanding of fundamental biological processes, and has empowered our discoveries, particularly in the field of neuroscience. In recent years, Drosophila has emerged as a model organism for human neurodegenerative and neuromuscular disorders. In this review, we highlight a number of recent studies that utilized the Drosophila model to study repeat-expansion associated diseases (READs), such as polyglutamine diseases, fragile X-associated tremor/ataxia syndrome (FXTAS), myotonic dystrophy type 1 (DM1) and type 2 (DM2), and C9ORF72-associated amyotrophic lateral sclerosis/frontotemporal dementia (C9-ALS/FTD). Discoveries regarding the possible mechanisms of RNA toxicity will be focused here. These studies demonstrate Drosophila as an excellent in vivo model system that can reveal novel mechanistic insights into human disorders, providing the foundation for translational research and therapeutic development.}
}        
@inproceedings{KoppenetalICANN-01,
       author = {M. Koppen and A. Soria-Frisch and R. Vicente-García},
        title = {NeuroHough: A Neural Network for Computing the Hough Transform},
    booktitle = {Proceedings International Conference on Artificial Neural Nets and Genetic Algorithms},
       editor = {Kurkova, V.},
     location = {Prague, Czech Republic},
    publisher = {Springer Computer Science},
         year = 2001,
        pages = {197-200},
} 
@article{KordingPLoS-11,
        title = {Of toasters and molecular ticker tapes},
       author = {Kording, Konrad},
      journal = {{PLoS} Computational Biology},
       volume = 7,
        issue = 12,
         year = 2011,
     abstract = {Experiments in systems neuroscience can be seen as consisting of three steps: (1) selecting the signals we are interested in, (2) probing the system with carefully chosen stimuli, and (3) getting data out of the brain. Here I discuss how emerging techniques in molecular biology are starting to improve these three steps. To estimate its future impact on experimental neuroscience, I will stress the analogy of ongoing progress with that of microprocessor production techniques. These techniques have allowed computers to simplify countless problems; because they are easier to use than mechanical timers, they are even built into toasters. Molecular biology may advance even faster than computer speeds and has made immense progress in understanding and designing molecules. These advancements may in turn produce impressive improvements to each of the three steps, ultimately shifting the bottleneck from obtaining data to interpreting it.}
}
@article{KorenetalNB-12,
        title = {Hybrid error correction and de novo assembly of single-molecule sequencing reads},
       author = {Koren, Sergey and Schatz, Michael C. and Walenz, Brian P. and Martin, Jeffrey  and Howard, Jason T. and  Ganapathy, Ganeshkumar and Wang, Zhong and Rasko, David A. and McCombie, W. Richard and Jarvis, Erich D. and Phillippy, Adam M.},
      journal = {Nature Biotechnology},
       volume = 30,
        issue = 7,
         year = 2012,
        pages = {693-700}
}
@inproceedings{Korf,
       author = {Korf, Richard E.},
        title = {Real-Time Path Planning},
    booktitle = {Proceedings DARPA Knowledge-Based Planning Workshop},
     location = {Austin, Texas},
  institution = {DARPA},
         year = 1987,
}
@inproceedings{KorfAAAI-88,
       author = {Korf, Richard},
        title = {Real-Time Heuristic Search: New Results},
    booktitle = {Proceedings AAAI-88},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
        pages = {139-144},
         year = 1988,
}
@article{KorfAIJ-85a,
       author = {Korf, Richard},
        title = {Depth-first iterative deepening: an optimal admissible tree search},
      journal = AIJ,
       volume = 29,
         year = 1985,
        pages = {97-109},
}
@article{KorfAIJ-85b,
       author = {Korf, Richard},
        title = {Macro-Operators: A Weak Method for Learning},
      journal = AIJ,
       volume = 26,
         year = 1985,
        pages = {35-77},
}
@article{KorfAIJ-87,
       author = {Korf, Richard},
        title = {Planning as Search: A Quantitative Approach},
      journal = AIJ,
       volume = 33,
        issue = 1,
         year = 1987,
        pages = {65-88},
}
@article{KoricheSBS-15,
        title = {The Emergence of a New Discourse for Business Communication, 'A Case Study of e-mails in Shipping Company'},
       author = {Hassiba Koriche},
      journal = {Procedia - Social and Behavioral Sciences},
       volume = {199},
        pages = {539-547},
         year = {2015},
     abstract = {E-mail, as an example of electronic medium, has assumed functions which are in some cases associated with spoken language and in other cases associated with formal writing. David Crystal (2001) considers that online language as a new species of interaction, a genuine "third medium". Business professionals in Algeria have to improve their employees' English competence in order to be efficient in international trade. In order to prove that e-mail is not only a medium, but a new style of communication as well; 175 business e-mails exchanged between the employees of the shipping company concerned were sampled to identify the rhetorical structure. The analysis based on the corpus reveals features that make up the pattern of business e-mail; it also shows that the electronic channel has an impact on the choice of the language exponents. A wide range of abbreviations represented in letter and number homophones is frequent. Moreover, syntactic reductions are used. The e-mail as a new emerging medium is a new type of discourse. It is a discourse that is developing its own language. In order to bridge the gap between theory and practice, tasks in business communication course should involve students in real business situation to familiarize them with the social features and interactional aspect of the language.}
}
@article{KorlachetalPNAS-08,
        title = {Selective aluminum passivation for targeted immobilization of single DNA polymerase molecules in zero-mode waveguide nanostructures},
       author = {Korlach, J. and Marks, P.J. and Cicero, R.L. and Gray, J.J. and Murphy, D.L. and Roitman, D.B. and Pham, T.T. and Otto, G.A. and Foquet, M. and Turner, S.W.},
      journal = {Proceedings of the National Academy of Science},
       volume = 105,
        issue = 4,
        pages = {1176-1181},
         year = 2008,
     abstract = {Optical nanostructures have enabled the creation of subdiffraction detection volumes for single-molecule fluorescence microscopy. Their applicability is extended by the ability to place molecules in the confined observation volume without interfering with their biological function. Here, we demonstrate that processive DNA synthesis thousands of bases in length was carried out by individual DNA polymerase molecules immobilized in the observation volumes of zero-mode waveguides (ZMWs) in high-density arrays. Selective immobilization of polymerase to the fused silica floor of the ZMW was achieved by passivation of the metal cladding surface using polyphosphonate chemistry, producing enzyme density contrasts of glass over aluminum in excess of 400:1. Yields of single-molecule occupancies of approximately 30\% were obtained for a range of ZMW diameters (70-100 nm). Results presented here support the application of immobilized single DNA polymerases in ZMW arrays for long-read-length DNA sequencing.}
}
@article{KornfeldetalBIORXIV-20,
       author = {Kornfeld, Joergen M and Januszewski, Micha{\l} and Schubert, Philipp Johannes and Jain, Viren and Denk, Winfried and Fee, Michale S},
        title = {An anatomical substrate of credit assignment in reinforcement learning},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Laboratory},
         year = {2020},
     abstract = {Learning turns experience into better decisions. A key problem in learning is credit assignment-knowing how to change parameters, such as synaptic weights deep within a neural network, in order to improve behavioral performance. Artificial intelligence owes its recent bloom largely to the error-backpropagation algorithm1, which estimates the contribution of every synapse to output errors and allows rapid weight adjustment. Biological systems, however, lack an obvious mechanism to backpropagate errors. Here we show, by combining high-throughput volume electron microscopy 2 and automated connectomic analysis, that the synaptic architecture of songbird basal ganglia supports local credit assignment using a variant of the node perturbation algorithm proposed in a model of songbird reinforcement learning. We find that key predictions of the model hold true: first, cortical axons that encode exploratory motor variability terminate predominantly on dendritic shafts of striatal spiny neurons, while cortical axons that encode song timing terminate almost exclusively on spines. Second, synapse pairs that share a presynaptic cortical timing axon and a postsynaptic spiny dendrite are substantially more similar in size than expected, indicating Hebbian plasticity. Combined with numerical simulations, these findings provide strong evidence for a biologically plausible credit assignment mechanism6.},
}
@article{KorobovaandSvitkinaMBC-10,
       author = {Korobova, Farida and Svitkina, Tatyana},
        title = {Molecular Architecture of Synaptic Actin Cytoskeleton in Hippocampal Neurons Reveals a Mechanism of Dendritic Spine Morphogenesis},
      journal = {Molecular Biology of the Cell},
    publisher = {The American Society for Cell Biology},
       volume = 21,
        issue = 1,
         year = 2010,
        pages = {165-176},
     abstract = {The high resolution structure of the cytoskeleton in dendritic spines and their precursors, dendritic filopodia, was characterized by platinum replica electron microscopy. The unexpected features of the actin architecture of these protrusions dramatically change the existing paradigm of cytoskeletal organization of elongated membrane protrusions.},
}
@article{KosinskietalPNAS-13,
        title = {Private traits and attributes are predictable from digital records of human behavior},
       author = {Kosinski, Michal and Stillwell, David and Graepel, Thore},
      journal = {Proceedings of the National Academy of Sciences},
       volume = 110,
        issue = 15,
         year = 2013,
        pages = {5802-5805},
     abstract = {We show that easily accessible digital records of behavior, Facebook Likes, can be used to automatically and accurately predict a range of highly sensitive personal attributes including: sexual orientation, ethnicity, religious and political views, personality traits, intelligence, happiness, use of addictive substances, parental separation, age, and gender. The analysis presented is based on a dataset of over 58,000 volunteers who provided their Facebook Likes, detailed demographic profiles, and the results of several psychometric tests. The proposed model uses dimensionality reduction for preprocessing the Likes data, which are then entered into logistic/linear regression to predict individual psychodemographic profiles from Likes. The model correctly discriminates between homosexual and heterosexual men in 88\% of cases, African Americans and Caucasian Americans in 95\% of cases, and between Democrat and Republican in 85\% of cases. For the personality trait "Openness," prediction accuracy is close to the test-retest accuracy of a standard personality test. We give examples of associations between attributes and Likes and discuss implications for online personalization and privacy.},
}
@inproceedings{KosinskietalWEBSCI-12,
        title = {Personality and Website Choice},
       author = {Michal Kosinski and David Stillwell and Pushmeet Kohli and Yoram Bachrach and Thore Graepel},
    booktitle = {{ACM} Conference on Web Sciences},
         year = 2012,
    publisher = {ACM},
     abstract = {We find that preference for websites, like preference for objects in the offline world, is influenced by personality. We combine personality profiles and website choices of more than 160.000 users and investigate whether different websites attract audience of different personality. Using two independent sources of website choices, we show that website audiences often have distinct personality profiles, that there is a psychologically meaningful relationship between personality and preferences related to website and website categories, and that results are stable across independent data sources. Our findings are useful for researchers interested in website content personalization, text search, search result optimization and online marketing.},
} 
@incollection{KostovicandJudasTCN-09,
       author = {Kostovi\'{c}, Ivica and Juda\v{s}, Milo\v{s}},
        title = {Early development of neuronal circuitry of the human prefrontal cortex},
    booktitle = {The Cognitive Neurosciences, 4th Edition},
    publisher = {The MIT Press},
      address = {Cambridge, MA},
       editor = {Gazzaniga, Michael S.},
         year = {2009},
        pages = {29-48},
}
@article{KotovetalADV-MAT-09,
       author = {Nicholas A. Kotov and Jessica O. Winter and Isaac P. Clements and Edward Jan and Brian P. Timko and  St\'{e}phane Campidelli and Smita Pathak and Andrea Mazzatenta and Charles M. Lieber and Maurizio Prato and  Ravi V. Bellamkonda and Gabriel A. Silva and Nadine Wong Shi Kam and Fernando Patolsky and Laura Ballerini},
        title = {Nanomaterials for Neural Interfaces},
      journal = {Advanced Materials},
       volume = 21,
         year = 2009,
        pages = {3970-4004},
}
@article{KouideretalSCIENCE-13,
       author = {Kouider, S. and Stahlhut, C. and Gelskov, S. V. and Barbosa, L. S. and Dutat, M. and de Gardelle, V. and Christophe, A. and Dehaene, S. and Dehaene-Lambertz, G. },
        title = {A neural marker of perceptual consciousness in infants},
      journal = {Science},
         year = {2013},
       volume = {340},
       number = {6130},
        pages = {376-380},
     abstract = {Infants have a sophisticated behavioral and cognitive repertoire suggestive of a capacity for conscious reflection. Yet, demonstrating conscious access in infants remains challenging, mainly because they cannot report their thoughts. Here, to circumvent this problem, we studied whether an electrophysiological signature of consciousness found in adults, corresponding to a late nonlinear cortical response [~300 milliseconds (ms)] to brief pictures, already exists in infants. We recorded event-related potentials while 5-, 12-, and 15-month-old infants (N = 80) viewed masked faces at various levels of visibility. In all age groups, we found a late slow wave showing a nonlinear profile at the expected perceptual thresholds. However, this late component shifted from a weak and delayed response in 5-month-olds (starting around 900 ms) to a more sustained and faster response in older infants (around 750 ms). These results reveal that the brain mechanisms underlying the threshold for conscious perception are already present in infancy but undergo a slow acceleration during development.},
}
@inproceedings{KoutniketalICML-14,
       author = {Jan Koutn\'{i}k and Klaus Greff and Faustino Gomez and J\"{u}rgen Schmidhuber},
        title = {Clockwork {RNN}},
    booktitle = {Proceedings of the 31th International Conference on Machine Learning},
       volume = 32,
        pages = {1863-1871},
         year = 2014,
     abstract = {Sequence prediction and classification are ubiquitous and challenging problems in machine learning that can require identifying complex dependencies between temporally distant inputs. Recurrent Neural Networks (RNNs) have the ability, in theory, to cope with these temporal dependencies by virtue of the short-term memory implemented by their recurrent (feedback) connections. However, in practice they are difficult to train successfully when long-term memory is required. This paper introduces a simple, yet powerful modification to the simple RNN (SRN) architecture, the Clockwork RNN (CW-RNN), in which the hidden layer is partitioned into separate modules, each processing inputs at its own temporal granularity, making computations only at its prescribed clock rate. Rather than making the standard RNN models more complex, CW-RNN reduces the number of SRN parameters, improves the performance significantly in the tasks tested, and speeds up the network evaluation. The network is demonstrated in preliminary experiments involving three tasks: audio signal generation, TIMIT spoken word classification, where it outperforms both SRN and LSTM networks, and online handwriting recognition, where it outperforms SRNs.}
}
@article{KoutsoupiasandPapadimitriouIPL-92,
       author = {Koutsoupias, Elias and Papadimitriou, Christos H.},
        title = {On the greedy algorithm for satisfiability},
      journal = {Information Processing Letters},
    publisher = {North-Holland},
       volume = 43,
         year = 1992,
        pages = {53-55},
}
@article{KowadloetalCoRR-19,
        title = {{AHA! an 'Artificial Hippocampal Algorithm' for Episodic Machine Learning}},
       author = {Gideon Kowadlo and Abdelrahman Ahmed and David Rawlinson},
      journal = {CoRR},
         year = {2019},
       volume = {arXiv:1909.10340},
     abstract = {The majority of ML research concerns slow, statistical learning of i.i.d. samples from large, labelled datasets. Animals do not learn this way. An enviable characteristic of animal learning is 'episodic' learning - the ability to rapidly memorize a specific experience as a composition of existing concepts, without provided labels. The new knowledge can then be used to distinguish between similar experiences, to generalize between classes, and to selectively consolidate to long-term memory. The Hippocampus is known to be vital to these abilities. AHA is a biologically-plausible computational model of the Hippocampus. Unlike most machine learning models, AHA is trained without any external labels and uses only local and immediate credit assignment. We demonstrate AHA in a superset of the Omniglot classification benchmark. The extended benchmark covers a wider range of known Hippocampal functions by testing pattern separation, completion, and reconstruction of original input. These functions are all performed within a single configuration of the computational model. Despite these constraints, results are comparable to state-of-the-art deep convolutional ANNs. In addition to the demonstrated high degree of functional overlap with the Hippocampal region, AHA is remarkably aligned to current macro-scale biological models and uses biologically plausible micro-scale learning rules.},
}
@book{Kowalski79,
       author = {Kowalski, Robert},
        title = {A Logic for Problem Solving},
    publisher = {North-Holland},
      address = {New York},
         year = 1979,
}
@misc{Kowalski86,
       author = {Kowalski, Robert},
        title = {Database Updates in the Event Calculus},
         year = 1986,
}
@article{KowalskiandSergot86,
       author = {Kowalski, Robert and Sergot, M. J.},
        title = {A Logic-based Calculus of Events},
      journal = {New Generation Computing},
       volume = 4,
    publisher = {Ohmsha Limited and Springer-Verlag},
         year = 1986,
        pages = {67-95},
}
@article{KozioletalCEREBELLUM-14,
       author = {Koziol, Leonard F. and Budding, Deborah and Andreasen, Nancy and D'Arrigo, Stefano and Bulgheroni, Sara and Imamizu, Hiroshi and Ito, Masao and Manto, Mario and Marvel, Cherie and Parker, Krystal and Pezzulo, Giovanni and Ramnani, Narender and Riva, Daria and Schmahmann, Jeremy and Vandervert, Larry and Yamazaki, Tadashi},
        title = {Consensus paper: the cerebellum's role in movement and cognition},
      journal = {Cerebellum (London, England)},
       volume = {13},
        issue = {1},
         year = {2014},
        pages = {151-177},
     abstract = {While the cerebellum's role in motor function is well recognized, the nature of its concurrent role in cognitive function remains considerably less clear. The current consensus paper gathers diverse views on a variety of important roles played by the cerebellum across a range of cognitive and emotional functions. This paper considers the cerebellum in relation to neurocognitive development, language function, working memory, executive function, and the development of cerebellar internal control models and reflects upon some of the ways in which better understanding the cerebellum's status as a ``supervised learning machine'' can enrich our ability to understand human function and adaptation. As all contributors agree that the cerebellum plays a role in cognition, there is also an agreement that this conclusion remains highly inferential. Many conclusions about the role of the cerebellum in cognition originate from applying known information about cerebellar contributions to the coordination and quality of movement. These inferences are based on the uniformity of the cerebellum's compositional infrastructure and its apparent modular organization. There is considerable support for this view, based upon observations of patients with pathology within the cerebellum.},
}
@book{Kozmaetal2012,
        title = {Advances in Neuromorphic Memristor Science and Applications},
       author = {Kozma, R. and Pino, R.E. and Pazienza, G.E.},
       series = {Springer Series in Cognitive and Neural Systems},
         year = {2012},
    publisher = {Springer Netherlands},
     abstract = {Physical implementation of the memristor at industrial scale sparked the interest from various disciplines, ranging from physics, nanotechnology, electrical engineering, neuroscience, to intelligent robotics. As any promising new technology, it has raised hopes and questions; it is an extremely challenging task to live up to the high expectations and to devise revolutionary and feasible future applications for memristive devices. The possibility of gathering prominent scientists in the heart of the Silicon Valley given by the 2011 International Joint Conference on Neural Networks held in San Jose, CA, has offered us the unique opportunity of organizing a series of special events on the present status and future perspectives in neuromorphic memristor science. This book presents a selection of the remarkable contributions given by the leaders of the field and it may serve as inspiration and future reference to all researchers that want to explore the extraordinary possibilities given by this revolutionary concept.},
}
@article{KraljetalNATURE-12,
        title = {Optical recording of action potentials in mammalian neurons using a microbial rhodopsin},
       author = {Kralj, J.M. and Douglass, A.D. and Hochbaum, D.R. and Maclaurin, D. and Cohen, A.E.},
      journal = {Nature Methods},
       volume = 9,
        issue = 1,
        pages = {90-5},
         year = 2012,
     abstract = {Reliable optical detection of single action potentials in mammalian neurons has been one of the longest-standing challenges in neuroscience. Here we achieved this goal by using the endogenous fluorescence of a microbial rhodopsin protein, Archaerhodopsin 3 (Arch) from Halorubrum sodomense, expressed in cultured rat hippocampal neurons. This genetically encoded voltage indicator exhibited an approximately ten-fold improvement in sensitivity and speed over existing protein-based voltage indicators, with a roughly linear twofold increase in brightness between -150 mV and +150 mV and a sub-millisecond response time. Arch detected single electrically triggered action potentials with an optical signal-to-noise ratio >10. Arch(D95N) lacked endogenous proton pumping and had 50\% greater sensitivity than wild type but had a slower response (41 ms). Nonetheless, Arch(D95N) also resolved individual action potentials. Microbial rhodopsin-based voltage indicators promise to enable optical interrogation of complex neural circuits and electrophysiology in systems for which electrode-based techniques are challenging.}
}
@article{KraljetalNATURE-METHODS-11,
       author = {Kralj, J. M.  and Douglass, A. D.  and Hochbaum, D. R.  and Maclaurin, D.  and Cohen, A. E.},
        title = {Optical recording of action potentials in mammalian neurons using a microbial rhodopsin},
      journal = {Nat. Methods},
         year = {2011},
       volume = {9},
       number = {1},
        pages = {90-95},
     abstract = {Reliable optical detection of single action potentials in mammalian neurons has been one of the longest-standing challenges in neuroscience. Here we achieved this goal by using the endogenous fluorescence of a microbial rhodopsin protein, Archaerhodopsin 3 (Arch) from Halorubrum sodomense, expressed in cultured rat hippocampal neurons. This genetically encoded voltage indicator exhibited an approximately tenfold improvement in sensitivity and speed over existing protein-based voltage indicators, with a roughly linear twofold increase in brightness between -150 mV and +150 mV and a sub-millisecond response time. Arch detected single electrically triggered action potentials with an optical signal-to-noise ratio >10. Arch(D95N) lacked endogenous proton pumping and had 50\% greater sensitivity than wild type but had a slower response (41 ms). Nonetheless, Arch(D95N) also resolved individual action potentials. Microbial rhodopsin-based voltage indicators promise to enable optical interrogation of complex neural circuits and electrophysiology in systems for which electrode-based techniques are challenging.}
}
@article{KravitzetalTiCS-13,
       author = {Kravitz, D. J. and Saleem, K. S. and Baker, C. I. and Ungerleider, L. G. and Mishkin, M.},
        title = {{{T}he ventral visual pathway: an expanded neural framework for the processing of object quality}},
      journal = {Trends in Cognitive Science)},
         year = {2013},
       volume = {17},
       number = {1},
        pages = {26-49},
     abstract = {Since the original characterization of the ventral visual pathway, our knowledge of its neuroanatomy, functional properties, and extrinsic targets has grown considerably. Here we synthesize this recent evidence and propose that the ventral pathway is best understood as a recurrent occipitotemporal network containing neural representations of object quality both utilized and constrained by at least six distinct cortical and subcortical systems. Each system serves its own specialized behavioral, cognitive, or affective function, collectively providing the raison d'etre for the ventral visual pathway. This expanded framework contrasts with the depiction of the ventral visual pathway as a largely serial staged hierarchy culminating in singular object representations and more parsimoniously incorporates attentional, contextual, and feedback effects.}
}
@inproceedings{KrebsbachetalAIPS-92,
       author = {Krebsbach, K. and Olawsky, D. and Gini, M.},
        title = {An Empirical Studing of Sensing and Defaulting in Planning},
    booktitle = {Proceedings of the First International Conference on Artificial Intelligence Planning Systems},
        pages = {136-144},
         year = 1992,
}
@techreport{KremppetalTR-02,
       author = {S. Krempp and D. Geman and Y. Amit},
        title = {Sequential learning of reusable parts for object detection},
  institution = {Johns Hopkins University},
        issue = {Technical Report CS 2002},
         year = 2002,
}
@article{KreshuketalPLoS-14,
       author = {Kreshuk, A. and Koethe, U. and Pax, E. and Bock, D. D. and Hamprecht, F. A.},
        title = {{A}utomated detection of synapses in serial section transmission electron microscopy image stacks},
      journal = {{PLoS} ONE},
         year = 2014,
       volume = 9,
        issue = 2,
        pages = {e87351},
     abstract = {We describe a method for fully automated detection of chemical synapses in serial electron microscopy images with highly anisotropic axial and lateral resolution, such as images taken on transmission electron microscopes. Our pipeline starts from classification of the pixels based on {3D} pixel features, which is followed by segmentation with an Ising model MRF and another classification step, based on object-level features. Classifiers are learned on sparse user labels; a fully annotated data subvolume is not required for training. The algorithm was validated on a set of 238 synapses in 20 serial 719767351 pixel images (4.564.5645 nm resolution) of mouse visual cortex, manually labeled by three independent human annotators and additionally re-verified by an expert neuroscientist. The error rate of the algorithm (12\% false negative, 7\% false positive detections) is better than state-of-the-art, even though, unlike the state-of-the-art method, our algorithm does not require a prior segmentation of the image volume into cells.}
}
@inproceedings{KrichmaretalICCAD-11,
       author = {Krichmar, Jeffrey L. and Dutt, Nikil and Nageswaran, Jayram M. and Richert, Micah},
        title = {Neuromorphic modeling abstractions and simulation of large-scale cortical networks},
    booktitle = {Proceedings of the International Conference on Computer-Aided Design},
    publisher = {IEEE Press},
      address = {Piscataway, NJ},
         year = 2011,
        pages = {334-338},
     abstract = {Biological neural systems are well known for their robust and power-efficient operation in highly noisy environments. We outline key modeling abstractions for the brain and focus on spiking neural network models. We discuss aspects of neuronal processing and computational issues related to modeling these processes. Although many of these algorithms can be efficiently realized in specialized hardware, we present a case study of simulation of the visual cortex using a GPU based simulation environment that is readily usable by neuroscientists and computer scientists and efficient enough to construct very large networks comparable to brain networks.},
}
@article{KrieteetalPNAS-13,
       author = {Kriete, Trenton and Noelle, David C. and Cohen, Jonathan D. and O'Reilly, Randall C.},
        title = {Indirection and symbol-like processing in the prefrontal cortex and basal ganglia},
      journal = {Proceedings of the National Academy of Sciences},
         year = 2013,
     abstract = {The ability to flexibly, rapidly, and accurately perform novel tasks is a hallmark of human behavior. In our everyday lives we are often faced with arbitrary instructions that we must understand and follow, and we are able to do so with remarkable ease. It has frequently been argued that this ability relies on symbol processing, which depends critically on the ability to represent variables and bind them to arbitrary values. Whereas symbol processing is a fundamental feature of all computer systems, it remains a mystery whether and how this ability is carried out by the brain. Here, we provide an example of how the structure and functioning of the prefrontal cortex/basal ganglia working memory system can support variable binding, through a form of indirection (akin to a pointer in computer science). We show how indirection enables the system to flexibly generalize its behavior substantially beyond its direct experience (i.e., systematicity). We argue that this provides a biologically plausible mechanism that approximates a key component of symbol processing, exhibiting both the flexibility, but also some of the limitations, that are associated with this ability in humans.},
}
@article{KrimholtzetalELECTRONICS-LETTERS-70,
       author = {Krimholtz, R. and Leedom, D.A. and Matthaei, G.L.},
        title = {New equivalent circuits for elementary piezoelectric transducers},
      journal = {Electronics Letters},
       volume = 6,
        issue = 13,
         year = 1970,
        pages = {398-399},
}
@book{Kripke1982wittgenstein,
        title = {Wittgenstein on Rules and Private Language: {A}n Elementary Exposition},
       author = {Kripke, Saul A.},
    publisher = {Harvard University Press},
         year = {1982},
}
@book{Kripke71,
       author = {Kripke, Saul A.},
        title = {Semantical Considerations on Modal Logic},
    publisher = {Oxford University Press},
     location = {Oxford},
         year = 1971,
}
@inproceedings{KrizhevskyetalNIPS-12,
      title = {ImageNet Classification with Deep Convolutional Neural Networks},
     author = {Alex Krizhevsky and Ilya Sutskever and Geoff Hinton},
  booktitle = {Advances in Neural Information Processing Systems 25},
     editor = {P. Bartlett and F.C.N. Pereira and C.J.C. Burges and L. Bottou and K.Q. Weinberger},
       year = 2012,
      pages = {1106-1114},
}
@inproceedings{KruegerandDayanCOSYNE-07,
        title = {Flexible shaping: how learning in small steps helps},
       author = {Krueger, K. A. and Dayan, P.},
    booktitle = {Computational and Systems Neuroscience},
         year = 2007,
        pages = {II-41}
}
@article{KruegeretalCOIN-12,
        title = {The role of neurexins and neuroligins in the formation, maturation, and function of vertebrate synapses},
       author = {Dilja D Krueger and Liam P Tuffy and Theofilos Papadopoulos and Nils Brose},
      journal = {Current Opinion in Neurobiology},
       volume = {22},
        issue = {3},
         year = {2012},
        pages = {412-422},
     abstract = {Neurexins (NXs) and neuroligins (NLs) are transsynaptically interacting cell adhesion proteins that play a key role in the formation, maturation, activity-dependent validation, and maintenance of synapses. As complex alternative splicing processes in nerve cells generate a large number of \{NX\} and \{NLs\} variants, it has been proposed that a combinatorial interaction code generated by these variants may determine synapse identity and network connectivity during brain development. The functional importance of \{NXs\} and \{NLs\} is exemplified by the fact that mutations in \{NX\} and \{NL\} genes are associated with several neuropsychiatric disorders, most notably with autism. Accordingly, major research efforts have focused on the molecular mechanisms by which \{NXs\} and \{NLs\} operate at synapses. In this review, we summarize recent progress in this field and discuss emerging topics, such as the role of alternative interaction partners of \{NXs\} and \{NLs\} in synapse formation and function, and their relevance for synaptic plasticity in the mature brain. The novel findings highlight the fundamental importance of NX-NL interactions in a wide range of synaptic functions.},
}
@inproceedings{KruegeretalCS-17,
        title = {Enhancing metacognitive reinforcement learning using reward structures and feedback},
       author = {Paul Krueger and Falk Lieder and Tom Griffiths},
    booktitle = {Proceedings of the 2017 Annual Meeting of the Cognitive Science Society},
         year = {2017},
     abstract = {How do we learn to think better, and what can we do to promote such metacognitive learning? Here, we propose that cognitive growth proceeds through metacognitive reinforcement learning. We apply this theory to model how people learn how far to plan ahead and test its predictions about the speed of metacognitive learning in two experiments. In the first experiment, we find that our model can discern a reward structure that promotes metacognitive reinforcement learning from one that hinders it. In the second experiment, we show that our model can be used to design a feedback mechanism that enhances metacognitive reinforcement learning in an environment that hinders learning. Our results suggest that modeling metacognitive learning is a promising step towards promoting cognitive growth.},
}
@article{KruegeretalPMC-08,
       author = {Krueger, Frank and Spampinato, Maria Vittoria and Pardini, Matteo and Pajevic, Sinisa and Wood, Jacqueline N. and Weiss, George H. and Landgraf, Steffen and Grafman, Jordan},
        title = {Integral calculus problem solving: an fMRI investigation},
      journal = {Neuroreport},
         year = {2008},
       volume = {19},
        issue = {11},
        pages = {1095-1099},
     abstract = {Only a subset of adults acquires specific advanced mathematical skills, such as integral calculus. The representation of more sophisticated mathematical concepts probably evolved from basic number systems; however its neuroanatomical basis is still unknown. Using fMRI, we investigated the neural basis of integral calculus while healthy participants were engaged in an integration verification task. Solving integrals activated a left-lateralized cortical network including the horizontal intraparietal sulcus, posterior superior parietal lobe, posterior cingulate gyrus, and dorsolateral prefrontal cortex. Our results indicate that solving of more abstract and sophisticated mathematical facts, such as calculus integrals, elicits a pattern of brain activation similar to the cortical network engaged in basic numeric comparison, quantity manipulation, and arithmetic problem solving.}
}
@article{KrupaetalPNAS-99,
       author = {Krupa, David J. and Ghazanfar, Asif A. and Nicolelis, Miguel A. L.},
        title = {Immediate thalamic sensory plasticity depends on corticothalamic feedback},
      journal = {Proceedings of the National Academy of Sciences},
       volume = 96,
        issue = 14,
        pages = {8200-8205},
         year = 1999,
      comment = {http://www.pnas.org/content/96/14/8200.figures-only},
     abstract = {Multiple neuron ensemble recordings were obtained simultaneously from both the primary somatosensory (SI) cortex and the ventroposterior medial thalamus (VPM) before and during the combined administration of reversible inactivation of the SI cortex and a reversible subcutaneous block of peripheral trigeminal nerve fibers. This procedure was performed to quantify the contribution of descending corticofugal projections on (i) the normal organization of thalamic somatosensory receptive fields and (ii) the thalamic somatosensory plastic reorganization that immediately follows a peripheral deafferentation. Reversible inactivation of SI cortex resulted in immediate changes in receptive field properties throughout the VPM. Cortical inactivation also significantly reduced but did not completely eliminate the occurrence of VPM receptive field reorganization resulting from the reversible peripheral deafferentation. This result suggests that the thalamic plasticity that is seen immediately after a peripheral deafferentation is dependent upon both descending corticofugal projections and ascending trigeminothalamic projections.},
}
@article{KrzysztofetalCoRR-16,
       author = {Krzysztof J. Geras and Abdel-rahman Mohamed and Rich Caruana and Gregor Urban and Shengjie Wang and \"{O}zlem Aslan and Matthai Philipose and Matthew Richardson and Charles A. Sutton},
        title = {Compressing {LSTMs} into {CNNs}},
      journal = {CoRR},
       volume = {arXiv:1511.06433},
         year = {2015},
}
@article{KubiliusetalPLoS-16,
       author = {Kubilius, Jonas and Bracci, Stefania and Op de Beeck, Hans P.},
      journal = {PLoS Computational Biology},
    publisher = {Public Library of Science},
        title = {Deep Neural Networks as a Computational Model for Human Shape Sensitivity},
       volume = {12},
       number = {4},
         year = {2016},
        pages = {1-26},
     abstract = {Shape plays an important role in object recognition. Despite years of research, no models of vision could account for shape understanding as found in human vision of natural images. Given recent successes of deep neural networks (DNNs) in object recognition, we hypothesized that DNNs might in fact learn to capture perceptually salient shape dimensions. Using a variety of stimulus sets, we demonstrate here that the output layers of several DNNs develop representations that relate closely to human perceptual shape judgments. Surprisingly, such sensitivity to shape develops in these models even though they were never explicitly trained for shape processing. Moreover, we show that these models also represent categorical object similarity that follows human semantic judgments, albeit to a lesser extent. Taken together, our results bring forward the exciting idea that DNNs capture not only objective dimensions of stimuli, such as their category, but also their subjective, or perceptual, aspects, such as shape and semantic similarity as judged by humans.},
}
@incollection{KubovyandGepshteinCMSC-03,
       author = {Kubovy, M. and Gepshtein, S.},
        title = {Perceptual grouping in space and in space-time {A}n exercise in phenomenological psychophysics},
    booktitle = {Perceptual Organization in Vision: Behavioral and Neural Perspectives},
       editor = {Behrmann, M. and Kimchi, R. and Olson, C. R.},
    publisher = {Lawrence Erlbaum},
      address = {Mahwah, NJ},
        pages = {45-85},
         year = 2003,
}
@article{KubyshkinBIOSYSTEMS-18,
       author = {Vladimir Kubyshkin and Carlos G. Acevedo-Rocha and Nediljko Budisa},
        title = {On universal coding events in protein biogenesis},
      journal = {Biosystems},
       volume = {164},
         year = {2018},
        pages = {16-25},
     abstract = {The complete ribosomal protein synthesis cycle and codon-amino acids associations are universally preserved in all life taxa on Earth. This process is accompanied by a set of hierarchically organized recognition and controlling events at different complexity levels. It starts with amino acid activation by aminoacyl tRNA synthetases (aaRS) followed by matching with the acceptor units of their cognate tRNAs ("operational RNA code") and ribosomal codon-anticodon pairing of messenger RNA ("triplet code"). However, this codon-anticodon matching is possible only when protein translation machinery (translation factors, ribosome) accepts an esterified amino acid. This capacity ("charge code") correlates mainly with the amino acid nature and the identity elements in the tRNA 3D structure. A fourth potential "folding code" (also referred as "stereochemical code") between the translation dynamics, sequence composition and folding of the resulting protein can also be defined in the frame of the 'Anfinsen dogma' followed by post-translational modifications. All these coding events as well as the basic chemistry of life are deemed invariant across biological taxa due to the horizontal gene transfer (HGT) making the 'universal genetic code' the 'lingua franca' of life of earth. When cells (or organelles) are prevented from transmitting genetic information (i.e., HGT) the deviations in the above-mentioned coding events become inevitable. A better understanding of these codes, in particular the mechanisms of their conservation in the context of HGT could provide a guide for the experimental engineering "What I cannot create − I do not understand" − Richard Feynman of the ribosomal protein biosynthesis machinery. This is highly relevant, among others, in attempts to create synthetic life forms in genetic isolation by using tailored "minimal genomes" and may explain the necessity for multiple coding events in nature.}
}
@article{KudrimotietalJoN-99,
       author = {Kudrimoti, H. S. and Barnes, C. A. and McNaughton, B. L.},
        title = {Reactivation of hippocampal cell assemblies: effects of behavioral state, experience, and {EEG} dynamics}}
@article{KuetalNATURE-BIOTECHNOLOGY-16,
       author = {Ku, Taeyun and Swaney, Justin and Park, Jeong-Yoon and Albanese, Alexandre and Murray, Evan and Cho, Jae Hun and Park, Young-Gyun and Mangena, Vamsi and Chen, Jiapei and Chung, Kwanghun},
        title = {Multiplexed and scalable super-resolution imaging of three-dimensional protein localization in size-adjustable tissues},
      journal = {Nature Biotechnology},
         year = {2016},
    publisher = {Nature Publishing Group},
     abstract = {The biology of multicellular organisms is coordinated across multiple size scales, from the subnanoscale of molecules to the macroscale, tissue-wide interconnectivity of cell populations. Here we introduce a method for super-resolution imaging of the multiscale organization of intact tissues. The method, called magnified analysis of the proteome (MAP), linearly expands entire organs fourfold while preserving their overall architecture and three-dimensional proteome organization. MAP is based on the observation that preventing crosslinking within and between endogenous proteins during hydrogel-tissue hybridization allows for natural expansion upon protein denaturation and dissociation. The expanded tissue preserves its protein content, its fine subcellular details, and its organ-scale intercellular connectivity. We use off-the-shelf antibodies for multiple rounds of immunolabeling and imaging of a tissue's magnified proteome, and our experiments demonstrate a success rate of 82\% (100/122 antibodies tested). We show that specimen size can be reversibly modulated to image both inter-regional connections and fine synaptic architectures in the mouse brain.},
}
@article{KuhletalPNAS-14,
       author = {Kuhl, Patricia K. and Ram\'{i}rez, Rey R. and Bosseler, Alexis and Lin, Jo-Fu Lotus and Imada, Toshiaki},
        title = {Infants' brain responses to speech suggest Analysis by Synthesis},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
         year = {2014},
       volume = {111},
       number = {31},
        pages = {11238-11245},
     abstract = {Infants discriminate speech sounds universally until 8 months of age, then native discrimination improves and nonnative discrimination declines. Using magnetoencephalography, we investigate the contribution of auditory and motor brain systems to this developmental transition. We show that 7-month-old infants activate auditory and motor brain areas similarly for native and nonnative sounds; by 11-12 month, greater activation in auditory brain areas occurs for native sounds, whereas greater activation in motor brain areas occurs for nonnative sounds, matching the adult pattern. We posit that hearing speech invokes an Analysis by Synthesis process: auditory analysis of speech is coupled with synthesis that predicts the motor plans necessary to produce it. Both brain systems contribute to the developmental transition in infant speech perception. Historic theories of speech perception (Motor Theory and Analysis by Synthesis) invoked listeners knowledge of speech production to explain speech perception. Neuroimaging data show that adult listeners activate motor brain areas during speech perception. In two experiments using magnetoencephalography (MEG), we investigated motor brain activation, as well as auditory brain activation, during discrimination of native and nonnative syllables in infants at two ages that straddle the developmental transition from language-universal to language-specific speech perception. Adults are also tested in Exp. 1. MEG data revealed that 7-mo-old infants activate auditory (superior temporal) as well as motor brain areas (Broca's area, cerebellum) in response to speech, and equivalently for native and nonnative syllables. However, in 11- and 12-month-old infants, native speech activates auditory brain areas to a greater degree than nonnative, whereas nonnative speech activates motor brain areas to a greater degree than native speech. This double dissociation in 11- to 12-month-old infants matches the pattern of results obtained in adult listeners. Our infant data are consistent with Analysis by Synthesis: auditory analysis of speech is coupled with synthesis of the motor plans necessary to produce the speech signal. The findings have implications for: (i) perception-action theories of speech perception, (ii) the impact of 'motherese' on early language learning, and (iii) the 'social-gating' hypothesis and humans' development of social understanding.},
}
@book{KuhnandJohnson2013applied,
        title = {Applied Predictive Modeling},
       author = {Kuhn, M. and Johnson, K.},
       series = {SpringerLink : B{\"{u}}cher},
         year = {2013},
    publisher = {Springer New York}
}
@inproceedings{KuipersAAAI-88,
       author = {Kuipers, Benjamin J. and Byun, Yung-Tai},
        title = {A Robust, Qualitative Method for Robot Spatial Reasoning},
    booktitle = {Proceedings AAAI-88},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
        pages = {774-779},
         year = 1988,
}
@article{KuipersCS-78,
       author = {Kuipers, Benjamin},
        title = {Modeling Spatial Knowledge},
      journal = {Cognitive Science},
       volume = 2,
         year = 1978,
        pages = {129-153},
}
@incollection{KuipersQR-85,
       author = {Kuipers, Benjamin J.},
        title = {Commonsense Reasoning About Causality: Deriving Behavior from Structure},
     crossref = {QR-85},
        pages = {169-203},
}
@article{KuipersandLevittAIM-88,
       author = {Kuipers, Benjamin and Levitt, Tod S.},
        title = {Navigation and Mapping in Large-Scale Space},
      journal = {AI Magazine},
       volume = 9,
         year = 1988,
        pages = {25-43},
}
@article{KuipersetalAIM-17,
       author = {Benjamin Kuipers and Edward A. Feigenbaum and Peter E. Hart and Nils J. Nilsson},
        title = {Shakey: From Conception to History},
      journal = {AI Magazine},
         year = {2017},
        pages = {88-103},
     abstract = {Shakey the Robot, conceived 50 years ago, was a seminal contribution to AI. Shakey perceived its world, planned how to achieve a goal, and acted to carry out that plan. This was revolutionary. At the 29th AAAI Conference on Artifcial Intelligence, attendees gathered to celebrate Shakey and to gain insights into how the AI revolution moves ahead. The celebration included a panel that was chaired by Benjamin Kuipers and featured AI pioneers Ed Feigenbaum, Peter Hart, and Nils Nilsson. This article includes written versions of the contributions of those panelists.}
}
@book{Kuldell2015biobuilder,
        title = {BioBuilder},
       author = {Kuldell, N. and Bernstein, R. and Hart, K.M. and Ingram, K.},
         year = {2015},
    publisher = {O'Reilly}
}
@article{KulikowskiandBishop81,
       author = {J. J. Kulikowski and P. O. Bishop},
        title = {Fourier analysis and spatial representation in the visual cortex},
      journal = {Experientia},
       volume = 37,
        pages = {160-163},
         year = 1981,
}
@inproceedings{KulisandGraumanICCV-09,
       author = {Brian Kulis and Kristen Grauman},
        title = {Kernelized locality-sensitive hashing for scalable image search},
    booktitle = {{IEEE} International Conference on Computer Vision},
         year = 2009
}
@article{KulkarnietalCoRR-16,
       author = {Tejas D. Kulkarni and  Ardavan Saeedi and  Simanta Gautam and  Samuel J. Gershman},
        title = {Deep Successor Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1606.02396},
         year = {2016},
     abstract = {Learning robust value functions given raw observations and rewards is now possible with model-free and model-based deep reinforcement learning algorithms. There is a third alternative, called Successor Representations (SR), which decomposes the value function into two components -- a reward predictor and a successor map. The successor map represents the expected future state occupancy from any given state and the reward predictor maps states to scalar rewards. The value function of a state can be computed as the inner product between the successor map and the reward weights. In this paper, we present DSR, which generalizes SR within an end-to-end deep reinforcement learning framework. DSR has several appealing properties including: increased sensitivity to distal reward changes due to factorization of reward and world dynamics, and the ability to extract bottleneck states (subgoals) given successor maps trained under a random policy. We show the efficacy of our approach on two diverse environments given raw pixel observations -- simple grid-world domains (MazeBase) and the Doom game engine.},
}
@article{KulkarnietalCoRR-16b,
       author = {Tejas D. Kulkarni and Karthik Narasimhan and Ardavan Saeedi and Joshua B. Tenenbaum},
        title = {Hierarchical Deep Reinforcement Learning: Integrating Temporal Abstraction and Intrinsic Motivation},
      journal = {CoRR},
       volume = {arXiv:1604.06057},
         year = {2016},
     abstract = {Learning goal-directed behavior in environments with sparse feedback is a major challenge for reinforcement learning algorithms. The primary difficulty arises due to insufficient exploration, resulting in an agent being unable to learn robust value functions. Intrinsically motivated agents can explore new behavior for its own sake rather than to directly solve problems. Such intrinsic behaviors could eventually help the agent solve tasks posed by the environment. We present hierarchical-DQN (h-DQN), a framework to integrate hierarchical value functions, operating at different temporal scales, with intrinsically motivated deep reinforcement learning. A top-level value function learns a policy over intrinsic goals, and a lower-level function learns a policy over atomic actions to satisfy the given goals. h-DQN allows for flexible goal specifications, such as functions over entities and relations. This provides an efficient space for exploration in complicated environments. We demonstrate the strength of our approach on two problems with very sparse, delayed feedback: (1) a complex discrete stochastic decision process, and (2) the classic ATARI game `Montezuma's Revenge'}
}
@inproceedings{KulkarnietalNIPS-16,
       author = {Tejas D. Kulkarni and Karthik Narasimhan and Ardavan Saeedi and Josh Tenenbaum},
        title = {Hierarchical Deep Reinforcement Learning: Integrating Temporal Abstraction and Intrinsic Motivation},
    booktitle = {Advances in Neural Information Processing Systems 29},
     location = {Barcelona, Spain},
        pages = {3675-3683},
         year = {2016},
     abstract = {Learning goal-directed behavior in environments with sparse feedback is a major challenge for reinforcement learning algorithms. One of the key difficulties is insufficient exploration, resulting in an agent being unable to learn robust policies. Intrinsically motivated agents can explore new behavior for their own sake rather than to directly solve external goals. Such intrinsic behaviors could eventually help the agent solve tasks posed by the environment. We present hierarchicalDQN (h-DQN), a framework to integrate hierarchical action-value functions, operating at different temporal scales, with goal-driven intrinsically motivated deep reinforcement learning. A top-level q-value function learns a policy over intrinsic goals, while a lower-level function learns a policy over atomic actions to satisfy the given goals. h-DQN allows for flexible goal specifications, such as functions over entities and relations. This provides an efficient space for exploration in complicated environments. We demonstrate the strength of our approach on two problems with very sparse and delayed feedback: (1) a complex discrete stochastic decision process with stochastic transitions, and (2) the classic ATARI game - 'Montezuma's Revenge'.}
}
@article{KullbackLeibler51,
       author = {S. Kullback and R. A. Leibler},
        title = {On Information and Sufficiency},
      journal = {Annals of Mathematical Statistics},
       volume = 22,
        issue = 1,
         year = 1951,
        pages = {79-86},
}
@inproceedings{Kumar99,
       author = {S.R. Kumar and P. Raghavan and S. Rajagopalan and A. Tomkins},
        title = {Extracting large scale knowledge bases from the {W}eb},
    booktitle = {IEEE International conference on Very Large Databases (VLDB)},
         year = 1999
} 
@article {KumaranandMaguireJoN-05,
       author = {Kumaran, Dharshan and Maguire, Eleanor A.},
        title = {The Human Hippocampus: Cognitive Maps or Relational Memory?},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {25},
       number = {31},
         year = {2005},
        pages = {7254-7259},
     abstract = {The hippocampus is widely accepted to play a pivotal role in memory. Two influential theories offer competing accounts of its fundamental operating mechanism. The cognitive map theory posits a special role in mapping large-scale space, whereas the relational theory argues it supports amodal relational processing. Here, we pit the two theories against each other using a novel paradigm in which the relational processing involved in navigating in a city was matched with similar navigational and relational processing demands in a nonspatial (social) domain. During functional magnetic resonance imaging, participants determined the optimal route either between friends homes or between the friends themselves using social connections. Separate brain networks were engaged preferentially during the two tasks, with hippocampal activation driven only by spatial relational processing. We conclude that the human hippocampus appears to have a bias toward the processing of spatial relationships, in accordance with the cognitive map theory. Our results both advance our understanding of the nature of the hippocampal contribution to memory and provide insights into how social networks are instantiated at the neural level.},
}
@article{KumaranandMcClellandPR-12,
       author = {Kumaran, Dharshan and McClelland, James L.},
        title = {Generalization Through the Recurrent Interaction of Episodic Memories: A Model of the Hippocampal System},
      journal = {Psychology Review},
    publisher = {American Psychological Association},
       volume = {119},
        issue = {3},
         year = {2012},
        pages = {573-616},
     abstract = {In this article, we present a perspective on the role of the hippocampal system in generalization, instantiated in a computational model called REMERGE (recurrency and episodic memory results in generalization). We expose a fundamental, but neglected, tension between prevailing computational theories that emphasize the function of the hippocampus in pattern separation (Marr, 1971; McClelland, McNaughton, \& O'Reilly, 1995), and empirical support for its role in generalization and flexible relational memory (Cohen \& Eichenbaum, 1993; Eichenbaum, 1999). Our account provides a means by which to resolve this conflict, by demonstrating that the basic representational scheme envisioned by complementary learning systems theory (McClelland et al., 1995), which relies upon orthogonalized codes in the hippocampus, is compatible with efficient generalization - as long as there is recurrence rather than unidirectional flow within the hippocampal circuit or, more widely, between the hippocampus and neocortex. We propose that recurrent similarity computation, a process that facilitates the discovery of higher-order relationships between a set of related experiences, expands the scope of classical exemplar-based models of memory (e.g., Nosofsky, 1984) and allows the hippocampus to support generalization through interactions that unfold within a dynamically created memory space.},
}
@inproceedings{KumarandHebertCVPR-03,
       author = {Sanjiv Kumar and Martial Hebert},
        title = {Man-Made Structure Detection in Natural Images Using a Causal Multiscale Random Field},
    booktitle = {Proceedings of {IEEE} Computer Vision and Pattern Recognition},
     location = {Madison, Wisconsin},
       volume = 1,
        pages = {119-126},
         year = 2003,
}
@article{KumarandHebertIJCV-06,
       author = {Sanjiv Kumar and Martial Hebert},
        title = {Discriminative Random Fields},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 68,
        issue = 2,
        pages = {179-201},
         year = 2006,
}
@article{KumarandMehtaFCN-11,
       author = {Arvind Kumar and Mayank R. Mehta},
        title = {Frequency dependent changes in {NMDAR}-dependent synaptic plasticity},
      journal = {Frontiers in Computational Neuroscience},
       volume = 5,
         year = 2011,
        issue = 0,
     abstract = {The NMDAR-dependent synaptic plasticity is thought to mediate several forms of learning, and can be induced by spike trains containing a small number of spikes occurring with varying rates and timing, as well as with oscillations. We computed the influence of these variables on the plasticity induced at a single NMDAR containing synapse using a reduced model that was analytically tractable, and these findings were confirmed using detailed, multi-compartment model. In addition to explaining diverse experimental results about the rate and timing dependence of synaptic plasticity, the model made several novel and testable predictions. We found that there was a preferred frequency for inducing long-term potentiation (LTP) such that higher frequency stimuli induced lesser LTP, decreasing as 1/f when the number of spikes in the stimulus was kept fixed. Among other things, the preferred frequency for inducing LTP varied as a function of the distance of the synapse from the soma. In fact, same stimulation frequencies could induce LTP or LTD depending on the dendritic location of the synapse. Next, we found that rhythmic stimuli induced greater plasticity then irregular stimuli. Furthermore, brief bursts of spikes significantly expanded the timing dependence of plasticity. Finally, we found that in the \tilde{}5-15Hz frequency range both rate- and timing-dependent plasticity mechanisms work synergistically to render the synaptic plasticity most sensitive to spike-timing. These findings provide computational evidence that oscillations can have a profound influence on the plasticity of an NMDAR-dependent synapse, and show a novel role for the dendritic morphology in this process.},
}
@article{KumaranetalTiCS-16,
       author = {Dharshan Kumaran and Demis Hassabis and James L. McClelland},
        title = {What Learning Systems do Intelligent Agents Need? {C}omplementary Learning Systems Theory Updated},
      journal = {Trends in Cognitive Sciences},
       volume = {20},
       number = {7},
         year = {2016},
        pages = {512-534},
     abstract = {We update complementary learning systems (CLS) theory, which holds that intelligent agents must possess two learning systems, instantiated in mammalians in neocortex and hippocampus. The first gradually acquires structured knowledge representations while the second quickly learns the specifics of individual experiences. We broaden the role of replay of hippocampal memories in the theory, noting that replay allows goal-dependent weighting of experience statistics. We also address recent challenges to the theory and extend it by showing that recurrent activation of hippocampal traces can support some forms of generalization and that neocortical learning can be rapid for information that is consistent with known structure. Finally, we note the relevance of the theory to the design of artificial intelligent agents, highlighting connections between neuroscience and machine learning.}
}
@inproceedings{Kumaretal99,
       author = {R. Kumar and P. Raghavan and S. Rajagopalan and A. Tomkins},
        title = {Extracting large scale knowledge bases from the {W}eb},
    booktitle = {Proceedings of the 25th VLDB Conference},
         year = 1999,
}
@article{KumaretalCoRR-15,
       author = {Ankit Kumar and Ozan Irsoy and Jonathan Su and James Bradbury and Robert English and Brian Pierce and Peter Ondruska and Ishaan Gulrajani and Richard Socher},
        title = {Ask Me Anything: Dynamic Memory Networks for Natural Language Processing},
      journal = {CoRR},
       volume = {arXiv:1506.07285},
         year = {2015},
     abstract = {Most tasks in natural language processing can be cast into question answering (QA) problems over language input. We introduce the dynamic memory network (DMN), a neural network architecture which processes input sequences and questions, forms episodic memories, and generates relevant answers. Questions trigger an iterative attention process which allows the model to condition its attention on the inputs and the result of previous iterations. These results are then reasoned over in a hierarchical recurrent sequence model to generate answers. The DMN can be trained end-to-end and obtains state-of-the-art results on several types of tasks and datasets: question answering (Facebook's bAbI dataset), text classification for sentiment analysis (Stanford Sentiment Treebank) and sequence modeling for part-of-speech tagging (WSJ-PTB). The training for these different tasks relies exclusively on trained word vector representations and input-question-answer triplets.}
}
@inproceedings{KumaretalFOCS-00,
       author = {S.R. Kumar and P. Raghavan and S. Rajagopalan and D. Sivakumar and A. Tomkins and Eli Upfal},
        title = {Stochastic models for the {W}eb graph},
    booktitle = {Proceedings of the 41st IEEE Symposium on the Foundations of Computer Science},
         year = 2000,
}
@inproceedings{KumaretalICML-09,
       author = {Sanjiv Kumar and Mehryar Mohri and Ameet Talwalkar},
        title = {On sampling-based approximate spectral decomposition},
       editor = {Andrea Pohoreckyj Danyluk and L{\'e}on Bottou and Michael L. Littman},
    booktitle = {Proceedings of the 26th Annual International Conference on Machine Learning, ICML 2009},
    publisher = {ACM},
       volume = 382,
         year = 2009,
        pages = 70,
}
@article{KumaretalICN-19,
       author = {Kumar, Ashutosh and Pareek, Vikas and Faiq, Muneeb A. and Ghosh, Sanjib K. and Kumari, Chiman},
        title = {Adult Neurogenesis in Humans: A Review of Basic Concepts, History, Current Research, and Clinical Implications},
      journal = {Innovations in Clinical Neuroscience},
         year = {2019},
    publisher = {Matrix Medical Communications},
       volume = {16},
        issue = {5-6},
        pages = {30-37},
     abstract = {Neurogenesis in adult humans remains a controversial area of research among neuroscientists. Methodological challenges have hampered investigators from conducting high-quality, in-vivo studies that can help elucidate the presence and/or activity of neurogenesis in human brains. Additionally, the studies that have been done in humans report conflicting results, further adding to the ambiguity surrounding the concept of adult neurogenesis in humans. In this review article, the authors seek to help clarify the concept of adult neurogenesis by providing an overview of the basic concept, as we currently understand it, including its historical birth and evolution. The authors also review and discuss current key studies (pro and con) on adult neurogenesis in humans and animals, as well as research challenges with potential solutions. Finally, the authors discuss the clinical implications of adult neurogenesis in humans, based on what we know so far, including its potential use as a drug target in the development of pharmacological treatments for various neuropsychiatric disorders.},
}
@article{KumaretalJMLR-09,
        title = {Sampling Techniques for the Nystr{\"o}m Method},
       author = {Kumar, Sanjiv and Mohri, Mehryar and Talwalkar, Ameet},
      journal = {Journal of Machine Learning Research},
       volume = 5,
         year = 2009,
        pages = {981},
}
@inproceedings{KumaretalPODS-00,
     author = {S.R. Kumar and P. Raghavan and S. Rajagopalan and D. Sivakumar and A. Tomkins and Eli Upfal},
      title = {The {W}eb as a graph},
  booktitle = {Proceedings of the 19th ACM Symposium on the Principles of Database Systems},
       year = 2000,
      pages = {1-10},
}
@article{KunertetalCoRR-14,
       author = {James Kunert and Eli Shlizerman and J. Nathan Kutz},
        title = {Low-dimensional functionality of complex network dynamics: Neuro-sensory integration in the Caenorhabditis elegans connectome},
      journal = {CoRR},
       volume = {arXiv:1310.6689},
         year = 2014,
          url = {http://arxiv.org/abs/1310.6689},
     abstract = {We develop a biophysical model of neuro-sensory integration in the model organism Caenorhabditis elegans. Building on recent experimental findings of the neuron conductances and their resolved connectome, we posit the first full dynamic model of the neural voltage excitations that allows for a characterization of input stimuli to behavioral responses. Thus a clear connection between receptory cell inputs to downstream motor-responses is illustrated, showing that robust, low-dimensional bifurcation structures dominate neural pathways of activity. The underlying bifurcation structures discovered, i.e. an induced Hopf bifurcation, are critical in explaining behavioral responses such as swimming and crawling. More broadly, we demonstrate that complex dynamical networks can produce robust functionality from underlying low-dimensional bifurcations.}
}
@article{KunjapuretalACIE-18,
        title = {From Designing the Molecules of Life to Designing Life: Future Applications Derived from Advances in DNA Technologies},
       author = {Kohman, R. and Kunjapur, A. and Hysolli, E. and Wang, Y. and Church, G.},
      journal = {Angewandte Chemie International Edition},
    publisher = {John Wiley and Sons},
       volume = 57,
        issue = 16,
         year = 2018,
        pages = {4313-4328},
     abstract = {Since the elucidation of its structure, DNA has been at the forefront of biological research. In the past half century, an explosion of DNA‐based technology development has occurred with the most rapid advances being made for DNA sequencing. In parallel, dramatic improvements have also been made in the synthesis and editing of DNA from the oligonucleotide to the genome scale. In this Review, we will summarize four different subfields relating to DNA technologies following this trajectory of smaller to larger scale. We begin by talking about building materials out of DNA which in turn can act as delivery vehicles in vivo. We then discuss how altering microbial genomes can lead to novel methods of production for industrial biologics. Next, we talk about the future of writing whole genomes as a method of studying evolution. Lastly, we highlight the ways in which barcoding biological systems will allow for their three‐dimensional analysis in a highly multiplexed fashion. : DNA‐based technologies continue to be developed at a rapid pace. Spurred by advances in the DNA synthesis and editing, future application derived from DNA technologies ranging from designing new nanomaterials to designing new organisms are discussed.}
}
@article{KunkeletalFiN-14,
       author = {Kunkel, S. and Schmidt, M. and Eppler, J. M. and Plesser, H. E. and Masumoto, G. and Igarashi, J. and Ishii, S. and Fukai, T. and Morrison, A. and Diesmann, M. and Helias, M.},
        title = {Spiking network simulation code for petascale computers},
      journal = {Frontiers in Neuroinformatics},
       volume = {8},
        issue = {78},
         year = {2014},
     abstract = {Brain-scale networks exhibit a breathtaking heterogeneity in the dynamical properties and parameters of their constituents. At cellular resolution, the entities of theory are neurons and synapses and over the past decade researchers have learned to manage the heterogeneity of neurons and synapses with efficient data structures. Already early parallel simulation codes stored synapses in a distributed fashion such that a synapse solely consumes memory on the compute node harboring the target neuron. As petaflop computers with some 100,000 nodes become increasingly available for neuroscience, new challenges arise for neuronal network simulation software: Each neuron contacts on the order of 10,000 other neurons and thus has targets only on a fraction of all compute nodes; furthermore, for any given source neuron, at most a single synapse is typically created on any compute node. From the viewpoint of an individual compute node, the heterogeneity in the synaptic target lists thus collapses along two dimensions: the dimension of the types of synapses and the dimension of the number of synapses of a given type. Here we present a data structure taking advantage of this double collapse using metaprogramming techniques. After introducing the relevant scaling scenario for brain-scale simulations, we quantitatively discuss the performance on two supercomputers. We show that the novel architecture scales to the largest petascale supercomputers available today.}
}
@book{Kurzweil2012,
       author = {Ray Kurzweil},
        title = {How to Create a Mind: The Secret of Human Thought Revealed},
    publisher = {Viking Press},
      address = {New York, NY},
         year = 2012
} 
@book{Kurzweil92,
       author = {Ray Kurzweil},
        title = {The Age of Intelligent Machines},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1992,
}
@book{Kurzweil99,
       author = {Ray Kurzweil},
        title = {The Age of Spiritual Machines: When Computers Exceed Human Intelligence},
    publisher = {Viking},
      address = {New York, NY},
         year = 1999,
}
@article{KushchayevetalBRAIN-12,
       author = {Kushchayev, Sergiy V. and Moskalenko, Vitaliy F. and Wiener, Philip C. and Tsymbaliuk, Vitaliy I. and Cherkasov, Viktor G. and Dzyavulska, Irina V. and Kovalchuk, Oleksander I. and Sonntag, Volker K. H. and Spetzler, Robert F. and Preul, Mark C.},
        title = {{The discovery of the pyramidal neurons: Vladimir Betz and a new era of neuroscience}},
      journal = {Brain},
       volume = {135},
       number = {1},
         year = {2011},
        pages = {285-300},
     abstract = {As a consequence of nascent technology, the 19th century witnessed a profound change in orientation to the nervous system. For example, improved microscopy in the first half of the 19th century allowed high magnification without blurring. The subsequent observation of nucleated cells led to the identification of individual brain cells. Philosophical changes in approach to the natural sciences took their lead from those applied to physical observations. The Ukrainian anatomist and histologist, Vladimir Alekseyevich Betz (1834-94) played a pivotal role in reshaping scientific and philosophical approaches to the brain, connecting cerebral localization, function and brain microstructure. Betz revolutionized methods of cell fixation and staining. Sometimes his efforts yielded enormously complicated technological improvements. Betz's greatest contribution, however, was connecting his discovery of the function of giant pyramidal neurons of the primary motor cortex (cells of Betz) with the cortical organization. Considering cortical cytoarchitectonics in relation with physiological function, Betz recognized this organization in two areas: motor and sensory. He defined a functional area on histological grounds and thereby opened the way to study precise cortical areas. Betz participated in the scientific transformation of cytoarchitectonics based on macro- and microscopic studies of the cortical surface, enabling him to view the paths of nerve cells in the brain. Betz's influence allowed systemization of scattered scientific findings. The discovery of pyramidal cells was a turning point in the prevailing philosophical and scientific approach to the brain, linking cytoarchitecture, neurophysiology and cerebral localization.},
}
@inproceedings{KushmanandBarzilayACL-13,
       author = {Kushman, Nate and Barzilay, Regina},
        title = {Using Semantic Unification to Generate Regular Expressions from Natural Language},
    booktitle = {Proceedings of the 2013 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies},
         year = {2013},
    publisher = {Association for Computational Linguistics},
        pages = {826-836},
     location = {Atlanta, Georgia},
     abstract = {We consider the problem of translating natural language text queries into regular expressions which represent their meaning. The mismatch in the level of abstraction between the natural language representation and the regular expression representation make this a novel and challenging problem. However, a given regular expression can be written in many semantically equivalent forms, and we exploit this flexibility to facilitate translation by finding a form which more directly corresponds to the natural language. We evaluate our technique on a set of natural language queries and their associated regular expressions which we gathered from Amazon Mechanical Turk. Our model substantially outperforms a stateof-the-art semantic parsing baseline, yielding a 29\% absolute improvement in accuracy}
}
@inproceedings{KushmericketalAAAI-94,
       author = {Nicholas Kushmerick and Steve Hanks and Daniel Weld},
        title = {An Algorithm for Probabilistic Planning},
    booktitle = {Proceedings AAAI-94},
     location = {Seattle, Washington},
 organization = {AAAI},
         year = 1994,
}
@article{KushmericketalAIJ-95,
       author = {Nicholas Kushmerick and Steve Hanks and Daniel Weld},
        title = {An Algorithm for Probabilistic Planning},
       volume = 76,
        issue = {1-2},
      journal = AIJ,
         year = 1995,
}
@article{KushnerandChenTAC-74,
       author = {Harold J. Kushner and Ching-Hui Chen},
        title = {Decomposition of systems governed by {M}arkov chains},
      journal = {{IEEE} Transactions on Automatic Control},
         year = 1974,
       volume = 19,
        issue = 5,
        pages = {501-507},
}
@techreport{KushnerandChenTR-74,
       author = {Kushner, H. J. and Chen, Ching-Hui},
        title = {Decomposition of Systems Governed by {M}arkov Chains},
  institution = {Brown University, Department of Applied Mathematics},
         year = 1974,
}
@article{KushnerandKleinmanIJC-71,
       author = {Kushner, H. J. and Kleinman, A. J.},
        title = {Mathematical Programming and the control of {M}arkov Chains},
      journal = {International Journal on Control},
       volume = 13,
        issue = 5,
         year = 1971,
        pages = {801-820},
}
@article{KuwajimaetalDEVELOPMENT-13,
       author = {Takaaki Kuwajima and Austen A. Sitko and Punita Bhansali and Chris Jurgens and William Guido and Carol Mason},
        title = {{Clear$^{\mbox{\rm{T}}}$}: a detergent- and solvent-free clearing method for neuronal and non-neuronal tissue},
      journal = {Development},
       volume = 140,
         year = 2013,
        pages = {1364-1368},
     abstract = {We describe a clearing method for enhanced visualization of cell morphology and connections in neuronal and non-neuronal tissue. Using Clear^{T} or Clear^{T2}, which are composed of formamide or formamide/polyethylene glycol, respectively, embryos, whole mounts and thick brain sections can be rapidly cleared with minimal volume changes. Unlike other available clearing techniques, these methods do not use detergents or solvents, and thus preserve lipophilic dyes, fluorescent tracers and immunohistochemical labeling, as well as fluorescent-protein labeling.}
}
@misc{KuznetsovandMohriNIPS-16,
        title = {Theory and Algorithms for Forecasting Non-Stationary Time Series: {T}utorial at {NIPS} 2016},
       author = {Vitaly Kuznetsov and Mehryar Mohri},
         year = {2016},
 howpublished = {https://cs.nyu.edu/~mohri/talks/NIPSTutorial2016.pdf},
}
@article{KvatinskyetalCSM-13,
       author = {Kvatinsky, S. and Friedman, E.G. and Kolodny, A. and Weiser, U.C.},
      journal = {Circuits and Systems Magazine, {IEEE}},
        title = {The Desired Memristor for Circuit Designers},
         year = 2013,
       volume = 13,
        issue = 2,
        pages = {17-22},
     abstract = {Memristors are two-terminal devices with varying resistance, where the behavior is dependent on the history of the device. In recent years, different physical phenomena of resistive switching have been linked with the theoretical concept of a memristor, and several emerging memory devices (e.g., Phase Change Memory, Resistive RAM, STT-MRAM) are now considered as memristors. Memristors hold promise for use in diverse applications such as memory, digital logic, analog circuits, and neuromorphic systems. Important characteristics of memristors include high speed, low power, good scalability, data retention, endurance, and compatibility with conventional CMOS in terms of manufacturing and operating voltages. One interesting property of some memristors is a nonlinear response to current or voltage. Nonlinear memristors exhibit a current or voltage threshold, such that the resistance is affected only by currents or voltages which exceed the threshold, while the resistance of a linear memristor changes with small perturbations in device current. Different applications exploit different characteristics of a memristor. In this article, the desired characteristics for different applications are presented from the viewpoint of an integrated circuit designer. Understanding the desired characteristics for different applications can assist device and material engineers in providing the appropriate behavior when developing memristive devices, thereby optimizing these devices for different applications.},
}
@inproceedings{KvatinskyetalIEEE-12,
       author = {Kvatinsky, S. and Talisveyberg, K. and Fliter, D. and Kolodny, A. and Weiser, U.C. and Friedman, E.G.},
    booktitle = {IEEE 27th Convention of Electrical Electronics Engineers in Israel},
        title = {Models of memristors for {SPICE} simulations},
         year = 2012,
        pages = {1-5},
     abstract = {Memristors are novel devices which can be used in applications such as memory, logic, analog circuits, and neuromorphic systems. Several memristor technologies have been developed such as ReRAM (Resistive RAM), MRAM (Magnetoresistance RAM), and PCM (Phase Change Memory). To design circuits with memristors, the behavior of the memristor needs to be described by a mathematical model. While the model for memristors should be sufficiently accurate as compared to the behavior of physical devices, the model must also be computationally efficient. Several models for memristors have been proposed - the linear ion drift model, the nonlinear ion drift model, the Simmons tunnel barrier model, and the ThrEshold Adaptive Memristor (TEAM) model. In this paper, the different memristor models are described and a Verilog-A implementation for these models, including the relevant window functions, are presented. These models are suitable for EDA tools such as SPICE.},
}
@misc{KwonCEREBELLUM-20,
        title = {The mysterious, multifaceted cerebellum},
       author = {Diana Kwon},
 howpublished = {\url{https://knowablemagazine.org/article/mind/2020/what-does-the-cerebellum-do}},
         year = {2020},
}
@unpublished{LIBSVM,
       author = {Chih-Chung Chang and Chih-Jen Lin},
        title = {{LIBSVM}: a Library for Support Vector Machines (Version 2.36)},
          url = {http://www.csie.ntu.edu.tw/~cjlin/libsvm/},
         year = 2002,
}
@book{LaarhovenandAarts87,
       author = {Laarhoven, P. J. M. van, and Aarts, E. H. L.},
        title = {Simulated Annealing: Theory and Applications},
    publisher = {D. Reidel Publishing Company},
      address = {Holland},
         year = 1987,
}
@article{LacyandStarkNEUROSCIENCE-13,
       author = {Lacy, J. W. and Stark, C. E. L.},
        title = {The neuroscience of memory: implications for the courtroom},
      journal = {Nature Reviews Neuroscience},
       volume = {14},
       number = {9},
         year = {2013},
        pages = {649-658},
     abstract = {Although memory can be hazy at times, it is often assumed that memories of violent or otherwise stressful events are so well encoded that they are effectively indelible and that confidently retrieved memories are almost certainly accurate. However, findings from basic psychological research and neuroscience studies indicate that memory is a reconstructive process that is susceptible to distortion. In the courtroom, even minor memory distortions can have severe consequences that are partly driven by common misunderstandings about memory--for example, that memory is more veridical than it may actually be.}
}
@inproceedings{LadkinIJCAI-87,
       author = {Ladkin, Peter},
        title = {The Completeness of a Natural System for Reasoning with Time Intervals},
    booktitle = {Proceedings of the Tenth International Joint Conference on Artificial Intelligence},
     location = {Milan, Italy},
         year = 1987,
        pages = {462-467},
}
@inproceedings{LaffertyetalICML-01,
       author = {John D. Lafferty and Andrew McCallum and Fernando C. N. Pereira},
        title = {Conditional Random Fields: Probabilistic Models for Segmenting and Labeling Sequence Data},
    booktitle = {Proceedings of the Eighteenth International Conference on Machine Learning},
    publisher = {Morgan Kaufmann Publishers Inc.},
      address = {San Francisco, CA, USA},
        pages = {282-289},
         year = 2001,
}
@article{LaiandBreakefieldPHYSIOLOGY-12,
        title = {Role of Exosomes/Microvesicles in the Nervous System and Use in Emerging Therapies},
       author = {Lai, Charles Pin-Kuang and Breakefield, Xandra Owen},
      journal = {Frontiers in Physiology},
       volume = 3,
        issue = 228,
         year = 2012,
     abstract = {Extracellular membrane vesicles (EMVs) are nanometer sized vesicles, including exosomes and  microvesicles capable of transferring DNAs, mRNAs, microRNAs, non-coding RNAs, proteins and lipids among cells without direct cell-to-cell contact, thereby representing a novel form of intercellular communication. Many cells in the nervous system have been shown to release EMVs, implicating their active roles in development, function and pathologies of this system. While substantial progress has been made in understanding the biogenesis, biophysical properties and involvement of EMVs in diseases, relatively less information is known about their biological function in the normal nervous system. In addition, since EMVs are endogenous vehicles with low immunogenicity, they have also been actively investigated for the delivery of therapeutic genes/molecules in treatment of cancer and neurological diseases. The present review summarizes current knowledge about EMV functions in the nervous system under both physiological and pathological conditions, as well as emerging EMV-based therapies that could be applied to the nervous system in the foreseeable future.}
}
@article{LaiandJanNATURE-06,
       author = {Lai, Helen C. and Jan, Lily Y.},
        title = {The distribution and targeting of neuronal voltage-gated ion channels},
      journal = {Nature Reviews Neuroscience},
         year = 2006,
       volume = 7,
        issue = 7,
        pages = {548-562},
     abstract = {Voltage-gated ion channels have to be at the right place in the right number to endow individual neurons with their specific character. Their biophysical properties together with their spatial distribution define the signalling characteristics of a neuron. Improper channel localization could cause communication defects in a neuronal network. This review covers recent studies of mechanisms for targeting voltage-gated ion channels to axons and dendrites, including trafficking, retention and endocytosis pathways for the preferential localization of particular ion channels. We also discuss how the spatial localization of these channels might contribute to the electrical excitability of neurons, and consider the need for future work in this emerging field.}
} 
@article{LairdAIJ-87,
       author = {Laird, John E. and Newell, Allen and Rosenbloom, Paul S.},
        title = {{SOAR}: An Architecture for General Intelligence},
      journal = AIJ,
       volume = 33,
         year = 1987,
        pages = {1-64},
}
@article{LairdandSaulML-94,
       author = {Philip D. Laird and Ronald Saul},
        title = {Discrete Sequence Prediction and Its Applications},
      journal = {Machine Learning},
       volume = 15,
       number = 1,
        pages = {43-68},
         year = 1994,
}
@article{LajeandBuonomanoNATURE-NEUROSCIENCE-13,
       author = {Laje, R. and Buonomano, D. V.},
        title = {Robust timing and motor patterns by taming chaos in recurrent neural networks},
      journal = {Nature Neuroscience},
       volume = {16},
       number = {7},
         year = {2013},
        pages = {925-933},
     abstract = {The brain's ability to tell time and produce complex spatiotemporal motor patterns is critical for anticipating the next ring of a telephone or playing a musical instrument. One class of models proposes that these abilities emerge from dynamically changing patterns of neural activity generated in recurrent neural networks. However, the relevant dynamic regimes of recurrent networks are highly sensitive to noise; that is, chaotic. We developed a firing rate model that tells time on the order of seconds and generates complex spatiotemporal patterns in the presence of high levels of noise. This is achieved through the tuning of the recurrent connections. The network operates in a dynamic regime that exhibits coexisting chaotic and locally stable trajectories. These stable patterns function as 'dynamic attractors' and provide a feature that is characteristic of biological systems: the ability to 'return' to the pattern being generated in the face of perturbations.}
}
@article{LakeetalBBS-17,
        title = {Building machines that learn and think like people},
       author = {Lake, Brenden M. and Ullman, Tomer D. and Tenenbaum, Joshua B. and Gershman, Samuel J.},
      journal = {Behavioral and Brain Sciences},
       volume = {40},
    publisher = {Cambridge University Press},
         year = {2017},
        pages = {e253},
     abstract = {Recent progress in artificial intelligence has renewed interest in building systems that learn and think like people. Many advances have come from using deep neural networks trained end-to-end in tasks such as object recognition, video games, and board games, achieving performance that equals or even beats that of humans in some respects. Despite their biological inspiration and performance achievements, these systems differ from human intelligence in crucial ways. We review progress in cognitive science suggesting that truly human-like learning and thinking machines will have to reach beyond current engineering trends in both what they learn and how they learn it. Specifically, we argue that these machines should (1) build causal models of the world that support explanation and understanding, rather than merely solving pattern recognition problems; (2) ground learning in intuitive theories of physics and psychology to support and enrich the knowledge that is learned; and (3) harness compositionality and learning-to-learn to rapidly acquire and generalize knowledge to new tasks and situations. We suggest concrete challenges and promising routes toward these goals that can combine the strengths of recent neural network advances with more structured cognitive models.}
} 
@inproceedings{LakeetalCSS-15,
        title = {Deep neural networks predict category typicality ratings for images},
       author = {Brenden Lake and Wojciech Zaremba and R Fergus and Todd Gureckis},
       editor = {R. Dale and C. Jennings and P. Maglio and T. Matlock and D. Noelle and A. Warlaumont and J. Yoshimi},
    booktitle = {Proceedings of the 37th Annual Conference of the Cognitive Science Society},
    publisher = {Cognitive Science Society},
         year = {2015},
     abstract = {The latest generation of neural networks has made major performance advances in object categorization from raw images. In particular, deep convolutional neural networks currently outperform alternative approaches on standard benchmarks by wide margins and achieve human-like accuracy on some tasks. These engineering successes present an opportunity to explore long-standing questions about the nature of human concepts by putting psychological theories to test at an unprecedented scale. This paper evaluates deep convolutional networks trained for classification on their ability to predict category typicality -- a variable of paramount importance in the psychology of concepts -- from the raw pixels of naturalistic images of objects. We find that these models have substantial predictive power, unlike simpler features computed from the same massive dataset, showing how typicality might emerge as a byproduct of a complex model trained to maximize classification performance.}
}
@article{LakeetalCoRR-16,
       author = {Brenden M. Lake and Tomer D. Ullman and Joshua B. Tenenbaum and Samuel J. Gershman},
        title = {Building Machines That Learn and Think Like People},
      journal = {CoRR},
       volume = {arXiv:1604.00289},
         year = {2016},
          url = {http://arxiv.org/abs/1604.00289},
        pages = {139-142},
     abstract = {Recent progress in artificial intelligence (AI) has renewed interest in building systems that learn and think like people. Many advances have come from using deep neural networks trained end-to-end in tasks such as object recognition, video games, and board games, achieving performance that equals or even beats humans in some respects. Despite their biological inspiration and performance achievements, these systems differ from human intelligence in crucial ways. We review progress in cognitive science suggesting that truly human-like learning and thinking machines will have to reach beyond current engineering trends in both what they learn, and how they learn it. Specifically, we argue that these machines should (a) build causal models of the world that support explanation and understanding, rather than merely solving pattern recognition problems; (b) ground learning in intuitive theories of physics and psychology, to support and enrich the knowledge that is learned; and (c) harness compositionality and learning-to-learn to rapidly acquire and generalize knowledge to new tasks and situations. We suggest concrete challenges and promising routes towards these goals that can combine the strengths of recent neural network advances with more structured cognitive models.},
}
@article{LakeetalSCIENCE-15,
       author = {Lake, Brenden M. and Salakhutdinov, Ruslan and Tenenbaum, Joshua B.},
        title = {Human-level concept learning through probabilistic program induction},
      journal = {Science},
       volume = {350},
       number = {6266},
         year = {2015},
        pages = {1332-1338},
     abstract = {People learning new concepts can often generalize successfully from just a single example, yet machine learning algorithms typically require tens or hundreds of examples to perform with similar accuracy. People can also use learned concepts in richer ways than conventional algorithms--for action, imagination, and explanation. We present a computational model that captures these human learning abilities for a large class of simple visual concepts: handwritten characters from the world's alphabets. The model represents concepts as simple programs that best explain observed examples under a Bayesian criterion. On a challenging one-shot classification task, the model achieves human-level performance while outperforming recent deep learning approaches. We also present several "visual Turing tests" probing the model's creative generalization abilities, which in many cases are indistinguishable from human behavior.},
}
@article{LakoffSYNTHESE-70,
       author = {George Lakoff},
        title = {Linguistics and Natural Logic},
      journal = {Synthese},
       volume = 22,
         year = 1970,
        pages = {151-271},
}
@article{LamandBacchusCIJ-94,
       author = {Lam, Wai and Bacchus, Fahiem},
        title = {Learning {B}ayesian Belief Networks: An Approach Based on the {MDL} Principle},
      journal = CIJ,
       volume = 10,
        issue = 4,
         year = 1994,
}
@article{LamertetalPAMI-09,
        title = {Efficient Subwindow Search: A Branch and Bound Framework for Object Localization},
       author = {Lampert, C.H. and Blaschko, M.B. and Hofmann, T.},
      journal = {Pattern Analysis and Machine Intelligence, IEEE Transactions on},
    publisher = {IEEE Computer Society},
       volume = 31,
        issue = 12,
         year = 2009,
        pages = {2129-2142},
}
@article{LammeTiCS-06,
       author = {Lamme, V. A.},
        title = {Towards a true neural stance on consciousness},
      journal = {Trends in Cognitive Science},
         year = {2006},
       volume = {10},
       number = {11},
        pages = {494-501},
     abstract = {Consciousness is traditionally defined in mental or psychological terms. In trying to find its neural basis, introspective or behavioral observations are considered the gold standard, to which neural measures should be fitted. I argue that this poses serious problems for understanding the mind-brain relationship. To solve these problems, neural and behavioral measures should be put on an equal footing. I illustrate this by an example from visual neuroscience, in which both neural and behavioral arguments converge towards a coherent scientific definition of visual consciousness. However, to accept this definition, we need to let go of our intuitive or psychological notions of conscious experience and let the neuroscience arguments have their way. Only by moving our notion of mind towards that of brain can progress be made.}
}
@inproceedings{LampertetalCVPR-08,
       title = {Beyond Sliding Windows: Object Localization by Efficient Subwindow Search}, 
      author = {Lampert, Christoph and Blaschko, Matthew and Hofmann, Thomas}, 
   booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
   publisher = {IEEE},
        year = 2008,
       pages = {1-8},
}
@inproceedings{LampinenetalCOGSCI-17,
       author = {Andrew Kyle Lampinen and Shaw Hsu and James L. McClelland},
        title = {Analogies Emerge from Learning Dyamics in Neural Networks},
    booktitle = {Proceedings of the 39th Annual Meeting of the Cognitive Science Society},
         year = {2017},
     abstract = {When a neural network is trained on multiple analogous tasks, previous research has shown that it will often generate representations that reflect the analogy. This may explain the value of multi-task training, and also may underlie the power of human analogical reasoning \emdash{} awareness of analogies may emerge naturally from gradient-based learning in neural networks. We explore this issue by generalizing linear analysis techniques to explore two sets of analogous tasks, show that analogical structure is commonly extracted, and address some potential implications.}
}
@inproceedings{LampinenetalCOGSCI-21,
       author = {Andrew Lampinen and Shaw Hsu and James L. McClelland},
       editor = {Tecumseh Fitch and Claus Lamm and Helmut Leder and Kristin Tessmar},
        title = {Analogies Emerge from Learning Dyamics in Neural Networks},
    booktitle = {Proceedings of the 43rd Annual Meeting of the Cognitive Science Society},
    publisher = {Cognitive Science Society},
         year = {2020},
        pages = {2513-2517},
     abstract = {When a neural network is trained on multiple analogous tasks, previous research has shown that it will often generate rep- resentations that reflect the analogy. This may explain the value of multi-task training, and also may underlie the power of human analogical reasoning - awareness of analogies may emerge naturally from gradient-based learning in neural net- works. We explore this issue by generalizing linear analysis techniques to explore two sets of analogous tasks, show that analogical structure is commonly extracted, and address some potential implications.},
}
@inproceedings{LampinenetalCS-17,
        title = {Analogies Emerge from Learning Dyamics in Neural Networks},
       author = {Andrew Lampinen and Shaw Hsu and James L. McClelland},
    booktitle = {Proceedings of the 39th Annual Conference of the Cognitive Science Society,},
       editor = {G. Gunzelmann and A. Howes and T. Tenbrink and E. J. Davelaar},
        pages = {2512-2517},
         year = {2017},
     abstract = {When a neural network is trained on multiple analogous tasks, previous research has shown that it will often generate representations that reflect the analogy. This may explain the value of multi-task training, and also may underlie the power of human analogical reasoning \emdash{} awareness of analogies may emerge naturally from gradient-based learning in neural networks. We explore this issue by generalizing linear analysis techniques to explore two sets of analogous tasks, show that analogical structure is commonly extracted, and address some potential implications.}
}
@article{LampinenetalCoRR-19,
       author = {Sebastien Racaniere and Andrew K. Lampinen and Adam Santoro and David P. Reichert and Vlad Firoiu and Timothy P. Lillicrap},
        title = {Automated curricula through setter-solver interactions},
      journal = {CoRR},
       volume = {arXiv:1909.12892},
         year = {2019},
     abstract = {Reinforcement learning algorithms use correlations between policies and rewards to improve agent performance. But in dynamic or sparsely rewarding environments these correlations are often too small, or rewarding events are too infrequent to make learning feasible. Human education instead relies on curricula--the breakdown of tasks into simpler, static challenges with dense rewards--to build up to complex behaviors. While curricula are also useful for artificial agents, hand-crafting them is time consuming. This has lead researchers to explore automatic curriculum generation. Here we explore automatic curriculum generation in rich, dynamic environments. Using a setter-solver paradigm we show the importance of considering goal validity, goal feasibility, and goal coverage to construct useful curricula. We demonstrate the success of our approach in rich but sparsely rewarding 2D and 3D environments, where an agent is tasked to achieve a single goal selected from a set of possible goals that varies between episodes, and identify challenges for future work. Finally, we demonstrate the value of a novel technique that guides agents towards a desired goal distribution. Altogether, these results represent a substantial step towards applying automatic task curricula to learn complex, otherwise unlearnable goals, and to our knowledge are the first to demonstrate automated curriculum generation for goal-conditioned agents in environments where the possible goals vary between episodes.}
}
@article{LanciegoetalCSHPM-12,
       author = {Lanciego, J. L. and Luquin, N. and Obeso, J. A.},
        title = {Functional neuroanatomy of the basal ganglia},
      journal = {Cold Spring Harbor Perspectives in Medicine},
       volume = {2},
       number = {12},
         year = {2012},
        pages = {a009621},
     abstract = {The "basal ganglia" refers to a group of subcortical nuclei responsible primarily for motor control, as well as other roles such as motor learning, executive functions and behaviors, and emotions. Proposed more than two decades ago, the classical basal ganglia model shows how information flows through the basal ganglia back to the cortex through two pathways with opposing effects for the proper execution of movement. Although much of the model has remained, the model has been modified and amplified with the emergence of new data. Furthermore, parallel circuits subserve the other functions of the basal ganglia engaging associative and limbic territories. Disruption of the basal ganglia network forms the basis for several movement disorders. This article provides a comprehensive account of basal ganglia functional anatomy and chemistry and the major pathophysiological changes underlying disorders of movement. We try to answer three key questions related to the basal ganglia, as follows: What are the basal ganglia? What are they made of? How do they work? Some insight on the canonical basal ganglia model is provided, together with a selection of paradoxes and some views over the horizon in the field.}
} 
@book{LandandNilsson2002,
        title = {Animal Eyes},
       author = {Michael F. Land and Dan-Eric Nilsson},
    publisher = {Oxford University Press},
      address = {Oxford, UK},
         year = 2002,
}
@article{LanderCELL-16,
       author = {Eric S. Lander},
        title = {The Heroes of \{CRISPR\}},
      journal = {Cell},
       volume = {164},
       number = {1-2},
         year = {2016},
        pages = {18-28},
     abstract = {Three years ago, scientists reported that \{CRISPR\} technology can enable precise and efficient genome editing in living eukaryotic cells. Since then, the method has taken the scientific community by storm, with thousands of labs using it for applications from biomedicine to agriculture. Yet, the preceding 20-year journey--the discovery of a strange microbial repeat sequence; its recognition as an adaptive immune system; its biological characterization; and its repurposing for genome engineering--remains little known. This Perspective aims to fill in this backstory--the history of ideas and the stories of pioneers--and draw lessons about the remarkable ecosystem underlying scientific discovery.}
}
@inproceedings{LanetalICCV-11,
       author = {Tian Lan and Yang Wang and Greg Mori},
        title = {Discriminative Figure-Centric Models for Joint Action Localization and Recognition},
    booktitle = {International Conference on Computer Vision},
         year = 2011,
        pages = {2003-2010}
}
@inproceedings{LangACL-09,
       author = {Percy Liang and Michael I. Jordan and Dan Klein},
        title = {Learning Semantic Correspondences with Less Supervision},
    booktitle = {Proceedings of the 47th Annual Meeting of the Association for Computational 
                 Linguistics and the 4th International Joint Conference on Natural Language Processing},
       editor = {Keh-Yih Su and Jian Su and Janyce Wiebe},
         year = 2009,
        pages = {91-99},
}
@article{Langlotz,
       author = {Langlotz, Curtis P. and Fagan, Lawrence M. and Tu, Samson W. and Sikic, Branimir I. and Shortliffe, Edward H.},
        title = {A Therapy Planning Architecture That Combines Decision Theory and Artificial Intelligence Techniques},
      journal = {Computers and Biomedical Research},
       volume = 20,
         year = 1987,
        pages = {279-303},
}
@incollection{Langton89,
      author = {C. G. Langton},
       title = {Artificial Life},
   booktitle = {Artificial Life {I}, Santa Fe Institut Studies in the Sciences of Complexity},
      editor = {C. G. Langton},
   publisher = {Addison-Wesley},
     address = {Reading, MA},
       pages = {1-44},
        year = 1989,
}
@article{LankaranyetalPNAS-19,
       author = {Lankarany, Milad and Al-Basha, Dhekra and Ratt{\'e}, St{\'e}phanie and Prescott, Steven A.},
        title = {Differentially synchronized spiking enables multiplexed neural coding},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {116},
       number = {20},
         year = {2019},
	pages = {10097-10102},
     abstract = {The nervous system processes phenomenal amounts of information. This processing must be conducted efficiently. In telecommunications systems, efficiency is increased by transmitting multiple signals through a single communication channel, or multiplexing. Neurons also multiplex. Here, we demonstrate a strategy for multiplexing different features of aperiodic stimuli: Cortical neurons use the rate of asynchronous spiking to encode stimulus intensity while using the timing of synchronous spikes to encode abrupt changes in stimulus intensity. This is possible because high-contrast features (e.g., edges) evoke spikes that transiently synchronize across neurons, whereas low-contrast features evoke sustained asynchronous spiking whose rate is proportional to stimulus intensity. Differentially synchronized spiking evoked in the same neurons by different stimulus features enables the formation of multiplexed representations.Multiplexing refers to the simultaneous encoding of two or more signals. Neurons have been shown to multiplex, but different stimuli require different multiplexing strategies. Whereas the frequency and amplitude of periodic stimuli can be encoded by the timing and rate of the same spikes, natural scenes, which comprise areas over which intensity varies gradually and sparse edges where intensity changes abruptly, require a different multiplexing strategy. Recording in vivo from neurons in primary somatosensory cortex during tactile stimulation, we found that stimulus onset and offset (edges) evoked highly synchronized spiking, whereas other spikes in the same neurons occurred asynchronously. Stimulus intensity modulated the rate of asynchronous spiking, but did not affect the timing of synchronous spikes. From this, we hypothesized that spikes driven by high- and low-contrast stimulus features can be distinguished on the basis of their synchronization, and that differentially synchronized spiking can thus be used to form multiplexed representations. Applying a Bayesian decoding method, we verified that information about high- and low-contrast features can be recovered from an ensemble of model neurons receiving common input. Equally good decoding was achieved by distinguishing synchronous from asynchronous spikes and applying reverse correlation methods separately to each spike type. This result, which we verified with patch clamp recordings in vitro, demonstrates that neurons receiving common input can use the rate of asynchronous spiking to encode the intensity of low-contrast features while using the timing of synchronous spikes to encode the occurrence of high-contrast features. We refer to this strategy as synchrony-division multiplexing.},
}
@inproceedings{LansdelletalICLR-20,
        title = {Learning to solve the credit assignment problem},
       author = {Benjamin James Lansdell and Prashanth Ravi Prakash and Konrad Paul Kording},
    booktitle = {International Conference on Learning Representations},
         year = {2020},
     abstract = {Backpropagation is driving today's artificial neural networks (ANNs). However, despite extensive research, it remains unclear if the brain implements this algorithm. Among neuroscientists, reinforcement learning (RL) algorithms are often seen as a realistic alternative: neurons can randomly introduce change, and use unspecific feedback signals to observe their effect on the cost and thus approximate their gradient. However, the convergence rate of such learning scales poorly with the number of involved neurons. Here we propose a hybrid learning approach. Each neuron uses an RL-type strategy to learn how to approximate the gradients that backpropagation would provide. We provide proof that our approach converges to the true gradient for certain classes of networks. In both feedforward and convolutional networks, we empirically show that our approach learns to approximate the gradient, and can match or the performance of exact gradient-based learning. Learning feedback weights provides a biologically plausible mechanism of achieving good performance, without the need for precise, pre-specified learning rules.},
}
@article{LanskyCIJ-88,
       author = {Lansky, Amy L.},
        title = {Localized Event-Based Reasoning for Multiagent Domains},
      journal = CIJ,
       volume = 4,
        issue = 4,
         year = 1988,
}
@inproceedings{LaptevandLindebergICCV-03,
     author = {Ivan Laptev and Tony Lindeberg},
      title = {Space-time interest points},
  booktitle = {Proceedings of the ninth {IEEE} International Conference on Computer Vision},
     volume = 1,
      pages = {432-439},
       year = 2003,
}
@article{LaptevandLindebergIVC-04,
     author = {Ivan Laptev and Tony Lindeberg},
      title = {Velocity adaptation of spatio-temporal receptive fields for direct 
               recognition of activities: an experimental study},
    journal = {Image and Vision Computing},
     volume = 22,
      pages = {105-116},
       year = 2004,
}
@article{LaraandWallisFiSN-15,
       author = {Lara, Antonio H. and Wallis, Jonathan D.},
        title = {The Role of Prefrontal Cortex in Working Memory: A Mini Review},
      journal = {Frontiers in System Neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = {9},
         year = {2015},
        pages = {173},
     abstract = {A prominent account of prefrontal cortex (PFC) function is that single neurons within the PFC maintain representations of task-relevant stimuli in working memory. Evidence for this view comes from studies in which subjects hold a stimulus across a delay lasting up to several seconds. Persistent elevated activity in the PFC has been observed in animal models as well as in humans performing these tasks. This persistent activity has been interpreted as evidence for the encoding of the stimulus itself in working memory. However, recent findings have posed a challenge to this notion. A number of recent studies have examined neural data from the PFC and posterior sensory areas, both at the single neuron level in primates, and at a larger scale in humans, and have failed to find encoding of stimulus information in the PFC during tasks with a substantial working memory component. Strong stimulus related information, however, was seen in posterior sensory areas. These results suggest that delay period activity in the PFC might be better understood not as a signature of memory storage per se, but as a top down signal that influences posterior sensory areas where the actual working memory representations are maintained.},
}
@article{LarkeyandLoveCOGNITIVE-SCIENCE-03,
        title = {CAB: Connectionist Analogy Builder},
       author = {Levi B. Larkey and Bradley C. Love},
      journal = {Cognitive Science},
       volume = {27},
       number = {5},
        pages = {781-794},
         year = {2003},
     abstract = {The ability to make informative comparisons is central to human cognition. Comparison involves aligning two representations and placing their elements into correspondence. Detecting correspondences is a necessary component of analogical inference, recognition, categorization, schema formation, and similarity judgment. Connectionist Analogy Builder (CAB) determines correspondences through a simple iterative computation that matches elements in one representation with elements playing compatible roles in the other representation while simultaneously enforcing structural constraints. CAB shows promise as a process model of comparison as its performance can be related to human performance (e.g., solution trajectory, error patterns, time-on-task). Furthermore, CAB's bounded working memory allows it to account for the inherent capacity limitations of human processing. CAB's strengths are its parsimony, transparency of operations, and ability to generate performance predictions. In this paper, CAB is evaluated against benchmark phenomena from the analogy literature.}
}
@article{LarkumCELL-12,
       author = {Larkum, M.},
        title = {{A} cellular mechanism for cortical associations: an organizing principle for the cerebral cortex},
      journal = {Trends Neuroscience},
         year = 2013,
       volume = 36,
        issue = 3,
        pages = {141-151},
     abstract = {A basic feature of intelligent systems such as the cerebral cortex is the ability to freely associate aspects of perceived experience with an internal representation of the world and make predictions about the future. Here, a hypothesis is presented that the extraordinary performance of the cortex derives from an associative mechanism built in at the cellular level to the basic cortical neuronal unit: the pyramidal cell. The mechanism is robustly triggered by coincident input to opposite poles of the neuron, is exquisitely matched to the large- and fine-scale architecture of the cortex, and is tightly controlled by local microcircuits of inhibitory neurons targeting subcellular compartments. This article explores the experimental evidence and the implications for how the cortex operates.},
}
@article{LarocheetalHIPPOCAMPUS-00,
       author = {Laroche, Serge and Davis, Sabrina and Jay, Th\'{e}r\`{e}se M.},
        title = {Plasticity at hippocampal to prefrontal cortex synapses: Dual roles in working memory and consolidation},
      journal = {Hippocampus},
       volume = {10},
       number = {4},
         year = {2000},
        pages = {438-446},
     abstract = {The involvement of the hippocampus and the prefrontal cortex in cognitive processes and particularly in learning and memory has been known for a long time. However, the specific role of the projection which connects these two structures has remained elusive. The existence of a direct monosynaptic pathway from the ventral CA1 region of the hippocampus and subiculum to specific areas of the prefrontal cortex provides a useful model for conceptualizing the functional operations of hippocampal-prefrontal cortex communication in learning and memory. It is known now that hippocampal to prefrontal cortex synapses are modifiable synapses and can express different forms of plasticity, including long-term potentiation, long-term depression, and depotentiation. Here we review these findings and focus on recent studies that start to relate synaptic plasticity in the hippocampo-prefrontal cortex pathway to two specific aspects of learning and memory, i.e., the consolidation of information and working memory. The available evidence suggests that functional interactions between the hippocampus and prefrontal cortex in cognition and memory are more complex than previously anticipated, with the possibility for bidirectional regulation of synaptic strength as a function of the specific demands of tasks.},
} 
@inproceedings{LarochelleandLaulyNIPS-12,
      title = {A Neural Autoregressive Topic Model},
     author = {Larochelle, Hugo and Lauly, Stanislas},
  booktitle = {Advances in Neural Information Processing Systems 25},
     editor = {Bartlett, Peter L. and Pereira, Fernando C. N. and Burges, Christopher J. C. and Bottou, L\'{e}on and Weinberger, Kilian Q.},
       year = 2012,
      pages = {2717-2725},
}
@inproceedings{LarochelleetalICML-07,
    author = {Hugo Larochelle and Dumitru Erhan and Aaron Courville and James Bergstra and Yoshua Bengio},
     title = {An Empirical Evaluation of Deep Architectures on Problems with Many Factors of Variation},
 booktitle = {Proceedings of the 24th International Conference on Machine learning},
    editor = {Zoubin Ghahramani},
 publisher = {ACM Press},
  location = {Corvalis, Oregon, USA},
     pages = {473-480},
      year = 2007,
}
@book{Lasdon70,
      title = {Optimization Theory for Large Systems},
     author = {Lasdon, Leon S.},
   location = {New York},
  publisher = {Macmillan Company},
       year = 1970,
}
@article{LaskeyetalCoRR-17,
       author = {Michael Laskey and Jonathan Lee and Wesley Yu{-}Shu Hsieh and Richard Liaw and Jeffrey Mahler and Roy Fox and Ken Goldberg},
        title = {Iterative Noise Injection for Scalable Imitation Learning},
      journal = {CoRR},
       volume = {arXiv:1703.09327},
         year = {2017},
     abstract = {One approach to Imitation Learning is Behavior Cloning, in which a robot observes a supervisor and infers a control policy. A known problem with this "off-policy" approach is that the robot's errors compound when drifting away from the supervisor's demonstrations. On-policy, techniques alleviate this by iteratively collecting corrective actions for the current robot policy. However, these techniques can be tedious for human supervisors, add significant computation burden, and may visit dangerous states during training. We propose an off-policy approach that injects noise into the supervisor's policy while demonstrating. This forces the supervisor to demonstrate how to recover from errors. We propose a new algorithm, DART (Disturbances for Augmenting Robot Trajectories), that collects demonstrations with injected noise, and optimizes the noise level to approximate the error of the robot's trained policy during data collection. We compare DART with DAgger and Behavior Cloning in two domains: in simulation with an algorithmic supervisor on the MuJoCo tasks (Walker, Humanoid, Hopper, Half-Cheetah) and in physical experiments with human supervisors training a Toyota HSR robot to perform grasping in clutter. For high dimensional tasks like Humanoid, DART can be up to 3x faster in computation time and only decreases the supervisor's cumulative reward by 5\% during training, whereas DAgger executes policies that have 80\% less cumulative reward than the supervisor. On the grasping in clutter task, DART obtains on average a 62\% performance increase over Behavior Cloning.},
}
@inproceedings{LassezetalAINAW-07,
       author = {J. L. Lassez and R. A. Rossi and A. E. Bernal},
    booktitle = {21st International Conference on Advanced Information Networking and Applications Workshops},
        title = {Crick's Hypothesis Revisited: The Existence of a Universal Coding Frame},
         year = {2007},
       volume = {1},
        pages = {745-751},
     abstract = {In 1957 Crick hypothesized that the genetic code was a comma free code. This property would imply the existence of a universal coding frame and make the set of coding sequences a locally testable language. As the link between nucleotides and amino acids became better understood, it appeared cleary that the genetic code was not comma free. Crick then adopted a radically different hypothesis: the "frozen accident". However, the notions of comma free codes and locally testable languages are now playing a role in DNA Computing, while circular codes have been found as subsets of the genetic code. We revisit Crick's 1957 hypothesis in that context. We show that coding sequences from a wide variety of genes from the three domains, eukaryotes, prokaryotes and archaea, have a property of testable by fragments, which is an adaptation of the notion of local testability to DNA sequences. These results support the existence of a universal coding frame, as the frame of a coding sequence can be determined from one of its fragments, independently from the gene or the organism the coding sequence comes from.},
}
@article{LatchoumaneetalNEURON-17,
       author = {Latchoumane, C. V. and Ngo, H. V. and Born, J. and Shin, H. S. },
        title = {Thalamic {S}pindles {P}romote {M}emory {F}ormation during {S}leep through {T}riple {P}hase-{L}ocking of {C}ortical, {T}halamic, and {H}ippocampal {R}hythms},
      journal = {Neuron},
         year = {2017},
       volume = {95},
       number = {2},
        pages = {424-435},
     abstract = {While the interaction of the cardinal rhythms of non-rapid-eye-movement (NREM) sleep-the thalamo-cortical spindles, hippocampal ripples, and the cortical slow oscillations-is thought to be critical for memory consolidation during sleep, the role spindles play in this interaction is elusive. Combining optogenetics with a closed-loop stimulation approach in mice, we show here that only thalamic spindles induced in-phase with cortical slow oscillation up-states, but not out-of-phase-induced spindles, improve consolidation of hippocampus-dependent memory during sleep. Whereas optogenetically stimulated spindles were as efficient as spontaneous spindles in nesting hippocampal ripples within their excitable troughs, stimulation in-phase with the slow oscillation up-state increased spindle co-occurrence and frontal spindle-ripple co-occurrence, eventually resulting in increased triple coupling of slow oscillation-spindle-ripple events. In-phase optogenetic suppression of thalamic spindles impaired hippocampus-dependent memory. Our results suggest a causal role for thalamic sleep spindles in hippocampus-dependent memory consolidation, conveyed through triple coupling of slow oscillations, spindles, and ripples.}
}
@article{LathamandNirenbergNC-04,
       author = {Latham, P. E. and Nirenberg, S.},
        title = {Computing and stability in cortical networks},
      journal = {Neural Computation},
       volume = {16},
       number = {7},
         year = {2004},
        pages = {1385-1412},
     abstract = {Cortical neurons are predominantly excitatory and highly interconnected. In spite of this, the cortex is remarkably stable: normal brains do not exhibit the kind of runaway excitation one might expect of such a system. How does the cortex maintain stability in the face of this massive excitatory feedback? More importantly, how does it do so during computations, which necessarily involve elevated firing rates? Here we address these questions in the context of attractor networks-networks that exhibit multiple stable states, or memories. We find that such networks can be stabilized at the relatively low firing rates observed in vivo if two conditions are met: (1) the background state, where all neurons are firing at low rates, is inhibition dominated, and (2) the fraction of neurons involved in a memory is above some threshold, so that there is sufficient coupling between the memory neurons and the background. This allows "dynamical stabilization" of the attractors, meaning feedback from the pool of background neurons stabilizes what would otherwise be an unstable state. We suggest that dynamical stabilization may be a strategy used for a broad range of computations, not just those involving attractors.}
}
@article{LathametalJoN-00,
       author = {Latham, P. E. and Richmond, B. J. and Nelson, P. G. and Nirenberg, S.},
        title = {Intrinsic dynamics in neuronal networks. {I}. {T}heory},
      journal = {Journal of Neurophysiology},
       volume = {83},
       number = {2},
         year = {2000},
        pages = {808-827},
     abstract = {Many networks in the mammalian nervous system remain active in the absence of stimuli. This activity falls into two main patterns: steady firing at low rates and rhythmic bursting. How are these firing patterns generated? Specifically, how do dynamic interactions between excitatory and inhibitory neurons produce these firing patterns, and how do networks switch from one firing pattern to the other? We investigated these questions theoretically by examining the intrinsic dynamics of large networks of neurons. Using both a semianalytic model based on mean firing rate dynamics and simulations with large neuronal networks, we found that the dynamics, and thus the firing patterns, are controlled largely by one parameter, the fraction of endogenously active cells. When no endogenously active cells are present, networks are either silent or fire at a high rate; as the number of endogenously active cells increases, there is a transition to bursting; and, with a further increase, there is a second transition to steady firing at a low rate. A secondary role is played by network connectivity, which determines whether activity occurs at a constant mean firing rate or oscillates around that mean. These conclusions require only conventional assumptions: excitatory input to a neuron increases its firing rate, inhibitory input decreases it, and neurons exhibit spike-frequency adaptation. These conclusions also lead to two experimentally testable predictions: (1) isolated networks that fire at low rates must contain endogenously active cells and (2) a reduction in the fraction of endogenously active cells in such networks must lead to bursting.}
}
@book{Latombe90,
     author = {Latombe, Jean-Claude},
      title = {Robot Motion Planning},
  publisher = {Kluwer},
    address = {Boston, Massachusetts},
       year = 1990,
}
@inproceedings{LattimoreetalICML-13,
        title = {The Sample-Complexity of General Reinforcement Learning},
       author = {Tor Lattimore and Marcus Hutter and Peter Sunehag},
    booktitle = {Proceedings of the 30th International Conference on Machine Learning},
       editor = {Sanjoy Dasgupta and David McAllester},
       volume = {28},
    publisher = {PMLR},
         year = {2013},
        pages = {28-36},
     abstract = {We study the sample-complexity of reinforcement learning in a general setting without  assuming ergodicity or finiteness of the environment. Instead, we define a topology  on the space of environments and show that  if an environment class is compact with respect to this topology then finite sample-complexity bounds are possible and give an  algorithm achieving these bounds. We also  show the existence of environment classes  that are non-compact where finite sample-complexity bounds are not achievable. A  lower bound is presented that matches the upper bound except for logarithmic factors.}
}
@article{LaueretalJoVE-18,
       author = {Lauer, Simon M. and Schneeweib, Undine and Brecht, Michael and Ray, Saikat},
        title = {Visualization of Cortical Modules in Flattened Mammalian Cortices},
      journal = {Journal of visualized experiments : JoVE},
         year = {2018},
    publisher = {MyJove Corporation},
        issue = {131},
        pages = {56992},
     abstract = {The cortex of mammalian brains is parcellated into distinct substructures or modules. Cortical modules typically lie parallel to the cortical sheet, and can be delineated by certain histochemical and immunohistochemical methods. In this study, we highlight a method to isolate the cortex from mammalian brains and flatten them to obtain sections parallel to the cortical sheet. We further highlight selected histochemical and immunohistochemical methods to process these flattened tangential sections to visualize cortical modules. In the somatosensory cortex of various mammals, we perform cytochrome oxidase histochemistry to reveal body maps or cortical modules representing different parts of the body of the animal. In the medial entorhinal cortex, an area where grid cells are generated, we utilize immunohistochemical methods to highlight modules of genetically determined neurons which are arranged in a grid-pattern in the cortical sheet across several species. Overall, we provide a framework to isolate and prepare layer-wise flattened cortical sections, and visualize cortical modules using histochemical and immunohistochemical methods in a wide variety of mammalian brains.},
} 
@article{LaughlinNATURE-98,
     title = {The metabolic cost of neural information},
    author = {Laughlin, S.B. and de Ruyter van Steveninck, R.R. and Anderson, J.C.},
   journal = {Nature Neuroscience},
    volume = 1,
     issue = 1,
     pages = {36-41},
      year = 1998,
  abstract = {We derive experimentally based estimates of the energy used by neural mechanisms to code known quantities of information. Biophysical measurements from cells in the blowfly retina yield estimates of the ATP required to generate graded (analog) electrical signals that transmit known amounts of information. Energy consumption is several orders of magnitude greater than the thermodynamic minimum. It costs 10(4) ATP molecules to transmit a bit at a chemical synapse, and 10(6)-10(7) ATP for graded signals in an interneuron or a photoreceptor, or for spike coding. Therefore, in noise-limited signaling systems, a weak pathway of low capacity transmits information more economically, which promotes the distribution of information among multiple pathways.}
}
@article{LaughlinandSejnowskiCNN-03,
       author = {Laughlin, Simon B. and Sejnowski, Terrence J.},
        title = {Communication in Neuronal Networks},
      journal = {Science},
       volume = 301,
        issue = 5641,
        pages = {1870-1874},
         year = 2003,
     abstract = {Brains perform with remarkable efficiency, are capable of prodigious computation, and are marvels of communication. We are beginning to understand some of the geometric, biophysical, and energy constraints that have governed the evolution of cortical networks. To operate efficiently within these constraints, nature has optimized the structure and function of cortical networks with design principles similar to those used in electronic networks. The brain also exploits the adaptability of biological systems to reconfigure in response to changing needs.},
}
@inproceedings{LaumondIJCAI-83,
     author = {Laumond, J.-P.},
      title = {Model Structuring and Concept Recognition: Two Aspects ofLearning for a Mobile Robot},
  booktitle = {Proceedings {IJCAI} 8, Karlsruhe, West Germany},
organization = {IJCAII},
       year = 1983,
      pages = {839-841},
}
@inproceedings{LaurenceandMargolis2013,
       author = {Stephen Laurence and Eric Margolis},
        title = {Concepts and Cognitive Science},
    booktitle = {Concepts: Core Readings},
       editor = {E. Margolis and S. Laurence},
    publisher = {MIT Press},
         year = {2013}
}
@techreport{Lauritzen88,
     author = {Steffen L. Lauritzen},
      title = {Mixed Graphical Association Models},
institution = {Institute for Electronic Systems, Department of Mathematics and Computer Science, University of Aalborg},
      issue = {R-88-21},
       year = 1988,
}
@techreport{Lauritzen88a,
     author = {Steffen L. Lauritzen and A.P. Dawid and B.N. Larsen and H.-G. Leimer},
      title = {Independence properties of directed {M}arkov fields},
institution = {Institute for Electronic Systems, Department of Mathematics and Computer Science, University of Aalborg},
      issue = {R-88-32},
       year = 1988,
}
@article{LauritzenJASA-92,
     author = {Steffen L. Lauritzen},
      title = {Propagation of Probabilities, Means, and Variances in Mixed Graphical Association Models},
    journal = {Journal of the American Statistical Association},
     volume = 87,
      issue = 420,
      pages = {1098-1108},
       year = 1992,
}
@article{LauritzenandJensenSC-01,
     author = {Steffen L. Lauritzen and Finn Jensen},
      title = {Stable local computation with conditional {G}aussian distributions},
    journal = {Statistics and Computing}, 
     volume = 11, 
      pages = {191-203}, 
       year = 2001,
}
@article{LauritzenandSpiegelhalterJRSS-88,
   author =  {Steffen L. Lauritzen and David J. Spiegelhalter},
   title =   {Local computations with probabilities on graphical structures and
              their application to expert systems},
   journal = JRSS,
   volume =  50,
    issue =  2,
   pages =   {157-194},
   year =    1988,
}        
@article{LavenexandAmaralHIPPOCAMPUS-00,
       author = {Lavenex, Pierre and Amaral, David G.},
        title = {Hippocampal-neocortical interaction: A hierarchy of associativity},
      journal = {Hippocampus},
       volume = {10},
       number = {4},
         year = {2000},
        pages = {420-430},
     abstract = {The structures forming the medial temporal lobe appear to be necessary for the establishment of long-term declarative memory. In particular, they may be involved in the "consolidation" of information in higher-order associational cortices, perhaps through feedback projections. This review highlights the fact that the medial temporal lobe is organized as a hierarchy of associational networks. Indeed, associational connections within the perirhinal, parahippocampal, and entorhinal cortices enables a significant amount of integration of unimodal and polymodal inputs, so that only highly integrated information reaches the remainder of the hippocampal formation. The feedback efferent projections from the perirhinal and parahippocampal cortices to the neocortex largely reciprocate the afferent projections from the neocortex to these areas. There are, however, noticeable differences in the degree of reciprocity of connections between the perirhinal and parahippocampal cortices and certain areas of the neocortex, in particular in the frontal and temporal lobes. These observations are particularly important for models of hippocampal-neocortical interaction and long-term storage of information in the neocortex. Furthermore, recent functional studies suggest that the perirhinal and parahippocampal cortices are more than interfaces for communication between the neocortex and the hippocampal formation. These structures participate actively in memory processes, but the precise role they play in the service of memory or other cognitive functions is currently unclear.},
}
@book{LawlerTSP,
   author =    {Lawler, E.L. and Lenstra, J.K. and Rinnooy Kan, A.H.G. and Shmoys, D.B.},
   title =     {The Travelling Salesman Problem},
   publisher = {Wiley},
   address =   {New York},
   year =      1985,
}
@article{LawlorandZuckerCoRR-14,
       author = {Matthew Lawlor, Steven Zucker},
        title = {An Online Algorithm for Learning Selectivity to Mixture Means},
      journal = {CoRR},
       volume = {arXiv:1410.8580},
         year = {2014},
     abstract = {We develop a biologically-plausible learning rule called Triplet BCM that provably converges to the class means of general mixture models. This rule generalizes the classical BCM neural rule, and provides a novel interpretation of classical BCM as performing a kind of tensor decomposition. It achieves a substantial generalization over classical BCM by incorporating triplets of samples from the mixtures, which provides a novel information processing interpretation to spike-timing-dependent plasticity. We provide complete proofs of convergence of this learning rule, and an extended discussion of the connection between BCM and tensor learning.}
}
@incollection{LawlorandZuckerHDNIPS-13,
    author = {Matthew Lawlor and Steven W. Zucker},
     title = {Tensor decomposition by modified BCM neurons finds mixture means through input triplets},
 booktitle = {Workship on High-Dimensional Statistical Inference in the Brain},
}
@inproceedings{LawlorandZuckerNIPS-13,
       author = {Lawlor, Matthew and Zucker, Steven W.},
        title = {Third-Order Edge Statistics: Contour Continuation, Curvature, and Cortical Connections.},
    booktitle = {NIPS},
       editor = {Burges, Christopher J. C. and Bottou, Leon and Ghahramani, Zoubin and Weinberger, Kilian Q.},
         year = 2013,
        pages = {1763-1771},
     abstract = {Association field models have attempted to explain human contour grouping performance, and to explain the mean frequency of long-range horizontal connections across cortical columns in V1. However, association fields only depend on the pairwise statistics of edges in natural scenes. We develop a spectral test of the sufficiency of pairwise statistics and show there is significant higher order structure. An analysis using a probabilistic spectral embedding reveals curvature-dependent components.},
}
@inproceedings{LawlorandZuckerNIPS-14,
       author = {Lawlor, Matthew and Zucker, Steven W.},
        title = {Feedforward Learning of Mixture Models},
    booktitle = {Advances in Neural Information Processing Systems 27},
         year = 2014,
}
@article{LawrenceandGiles99,
       author = {S. Lawrence and C. L. Giles},
        title = {Accessibility of Information on the Web},
      journal = {Nature},
       volume = 400,
        pages = {107-109},
         year = 1999,
}
@article{LazarevicandObradovicDPB-02,
       author = {Aleksandar Lazarevic and Zoran Obradovicp},
        title = {Boosting Algorithms for Parallel and Distributed Learning},
      journal = {Distributed and Parallel Databases},
       volume = 11,
       number = 2,
         year = 2002,
        pages = {203-229},
}
@inproceedings{LazaridouetalNAACL-15,
        title = {Combining Language and Vision with a Multimodal Skip-gram Model},
       author = {Angeliki Lazaridou and Nghia The Pham and Marco Baroni},
    booktitle = {Proceedings of the North American Chapter of the Association for Computational Linguistics - Human Language Technologies},
         year = 2015,
    publisher = {Association for Computational Linguistics},
     abstract = {We extend the SKIP-GRAM model of Mikolov et al. (2013a) by taking visual information into account. Like SKIP-GRAM, our multimodal models (MMSKIP-GRAM) build vector-based word representations by learning to predict linguistic contexts in text corpora. However, for a restricted set of words, the models are also exposed to visual representations of the objects they denote (extracted from natural images), and must predict linguistic and visual features jointly. The MMSKIP-GRAM models achieve good performance on a variety of semantic benchmarks. Moreover, since they propagate visual information to all words, we use them to improve image labeling and retrieval in the zero-shot setup, where the test concepts are never seen during model training. Finally, the MMSKIP-GRAM models discover intriguing visual properties of abstract words, paving the way to realistic implementations of embodied theories of meaning.}
}        
@article{Lazaro-GredillaetalCoRR-18,
       author = {Miguel L{\'{a}}zaro{-}Gredilla and Dianhuan Lin and J. Swaroop Guntupalli and Dileep George},
        title = {Beyond imitation: Zero-shot task transfer on robots by learning concepts as cognitive programs},
      journal = {CoRR},
       volume = {arXiv:1812.02788},
         year = {2018},
     abstract = {Humans can infer concepts from image pairs and apply those in the physical world in a completely different setting, enabling tasks like IKEA assembly from diagrams. If robots could represent and infer high-level concepts, it would significantly improve their ability to understand our intent and to transfer tasks between different environments. To that end, we introduce a computational framework that replicates aspects of human concept learning. Concepts are represented as programs on a novel computer architecture consisting of a visual perception system, working memory, and action controller. The instruction set of this "cognitive computer" has commands for parsing a visual scene, directing gaze and attention, imagining new objects, manipulating the contents of a visual working memory, and controlling arm movement. Inferring a concept corresponds to inducing a program that can transform the input to the output. Some concepts require the use of imagination and recursion. Previously learned concepts simplify the learning of subsequent more elaborate concepts, and create a hierarchy of abstractions. We demonstrate how a robot can use these abstractions to interpret novel concepts presented to it as schematic images, and then apply those concepts in dramatically different situations. By bringing cognitive science ideas on mental imagery, perceptual symbols, embodied cognition, and deictic mechanisms into the realm of machine learning, our work brings us closer to the goal of building robots that have interpretable representations and commonsense.}
}
@inproceedings{LazebniketalBMVC-04,
    author = {S. Lazebnik and C. Schmid and J. Ponce},
     title = {Semi-Local Affine Parts for Object Recognition},
 booktitle = {Proceedings of the {B}ritish Machine Vision Conference, ({BMVC})},
    volume = 2,
     pages = {779-788},
      year = 2004,
}        
@inproceedings{LazebniketalCVPR-06,
    author = {S. Lazebnik and C. Schmid and J. Ponce},
     title = {Beyond Bags of Features: Spatial Pyramid Matching 
              for Recognizing Natural Scene Categories},
 booktitle = {Proceedings of the 2006 IEEE Computer Society 
              Conference on Computer Vision and Pattern Recognition},
 publisher = {IEEE Computer Society},
   address = {Washington, DC, USA},
    Volume = 2,
     pages = {2161-2168},
      year = 2006,
}
@book{Le2012,
     title = {Orthogonal Frequency Division Multiplexing with Diversity for Future Wireless Systems},
    author = {Le, K.N.},
      year = {2012},
 publisher = {Bentham Science Publishers}
}
@book{LeCamandNeyman67,
       editor = {LeCam, L. M. and Neyman, J.},
        title = {Proceedings of the Fifth Berkeley Symposium on Mathematics, Statistics, and Probablity},
    publisher = {University of California Press},
      address = {Berkeley, California},
         year = 1967,
}
@article{LeCunBengioHintonNATURE-15,
       author = {LeCun, Yann and Bengio, Yoshua and Hinton, Geoffrey},
        title = {Deep learning},
      journal = {Nature},
         year = 2015,
    publisher = {Nature Publishing Group},
       volume = 521,
        issue = 7553,
        pages = {436-444},
     abstract = {Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.},
}
@incollection{LeCunandBengio95,
       author = {Yann LeCun and Yoshua Bengio},
        title = {Convolutional networks for images, speech, and time-series},
    booktitle = {The Handbook of Brain Theory and Neural Networks},
       editor = {Michael Arbib},
    publisher = {Bradford Books/MIT Press},
         year = 1995,
}
@incollection{LeCunetal06,
    author = {Y. LeCun and S. Chopra and M. Ranzato and F. J. Huang},
     title = {A Tutorial on Energy-Based Learning},
 booktitle = {Predicting Structured Outputs},
    editor = {G. Bakir and T. Hofmann and B. Sch\"{o}lkopf and A. Smola and B. Taskar},
   address = {Cambridge, MA},
 publisher = {MIT Press},
      year = 2006,
}
@inproceedings{LeCunetal98,
    author = {Yann LeCun and L\'{e}on Bottou and Genevieve Orr and Klause-Robert M\"{u}ller},
     title = {Efficient BackProp},
 booktitle = {Neural Networks: Tricks of the trade},
    editor = {Genevieve Orr and Klause-Robert M\"{u}ller},
 publisher = {Springer},
      year = 1998,
}
@inproceedings{LeCunetalCVPR-04,
       author = {LeCun, Yann and Huang, Fu-Jie and Bottou, Leon},
        title = {Learning Methods for Generic Object Recognition with Invariance to Pose and Lighting},
    booktitle = {Proceedings of {IEEE} Computer Vision and Pattern Recognition},
    publisher = {IEEE Press},
         year = 2004,
}
@inproceedings{LeCunetalICAS-10,
        title = {Convolutional networks and applications in vision},
       author = {LeCun, Y. and Kavukcuoglu, K. and Farabet, C.},
    booktitle = {Proceedings of 2010 IEEE International Symposium on Circuits and Systems},
         year = 2010,
        pages = {253-256},
}
@inproceedings{LeCunetalICPR-90,
       author = {LeCun, Y. and Matan, O. and Boser, B. and Denker, J. S. and Henderson, D. and Howard, R. E. and Hubbard, W. and Jackel, L. D. and Baird, H. S.},
        title = {Handwritten Zip Code Recognition with Multilayer Networks},
    booktitle = {Proceedings of the International Conference on Pattern Recognition},
     location = {Atlantic City},
    publisher = {IEEE Press},
        pages = {35-40},
       volume = {{II}},
         year = 1990,
}
@article{LeCunetalNATURE-15,
       author = {LeCun, Yann and Bengio, Yoshua and Hinton, Geoffrey},
        title = {Deep learning},
      journal = {Nature},
       volume = {521},
        issue = {7553},
         year = {2015},
        pages = {436-444},
     abstract = {Deep learning allows computational models that are composed of multiple processing layers to learn representations of data with multiple levels of abstraction. These methods have dramatically improved the state-of-the-art in speech recognition, visual object recognition, object detection and many other domains such as drug discovery and genomics. Deep learning discovers intricate structure in large data sets by using the backpropagation algorithm to indicate how a machine should change its internal parameters that are used to compute the representation in each layer from the representation in the previous layer. Deep convolutional nets have brought about breakthroughs in processing images, video, speech and audio, whereas recurrent nets have shone light on sequential data such as text and speech.},
}
@article{LePaineCoRR-16,
       author = {Tom Le Paine and Pooya Khorrami and Shiyu Chang and Yang Zhang and Prajit Ramachandran and Mark A. Hasegawa{-}Johnson and Thomas S. Huang},
        title = {Fast Wavenet Generation Algorithm},
      journal = {CoRR},
       volume = {arXiv:1611.09482},
         year = {2016},
     abstract = {This paper presents an efficient implementation of the Wavenet generation process called Fast Wavenet. Compared to a naive implementation that has complexity O(2^L) (L denotes the number of layers in the network), our proposed approach removes redundant convolution operations by caching previous calculations, thereby reducing the complexity to O(L) time. Timing experiments show significant advantages of our fast implementation over a naive one. While this method is presented for Wavenet, the same scheme can be applied anytime one wants to perform autoregressive generation or online prediction using a model with dilated convolution layers. The code for our method is publicly available.}
}
@article{LeandMikolovCoRR-14,
       author = {Quoc Le and Tom\`{a}s Mikolov},
        title = {Distributed Representations of Sentences and Documents},
      journal = {CoRR},
       volume = {arXiv:1405.4053v2},
          url = {http://arxiv.org/pdf/1405.4053v2.pdf},
         year = 2014,
     abstract = {Many machine learning algorithms require the input to be represented as a fixed length feature vector. When it comes to texts, one of the most common representations is bag-of-words. Despite their popularity, bag-of-words models have two major weaknesses: they lose the ordering of the words and they also ignore semantics of the words. For example, "powerful," "strong" and "Paris" are equally distant. In this paper, we propose an unsupervised algorithm that learns vector representations of sentences and text documents. This algorithm represents each document by a dense vector which is trained to predict words in the document. Its construction gives our algorithm the potential to overcome the weaknesses of bag-of-words models. Empirical results show that our technique outperforms bagof-words models as well as other techniques for text representations. Finally, we achieve new state-of-the-art results on several text classification and sentiment analysis tasks.}
}
@inproceedings{LeandMikolovICML-14,
       author = {Le, Quoc and Mikolov, Tomas},
        title = {Distributed Representations of Sentences and Documents},
    booktitle = {Proceedings of the 31st International Conference on International Conference on Machine Learning - Volume 32},
    publisher = {JMLR.org},
     location = {Beijing, China},
         year = {2014},
        pages = {II-1188--II-1196},
     abstract = {Many machine learning algorithms require the input to be represented as a fixed-length feature vector. When it comes to texts, one of the most common fixed-length features is bag-of-words. Despite their popularity, bag-of-words features have two major weaknesses: they lose the ordering of the words and they also ignore semantics of the words. For example, "powerful," "strong" and "Paris" are equally distant. In this paper, we propose Paragraph Vector, an unsupervised algorithm that learns fixed-length feature representations from variable-length pieces of texts, such as sentences, paragraphs, and documents. Our algorithm represents each document by a dense vector which is trained to predict words in the document. Its construction gives our algorithm the potential to overcome the weaknesses of bag-of-words models. Empirical results show that Paragraph Vectors outperforms bag-of-words models as well as other techniques for text representations. Finally, we achieve new state-of-the-art results on several text classification and sentiment analysis tasks.}
}
@inproceedings{LebiereandAndersonCSS-93,
        title = {A Connectionist Implementation of the ACT-R Production System},
       author = {Lebiere, Christian and Anderson, John},
    booktitle = {Proceedings of the Fifteenth Annual Conference of the Cognitive Science Society},
    publisher = {Cognitive Science Society},
         year = {1993},
        pages = {635-640},
}
@article{LecoqetalNATURE-NEUROSCIENCE-14,
       author = {Lecoq, J. and Savall, J. and Vucinic, D. and Grewe, B. F. and Kim, H. and Li, J. Z. and Kitch, L. J. and Schnitzer, M. J.},
        title = {{{V}isualizing mammalian brain area interactions by dual-axis two-photon calcium imaging}},
      journal = {Nature Neuroscience},
         year = {2014},
       volume = {17},
       number = {12},
        pages = {1825-1829},
     abstract = {Fluorescence Ca(2+) imaging enables large-scale recordings of neural activity, but collective dynamics across mammalian brain regions are generally inaccessible within single fields of view. Here we introduce a two-photon microscope possessing two articulated arms that can simultaneously image two brain areas (\approx{}0.38 mm(2) each), either nearby or distal, using microendoscopes. Concurrent Ca(2+) imaging of \approx{}100-300 neurons in primary visual cortex (V1) and lateromedial (LM) visual area in behaving mice revealed that the variability in LM neurons' visual responses was strongly dependent on that in V1, suggesting that fluctuations in sensory responses propagate through extended cortical networks.}
}
@article{LecoqetalNEUROSCIENE-09,
     title = {Odor-evoked oxygen consumption by action potential and synaptic transmission in the olfactory bulb},
    author = {Lecoq, J. and Tiret, P. and Najac, M. and Shepherd, G.M. and Greer, C.A. and Charpak, S.},
   journal = {Journal Neuroscience},
    volume = 29,
     issue = 5,
      year = 2009,
     pages = {1424-1433},
  abstract = {The relationship between metabolism of neuronal activity, microvascular organization, and blood flow dynamics is critical for interpreting functional brain imaging. Here we used the rat dorsal olfactory bulb as a model to determine in vivo the correlation between action potential propagation, synaptic transmission, oxygen consumption, and capillary density during odor stimulation. We find that capillary lumen occupies approximately 3\% of the glomerular volume, where synaptic transmission occurs, and only 0.1\% of the overlying nerve layer. In glomeruli, odor triggers a local early decrease in tissue oxygen partial pressure that results principally from dendritic activation rather than from firing of axon terminals, transmitter release or astrocyte activation. In the nerve layer, action potential propagation does not generate local changes in tissue oxygen partial pressure. We conclude that capillary density is tightly correlated with the oxidative metabolism of synaptic transmission, and suggest that action potential propagation operates mainly anaerobically.}
}
@article {LecunetalIEEE-98,
       author = {LeCun, Y. and Bottou, L. and Bengio, Y. and Haffner, P.},
        title = {Gradient-Based Learning Applied to Document Recognition},
      journal = {Proceedings of the IEEE},
       volume = {86},
        issue = {11},
         year = 1998,
        pages = {2278-2324},
}
@inproceedings{LecunetalISCAS-10,
      title = {Convolutional Networks and Applications in Vision},
     author = {LeCun, Yann and Kavukcuoglu, Koray and Farabet, {Cl\'ement}}, 
  booktitle = {Proceedings of International Symposium on Circuits and Systems (ISCAS'10)}, 
  publisher = {IEEE}, 
       year = 2010,
}
@techreport{LeeMSc,
       author = {Lee, Jin Joo},
        title = {Localization With Kalman Filtering},
 type         = {{M.Sc.} Thesis},
  institution = {Brown University},
         year = 1990,
}
@inproceedings{LeeandGraumanBMVC-08,
     title = {Foreground Focus: Finding Meaningful Features in Unlabeled Images},
    author = {Yong Jae Lee and Kristen Grauman},
 booktitle = {Proceedings of the British Machine Vision Conference},
      year = {2008},
}
@article{LeeandMumfordJOSA-03,
       author = {Lee, Tai Sing and Mumford, David},
        title = {Hierarchical {B}ayesian Inference in the Visual Cortex},
      journal = {Journal of the Optical Society of America},
    publisher = {Optical Society of America},
       volume = 2,
        issue = 7,
        pages = {1434-1448},
         year = 2003,
}
@article{LeeandReeveCABN-17,
       author = {Lee, Woogul and Reeve, Johnmarshall},
        title = {Identifying the neural substrates of intrinsic motivation during task performance},
      journal = {Cognitive, Affective, {\&} Behavioral Neuroscience},
       volume = {17},
       number = {5},
         year = {2017},
        pages = {939-953},
     abstract = {Intrinsic motivation is the inherent tendency to seek out novelty and challenge, to explore and investigate, and to stretch and extend one's capacities. When people imagine performing intrinsically motivating tasks, they show heightened anterior insular cortex (AIC) activity. To fully explain the neural system of intrinsic motivation, however, requires assessing neural activity while people actually perform intrinsically motivating tasks (i.e., while answering curiosity-inducing questions or solving competence-enabling anagrams). Using event-related functional magnetic resonance imaging, we found that the neural system of intrinsic motivation involves not only AIC activity, but also striatum activity and, further, AIC--striatum functional interactions. These findings suggest that subjective feelings of intrinsic satisfaction (associated with AIC activations), reward processing (associated with striatum activations), and their interactions underlie the actual experience of intrinsic motivation. These neural findings are consistent with the conceptualization of intrinsic motivation as the pursuit and satisfaction of subjective feelings (interest and enjoyment) as intrinsic rewards.},
}
@inproceedings{LeeandSeungNIPS-00,
       author = {Lee, Daniel D. and Seung, H. Sebastian},
        title = {Algorithms for Non-negative Matrix Factorization},
    booktitle = {Proceedings of the 13th International Conference on Neural Information Processing Systems},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
     location = {Denver, CO},
         year = {2000},
        pages = {535-541},
     abstract = {Non-negative matrix factorization (NMF) has previously been shown to be a useful decomposition for multivariate data. Two different multiplicative algorithms for NMF are analyzed. They differ only slightly in the multiplicative factor used in the update rules. One algorithm can be shown to minimize the conventional least squares error while the other minimizes the generalized Kullback-Leibler divergence. The monotonic convergence of both algorithms can be proven using an auxiliary function analogous to that used for proving convergence of the Expectation-Maximization algorithm. The algorithms can also be interpreted as diagonally rescaled gradient descent, where the rescaling factor is optimally chosen to ensure convergence.}
}
@article{LeeandSpelkeEBR-10,
       author = {Lee, S. A. and Spelke, E. S.},
        title = {Two systems of spatial representation underlying navigation},
      journal = {Experimental Brain Research},
         year = {2010},
       number = {2},
       volume = {206},
        pages = {179-188},
     abstract = {We review evidence for two distinct cognitive processes by which humans and animals represent the navigable environment. One process uses the shape of the extended 3D surface layout to specify the navigator's position and orientation. A second process uses objects and patterns as beacons to specify the locations of significant objects. Although much of the evidence for these processes comes from neurophysiological studies of navigating animals and neuroimaging studies of human adults, behavioral studies of navigating children shed light both on the nature of these systems and on their interactions.}
}
@article{LeeandYannakakisIEEE-94,
       author = {Lee, David and Yannakakis, Mihalis},
        title = {Testing Finite State Machines: State Identification and Verification},
      journal = {{IEEE} Transactions on Computers},
       volume = 43,
        issue = 3,
         year = 1994,
}
@inproceedings{LeeandYannakakisSTOC-92,
       author = {Lee, David and Yannakakis, Mihalis},
        title = {Online Minimization of Transition Systems},
    booktitle = {Proceedings of 24th Annual {ACM} Symposium on the Theory of Computing},
     location = {Victoria, Canada},
         year = 1992,
}
@article {LeeetalBIORXIV-18,
       author = {Lee, Henry H and Kalhor, Reza and Goela, Naveen and Bolot, Jean and Church, George M},
        title = {Enzymatic DNA synthesis for digital information storage},
      journal = {bioRxiv},
         year = {2018},
    publisher = {Cold Spring Harbor Laboratory},
     abstract = {DNA is an emerging storage medium for digital data but its adoption is hampered by limitations of phosphoramidite chemistry, which was developed for single-base accuracy required for biological functionality. Here, we establish a de novo enzymatic DNA synthesis strategy designed from the bottom-up for information storage. We harness a template-independent DNA polymerase for controlled synthesis of sequences with user-defined information content. We demonstrate retrieval of 144-bits, including addressing, from perfectly synthesized DNA strands using batch-processed Illumina and real-time Oxford Nanopore sequencing. We then develop a codec for data retrieval from populations of diverse but imperfectly synthesized DNA strands, each with a ~30\% error tolerance. With this codec, we experimentally validate a kilobyte-scale design which stores 1 bit per nucleotide. Simulations of the codec support reliable and robust storage of information for large-scale systems. This work paves the way for alternative synthesis and sequencing strategies to advance information storage in DNA.},
}
@incollection{LeeetalCN-97,
        title = {The Role of {V1} in Shape Representation},
       author = {Tai Sing Lee and David Mumford and Song Chun Zhu and Victor Lamme},
    booktitle = {Computational Neuroscience},
       editor = {Bower},
    publisher = {Plenum Press},
      address = {New York},
         year = 1997,
        pages = {697-703},
     abstract = {The analysis of higher level global visual attributes such as geometric shapes is thought to be a function of specialist modules in the extrastriate cortices. However, V1 is unique because only in this area can one find cells tuned to orientation and other local features with the spatial precision needed for representing high resolution and geometric aspects of an image. Geometric computation that demands such resolution may need to involve V1 via feedback. VI neurons are known to exhibit very different types of responses in the short latency (40-80 milliseconds post-stimulus time window) versus longer latencies (80-200 milliseconds). Here we report that the later part of V1 neurons' responses is sensitive to geometric attributes of globally defined shapes. We propose a computational interpretation of these results.}
}
@article{LeeetalCOGNITION-13,
       author = {Lee, Eun-Kyung and Brown-Schmidt, Sarah and Watson, Duane G.},
        title = {Ways of looking ahead: {H}ierarchical planning in language production},
      journal = {Cognition},
       volume = {129},
        issue = {3},
         year = {2013},
        pages = {544-562},
     abstract = {It is generally assumed that language production proceeds incrementally, with chunks of linguistic structure planned ahead of speech. Extensive research has examined the scope of language production and suggests that the size of planned chunks varies across contexts (Ferreira \& Swets, 2002; Wagner \& Jescheniak, 2010). By contrast, relatively little is known about the structure of advance planning, specifically whether planning proceeds incrementally according to the surface structure of the utterance, or whether speakers plan according to the hierarchical relationships between utterance elements. In two experiments, we examine the structure and scope of lexical planning in language production using a picture description task. Analyses of speech onset times and word durations show that speakers engage in hierarchical planning such that structurally dependent lexical items are planned together and that hierarchical planning occurs for both direct and indirect dependencies.},
}
@article{LeeetalCOIN-11,
       author = {Lee, Wei-Chung Allen and Reid, R. Clay},
        title = {Specificity and randomness: structure-function relationships in neural circuits},
      journal = {Current Opinion in Neurobiology},
       volume = 21,
         year = 2011,
        pages = {801-807},
     abstract = {A fundamental but unsolved problem in neuroscience is how connections between neurons might underlie information processing in central circuits. Building wiring diagrams of neural networks may accelerate our understanding of how they compute. But even if we had wiring diagrams, it is critical to know what neurons in a circuit are doing: their physiology. In both the retina and cerebral cortex, a great deal is known about topographic specificity, such as lamination and cell-type specificity of connections. Little, however, is known about connections as they relate to function. Here, we review how advances in functional imaging and electron microscopy have recently allowed the examination of relationships between sensory physiology and synaptic connections in cortical and retinal circuits.},
}
@article{LeeetalCoRR-14,
       author = {Chen-Yu Lee and Saining Xie and Patrick Gallagher and Zhengyou Zhang and Zhuowen Tu},
        title = {Deeply-Supervised Nets},
      journal = {CoRR},
       volume = {arXiv:1409.5185},
          url = {http://arxiv.org/abs/1409.5185},
         year = 2014,
     abstract = {Our proposed deeply-supervised nets (DSN) method simultaneously minimizes classification error while making the learning process of hidden layers direct and transparent. We make an attempt to boost the classification performance by studying a new formulation in deep networks. Three aspects in convolutional neural networks (CNN) style architectures are being looked at: (1) transparency of the intermediate layers to the overall classification; (2) discriminativeness and robustness of learned features, especially in the early layers; (3) effectiveness in training due to the presence of the exploding and vanishing gradients. We introduce "companion objective" to the individual hidden layers, in addition to the overall objective at the output layer (a different strategy to layer-wise pre-training). We extend techniques from stochastic gradient methods to analyze our algorithm. The advantage of our method is evident and our experimental result on benchmark datasets shows significant performance gain over existing methods, e.g. all state-of-the-art results on MNIST, CIFAR-10, CIFAR-100, and SVHN.}
} 
@article{LeeetalCoRR-19,
       author = {Junyeop Lee and Sungrae Park and Jeonghun Baek and Seong Joon Oh and Seonghyeon Kim and Hwalsuk Lee},
        title = {On Recognizing Texts of Arbitrary Shapes with {2-D} Self-Attention},
      journal = {CoRR},
       volume = {arXiv:1910.04396},
         year = {2019},
     abstract = {Scene text recognition (STR) is the task of recognizing character sequences in natural scenes. While there have been great advances in STR methods, current methods still fail to recognize texts in arbitrary shapes, such as heavily curved or rotated texts, which are abundant in daily life (e.g. restaurant signs, product labels, company logos, etc). This paper introduces a novel architecture to recognizing texts of arbitrary shapes, named Self-Attention Text Recognition Network (SATRN), which is inspired by the Transformer. SATRN utilizes the self-attention mechanism to describe two-dimensional (2-D) spatial dependencies of characters in a scene text image. Exploiting the full-graph propagation of self-attention, SATRN can recognize texts with arbitrary arrangements and large inter-character spacing. As a result, SATRN outperforms existing STR models by a large margin of 5.7 pp on average in "irregular text" benchmarks. We provide empirical analyses that illustrate the inner mechanisms and the extent to which the model is applicable (e.g. rotated and multi-line text). We will open-source the code.}
}
@inproceedings{LeeetalICML-09,
    author = {Lee, Honglak and Grosse, Roger and Ranganath, Rajesh and Ng, Andrew Y.},
     title = {Convolutional deep belief networks for scalable unsupervised learning of hierarchical representations},
 booktitle = {ICML '09: Proceedings of the 26th Annual International Conference on Machine Learning},
  location = {Montreal, Quebec, Canada},
 publisher = {ACM},
   address = {New York, NY},
     pages = {609-616},
      year = 2009,
}
@article{LeeetalIJCV-00,
       author = {A.B. Lee and D. Mumford and J. Huang},
        title = {Occlusion Models for Natural Images: A Statistical Study of a Scale-Invariant Dead Leaves Model},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 41,
        issue = {(1/2)},
        pages = {35-59},
         year = 2001,
}
@article{LeeetalJN-13,
    author = {Lee, D. J. and Gurkoff, G. G. and Izadi, A. and Berman, R. F. and Ekstrom, A. D. and Muizelaar, J. P. and Lyeth, B. G. and Shahlaie, K.},
     title = {{M}edial septal nucleus theta frequency deep brain stimulation improves spatial working memory after traumatic brain injury},
   journal = {Journal Neurotrauma},
    volume = 30,
     issue = 2,
      year = 2013,
     pages = {131-139},
  abstract = {More than 5,000,000 survivors of traumatic brain injury (TBI) live with persistent cognitive deficits, some of which likely derive from hippocampal dysfunction. Oscillatory activity in the hippocampus is critical for normal learning and memory functions, and can be modulated using deep brain stimulation techniques. In this pre-clinical study, we demonstrate that lateral fluid percussion TBI results in the attenuation of hippocampal theta oscillations in the first 6 days after injury, which correlate with deficits in the Barnes maze spatial working memory task. Theta band stimulation of the medial septal nucleus (MSN) results in a transient increase in hippocampal theta activity, and when delivered 1 minute prior to training in the Barnes maze, it significantly improves spatial working memory. These results suggest that MSN theta stimulation may be an effective neuromodulatory technique for treatment of persistent learning and memory deficits after TBI.}
}
@article{LeeetalNATURE-16,
       author = {Lee, Wei-Chung Allen and Bonin, Vincent and Reed, Michael and Graham, Brett J. and Hood, Greg and Glattfelder, Katie and Reid, R. Clay},
        title = {Anatomy and function of an excitatory network in the visual cortex},
    publisher = {Nature Publishing Group},
      journal = {Nature},
       volume = {532}, 
         year = {2016},
        pages = {370-374}
}
@article{LeeetalNEJM-92,
       author = {Lee, William M. and Galbraith, Robert M.},
        title = {The Extracellular Actin-Scavenger System and Actin Toxicity},
      journal = {New England Journal of Medicine},
       volume = 326,
        issue = 20,
         year = 1992,
        pages = {1335-1341},
     abstract = {Actin is the most abundant protein in mammalian cells. Cell motility and change in the size and shape of cells depend on the ability of monomers of actin to polymerize to actin filaments. When actin is released from cells, however, its strong tendency to polymerize can become a liability, and the presence of filaments of actin in blood vessels can be fatal. Cell necrosis is associated with the release of actin in a variety of clinical situations, including hepatic necrosis, septic shock, the adult respiratory distress syndrome, and certain disorders of pregnancy. A newly described homeostatic mechanism, termed the actin-scavenger system [...]},
}
@incollection{LeeetalNIPS-06,
      title = {Efficient sparse coding algorithms},
     author = {Honglak Lee and Alexis Battle and Rajat Raina and Andrew Y. Ng},
  booktitle = {Advances in Neural Information Processing Systems 19},
     editor = {B. Sch\"{o}lkopf and J. Platt and T. Hofmann},
  publisher = {MIT Press},
    address = {Cambridge, MA},
      pages = {801-808},
       year = 2007
}
@incollection{LeeetalNIPS-08,
     title = {Sparse deep belief net model for visual area {V2}},
    author = {Honglak Lee and Chaitanya Ekanadham and Andrew Ng},
 booktitle = {Advances in Neural Information Processing Systems 20},
    editor = {J.C. Platt and D. Koller and Y. Singer and S. Roweis},
 publisher = {MIT Press},
   address = {Cambridge, MA},
     pages = {873-880},
      year = {2008}
}
@inproceedings{LeeetalNIPS-16,
       author = {Kisuk Lee and Jonathan Zung and Peter Li and Viren Jain and H. Sebastian Seung},
        title = {Superhuman Accuracy on the SNEMI3D Connectomics Challenge},
    booktitle = {Proceedings of the 30th International Conference on Neural Information Processing Systems},
    publisher = {Curran Associates Inc.},
         year = {2016},
     location = {Barcelona, Spain},
         year = {2017},
     abstract = {For the past decade, convolutional networks have been used for 3D reconstruction of neurons from electron microscopic (EM) brain images. Recent years have seen great improvements in accuracy, as evidenced by submissions to the SNEMI3D benchmark challenge. Here we report the first submission to surpass the estimate of human accuracy provided by the SNEMI3D leaderboard. A variant of 3D U- Net is trained on a primary task of predicting affinities between nearest neighbor voxels, and an auxiliary task of predicting long-range affinities. The training data is augmented by simulated image defects. The nearest neighbor affinities are used to create an oversegmentation, and then supervoxels are greedily agglomerated based on mean affinity. The resulting SNEMI3D score exceeds the estimate of human accuracy by a large margin. While one should be cautious about extrapolating from the SNEMI3D benchmark to real-world accuracy of large-scale neural circuit reconstruction, our submission inspires optimism that the goal of full automation may be realizable in the future.}
}
@article{LeeetalSCIENCE-10,
       author = {Lee, Te-Hao and Bhunia, Swarup and Mehregany, Mehran},
        title = {Electromechanical Computing at 500\hmdeg{}C with Silicon Carbide},
      journal = {Science},
       volume = 329,
        issue = 5997,
         year = 2010,
        pages = {1316-1318},
     abstract = {Logic circuits capable of operating at high temperatures can alleviate expensive heat-sinking and thermal-management requirements of modern electronics and are enabling for advanced propulsion systems. Replacing existing complementary metal-oxide semiconductor field-effect transistors with silicon carbide (SiC) nanoelectromechanical system (NEMS) switches is a promising approach for low-power, high-performance logic operation at temperatures higher than 300\hmdeg{}C, beyond the capability of conventional silicon technology. These switches are capable of achieving virtually zero off-state current, microwave operating frequencies, radiation hardness, and nanoscale dimensions. Here, we report a microfabricated electromechanical inverter with SiC complementary NEMS switches capable of operating at 500\hmdeg{}C with ultralow leakage current.},
}
@article{LeeetalSCIENCE-14,
       author = {Lee, Je H. and Daugharthy, Evan R. and Scheiman, Jonathan and Kalhor, Reza and Yang, Joyce L. and Ferrante, Thomas C. and Terry, Richard and Jeanty, Sauveur S. F. and Li, Chao and Amamoto, Ryoji and Peters, Derek T. and Turczyk, Brian M. and Marblestone, Adam H. and Inverso, Samuel A. and Bernard, Amy and Mali, Prashant and Rios, Xavier and Aach, John and Church, George M.},
      journal = {Science},
       volume = 343,
        issue = 6177,
    publisher = {American Association for the Advancement of Science},
        title = {{Highly Multiplexed Subcellular {RNA} Sequencing in Situ}},
         year = 2014,
        pages = {1360-1363},
     abstract = {{Understanding the spatial organization of gene expression with single-nucleotide resolution requires localizing the sequences of expressed RNA transcripts within a cell in situ. Here, we describe fluorescent in situ RNA sequencing (FISSEQ), in which stably cross-linked complementary DNA (cDNA) amplicons are sequenced within a biological sample. Using 30-base reads from 8102 genes in situ, we examined RNA expression and localization in human primary fibroblasts with a simulated wound-healing assay. FISSEQ is compatible with tissue sections and whole-mount embryos and reduces the limitations of optical resolution and noisy signals on single-molecule detection. Our platform enables massively parallel detection of genetic elements, including gene transcripts and molecular barcodes, and can be used to investigate cellular phenotype, gene regulation, and environment in situ.}},
}
@article{LeeetalSIGARCH-10,
     author = {Lee, Victor W. and Kim, Changkyu and Chhugani, Jatin and Deisher, Michael and Kim, Daehyun and Nguyen, Anthony D. and Satish, Nadathur and Smelyanskiy, Mikhail and Chennupaty, Srinivas and Hammarlund, Per and Singhal, Ronak and Dubey, Pradeep},
      title = {Debunking the {100X} {GPU} versus {CPU} myth: an evaluation of throughput computing on {CPU} and {GPU}},
    journal = {SIGARCH Computer Architecture News},
     volume = 38,
      issue = 3,
      pages = {451-460},
  publisher = {ACM},
    address = {New York, NY, USA},
       year = 2010,
}
@article{LeeetalVR-98,
       author = {T.S. Lee and D. Mumford and R.Romero and V.Lamme},
        title = {The Role of Primary Visual Cortex in Higher Level Vision},
      journal = {Vision Research},
       volume = 38,
         year = 1998,
        pages = {2429-2454},
     abstract = {In the classical feed-forward, modular view of visual processing, the primary visual cortex (area V1) is a module that serves to extract local features such as edges and bars. Representation and recognition of objects are thought to be functions of higher extrastriate cortical areas. This paper presents neurophysiological data that show the later part of V1 neurons responses reflecting higher order perceptual computations related to Ullman's (Cognition, 1984) visual routines and Marr's (Vision, 1982) full primal sketch, 2 1/2-D sketch and {3D} model. Based on theoretical reasoning and the experimental evidence, we propose a possible reinterpretation of the functional role of V1. In this framework, because of V1 neurons precise encoding of orientation and spatial information, higher level perceptual computations and representations that involve high resolution details, fine geometry and spatial precision would necessarily involve V1 and be reflected in the later part of its neurons activities.}
}
@article{LeekandJagerARSA-17,
       author = {Jeffrey Leek and Leah R. Jager},
        title = {Is Most Published Research Really False?},
      journal = {Annual Review of Statistics and Its Application},
       volume = {4},
         year = {2017},
        pages = {109-122},
     abstract = {There has been an increasing concern in both the scientific and lay communities that most published medical findings are false. But what does it mean to be false? Here we describe the range of definitions of false discoveries in the scientific literature. We summarize the philosophical, statistical, and experimental evidence for each type of false discovery. We discuss common underpinning problems with the scientific and data analytic practices and point to tools and behaviors that can be implemented to reduce the problems with published scientific results.}
}
@inproceedings{LeesonCIT-11,
        title = {Communicating at the nanoscale},
       author = {Leeson, Mark S.},
    booktitle = {Proceedings of the 5th WSEAS International Conference on Communications and information technology},
         year = 2011,
     location = {Corfu Island, Greece},
        pages = {288-293},
    publisher = {World Scientific and Engineering Academy and Society (WSEAS)},
} 
@inproceedings{LeetalCVPR-11,
     author = {Q.V. Le and W.Y. Zou and S.Y. Yeung and A.Y. Ng},
      title = {Learning hierarchical spatio-temporal features for action recognition with independent subspace analysis},
  booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
       year = 2011,
}
@article{LeetalCoRR-11,
       author = {Quoc V. Le and Rajat Monga and Matthieu Devin and Greg Corrado and Kai Chen and Marc'Aurelio Ranzato and Jeffrey Dean and Andrew Y. Ng},
        title = {Building high-level features using large scale unsupervised learning},
      journal = {CoRR},
       volume = {arXiv:1112.6209},
         year = {2011},
     abstract = {We consider the problem of building high-level, class-specific feature detectors from only unlabeled data. For example, is it possible to learn a face detector using only unlabeled images? To answer this, we train a 9-layered locally connected sparse autoencoder with pooling and local contrast normalization on a large dataset of images (the model has 1 billion connections, the dataset has 10 million 200x200 pixel images downloaded from the Internet). We train this network using model parallelism and asynchronous SGD on a cluster with 1,000 machines (16,000 cores) for three days. Contrary to what appears to be a widely-held intuition, our experimental results reveal that it is possible to train a face detector without having to label images as containing a face or not. Control experiments show that this feature detector is robust not only to translation but also to scaling and out-of-plane rotation. We also find that the same network is sensitive to other high-level concepts such as cat faces and human bodies. Starting with these learned features, we trained our network to obtain 15.8\% accuracy in recognizing 20,000 object categories from ImageNet, a leap of 70\% relative improvement over the previous state-of-the-art.}
}
@article{LeetalCoRR-18,
       author = {Hoang M. Le and Nan Jiang and Alekh Agarwal and Miroslav Dud{\'{\i}}k and Yisong Yue and Hal Daum{\'{e}} III},
        title = {Hierarchical Imitation and Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1803.00590},
         year = {2018},
     abstract = {We study how to effectively leverage expert feedback to learn sequential decision-making policies. We focus on problems with sparse rewards and long time horizons, which typically pose significant challenges in reinforcement learning. We propose an algorithmic framework, called hierarchical guidance, that leverages the hierarchical structure of the underlying problem to integrate different modes of expert interaction. Our framework can incorporate different combinations of imitation learning (IL) and reinforcement learning (RL) at different levels, leading to dramatic reductions in both expert effort and cost of exploration. Using long-horizon benchmarks, including Montezuma's Revenge, we demonstrate that our approach can learn significantly faster than hierarchical RL, and be significantly more label-efficient than standard IL. We also theoretically analyze labeling cost for certain instantiations of our framework.},
}
@inproceedings{LeetalICML-12,
       author = {Quoc Le and Marc'Aurelio Ranzato and Rajat Monga and Matthieu Devin and Kai Chen and Greg Corrado and Jeff Dean and Andrew Ng},
        title = {Building high-level features using large scale unsupervised learning},
    booktitle = {Proceedings of the 29th International Conference on Machine Learning},
       editor = {John Langford and Joelle Pineau},
     location = {Edinburgh, Scotland, GB},
         year = 2012,
        pages = {81-88},
}
@inproceedings{LehnertetalRLDM-18,
       author = {Lucas Lehnert and Michael Littman and Michael Frank},
    booktitle = {Proceedings of The 4th Multidisciplinary Conference on Reinforcement Learning and Decision Making},
        title = {Model-based Knowledge Representations},
         year = {2018},
     abstract = {One question central to reinforcement learning is which representations - including aspects of the state space, transition function and reward function - can be generalized or re-used across different tasks. Humans are adept at such flexible transfer but existing RL algorithms are much more limited. This paper presents model features, a latent representation that compresses the state space of a control problem by exploiting which states are equivalent in terms of both transition and reward functions. Because model features only extract these equivalences from the transition and reward functions, but are not tied to the functions themselves, this latent state representation generalizes across tasks that change in both their transition and reward functions. While transferring successor features between different tasks has been shown to improve learning speed, this representation is overly specific and hence needs to be re-learned when the optimal policy or transition function change. Model features link successor features to model reductions, facilitating the design of gradient-based optimization algorithms to approximate model reductions directly from transition data. Learning model features is akin to model-based RL, because from the learned model a linear action model can be extracted to predict future reward outcomes. This paper summarizes theoretical results from our extended paper and presents empirical simulation results showing that model features which minimize prediction errors on future reward outcomes serve as useful state abstractions that afford generalization across tasks that differ in both transition and reward functions. Because model features construct a latent state representation that supports predictions of future reward outcomes, the presented results motivate further experiments to investigate if humans or animals learn such a representation, and whether neural systems involved in state representation reflect the equivalence abstraction.}
}
@book{Lehrer2009,
       author = {Jonah Lehrer},
        title = {How We Decide},
    publisher = {Houghton Mifflin},
      address = {Boston, MA},
         year = 2009,
      comment = {http://fora.tv/2010/01/05/Jonah_Lehrer_How_We_Decide},
}
@inproceedings{LeietalACL-13,
       author = {Tao Lei, Fan Long, Regina Barzilay and Martin Rinard},
        title = {From Natural Language Specifications to Program Input Parsers},
    booktitle = {The 51st Annual Meeting of the Association for Computational Linguistics},
     location = {Sofia, Bulgaria},
         year = {2013},
     abstract = {We present a method for automatically generating input parsers from English specifications of input file formats. We use a Bayesian generative model to capture relevant natural language phenomena and translate the English specification into a specification tree, which is then translated into a C++ input parser. We model the problem as a joint dependency parsing and semantic role labeling task. Our method is based on two sources of information: (1) the correlation between the text and the specification tree and (2) noisy supervision as determined by the success of the generated C++ parser in reading input examples. Our results show that our approach achieves 80.0\% F-Score accuracy compared to an F-Score of 66.7\% produced by a state-of-the-art semantic parser on a dataset of input format speci- fications from the ACM International Collegiate Programming Contest (which were written in English for humans with no intention of providing support for automated processing).}
}
@article {LeinengaandGotzSCIENCE-TRANSLATIONAL-MEDICINE-15,
       author = {Leinenga, Gerhard and G{\"o}tz, J{\"u}rgen},
        title = {Scanning ultrasound removes amyloid-\hmBeta{} and restores memory in an Alzheimer{\textquoteright}s disease mouse model},
      journal = {Science Translational Medicine},
    publisher = {American Association for the Advancement of Science},
       volume = {7},
       number = {278},
         year = {2015},
     abstract = {Amyloid-\hmBeta{} (A\hmBeta{}) peptide has been implicated in the pathogenesis of Alzheimer{\textquoteright}s disease (AD). We present a nonpharmacological approach for removing A\hmBeta{} and restoring memory function in a mouse model of AD in which A\hmBeta{} is deposited in the brain. We used repeated scanning ultrasound (SUS) treatments of the mouse brain to remove A\hmBeta{}, without the need for any additional therapeutic agent such as anti-A\hmBeta{} antibody. Spinning disk confocal microscopy and high-resolution three-dimensional reconstruction revealed extensive internalization of A\hmBeta{} into the lysosomes of activated microglia in mouse brains subjected to SUS, with no concomitant increase observed in the number of microglia. Plaque burden was reduced in SUS-treated AD mice compared to sham-treated animals, and cleared plaques were observed in 75\% of SUS-treated mice. Treated AD mice also displayed improved performance on three memory tasks: the Y-maze, the novel object recognition test, and the active place avoidance task. Our findings suggest that repeated SUS is useful for removing A\hmBeta{} in the mouse brain without causing overt damage, and should be explored further as a noninvasive method with therapeutic potential in AD.},
}
@article{LeismanetalFiPH-16,
       author = {Leisman, Gerry and Moustafa, Ahmed A. and Shafir, Tal},
        title = {Thinking, Walking, Talking: Integratory Motor and Cognitive Brain Function},
      journal = {Frontiers in Public Health},
         year = {2016},
    publisher = {Frontiers Media S.A.},
       volume = {4},
        pages = {94},
     abstract = {In this article, we argue that motor and cognitive processes are functionally related and most likely share a similar evolutionary history. This is supported by clinical and neural data showing that some brain regions integrate both motor and cognitive functions. In addition, we also argue that cognitive processes coincide with complex motor output. Further, we also review data that support the converse notion that motor processes can contribute to cognitive function, as found by many rehabilitation and aerobic exercise training programs. Support is provided for motor and cognitive processes possessing dynamic bidirectional influences on each other.}
}
@article{LeismanetalFiSN-14,
       author = {Leisman, G. and Braun-Benjamin, O. and Melillo, R.},
        title = {Cognitive-motor interactions of the basal ganglia in development},
      journal = {Frontiers in Systems Neuroscience},
         year = {2014},
       volume = {8},
        pages = {16},
     abstract = {Neural circuits linking activity in anatomically segregated populations of neurons in subcortical structures and the neocortex throughout the human brain regulate complex behaviors such as walking, talking, language comprehension, and other cognitive functions associated with frontal lobes. The basal ganglia, which regulate motor control, are also crucial elements in the circuits that confer human reasoning and adaptive function. The basal ganglia are key elements in the control of reward-based learning, sequencing, discrete elements that constitute a complete motor act, and cognitive function. Imaging studies of intact human subjects and electrophysiologic and tracer studies of the brains and behavior of other species confirm these findings. We know that the relation between the basal ganglia and the cerebral cortical region allows for connections organized into discrete circuits. Rather than serving as a means for widespread cortical areas to gain access to the motor system, these loops reciprocally interconnect a large and diverse set of cerebral cortical areas with the basal ganglia. Neuronal activity within the basal ganglia associated with motor areas of the cerebral cortex is highly correlated with parameters of movement. Neuronal activity within the basal ganglia and cerebellar loops associated with the prefrontal cortex is related to the aspects of cognitive function. Thus, individual loops appear to be involved in distinct behavioral functions. Damage to the basal ganglia of circuits with motor areas of the cortex leads to motor symptoms, whereas damage to the subcortical components of circuits with non-motor areas of the cortex causes higher-order deficits. In this report, we review some of the anatomic, physiologic, and behavioral findings that have contributed to a reappraisal of function concerning the basal ganglia and cerebellar loops with the cerebral cortex and apply it in clinical applications to attention deficit/hyperactivity disorder (ADHD) with biomechanics and a discussion of retention of primitive reflexes being highly associated with the condition.}
}
@inproceedings{LeithingerUIST-14,
       author = {Leithinger, Daniel and Follmer, Sean and Olwal, Alex and Ishii, Hiroshi},
        title = {Physical Telepresence: Shape Capture and Display for Embodied, Computer-Mediated Remote Collaboration},
    publisher = {Association for Computing Machinery},
    booktitle = {Proceedings of the 27th Annual ACM Symposium on User Interface Software and Technology},
         year = {2014},
        pages = {461-470},
     abstract = {We propose a new approach to Physical Telepresence, based on shared workspaces with the ability to capture and remotely render the shapes of people and objects. In this paper, we de- scribe the concept of shape transmission, and propose interaction techniques to manipulate remote physical objects and physical renderings of shared digital content. We investigate how the representation of user's body parts can be altered to amplify their capabilities for teleoperation. We also describe the details of building and testing prototype Physical Telep- resence workspaces based on shape displays. A preliminary evaluation shows how users are able to manipulate remote objects, and we report on our observations of several different manipulation techniques that highlight the expressive nature of our system.}
}
@article{LejterandDeanEXPERT-96,
   author = {Lejter, Moises and Dean, Thomas},
    title = {A Framework for the Development of Multiagent Architectures},
  journal = {{IEEE} Expert},
   volume = 11,
    issue = 6,
    pages = {47-59},
     year = 1996,
} 
@article{Lelandetal94,
       author = {W. E. Leland and M. S. Taqqu and W. Willinger and D. V. Wilson},
        title = {On the self-similar nature of Ethernet traffic},
      journal = {IEEE/ACM Transactions on Networking},
       volume = {},
        pages = {1-15},
         year = 1994,
}
@incollection{LemmerUAI-88,
       author = {Lemmer, John F.},
        title = {Confidence Factors, Empiricism and the Dempster-Shafer Theory of Evidence},
     crossref = UAI-88,
}
@article{LengyelCOMPUTER-98,
     author = {Jed Lengyel},
      title = {The Convergence of Graphics and Vision},
    journal = {Computer},
     volume = 31,
      issue = 7,
      pages = {46-53},
       year = 1998,
  publisher = {IEEE Computer Society},
}
@article{LengyelandDayanNIPS-2008,
       author = {Lengyel, M and Dayan, P},
        title = {Hippocampal contributions to control: the Third way},
      journal = {Advances in Neural Information Processing Systems},
       volume = {20},
         year = {2008},
        pages = {889-896},
     abstract = {Recent experimental studies have focused on the specialization of different neural structures for different types of instrumental behavior. Recent theoretical work has provided normative accounts for why there should be more than one control system, and how the output of different controllers can be integrated. Two particlar controllers have been identified, one associated with a forward model and the prefrontal cortex and a second associated with computationally simpler, habitual, actor-critic methods and part of the striatum. We argue here for the normative appropriateness of an additional, but so far marginalized control system, associated with episodic memory, and involving the hippocampus and medial temporal cortices. We analyze in depth a class of simple environments to show that episodic control should be useful in a range of cases characterized by complexity and inferential noise, and most particularly at the very early stages of learning, long before habitization has set in. We interpret data on the transfer of control from the hippocampus to the striatum in the light of this hypothesis.}
}
@article{LeonBICA-16,
       author = {Carlos Le{\'{o}}n},
        title = {An architecture of narrative memory},
      journal = {Biologically Inspired Cognitive Architectures},
       volume = {16},
        pages = {19-33},
         year = {2016},
     abstract = {Narrative is ubiquitous. According to some models, this is due to the hypothesis that narrative is not only a successful way of communication, but a specific way of structuring knowledge. While most cognitive architectures acknowledge the importance of narrative, they usually do so from a functional point of view and not as a fundamental way of storing material in memory. The presented approach takes one step further towards the inclusion of narrative-aware structures in general cognitive architectures. In particular, the presented architecture studies how episodic memory and procedures in semantic memory can be redefined in terms of narrative structures. A formal definition of narrative for cognition and its constituents are presented, and the functions that an implementation of the architecture needs are described. The relative merits and the potential benefits with regard to general cognitive architectures are discussed and exemplified.}
} 
@techreport{LeonardandDurrant-Whyte,
       author = {Leonard, John J. and Durrant-Whyte, Hugh F.},
        title = {Active Sensor Control for Mobile Robotics},
        issue = {OUEL-1756/89},
  institution = {Oxford University Robotics Research Group},
         year = 1989,
}
@article{LerchnerandLathamCoRR-15,
       author = {Lerchner A. and Latham P.E.},
        title = {Cortical activity demystified: a unifying theory that explains state switching in cortex},
      journal = {CoRR},
       volume = {arXiv:1511.00411},
         year = {2015},
     abstract = {Activity in neocortex exhibits a range of behaviors, from irregular to temporally precise, and from weakly to strongly correlated. So far there has been no single theoretical framework that could explain all these behaviors, leaving open the possibility that they are a signature of radically different mechanisms. Here, we suggest that this is not the case. Instead, we show that a single theory can account for a broad spectrum of experimental observations, including specifics such as the fine temporal details of subthreshold cross-correlations. For the model underlying our theory, we need only assume a small number of well-established properties common to all local cortical networks. When these assumptions are combined with realistically structured input, they produce exactly the repertoire of behaviors that is observed experimentally, and lead to a number of testable predictions.}
} 
@article{LerchnerandLathamFiN-10,
       author = {Lerchner A. and Latham P.E.},
        title = {Cortical activity demystified: a unifying theory that explains state switching in cortex},
      journal = {Frontiers in Neuroscience, Conference Abstract: Computational and Systems Neuroscience},
    publisher = {Frontiers Media S.A.},
         year = 2010,
     abstract = {An increasing number of experiments show that neocortical activity not only exhibits different states, but can rapidly switch between them. Examples include up and down states [1], synchronized and de-synchronized activity [2], high and low spiking precision [3], and large and small subthreshold correlations [4]. Although there is agreement that these states have important implications for computations, it has not been clear what mechanisms underlie them. Here we provide a unifying explanation by showing that different inputs applied to the same cortical network can generate all the states, and transitions between the states, that have been observed experimentally. For the model underlying our theory, we need to assume only a small number of well-established properties that are shared by all local networks in the cortex. To analyze the model, we use an extended mean-field theory with temporal fluctuations. As with most mean-field theories of this type, we combine subthreshold activity with spiking activity to derive a self-consistent set of equations. Importantly, the equations contain only a small number of parameters, and even these have values that are constrained by biology, so the network can exhibit only a restricted range of behaviors. 

       The main outcome of the theory is that the activity of the network depends primarily on the external input, and far less so on single-neuron properties or connectivity. For constant input, our theory (and simulations) recovers previous results indicating that networks exhibit weak correlations and irregular spiking activity. Inputs with intermediate structures lead to states in which sub-threshold correlations increase during sudden stimulus increases, accompanied by more precise spike timing during such events. And finally, brief inputs that drive the network only occasionally result in a dynamic state exhibiting membrane-potential "bumps" that are strongly correlated across neurons. These bumps lead to almost perfect synchrony for both excitation and inhibition, but with a characteristic lag of inhibition behind excitation of a few milliseconds. In this regime, the theory predicts that membrane potentials are highly correlated, and the first spikes after stimulus onsets are precisely timed. Our results show that both single-neuron and population activity in cortex are constrained by fundamental dynamics of local cortical networks, which in turn depend on the structure of the network input in a non-trivial manner. The theory predicts that cortical networks can have only a small number of behaviors. It is nontrivial, then, that the behaviors predicted by our theory (and verified with detailed network simulations) are consistent with experimental observations. Beyond providing a mechanistic explanation for a wealth of existing data, our theory can guide design and data analysis for future experiments that aim to probe detailed function and micro-structure of cortical networks.

       References:
       [1] Lampl et al. (1999) Neuron 22, 361-374.
       [2] Curto et al. (2009) Journal of Neuroscience 29, 10600-10612
       [3] Buracas et al. (1998) Neuron 20, 959-969
       [4] Poulet and Petersen (2008) Nature 454, 881-885},
}
@article{LerneretalCELL-16,
       author = {Lerner, Talia N. and Ye, Li and Deisseroth, Karl},
        title = {Communication in Neural Circuits: Tools, Opportunities, and Challenges},
      journal = {Cell},
         year = {2016},
    publisher = {Elsevier},
       volume = {164},
        issue = {6},
        pages = {1136-1150},
     abstract = {Communication, the effective delivery of information, is fundamental to life across all scales and species. Nervous systems (by necessity) may be most specifically adapted among biological tissues for high rate and complexity of information transmitted, and thus, the properties of neural tissue and principles of its organization into circuits may illuminate capabilities and limitations of biological communication. Here, we consider recent developments in tools for studying neural circuits with particular attention to defining neuronal cell types by input and output information streams, i.e., by how they communicate. Complementing approaches that define cell types by virtue of genetic promoter/enhancer properties, this communication-based approach to defining cell types operationally by input/output (I/O) relationships links structure and function, resolves difficulties associated with single-genetic-feature definitions, leverages technology for observing and testing significance of precisely these I/O relationships in intact brains, and maps onto processes through which behavior may be adapted during development, experience, and evolution.},
}
@article{LesserAIM-83,
       author = {Lesser, V. R. and Corkill, D. D.},
        title = {The Distributed Vehicle Monitoring Testbed: A Tool for Investigating Distributed Problem Solving Networks},
      journal = {AI Magazine},
       volume = 4,
         year = 1983,
        pages = {15-33},
}
@article{LesserAIM-88,
       author = {Lesser, Victor R. and Pavlin, Jasmina and Durfee, Edmund},
        title = {Approximate Processing in Real-Time Problem Solving},
      journal = {AI Magazine},
       volume = 9,
        issue = 1,
         year = 1988,
        pages = {49-61},
}
@article{Lettvinetal59,
       author = {Lettvin, J. Y. and Maturana, H. R. and McCulloch, W. S. and Pitts, W. H.},
        title = {What the Frog's Eye Tells the Frog's Brain},
      journal = {Proceedings of the Institute for Radio Engineers},
       volume = 47,
          url = {http://jerome.lettvin.info/lettvin/Jerome/WhatTheFrogsEyeTellsTheFrogsBrain.pdf},
         year = 1959,
        pages = {1940-1951},
     abstract = {A frog hunts on land by vision. He escapes enemies mainly by seeing them. His eyes do not move, as do ours, to follow prey, attend suspicious events, or search for things of interest. If his body changes its position with respect to gravity or the whole visual world is rotated about him, then he shows compensatory eye movements. These movements enter his hunting and evading habits only, e.g., as he sits on a rocking lily pad. Thus his eyes are actively stabilized. He has no fovea, or region of greatest acuity in vision, upon which he must center a part of the image. He has only a single visual system, retina to colliculus, not a double one such as ours where the retina sends fibers not only to colliculus but to the lateral geniculate body which relays to cerebral cortex. Thus, we chose to work on the frog because of the uniformity of his retina, the normal lack of eye and head movements except for those which stabilize the retinal image, and the relative simplicity of the connection of his eye to his brain.
      The frog does not seem to see or, at any rate, is not concerned with the detail of stationary parts of the world around him. He will starve to death surrounded by food if it is not moving. His choice of food is determined only by size and movement. He will leap to capture any object the size of an insect or worm, providing it moves like one. He can be fooled easily not only by a bit of dangled meat but by any moving small object. His sex life is conducted by sound and touch. His choice of paths in escaping enemies does not seem to be governed by anything more devious than leaping to where it is darker. Since he is equally at home in water and on land, why should it matter where he lights after jumping or what particular direction he takes? He does remember a moving thing providing it stays within his field of vision and he is not distracted.}
}
@article{LeubaandKraftsikAE-94,
      title = {Changes in volume, surface estimate, three-dimensional shape and total number of neurons of the human primary visual cortex from midgestation until old age},
     author = {G. Leuba and R. Kraftsik},
    journal = {Anatomy Embryo1ogy},
     volume = 190,
      pages = {351-366},
       year = 1994,
}
@inproceedings{LeungandMalikICCV-99,
     author = {Thomas K. Leung and Jitendra Malik},
      title = {Recognizing Surfaces using Three-Dimensional Textons},
  booktitle = {Proceedings of the Seventh {IEEE} International Conference on Computer Vision}, 
  publisher = {IEEE Computer Society},
      pages = {1010-1017},
       year = 1999,
}
@article{LeuskiandTraumAIM-11,
    author = {Anton Leuski and David Traum},
     title = {{NPCEditor}: Creating Virtual Human Dialogue Using Information Retrieval Techniques},
   journal = {{AI Magazine}},
    volume = 32,
     issue = 2,
      year = 2011,
     pages = {42-56},
} 
@inproceedings{LeuskiandTraumASC-08,
    author = {Anton Leuski and David Traum},
     title = {A Statistical Approach for Text Processing in Virtual Humans},
 booktitle = {Proceedings of the Army Science Conference (26th)},
  location = {Orlando, Florida},
      year = 2008,
  abstract = {We describe a text classification approach based on statistical language modeling. We show how this approach can be used for several natural language processing tasks in a virtual human system. Specifically, we show it can applied to language understanding, language generation, and character response selection tasks. We illustrate these applications with some experimental results. The resulting system is composed of three primary components: 1. natural language understanding (NLU) module that interprets the text of the user's utterance and converts it into some internal representation; 2. dialog manager (DM) module that analyzes the interpretation and selects the appropriate response; 3. natural language generation (NLG) module that converts the internal representation to the text of the response.}
}
@inproceedings{LeuskiandTraumIAAI-10,
        title = {Practical Language Processing for Virtual Humans},
       author = {Anton Leuski and David R. Traum},
    booktitle = {Twenty-Second Annual Conference on Innovative Applications of Artificial Intelligence},
         year = 2010,
          url = {http://www.ict.usc.edu/\%7Eleuski/publications/papers/iaai-10-npceditor.pdf},
}
@article{Lev-RamNEURON-99,
       author = {Varda Lev-Ram and Tao Jiang and Jason Wood and David S Lawrence and Roger Y Tsien},
        title = {Synergies and Coincidence Requirements between NO, cGMP, and Ca2+ in the Induction of Cerebellar Long-Term Depression},
      journal = {Neuron},
       volume = 18,
        issue = 6,
         year = 1997,
        pages = {1025-1038},
     abstract = {Parallel fiber synapses onto Purkinje neurons in acute cerebellar slices undergo long-term depression (LTD) when presynaptic activity coincides with postsynaptic depolarization. These electrical inputs can be respectively replaced by nitric oxide (NO) and Ca2+ photolytically released inside the Purkinje neuron, showing that these two messengers are sufficient for \{LTD\} induction. \{NO\} acts via cGMP production because inhibitors of guanylate cyclase prevent \{LTD\} but can be circumvented by photoreleased cGMP combined with Ca2+ elevation. Three inhibitors of cGMP-dependent protein kinase, R p-8Br-PET-cGMPS, KT5823, and a novel pseudosubstrate peptide, all block LTD. \{LTD\} induction permits \&lt;10 ms gap between \{NO\} release and Ca2+ elevation, whereas 200-300 ms is allowed between uncaged cGMP and Ca2+ increase. This surprising difference in timing precision can be explained either by tighter localization and faster decay of cGMP when generated by \{NO\} rather than uncaging, or by two independent coincidence detectors in series.}
}
@article{Levesque86,
       author = {Levesque, Hector J.},
        title = {Making Believers out of Computers},
      journal = AIJ,
       volume = 30,
        pages = {81-108},
         year = 1986,
}
@article{Levin73,
       author = {Levin, L. A.},
        title = {Universal Sorting Problems},
      journal = {Problems of Information Transmission},
       volume = 9,
        pages = {265-266},
         year = 1973,
}
@inproceedings{LevinSTOC-84,
       author = {Levin, Leonid A.},
        title = {Problems complete in `average' instance},
    booktitle = {Proceedings of the Sixteenth Annual {ACM} Symposium on Theory of Computing},
     location = {Boulder, Colorado},
         year = 1984,
        pages = 465,
}
@article{LevineCoRR-18,
       author = {Sergey Levine},
        title = {Reinforcement Learning and Control as Probabilistic Inference: Tutorial and Review},
      journal = {CoRR},
       volume = {arXiv:1805.00909},
         year = {2018},
     abstract = {The framework of reinforcement learning or optimal control provides a mathematical formalization of intelligent decision making that is powerful and broadly applicable. While the general form of the reinforcement learning problem enables effective reasoning about uncertainty, the connection between reinforcement learning and inference in probabilistic models is not immediately obvious. However, such a connection has considerable value when it comes to algorithm design: formalizing a problem as probabilistic inference in principle allows us to bring to bear a wide array of approximate inference tools, extend the model in flexible and powerful ways, and reason about compositionality and partial observability. In this article, we will discuss how a generalization of the reinforcement learning or optimal control problem, which is sometimes termed maximum entropy reinforcement learning, is equivalent to exact probabilistic inference in the case of deterministic dynamics, and variational inference in the case of stochastic dynamics. We will present a detailed derivation of this framework, overview prior work that has drawn on this and related ideas to propose new reinforcement learning and control algorithms, and describe perspectives on future research.}
}
@article{LevineandDavidsonPNAS-05,
     title = {{Gene regulatory networks for development}},
    author = {Levine, Michael and Davidson, Eric H.},
   journal = {Proceedings of the National Academy of Sciences of the United States of America},
    volume = 102,
     issue = 14,
     pages = {4936-4942},
      year = 2005,
  abstract = {The genomic program for development operates primarily by the regulated expression of genes encoding transcription factors and components of cell signaling pathways. This program is executed by cis-regulatory DNAs (e.g., enhancers and silencers) that control gene expression. The regulatory inputs and functional outputs of developmental control genes constitute network-like architectures. In this PNAS Special Feature are assembled papers on developmental gene regulatory networks governing the formation of various tissues and organs in nematodes, flies, sea urchins, frogs  and mammals. Here, we survey salient points of these networks, by using as reference those governing specification of the endomesoderm in sea urchin embryos and dorsal-ventral patterning in the embryo.},
}
@article{LevinsonHMM,
       author = {Levinson, S. E. and Rabiner, L. R. and Sondhi, M. M.},
        title = {An Introduction to the Application of the Theory of Probabilistic Functions of a {M}arkov Process to Automatic Speech Recognition},
      journal = {The Bell System Technical Journal},
       volume = 62,
        issue = 4,
         year = 1983,
        pages = {1035-1074},
}
@article{LevinsonRabinerSondhi83,
       author = {S. E. Levinson and L. R. Rabiner and M. M. Sondhi},
        title = {An Introduction to the Application of the Theory of Probabilistic Functions of a {M}arkov Process to Automatic Speech Recognition},
      journal = {The Bell System Technical Journal},
       volume = 62,
        issue = 4,
         year = 1983,
        pages = {1035-1074},
}
@inproceedings{LevittAAAI-87,
       author = {Levitt, Tod S. and Lawton, Daryl T. and Chelberg, David M. and Nelson, Philip C.},
        title = {Qualitative Landmark-based Path Planning and Following},
    booktitle = {Proceedings AAAI-87},
     location = {Seattle, Washington},
 organization = {AAAI},
         year = 1987,
        pages = {689-694},
}
@article{LevittAIJ-90,
       author = {Levitt, Tod S. and Lawton, Daryl T.},
        title = {Qualitative Navigation for Mobile Robots},
      journal = AIJ,
       volume = 44,
        issue = 3,
        pages = {305-360},
         year = 1990,
}
@inproceedings{LevittUAI-88,
       author = {Levitt, Tod and Binford, Thomas and Ettinger, Gil and Gelband, Patrice},
        title = {Utility-Based Control for Computer Vision},
    booktitle = UAI-88,
     location = {Minneapolis, Minnesota},
         year = 1988,
}
@article{Levittetal-JoCN-93,
       author = {Levitt, Jonathan B. and Lewis, David A. and Yoshioka, Takashi and Lund, Jennifer S.},
        title = {Topography of pyramidal neuron intrinsic connections in macaque monkey prefrontal cortex (areas 9 and 46)},
      journal = {Journal of Comparative Neurology},
         year = {1993},
       volume = {338},
       number = {3},
        pages = {360-376},
     abstract = {Abstract An understanding of the normal organization of prefrontal cortex is essential to the recognition of pathology underlying human behavioral disorders believed to depend on this region. We have therefore studied the pattern of intrinsic intra- and interlaminar pyramidal neuron connectivity in prefrontal areas 9 and 46 (of Walker) in macaque monkey cerebral cortex (anterior to the arcuate sulcus between the principal sulcus and midline). We made focal (200 400 m) injections of biocytin and mapped the pattern of orthogradely transported label. Injections made into the superficial layers label wide-ranging lateral projections within the same areas of prefrontal cortex. Projections local to such small injections form a narrow band of terminals in layers 1-3 (200-400 $\mu$m wide, 2-4 mm long) centered on the injection site. Collateral fibers spread orthogonal to this terminal band, making frequent bifurcations, to establish a series of parallel bands of terminals with uninnervated bands between, spaced regularly across the cortex (center to center 500-600 $\mu$m). The entire pattern of terminal label is stripelike, with occasional narrower interbands and crosslinks between the bands, and can extend over 7-8 mm across the cortex. These projections arise from pyramidal neurons in layers 2, 3, and 5 and terminate in layers 1-3. The stripelike pattern contrasts with patchlike patterns in other cortical regions (V1, V2, V4, motor, somatosensory) and is smaller in scale than stripelike zones of corticocortical afferent terminals to this region, reported to be 300-750 $\mu$m wide and spaced 1.0-1.5 mm center to center.},
}
@article{LevittetalJoCN-93,
       author = {Levitt, Jonathan B. and Lewis, David A. and Yoshioka, Takashi and Lund, Jennifer S.},
        title = {Topography of pyramidal neuron intrinsic connections in macaque monkey prefrontal cortex (areas 9 and 46)},
      journal = {Journal of Comparative Neurology},
       volume = {338},
       number = {3},
         year = {1993},
        pages = {360-376},
     abstract = {An understanding of the normal organization of prefrontal cortex is essential to the recognition of pathology underlying human behavioral disorders believed to depend on this region. We have therefore studied the pattern of intrinsic intra- and interlaminar pyramidal neuron connectivity in prefrontal areas 9 and 46 (of Walker) in macaque monkey cerebral cortex (anterior to the arcuate sulcus between the principal sulcus and midline). We made focal (200-400 $\mu$m) injections of biocytin and mapped the pattern of orthogradely transported label. Injections made into the superficial layers label wide-ranging lateral projections within the same areas of prefrontal cortex. Projections local to such small injections form a narrow band of terminals in layers 1-3 (200-400 $\mu$m wide, 2-4 mm long) centered on the injection site. Collateral fibers spread orthogonal to this terminal band, making frequent bifurcations, to establish a series of parallel bands of terminals with uninnervated bands between, spaced regularly across the cortex (center to center 500-600 $\mu$m). The entire pattern of terminal label is stripelike, with occasional narrower interbands and crosslinks between the bands, and can extend over 7-8 mm across the cortex. These projections arise from pyramidal neurons in layers 2, 3, and 5 and terminate in layers 1-3. The stripelike pattern contrasts with patchlike patterns in other cortical regions (V1, V2, V4, motor, somatosensory) and is smaller in scale than stripelike zones of corticocortical afferent terminals to this region, reported to be 300-750 $\mu$m wide and spaced 1.0-1.5 mm center to center.},
}
@article{LevskyetalSCIENCE-02,
     title = {Single-cell gene expression profiling},
    author = {Levsky, J.M. and Shenoy, S.M. and Pezo, R.C. and Singer, R.H.},
   journal = {Science},
    volume = {297},
     issue = {5582},
     pages = {836-40},
      year = {2002},
  abstract = {A key goal of biology is to relate the expression of specific genes to a particular cellular phenotype. However, current assays for gene expression destroy the structural context. By combining advances in computational fluorescence microscopy with multiplex probe design, we devised technology in which the expression of many genes can be visualized simultaneously inside single cells with high spatial and temporal resolution. Analysis of 11 genes in serum-stimulated cultured cells revealed unique patterns of gene expression within individual cells. Using the nucleus as the substrate for parallel gene analysis, we provide a platform for the fusion of genomics and cell biology: "cellular genomics".}
}
@book{LevyLevyandSolomon2000,
        title = {Microscopic Simulation of Financial Markets: From Investor Behavior to Market Phenomena},
       author = {Moshe Levy, Haim Levy and Sorin Solomon},
    publisher = {Elsevier},
         year = {2000},
     abstract = {Microscopic Simulation (MS) uses a computer to represent and keep track of individual ("microscopic") elements in order to investigate complex systems which are analytically intractable. A methodology that was developed to solve physics problems, MS has been used to study the relation between microscopic behavior and macroscopic phenomena in systems ranging from those of atomic particles, to cars, animals, and even humans. In finance, MS can help explain, among other things, the effects of various elements of investor behavior on market dynamics and asset pricing. It is these issues in particular, and the value of an MS approach to finance in general, that are the subjects of this book. The authors not only put their work in perspective by surveying traditional economic analyses of investor behavior, but they also briefly examine the use of MS in fields other than finance.

Most models in economics and finance assume that investors are rational. However, experimental studies reveal systematic deviations from rational behavior. How can we determine the effect of investors' deviations from rational behavior on asset prices and market dynamics? By using Microscopic Simulation, a methodology originally developed by physicists for the investigation of complex systems, the authors are able to relax classical assumptions about investor behavior and to model it as empirically and experimentally observed. This rounded and judicious introduction to the application of MS in finance and economics reveals that many of the empirically-observed "puzzles" in finance can be explained by investors' quasi-rationality.

Researchers use the book because it models heterogeneous investors, a group that has proven difficult to model. Being able to predict how people will invest and setting asset prices accordingly is inherently appealing, and the combination of computing power and statistical mechanics in this book makes such modeling possible. Because many finance researchers have backgrounds in physics, the material here is accessible. Features: (i) Emphasizes investor behavior in determining asset prices and market dynamics, (ii) Introduces Microscopic Simulation within a simplified framework, and (iii) Offers ways to model deviations from rational decision-making.},
}
@incollection{LevyandBanajiAGEISM-02,
        title = {Implicit Ageism},
       author = {Becca R. Levy and Mahzarin R. Banaji},
    booktitle = {Ageism: Stereotyping and Prejudice against Older Persons},
       editor = {T.D. Nelson},
    publisher = {The MIT Press},
      address = {Cambridge, MA},
         year = {2002},
        pages = {49-75},
     abstract = {Beliefs about the elderly as unable to contribute to society, and hence as dispensable members of a community, and attitudes toward them of dislike and distancing are prevalent (Kite and Johnson 1988). Social scientists have focused on the effects of negative beliefs and attitudes to examine discrimination toward the aged in a variety of spheres, including
everyday conversations (Hummert and Ryan 1996; Williams and Giles 1998), politics (Sigelman and Sigelman 1982) arid the workplace (Butler 1980; Finkelstein, Burke, and Raju 1995). To take just one example of discriminatory treatment, many older Americans do not receive necessary care for common and treatable medical conditions, including heart disease (Asch et al. 2000; Bowling 1999; Guigliano et al. 1998; Hillerbrand and Shaw 1990). Although poverty, insofar as its effects are magnified in old age, is obviously involved, the reason for discrimination does not ap pear to be solely financial, because the vast majority of treatments are covered by Medicare (Asch et al. 2000). Instead, some believe that "ageism in clinical medicine and health policy reflects the ageism evident in wider society" (Bowling 1999, p. 1353). As health maintenance organizations and medical centers place pressure on physicians to spend increasingly less time with each patient, discrimination against the elderly is likely to increase. In time-pressured situations in which attention is called to a number of tasks at once, individuals may be more likely to engage in stereotyping (Gilbert and Hixon 1991; Blair and Banaji 1996; Jamieson and Zanna 1989; Pratto and Bargh 1991). In this chapter, we make two claims regarding ageism, which we define as an alteration in feeling, belief, or behavior in response to an individual's or group's perceived chronological age. First, one of the most insidious aspects of ageism is that it can operate without conscious awareness, control, or intention to harm. Although the idea of the implicit nature of stereotyping and prejudice is not new to social science research (Katz and Braly 1935; Crosby, Bromley, and Saxe 1980; Devine 1989), the idea of implicit ageism is unique in at least one way. There are no hate groups that target the elderly as there are hate groups that target members of religious and racial and ethnic groups. Even gender prejudice has produced the recognition that there are those who have explicit antipathy toward one or the other group (e.g., misogynists, male chauvinists, man haters). In contrast, social sanctions against expressions of negative attitudes and beliefs about the elderly are almost completely absent. In fact, the widespread occurrence of socially acceptable expressions of negativity toward the elderly has been well documented (Williams and Giles 1998). This state of affairs stands in contrast to other social groups, where, at least in public discourse, there has been a notable change in recognition of social disadvantage and the need for action to ameliorate its consequences.}
}
@inproceedings{LevyandGoldbergCONIL-14,
       author = {Levy, Omer  and  Goldberg, Yoav},
        title = {Linguistic Regularities in Sparse and Explicit Word Representations},
    booktitle = {Proceedings of the Eighteenth Conference on Computational Natural Language Learning},
         year = 2014,
      address = {Ann Arbor, Michigan},
    publisher = {Association for Computational Linguistics},
        pages = {171-180},
     abstract = {Recent work has shown that neural embedded word representations capture many relational similarities, which can be recovered by means of vector arithmetic in the embedded space. We show that Mikolov et al.'s method of first adding and subtracting word vectors, and then searching for a word similar to the result, is equivalent to searching for a word that maximizes a linear combination of three pairwise word similarities. Based on this observation, we suggest an improved method of recovering relational similarities, improving the state-of-the-art results on two recent word-analogy datasets. Moreover, we demonstrate that analogy recovery is not restricted to neural word embeddings, and that a similar amount of relational similarities can be recovered from traditional distributional word representations.}
} 
@book{Lewis1973,
       author = {Lewis, David},
        title = {Counterfactuals},
    publisher = {Harvard University Press},
         year = 1973,
}
@book{LewisOPTIMAL,
       author = {Lewis, Frank L.},
        title = {Optimal Control},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1986,
}
@book{LewisOPTIMAL-86,
       author = {Lewis, Frank L.},
        title = {Optimal Control},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1986,
}
@book{LewisandPapadimitriou,
       author = {Lewis, Harry R. and Papadimitriou, Christos H.},
        title = {Elements of the Theory of Computation},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1981,
}
@article{LewisetalELIFE-15,
       author = {Lewis, L.D. and Voigts, J. and Flores, F.J. and Schmitt, L.I. and Wilson, M.A. and Halassa, M.M. and Brown, E.N.},
        title = {Thalamic reticular nucleus induces fast and local modulation of arousal state},
      journal = {Elife},
        volume = {4},
         year = {2015},
       pages = {e08760},
     abstract = {During low arousal states such as drowsiness and sleep, cortical neurons exhibit rhythmic slow wave activity associated with periods of neuronal silence. Slow waves are locally regulated, and local slow wave dynamics are important for memory, cognition, and behaviour. While several brainstem structures for controlling global sleep states have now been well characterized, a mechanism underlying fast and local modulation of cortical slow waves has not been identified. Here, using optogenetics and whole cortex electrophysiology, we show that local tonic activation of thalamic reticular nucleus (TRN) rapidly induces slow wave activity in a spatially restricted region of cortex. These slow waves resemble those seen in sleep, as cortical units undergo periods of silence phase-locked to the slow wave. Furthermore, animals exhibit behavioural changes consistent with a decrease in arousal state during TRN stimulation. We conclude that TRN can induce rapid modulation of local cortical state.}
}
@article{LewisetalJNC-02,
       author = {Lewis, David A. and Melchitzky, Darlene S. and Burgos, Guillermo-Gonzalez},
        title = {Specificity in the functional architecture of primate prefrontal cortex},
      journal = {Journal of Neurocytology},
         year = {2002},
       volume = {31},
       number = {3},
        pages = {265-276},
     abstract = {Multiple lines of evidence indicate that the performance of complex cognitive processes, such as those involving working memory, depend upon the functional properties of the circuitry of the prefrontal cortex (PFC). In primates, working memory has been proposed to be dependent upon the sustained activity of specific populations of PFC pyramidal cells, with this activity regulated by certain types of GABAergic interneurons. Thus, knowledge of the connectivity between PFC pyramidal cells and interneurons is crucial to the understanding the neural mechanisms that subserve working memory. This paper reviews recent findings that reveal specificity in the spatial organization, synaptic targets and postnatal development of pyramidal cells and interneurons in the primate prefrontal cortex, and considers the relevance of these findings for the neural circuitry that subserves working memory.},
}
@article{LewisetalTiCS-18,
       author = {Lewis, P. A. and Knoblich, G. and Poe, G.},
        title = {{{H}ow {M}emory {R}eplay in {S}leep {B}oosts {C}reative {P}roblem-{S}olving}},
      journal = {Trends Cognitive Science},
       volume = {22},
       number = {6},
        pages = {491-503},
         year = {2018},
     abstract = {Creative thought relies on the reorganisation of existing knowledge. Sleep is known to be important for creative thinking, but there is a debate about which sleep stage is most relevant, and why. We address this issue by proposing that rapid eye movement sleep, or 'REM', and non-REM sleep facilitate creativity in different ways. Memory replay mechanisms in non-REM can abstract rules from corpuses of learned information, while replay in REM may promote novel associations. We propose that the iterative interleaving of REM and non-REM across a night boosts the formation of complex knowledge frameworks, and allows these frameworks to be restructured, thus facilitating creative thought. We outline a hypothetical computational model which will allow explicit testing of these hypotheses.},
}
@article{LiCoRR-17,
       author = {Yuxi Li},
        title = {Deep Reinforcement Learning: An Overview},
      journal = {CoRR},
       volume = {arXiv:1701.07274},
         year = {2017},
     abstract = {We give an overview of recent exciting achievements of deep reinforcement learning (RL). We discuss six core elements, six important mechanisms, and twelve applications. We start with background of machine learning, deep learning and reinforcement learning. Next we discuss core RL elements, including value function, in particular, Deep Q-Network (DQN), policy, reward, model, planning, and exploration. After that, we discuss important mechanisms for RL, including attention and memory, unsupervised learning, transfer learning, multi-agent RL, hierarchical RL, and learning to learn. Then we discuss various applications of RL, including games, in particular, AlphaGo, robotics, natural language processing, including dialogue systems, machine translation, and text generation, computer vision, neural architecture design, business management, finance, healthcare, Industry 4.0, smart grid, intelligent transportation systems, and computer systems. We mention topics not reviewed yet, and list a collection of RL resources. After presenting a brief summary, we close with discussions.}
}
@article{LiandMalikCoRR-16,
       author = {Ke Li, Jitendra Malik},
        title = {Learning to Optimize Neural Nets},
       volume = {arXiv:arXiv:1703.00441},
         year = {2016},
     abstract = {Learning to Optimize is a recently proposed framework for learning optimization algorithms using reinforcement learning. In this paper, we explore learning an optimization algorithm for training shallow neural nets. Such high-dimensional stochastic optimization problems present interesting challenges for existing reinforcement learning algorithms. We develop an extension that is suited to learning optimization algorithms in this setting and demonstrate that the learned optimization algorithm consistently outperforms other known optimization algorithms even on unseen tasks and is robust to changes in stochasticity of gradients and the neural net architecture. More specifically, we show that an optimization algorithm trained with the proposed method on the problem of training a neural net on MNIST generalizes to the problems of training neural nets on the Toronto Faces Dataset, CIFAR-10 and CIFAR-100.}
} 
@article{LiandOsherIPI-07,
      title = {Coordinate Descent Optimization for {$L_1$} Minimization with 
               Application to Compressed Sensing: a Greedy Algorithm},
     author = {Yingying Li and Stanley Osher},
    journal = {Inverse Problems and Imaging},
     volume = 3,
      issue = 3,
      pages = {487-503},
       year = 2007,
}
@inproceedings{LiandYamanishiACL-97,
       author = {Hang Li and Kenji Yamanishi},
        title = {Document Classification Using a Finite Mixture Model},
    booktitle = {Proceedings of the Thirty-Fifth Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics and Eighth Conference of the {E}uropean {C}hapter of the {A}ssociation for {C}omputational {L}inguistics},
    publisher = {Association for Computational Linguistics},
      address = {Somerset, New Jersey},
       editor = {Philip R. Cohen and Wolfgang Wahlster},
        pages = {39-47},
          url = {citeseer.ist.psu.edu/li97document.html},
         year = 1997,
}
@article{LiangetalCL-13,
       author = {Percy Liang and Michael I. Jordan and Dan Klein},
        title = {Learning Dependency-Based Compositional Semantics},
      journal = {Computational Linguistics},
         year = 2013,
       volume = 39,
        issue = 2,
        pages = {389-446},
} 
@article{LiangetalCoRR-16,
       author = {Chen Liang and Jonathan Berant and Quoc Le and Kenneth D. Forbus and Ni Lao},
        title = {Neural Symbolic Machines: Learning Semantic Parsers on Freebase with Weak Supervision},
      journal = {CoRR},
       volume = {arXiv:1611.00020},
         year = {2016},
     abstract = {Harnessing the statistical power of neural networks to perform language understanding and symbolic reasoning is difficult, when it requires executing efficient discrete operations against a large knowledge-base. In this work, we introduce a Neural Symbolic Machine (NSM), which contains (a) a neural "programmer", i.e., a sequence-to-sequence model that maps language utterances to programs and utilizes a key-variable memory to handle compositionality (b) a symbolic "computer", i.e., a Lisp interpreter that performs program execution, and helps find good programs by pruning the search space. We apply REINFORCE to directly optimize the task reward of this structured prediction problem. To train with weak supervision and improve the stability of REINFORCE we augment it with an iterative maximum-likelihood training process. NSM outperforms the state-of-the-art on the WEBQUESTIONSSP dataset when trained from question-answer pairs only, without requiring any feature engineering or domain-specific knowledge.}
}
@article{LiaoetalTRANSLATIONAL-PSYCHIATRY-19,
       author = {Liao, Yuhua and Xie, Bo and Zhang, Huimin and He, Qian and Guo, Lan and Subramaniapillai, M. and Fan, Beifang and Lu, Ciyong and Mclntyer, R. S.},
        title = {Efficacy of omega-3 PUFAs in depression: A meta-analysis},
      journal = {Translational Psychiatry},
         year = {2019},
       volume = {9},
        issue = {1},
        pages = {190},
     abstract = {We conducted this meta-analysis of double-blind randomized placebo-controlled trials to estimate the efficacy of omega-3 polyunsaturated fatty acids (PUFAs), especially docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), in the improvement of depression. We applied a systematic bibliographic search in PubMed and EMBASE for articles published prior to 20 December 2017. This meta-analysis was performed using RevMan 5.3 and R 3.4.3, and means and standard deviations were calculated in fixed- or random-effects models based on the results of the Q-test. A sensitivity analysis was also conducted to evaluate the stability of the results, and publication bias was evaluated by a funnel plot and Egger?s linear regression analysis. Our search resulted in 180 articles; we analyzed 26 studies, which included 2160 participants. The meta-analysis showed an overall beneficial effect of omega-3 polyunsaturated fatty acids on depression symptoms (SMD =  0.28, P = 0.004). Compared with placebo, EPA-pure (=100\% EPA) and EPA-major formulations ({\textyen}60\% EPA) demonstrated clinical benefits with an EPA dosage {\texteuro}1 g/d (SMD =  0.50, P = 0.003, and SMD =  1.03, P =  0.03, respectively), whereas DHA-pure and DHA-major formulations did not exhibit such benefits.},
}
@inproceedings{LiarokapisetalICAR-15,
       author = {M. V. {Liarokapis} and A. M. {Dollar} and K. J. {Kyriakopoulos}},
    booktitle = {International Conference on Advanced Robotics},
        title = {Humanlike, task-specific reaching and grasping with redundant arms and low-complexity hands},
         year = {2015},
        pages = {490-497},
     abstract = {In this paper, we propose a methodology for closed-loop, humanlike, task-specific reaching and grasping with redundant robot arms and low-complexity robot hands. Human demonstrations are utilized in a learn by demonstration fashion, in order to map human to humanlike robot motion. Principal Components Analysis (PCA) is used to transform the humanlike robot motion in a low-dimensional manifold, where appropriate Navigation Function (NF) models are trained. A series of grasp quality measures, as well as task compatibility indexes are employed to guarantee robustness of the computed grasps and task specificity of goal robot configurations. The final scheme provides anthropomorphic robot motion, task-specific robot arm configurations and hand grasping postures, optimized fingertips placement on the object surface (that results to robust grasps) and guaranteed convergence to the desired goals. The position and geometry of the objects are considered a-priori known. The efficiency of the proposed methods is assessed with simulations and experiments that involve different robot arm hand systems. The proposed scheme can be useful for various Human Robot Interaction (HRI) applications.},
}
@article{LiawandBergerHIPPOCAMPUS-96,
       author = {Liaw, Jim-Shih and Berger, Theodore W.},
        title = {Dynamic synapse: A new concept of neural representation and computation},
      journal = {Hippocampus},
       volume = {6},
       number = {6},
    publisher = {John Wiley \& Sons, Inc.},
         year = {1996},
        pages = {591-600},
     abstract = {Presynaptic mechanisms influencing the probability of neurotransmitter release from an axon terminal, such as facilitation, augmentation, and presynaptic feedback inhibition, are fundamental features of biological neurons and are cardinal physiological properties of synaptic connections in the hippocampus. The consequence of these presynaptic mechanisms is that the probability of release becomes a function of the temporal pattern of action potential occurrence, and hence, the strength of a given synapse varies upon the arrival of each action potential invading the terminal region. From the perspective of neural information processing, the capability of dynamically tuning the synaptic strength as a function of the level of neuronal activation gives rise to a significant representational and processing power of temporal spike patterns at the synaptic level. Furthermore, there is an exponential growth in such computational power when the specific dynamics of presynaptic mechanisms varies quantitatively across axon terminals of a single neuron, a recently established characteristic of hippocampal synapses. During learning, alterations in the presynaptic mechanisms lead to different pattern transformation functions, whereas changes in the postsynaptic mechanisms determine how the synaptic signals are to be combined. We demonstrate the computational capability of dynamic synapses by performing speech recognition from unprocessed, noisy raw waveforms of words spoken by multiple speakers with a simple neural network consisting of a small number of neurons connected with synapses incorporating dynamically determined probability of release. The dynamics included in the model are consistent with available experimental data on hippocampal neurons in that parameter values were chosen so as to be consistent with time constants of facilitative and inhibitory processes governing the dynamics of hippocampal synaptic transmission studied using nonlinear systems analytic procedures.},
} 
@article{Licklider60,
       author = {J. C. R. Licklider},
        title = {Man-Computer Symbiosis},
      journal = {{IRE} Transactions on Human Factors in Electronics},
       volume = 1,
        pages = {4-11},
         year = 1960,
}
@article{Licklider68,
       author = {J. C. R. Licklider and Robert W. Taylor},
        title = {The Computer as a Communications Device},
      journal = {Science and Technology},
       volume = 76,
         year = 1968,
}
@article{LiddelowetalNATURE-17,
       author = {Liddelow, Shane A. and Guttenplan, Kevin A. and Clarke, Laura E. and Bennett, Frederick C. and Bohlen, Christopher J. and Schirmer, Lucas and Bennett, Mariko L. and Munch, Alexandra E. and Chung, Won-Suk and Peterson, Todd C. and Wilton, Daniel K. and Frouin, Arnaud and Napier, Brooke A. and Panicker, Nikhil and Kumar, Manoj and Buckwalter, Marion S. and Rowitch, David H. and Dawson, Valina L. and Dawson, Ted M. and Stevens, Beth and Barres, Ben A.},
        title = {Neurotoxic reactive astrocytes are induced by activated microglia},
      journal = {Nature},
         year = {2017},
    publisher = {Macmillan Publishers Limited},
       volume = {541},
        pages = {481-487},
     abstract = {Reactive astrocytes are strongly induced by central nervous system (CNS) injury and disease, but their role is poorly understood. Here we show that a subtype of reactive astrocytes, which we termed A1, is induced by classically activated neuroinflammatory microglia. We show that activated microglia induce A1 astrocytes by secreting Il-1', TNF and C1q, and that these cytokines together are necessary and sufficient to induce A1 astrocytes. A1 astrocytes lose the ability to promote neuronal survival, outgrowth, synaptogenesis and phagocytosis, and induce the death of neurons and oligodendrocytes. Death of axotomized CNS neurons in vivo is prevented when the formation of A1 astrocytes is blocked. Finally, we show that A1 astrocytes are abundant in various human neurodegenerative diseases including Alzheimer's, Huntington's and Parkinson's disease, amyotrophic lateral sclerosis and multiple sclerosis. Taken together these findings help to explain why CNS neurons die after axotomy, strongly suggest that A1 astrocytes contribute to the death of neurons and oligodendrocytes in neurodegenerative disorders, and provide opportunities for the development of new treatments for these diseases.}
}
@book{Lieberman02,
       author = {Philip Lieberman},
        title = {Human language and our reptilian brain: The subcortical bases for speech, syntax and thought},
    publisher = {Harvard University Press},
      address = {Cambridge, MA},
         year = 2002,
}
@book{Lieberman2006toward,
        title = {Toward an Evolutionary Biology of Language},
       author = {Lieberman, Philip},
         year = {2006},
    publisher = {Belknap Press of Harvard University Press}
}
@book{Lieberman2013unpredictable,
        title = {The Unpredictable Species: What Makes Humans Unique},
       author = {Lieberman, Philip},
         year = {2013},
    publisher = {Princeton University Press}
}
@article{LiebermanAJPT-02,
     author = {Philip Lieberman},
      title = {On the nature and evolution of the neural bases of human language},
    Journal = {American Journal of Physical Anthropology},
      month = {December},
     volume = 119,
      issue = {S35},
      pages = {36-62},
       year = 2002,
   abstract = {The traditional theory equating the brain bases of language with Broca's and Wernicke's neocortical areas is wrong. Neural circuits linking activity in anatomically segregated populations of neurons in subcortical structures and the neocortex throughout the human brain regulate complex behaviors such as walking, talking, and comprehending the meaning of sentences. When we hear or read a word, neural structures involved in the perception or real-world associations of the word are activated as well as posterior cortical regions adjacent to Wernicke's area. Many areas of the neocortex and subcortical structures support the cortical-striatal-cortical circuits that confer complex syntactic ability, speech production, and a large vocabulary. However, many of these structures also form part of the neural circuits regulating other aspects of behavior. For example, the basal ganglia, which regulate motor control, are also crucial elements in the circuits that confer human linguistic ability and abstract reasoning. The cerebellum, traditionally associated with motor control, is active in motor learning. The basal ganglia are also key elements in reward-based learning. Data from studies of Broca's aphasia, Parkinson's disease, hypoxia, focal brain damage, and a genetically transmitted brain anomaly (the putative "language gene," family KE), and from comparative studies of the brains and behavior of other species, demonstrate that the basal ganglia sequence the discrete elements that constitute a complete motor act, syntactic process, or thought process. Imaging studies of intact human subjects and electrophysiologic and tracer studies of the brains and behavior of other species confirm these findings. As Dobzansky put it, "Nothing in biology makes sense except in the light of evolution" (cited in Mayr, 1982). That applies with as much force to the human brain and the neural bases of language as it does to the human foot or jaw. The converse follows: the mark of evolution on the brains of human beings and other species provides insight into the evolution of the brain bases of human language. The neural substrate that regulated motor control in the common ancestor of apes and humans most likely was modified to enhance cognitive and linguistic ability. Speech communication played a central role in this process. However, the process that ultimately resulted in the human brain may have started when our earliest hominid ancestors began to walk.},
} 
@article{LiebermanPLOS-15,
       author = {Lieberman, Philip},
      journal = {{PLOS} Biology},
    publisher = {Public Library of Science},
        title = {Language Did Not Spring Forth 100,000 Years Ago},
       volume = {13},
       number = {2},
         year = {2015},
        pages = {1-4},
     abstract = {In response to an Essay by Johan Bolhuis and co-authors, Phillip Lieberman contends that rather than arising from a key recent innovation ("merge"), language arose by gradual evolution of ancient capabilities.},
}
@inbook{LiebermanandMccarthy2013,
        title = {The Evolution of Speech and Language},
       author = {Lieberman, Philip and Mccarthy, Robert},
    booktitle = {Handbook of Paleoanthropology},
         year = {2013},
        pages = {1-41},
    publisher = {Springer-Verlag},
      address = {Berlin Heidelberg},
     abstract = {Human speech, language, and cognition derive from anatomy and neural mechanisms that have been shaped by the Darwinian process of natural selection acting on variation but that have roots present in other living species. Language did not suddenly arise 50,000-100,000 years ago through a mutation that yielded an innate "faculty of language" nor does the human brain include an organ devoted to language and language alone. Broca's area is not the center of language. Neural circuits linking local activity in different neural structures regulate complex behaviors. Neural circuits that were present in early mammal-like reptiles play a part in regulating laryngeal phonation, conveying both referential information and emotion. Speech plays a central role, enabling transmission of information at a rate that exceeds the auditory fusion frequency. The unique human tongue enhances the robustness of speech, but Neanderthals and other archaic hominins whose neck and skull proportions preclude their having an adult-like human tongue nevertheless could talk. Comparative studies of present-day apes suggest that hominin "protolanguage" lacking syntax never existed. The neural bases of human language are not domain-specific - in other words, they are not devoted to language alone. Mutations on the FOXP2 transcriptional gene shared by humans, Neanderthals, and at least one other archaic species enhanced synaptic plasticity in cortical-basal ganglia circuits that are implicated in motor behavior, cognitive flexibility, language, and associative learning. A selective sweep occurred about 200,000 years ago on a unique human version of this gene. Other transcriptional genes appear to be implicated in enhancing cortical-basal ganglia and other neural circuits.}
}
@inproceedings{LiederandGriffithsCS-16,
        title = {Helping people make better decisions using optimal gamification},
       author = {Falk Lieder and Tom Griffiths},
    booktitle = {Proceedings of the 2016 Annual Meeting of the Cognitive Science Society},
         year = {2016},
     abstract = {Game elements like points and levels are a popular tool to nudge and engage students and customers. Yet, no theory can  tell us which incentive structures work and how to design them.  Here we connect the practice of gamification to the theory of  reward shaping in reinforcement learning. We leverage this  connection to develop a method for designing effective incentive  structures and delineating when gamification will succeed  from when it will fail. We evaluate our method in two behavioral  experiments. The results of the first experiment demonstrate  that incentive structures designed by our method help  people make better, less short-sighted decisions and avoid the pitfalls of less principled approaches. The results of the second  experiment illustrate that such incentive structures can be  effectively implemented using game elements like points and  badges. These results suggest that our method provides a principled  way to leverage gamification to help people make better decisions.}
} 
@inproceedings{LieetalIJCAI-18,
       author = {Li, Jian and Wang, Yue and Lyu, Michael and King, Irwin},
        title = {Code Completion with Neural Attention and Pointer Networks},
    booktitle = {Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence},
         year = {2018},
        pages = {4159-4165},
     abstract = {Intelligent code completion has become an essential research task to accelerate modern software development. To facilitate effective code completion for dynamically-typed programming lan- guages, we apply neural language models by learn- ing from large codebases, and develop a tailored at- tention mechanism for code completion. However, standard neural language models even with attention mechanism cannot correctly predict the out-of-vocabulary (OoV) words that restrict the code completion performance. In this paper, inspired by the prevalence of locally repeated terms in program source code, and the recently proposed point- er copy mechanism, we propose a pointer mixture network for better predicting OoV words in code completion. Based on the context, the pointer mixture network learns to either generate a within-vocabulary word through an RNN component, or regenerate an OoV word from local context through a pointer component. Experiments on two bench- marked datasets demonstrate the effectiveness of our attention mechanism and pointer mixture net- work on the code completion task.},
}
@article{LienandScanzianiNATURE-NEUROSCIENCE-13,
       author = {Lien, Anthony D. and Scanziani, Massimo},
        title = {Tuned Thalamic Excitation is Amplified by Visual Cortical Circuits},
      journal = {Nature Neuroscience},
         year = 2013,
       volume = 16,
        issue = 9,
        pages = {1315-1323},
     abstract = {Cortical neurons in thalamic recipient layers receive excitation from the thalamus and the cortex. The relative contribution of these two sources of excitation to sensory tuning is poorly understood. Here we optogenetically silence the visual cortex of mice to isolate thalamic excitation onto layer 4 neurons during visual stimulation. Thalamic excitation contributes to a third of total excitation and is organized in spatially offset, yet overlapping ON and OFF receptive fields. This receptive field structure predicts the orientation tuning of thalamic excitation. Finally, thalamic and total excitation are similarly tuned to orientation and direction, and have the same temporal phase relationship to the visual stimulus. Our results indicate that tuning of thalamic excitation is unlikely to be imparted by direction or orientation selective thalamic neurons and that a principal role of cortical circuits is to amplify tuned thalamic excitation.},
}
@article{LietalBIORXIV-16,
       author = {Li, Yanjie and Ascoli, Giorgio and Mitra, Partha P and Wang, Yusu},
        title = {Metrics for comparing Neuronal Tree Shapes based on Persistent Homology},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Labs Journals},
         year = {2016},
     abstract = {The geometrical tree structures of axonal and dendritic processes play important roles in determining the architecture and capabilities of neuronal circuitry. Morphological features based on this tree structure have played a central role in classifying neurons for over a century. Yet geometrical trees are not automatically adapted to the basic mathematical tool used widely in data analysis, namely vector spaces and linear algebra, since tree geometries cannot be naturally added and subtracted. Current methods for analysis reduce trees to feature vectors in more or less ad hoc ways. A more natural mathematical object suited to characterizing neuronal tree geometries, is a metric space, where only distances between objects need be defined. In recent years, there have been significant developments in the fields of computational topology and geometry that promise to be useful for the analysis of neuronal geometries. In this paper, we adapt these tools to the problem of characterizing and analyzing neuronal morphology. As more and more neuroanatomical data are made available through efforts such as NeuroMorpho.org and FlyCircuit.org, the need to develop computational tools to facilitate automatic knowledge discovery from such large datasets becomes more urgent. One fundamental question is how best to compare neuron structures, for instance to organize and classify large collection of neurons. We aim to develop a flexible yet powerful framework to support comparison and classification of large collection of neuron structures efficiently. Specifically we propose to use a topological persistence-based feature vectorization framework. Existing methods to vectorize a neuron (i.e, convert a neuron to a feature vector so as to support efficient comparison and/or searching) typically rely on statistics or summaries of morphometric information, such as the average or maximum local torque angle or partition asymmetry. These simple summaries have limited power in encoding global tree structures. Leveraging recent development in topological data analysis, we vectorize each neuron structure into a simple yet informative summary via the use of topological persistence. In particular, each type of information of interest can be represented as a descriptor function defined on the neuron tree, which is then mapped to a simple persistence-signature. Our framework can encode both local and global tree structure, as well as other information of interest (electrophysiological or dynamical measures), by considering multiple descriptor functions on the neuron. The resulting persistence-based signature is potentially more informative than simple statistical summaries (such as average/mean/max) of morphometric quantities -- Indeed, we show that using a certain descriptor function will give a persistence-based signature containing strictly more information than the classical Sholl analysis. At the same time, our framework retains the efficiency associated with treating neurons as points in a simple Euclidean feature space, which would be important for constructing efficient searching or indexing structures over them. We present preliminary experimental results to demonstrate the effectiveness of our persistence-based neuronal feature vectorization framework.},
}
@article{LietalBIORXIV-19,
       author = {Li, Peter H. and Lindsey, Larry F. and Januszewski, Micha{\l} and Zheng, Zhihao and Bates, Alexander Shakeel and Taisz, Istv{\'a}n and Tyka, Mike and Nichols, Matthew and Li, Feng and Perlman, Eric and Maitin-Shepard, Jeremy and Blakely, Tim and Leavitt, Laramie and Jefferis, Gregory S.X.E. and Bock, Davi and Jain, Viren},
        title = {Automated Reconstruction of a Serial-Section EM Drosophila Brain with Flood-Filling Networks and Local Realignment},
         year = {2019},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Laboratory},
     abstract = {Reconstruction of neural circuitry at single-synapse resolution is an attractive target for improving understanding of the nervous system in health and disease. Serial section transmission electron microscopy (ssTEM) is among the most prolific imaging methods employed in pursuit of such reconstructions. We demonstrate how Flood-Filling Networks (FFNs) can be used to computationally segment a forty-teravoxel whole-brain Drosophila ssTEM volume. To compensate for data irregularities and imperfect global alignment, FFNs were combined with procedures that locally re-align serial sections and dynamically adjust image content. The proposed approach produced a largely merger-free segmentation of the entire ssTEM Drosophila brain, which we make freely available. As compared to manual tracing using an efficient skeletonization strategy, the segmentation enabled circuit reconstruction and analysis workflows that were an order of magnitude faster.},
}
@article{LietalBRAIN-BEHAVIOR-18,
       author = {Li, Y. and Hu, X. and Yu, Y. and Zhao, K. and Saalmann, Y. B. and Wang, L. },
        title = {{{F}eedback from human posterior parietal cortex enables visuospatial category representations as early as primary visual cortex}},
      journal = {Brain and Behavior},
         year = {2018},
       volume = {8},
       number = {1},
        pages = {e00886},
     abstract = {Categorization is a fundamental cognitive process, whereby the brain assigns meaning to sensory stimuli. Previous studies have found category representations in prefrontal cortex and posterior parietal cortex (PPC). However, these higher-order areas lack the fine-scale spatial representations of early sensory areas, and it remains unclear what mechanisms enable flexible categorization based on fine-scale features. In this study, we decoded functional MRI signals and measured causal influences, across visual, parietal, and prefrontal cortex from participants performing categorization based on coarse- or fine-scale spatial information in thirteen healthy adults. We show that category information based on coarse discriminations was represented in the PPC, in the intraparietal sulcus region, IPS1/2, at an early stage of categorization trials, whereas representations of category information based on fine-scale discriminations formed later during interactions between IPS1/2 and primary visual cortex (V1). Specifically, when fine-scale discriminations were necessary, we decoded significant category information from V1 at an intermediate stage of trials and again from IPS1/2 at a late stage. IPS1/2 feedback was critical, because categorization performance improved as causal influence from IPS1/2 to V1 increased. Further, these mechanisms were plastic, as the selectivity of IPS1/2 and V1 responses shifted markedly with retraining to categorize the same stimuli into two new groups. Our findings suggest that reentrant processing between the PPC and visual cortex enables flexible abstraction of category information.}
}
@inproceedings{LietalCOSDI-14,
       author = {Li, Mu and Andersen, David G. and Park, Jun Woo and Smola, Alexander J. and Ahmed, Amr and Josifovski, Vanja and Long, James and Shekita, Eugene J. and Su, Bor-Yiing},
        title = {Scaling Distributed Machine Learning with the Parameter Server},
    booktitle = {Proceedings of the 11th {USENIX} Conference on Operating Systems Design and Implementation},
    publisher = {{USENIX} Association},
      address = {Berkeley, CA, USA},
         year = {2014},
        pages = {583-598},
     abstract = {We propose a parameter server framework for distributed machine learning problems. Both data and workloads are distributed over worker nodes, while the server nodes maintain globally shared parameters, represented as dense or sparse vectors and matrices. The framework manages asynchronous data communication between nodes, and supports flexible consistency models, elastic scalability, and continuous fault tolerance. To demonstrate the scalability of the proposed framework, we show experimental results on petabytes of real data with billions of examples and parameters on problems ranging from Sparse Logistic Regression to Latent Dirichlet Allocation and Distributed Sketching.},
}
@article{LietalCoRR-15,
       author = {Jiwei Li and Minh-Thang Luong and Dan Jurafsky},
        title = {A Hierarchical Neural Autoencoder for Paragraphs and Documents},
      journal = {CoRR},
       volume = {arXiv:1506.01057},
         year = {2015},
     abstract = {Natural language generation of coherent long texts like paragraphs or longer documents is a challenging problem for recurrent networks models. In this paper, we explore an important step toward this generation task: training an LSTM (Long-short term memory) auto-encoder to preserve and reconstruct multi-sentence paragraphs. We introduce an LSTM model that hierarchically builds an embedding for a paragraph from embeddings for sentences and words, then decodes this embedding to reconstruct the original paragraph. We evaluate the reconstructed paragraph using standard metrics like ROUGE and Entity Grid, showing that neural models are able to encode texts in a way that preserve syntactic, semantic, and discourse coherence. While only a first step toward generating coherent text units from neural models, our work has the potential to significantly impact natural language generation and summarization.}
}
@article{LietalCoRR-17,
       author = {Xiujun Li and Yun-Nung Chen and Lihong Li and Jianfeng Gao},
        title = {End-to-End Task-Completion Neural Dialogue Systems},
      journal = {CoRR},
       volume = {arXiv:1703.01008},
         year = {2017},
     abstract = {One of the major drawbacks of modularized task-completion dialogue systems is that each module is trained individually, which presents several challenges. For example, downstream modules are affected by earlier modules, and the performance of the entire system is not robust to the accumulated errors. This paper presents a novel end-to-end learning framework for task-completion dialogue systems to tackle such issues. Our neural dialogue system can directly interact with a structured database to assist users in accessing information and accomplishing certain tasks. The reinforcement learning based dialogue manager offers robust capabilities to handle noises caused by other components of the dialogue system. Our experiments in a movie-ticket booking domain show that our end-to-end system not only outperforms modularized dialogue system baselines for both objective and subjective evaluation, but also is robust to noises as demonstrated by several systematic experiments with different error granularity and rates specific to the language understanding module.}
} 
@article{LietalCoRR-17c,
       author = {Jian Li and Yue Wang and Irwin King and Michael R. Lyu},
        title = {Code Completion with Neural Attention and Pointer Networks},
      journal = {CoRR},
       volume = {arXiv:1711.09573},
         year = {2017},
     abstract = {Intelligent code completion has become an essential research task to accelerate modern software development. To facilitate effective code completion for dynamically-typed programming languages, we apply neural language models by learning from large codebases, and develop a tailored attention mechanism for code completion. However, standard neural language models even with attention mechanism cannot correctly predict the out-of-vocabulary (OoV) words that restrict the code completion performance. In this paper, inspired by the prevalence of locally repeated terms in program source code, and the recently proposed pointer copy mechanism, we propose a pointer mixture network for better predicting OoV words in code completion. Based on the context, the pointer mixture network learns to either generate a within-vocabulary word through an RNN component, or regenerate an OoV word from local context through a pointer component. Experiments on two benchmarked datasets demonstrate the effectiveness of our attention mechanism and pointer mixture network on the code completion task.}
}
@article{LietalCoRR-20,
        title = {Strong Generalization and Efficiency in Neural Programs},
       author = {Yujia Li and Felix Gimeno and Pushmeet Kohli and Oriol Vinyals},
       volume = {arXiv:2007.03629},
         year = {2020},
      journal = {CoRR},
     abstract = {We study the problem of learning efficient algorithms that strongly generalize in the framework of neural program induction. By carefully designing the input / output interfaces of the neural model and through imitation, we are able to learn models that produce correct results for arbitrary input sizes, achieving strong generalization. Moreover, by using reinforcement learning, we optimize for program efficiency metrics, and discover new algorithms that surpass the teacher used in imitation. With this, our approach can learn to outperform custom-written solutions for a variety of problems, as we tested it on sorting, searching in ordered lists and the NP-complete 0/1 knapsack problem, which sets a notable milestone in the field of Neural Program Induction. As highlights, our learned model can perform sorting perfectly on any input data size we tested on, with O(nlogn) complexity, whilst outperforming hand-coded algorithms, including quick sort, in number of operations even for list sizes far beyond those seen during training.}
}
@inproceedings{LietalICLR-18,
        title = {Learning Deep Generative Models of Graphs},
       author = {Yujia Li and Oriol Vinyals and Chris Dyer and Razvan Pascanu and Peter Battaglia},
    booktitle = {International Conference on Learning Representations},
         year = {2018},
     abstract = {Graphs are fundamental data structures required to model many important real-world data, from knowledge graphs, physical and social interactions to molecules and proteins. In this paper, we study the problem of learning generative models of graphs from a dataset of graphs of interest. After learning, these models can be used to generate samples with similar properties as the ones in the dataset.  Such models can be useful in a lot of applications, e.g. drug discovery and knowledge graph construction. The task of learning generative models of graphs, however, has its unique challenges. In particular, how to handle symmetries in graphs and ordering of its elements during the generation process are important issues. We propose a generic graph neural net based model that is capable of generating any arbitrary graph.  We study its performance on a few graph generation tasks compared to baselines that exploit domain knowledge.  We discuss potential issues and open problems for such generative models going forward. TL;DR: We study the graph generation problem and propose a powerful deep generative model capable of generating arbitrary graphs.},
}
@article{LietalNATURE-08,
       author = {Li, Mo and Pernice, W. H. P. and Xiong, C. and Baehr-Jones, T. and Hochberg, M. and Tang, H. X.},
        title = {Harnessing optical forces in integrated photonic circuits},
      journal = {Nature},
    publisher = {Macmillan Publishers Limited},
       volume = 456,
        issue = 7221,
         year = 2008,
        pages = {480-484},
}
@article{LietalNATURE-16,
       author = {Li, Nuo and Daie, Kayvon and Svoboda, Karel and Druckmann, Shaul},
        title = {Robust neuronal dynamics in premotor cortex during motor planning},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = {532},
        issue = {7600},
         year = {2016},
        pages = {459-464},
     abstract = {Neural activity maintains representations that bridge past and future events, often over many seconds. Network models can produce persistent and ramping activity, but the positive feedback that is critical for these slow dynamics can cause sensitivity to perturbations. Here we use electrophysiology and optogenetic perturbations in the mouse premotor cortex to probe the robustness of persistent neural representations during motor planning. We show that preparatory activity is remarkably robust to large-scale unilateral silencing: detailed neural dynamics that drive specific future movements were quickly and selectively restored by the network. Selectivity did not recover after bilateral silencing of the premotor cortex. Perturbations to one hemisphere are thus corrected by information from the other hemisphere. Corpus callosum bisections demonstrated that premotor cortex hemispheres can maintain preparatory activity independently. Redundancy across selectively coupled modules, as we observed in the premotor cortex, is a hallmark of robust control systems. Network models incorporating these principles show robustness that is consistent with data.},
}
@article{LietalNEURON-14,
     title = {Retinal Representation of the Elementary Visual Signal},
    author = {Peter H. Li and Greg D. Field and Martin Greschner and Daniel Ahn and Deborah E. Gunning and Keith Mathieson and Alexander Sher and Alan M. Litke and E.J. Chichilnisky},
   journal = {Neuron},
    volume = 81,
     issue = 1,
     pages = {130-139},
      year = 2014,
  abstract = {The propagation of visual signals from individual cone photoreceptors through parallel neural circuits was examined in the primate retina. Targeted stimulation of individual cones was combined with simultaneous recording from multiple retinal ganglion cells of identified types. The visual signal initiated by an individual cone produced strong responses with different kinetics in three of the four numerically dominant ganglion cell types. The magnitude and kinetics of light responses in each ganglion cell varied nonlinearly with stimulus strength but in a manner that was independent of the cone of origin after accounting for the overall input strength of each cone. Based on this property of independence, the receptive field profile of an individual ganglion cell could be well estimated from responses to stimulation of each cone individually. Together, these findings provide a quantitative account of how elementary visual inputs form the ganglion cell receptive field.}
}
@article{LietalNEUROSCIENCE-15,
       author = {Peter H. Li and Gauthier, J.L. and Schiff, M. and Sher, A, and Ahn, D. and Field, G.D. and Greschner, M.5. and Callaway, E.M. and Alan M. Litke and E.J. Chichilnisky},
        title = {Anatomical identification of extracellularly recorded cells in large-scale multielectrode recordings},
      journal = {Journal Neuroscience},
         year = 2015,
       volume = 35,
        issue = 11,
        pages = {4663-4675},
     abstract = {This study combines for the first time two major approaches to understanding the function and structure of neural circuits: large-scale multielectrode recordings, and confocal imaging of labeled neurons. To achieve this end, we develop a novel approach to the central problem of anatomically identifying recorded cells, based on the electrical image: the spatiotemporal pattern of voltage deflections induced by spikes on a large-scale, high-density multielectrode array. Recordings were performed from identified ganglion cell types in the macaque retina. Anatomical images of cells in the same preparation were obtained using virally transfected fluorescent labeling or by immunolabeling after fixation. The electrical image was then used to locate recorded cell somas, axon initial segments, and axon trajectories, and these signatures were used to identify recorded cells. Comparison of anatomical and physiological measurements permitted visualization and physiological characterization of numerically dominant ganglion cell types with high efficiency in a single preparation.}
}
@inproceedings{LietalNIPS-10,
     title = {Object Bank: A High-Level Image Representation for Scene Classification \& Semantic Feature Sparsification},
    author = {Li-Jia Li, Hao Su, Eric P. Xing and Li Fei-Fei},
 booktitle = {Neural Information Processing Systems (NIPS)},
      year = 2010,
   address = {Vancouver, Canada}
} 
@article{LietalNN-04,
    author = {Wu Li and Valentin Pi\"{e}ch and Charles D Gilbert},
     title = {Perceptual learning and top-down influences in primary visual cortex},
   journal = {Nature Neuroscience},
    volume = 7,
     issue = 6,
     pages = {651-657},
      year = 2004,
}
@article{LietalPLoS-10,
      title = {Visualizing the distribution of synapses from individual neurons in the mouse brain},
     author = {Li, L. and Tasic, B. and Micheva, K.D. and Ivanov, V.M. and Spletter, M.L. and Smith, S.J. and Luo, L.},
    journal = {{PLoS} Biology},
  publisher = {Public Library of Science},
     volume = 5,
      issue = 7,
       year = 2010,
      pages = {e11503},
}
@article{LietalPLoS-15,
       author = {Li, Qingjiang and Serb, Alexander and Prodromakis, Themistoklis and Xu, Hui},
       editor = {Cymbalyuk, Gennady},
        title = {A Memristor {SPICE} Model Accounting for Synaptic Activity Dependence},
      journal = {{PLoS} ONE},
         year = 2015,
    publisher = {Public Library of Science},
      address = {San Francisco, CA USA},
       volume = 10,
        issue = 3,
        pages = {e0120506},
     abstract = {In this work, we propose a new memristor SPICE model that accounts for the typical synaptic characteristics that have been previously demonstrated with practical memristive devices. We show that this model could account for both volatile and non-volatile memristance changes under distinct stimuli. We then demonstrate that our model is capable of supporting typical STDP with simple non-overlapping digital pulse pairs. Finally, we investigate the capability of our model to simulate the activity dependence dynamics of synaptic modification and present simulated results that are in excellent agreement with biological results.},
}
@inproceedings{LifschitzAAAI-86,
       author = {Lifschitz, Vladimir},
        title = {Pointwise Circumscription: preliminary report},
    booktitle = {Proceedings AAAI-86},
 organization = {AAAI},
     location = {Philadelphia, Pennsylvania},
         year = 1986,
        pages = {406-410},
}
@inproceedings{LifschitzACTION,
       author = {Lifschitz, Vladimir},
        title = {Formal Theories of Action},
    booktitle = {Proceedings of the 1987 Workshop on the Frame Problem in Artificial Intelligence},
     location = {Lawrence, Kansas},
         year = 1987,
}
@article{LifschitzAIJ-89,
       author = {Lifschitz, Vladimir and Rabinov, Arkady},
        title = {Miracles in Formal Theories of Action},
      journal = {Artificial Intelligence},
       volume = 38,
         year = 1989,
        pages = {225-237},
}
@inproceedings{LifschitzIJCAI-87,
       author = {Lifschitz, Vladimir},
        title = {Formal Theories of Action},
    booktitle = {Proceedings {IJCAI} 10},
     location = {Milan, Italy},
 organization = {IJCAII},
         year = 1987,
        pages = {966-972},
} 
@incollection{LifschitzPW-86,
       author = {Lifschitz, Vladimir},
        title = {On the Semantics of {STRIPS}},
     crossref = {GeorgeffandLanskyPW-86},
        pages = {1-10},
} 
@article{LillicrapetalCoRR-14,
       author = {Timothy P. Lillicrap and Daniel Cownden and Douglas B. Tweed and Colin J. Akerman},
        title = {Random feedback weights support learning in deep neural networks},
      journal = {CoRR},
       volume = {arXiv:1411.0247},
         year = {2014},
     abstract = {The brain processes information through many layers of neurons. This deep architecture is representationally powerful, but it complicates learning by making it hard to identify the responsible neurons when a mistake is made. In machine learning, the backpropagation algorithm assigns blame to a neuron by computing exactly how it contributed to an error. To do this, it multiplies error signals by matrices consisting of all the synaptic weights on the neuron's axon and farther downstream. This operation requires a precisely choreographed transport of synaptic weight information, which is thought to be impossible in the brain. Here we present a surprisingly simple algorithm for deep learning, which assigns blame by multiplying error signals by random synaptic weights. We show that a network can learn to extract useful information from signals sent through these random feedback connections. In essence, the network learns to learn. We demonstrate that this new mechanism performs as quickly and accurately as backpropagation on a variety of problems and describe the principles which underlie its function. Our demonstration provides a plausible basis for how a neuron can be adapted using error signals generated at distal locations in the brain, and thus dispels long-held assumptions about the algorithmic constraints on learning in neural circuits.}
} 
@article{LillicrapetalCoRR-15,
        title = {Continuous control with deep reinforcement learning},
       author = {Timothy P. Lillicrap and Jonathan J. Hunt and Alexander Pritzel and Nicolas Heess and Tom Erez and Yuval Tassa and David Silver and Daan Wierstra},
         year = {2015},
      journal = {CoRR},
       volume = {arXiv:1509.02971},
         year = {2015},
     abstract = {We adapt the ideas underlying the success of Deep Q-Learning to the continuous action domain. We present an actor-critic, model-free algorithm based on the deterministic policy gradient that can operate over continuous action spaces. Using the same learning algorithm, network architecture and hyper-parameters, our algorithm robustly solves more than 20 simulated physics tasks, including classic problems such as cartpole swing-up, dexterous manipulation, legged locomotion and car driving. Our algorithm is able to find policies whose performance is competitive with those found by a planning algorithm with full access to the dynamics of the domain and its derivatives. We further demonstrate that for many of the tasks the algorithm can learn policies end-to-end: directly from raw pixel inputs.},
}
@article{LillicrapetalNATURE-COMMUNICATIONS-16,
       author = {Lillicrap, Timothy P. and Cownden, Daniel and Tweed, Douglas B. and Akerman, Colin J.},
        title = {Random synaptic feedback weights support error backpropagation for deep learning},
    publisher = {Nature Publishing Group},
      journal = {Nature Communications},
       volume = {7},
         year = {2016},
        pages = {13276},
     abstract = {The brain processes information through multiple layers of neurons. This deep architecture is representationally powerful, but complicates learning because it is difficult to identify the responsible neurons when a mistake is made. In machine learning, the backpropagation algorithm assigns blame by multiplying error signals with all the synaptic weights on each neuron's axon and further downstream. However, this involves a precise, symmetric backward connectivity pattern, which is thought to be impossible in the brain. Here we demonstrate that this strong architectural constraint is not required for effective error propagation. We present a surprisingly simple mechanism that assigns blame by multiplying errors by even random synaptic weights. This mechanism can transmit teaching signals across multiple layers of neurons and performs as effectively as backpropagation on a variety of tasks. Our results help reopen questions about how the brain could use error signals and dispel long-held assumptions about algorithmic constraints on learning.}
}
@article{LillicrapetalNATURE_COMMUNICATIONS-16,
       author = {Lillicrap, Timothy P. and Cownden, Daniel and Tweed, Douglas B. and Akerman, Colin J.},
        title = {Random synaptic feedback weights support error backpropagation for deep learning},
      journal = {Nature Communications},
         year = {2016},
       volume = {7},
        issue = {1},
        pages = {13276},
     abstract = {The brain processes information through multiple layers of neurons. This deep architecture is representationally powerful, but complicates learning because it is difficult to identify the responsible neurons when a mistake is made. In machine learning, the backpropagation algorithm assigns blame by multiplying error signals with all the synaptic weights on each neuron's axon and further downstream. However, this involves a precise, symmetric backward connectivity pattern, which is thought to be impossible in the brain. Here we demonstrate that this strong architectural constraint is not required for effective error propagation. We present a surprisingly simple mechanism that assigns blame by multiplying errors by even random synaptic weights. This mechanism can transmit teaching signals across multiple layers of neurons and performs as effectively as backpropagation on a variety of tasks. Our results help reopen questions about how the brain could use error signals and dispel long-held assumptions about algorithmic constraints on learning.},
} 
@article{LimetalCELL-12,
       author = {Lim, W. A. and Alvania, R. and Marshall, W. F.},
        title = {Cell biology 2.0},
      journal = {Trends Cell Biololgy},
         year = 2012,
       volume = 22,
        issue = 12,
        pages = {611-612},
} 
@article{LimetalCELL-13,
       author = {Lim, W. A. and Lee, C. M. and Tang, C.},
        title = {Design principles of regulatory networks: searching for the molecular algorithms of the cell},
      journal = {Molecular Cell},
         year = 2013,
       volume = 49,
        issue = 2,
        pages = {202-212},
}
@article{Lin91,
       author = {Jianhua Lin},
        title = {Divergence measures based on the {S}hannon entropy},
      journal = {IEEE Transactions on Information Theory},
       volume = 37,
        issue = 1,
        pages = {145-151},
         year = 1991,
}
@inproceedings{LinAAAI-91,
       author = {Lin, Long-Ji},
        title = {Programming Robots Using Reinforcement Learning and Teaching},
    booktitle = {Proceedings AAAI-91},
     location = {Anaheim, California},
 organization = {AAAI},
         year = 1991,
}
@inproceedings{LinACL-98,
       author = {Lin, Dekang},
        title = {Automatic Retrieval and Clustering of Similar Words},
    booktitle = {Proceedings of the 36th Annual Meeting of the Association for Computational Linguistics and 17th International Conference on Computational Linguistics - Volume 2},
     location = {Montreal, Quebec, Canada},
    publisher = {Association for Computational Linguistics},
      address = {Stroudsburg, PA, USA},
         year = {1998},
        pages = {768-774},
     abstract = {Bootstrapping semantics from text is one of the greatest challenges in natural language learning. We first define a word similarity measure based on the distributional pattern of words. The similarity measure allows us to construct a thesaurus using a parsed corpus. We then present a new evaluation methodology for the automatically constructed thesaurus. The evaluation results show that the thesaurns is significantly closer to WordNet than Roget Thesaurus is.}
}
@article{LinML-92,
       author = {Lin, Long-Ji},
        title = {Self-improving reactive agents based on reinforcement learning, planning and teaching},
      journal = {Machine Learning},
         year = {1992},
       volume = {8},
       number = {3},
        pages = {293-321},
     abstract = {To date, reinforcement learning has mostly been studied solving simple learning tasks. Reinforcement learning methods that have been studied so far typically converge slowly. The purpose of this work is thus two-fold: 1) to investigate the utility of reinforcement learning in solving much more complicated learning tasks than previously studied, and 2) to investigate methods that will speed up reinforcement learning.},
}
@phdthesis{LinPhD,
       author = {Lin, Shieu-Hong},
        title = {Exploiting Structure for Planning and Control},
       school = {Department of Computer Science, Brown University},
         year = 1997,
}
@article{LinandDeanCIJ-96,
        title = {Localized Temporal Reasoning Using Subgoals and Abstract Events},
       author = {Lin, Shieu-Hong and Dean, Thomas},
      journal = CIJ,
       volume = {12},
        issue = {3},
        pages = {423-449},
         year = {1996},
} 
@inproceedings{LinandDeanEWSP-94,
       author = {Lin, Shieu-Hong and Dean, Thomas},
        title = {Exploiting Locality in Temporal Reasoning},
       editor = {Sandewall, E. and Backstrom, C.},
    booktitle = {Current Trends in {AI} Planning},
    publisher = {{IOS} Press},
      address = {Amsterdam},
         year = 1994,
}
@inproceedings{LinandDeanEWSP-95,
        title = {Generating Optimal Policies for High-Level Plans with Conditional Branches and Loops},
       author = {Lin, Shieu-Hong and Dean, Thomas},
    booktitle = {Proceedings of the 3rd European Workshop on Planning},
     location = {Assisi, Italy},
        pages = {205-218},
         year = 1995,
}
@inproceedings{LinandDeanTIME-94,
      author = {Lin, Shieu-Hong and Dean, Thomas},
       title = {Temporal Reasoning: {A} State-based approach},
      editor = {S.D. Goodwin and H.J. Hamilton},
   booktitle = {Proceedings of the First International Workshop on Temporal Representation and Reasoning ({TIME}-94)},
    location = {Pensacala, Florida},
       pages = {3-12},
       month = {May},
        year = 1994,
}
@article{LinandKernighanOR-73,
       author = {Lin, S. and Kernighan, B. W.},
        title = {An Effective Heuristic for the Travelling Salesman Problem},
      journal = {Operations Research},
       volume = 21,
         year = 1973,
        pages = {498-516},
} 
@inproceedings{LinandPantelKDD-01,
       author = {Dekang Lin and Patrick Pantel},
        title = {{DIRT} - Discovery of Inference Rules from Text},
    booktitle = {Proceedings of the {ACM SIGKDD} Conference on Knowledge Discovery and Data Mining},
         year = 2001,
        pages = {323-328}
} 
@article{LinandParikhCoRR-15,
       author = {Xiao Lin and Devi Parikh},
        title = {Don't Just Listen, Use Your Imagination: Leveraging Visual Common Sense for Non-Visual Tasks},
      journal = {CoRR},
       volume = {arXiv:1502.06108},
         year = {2015},
          url = {http://arxiv.org/abs/1502.06108},
     abstract = {Artificial agents today can answer factual questions. But they fall short on questions that require common sense reasoning. Perhaps this is because most existing common sense databases rely on text to learn and represent knowledge. But much of common sense knowledge is unwritten - partly because it tends not to be interesting enough to talk about, and partly because some common sense is unnatural to articulate in text. While unwritten, it is not unseen. In this paper we leverage semantic common sense knowledge learned from images - i.e. visual common sense - in two textual tasks: fill-in-the-blank and visual paraphrasing. We propose to "imagine" the scene behind the text, and leverage visual cues from the "imagined" scenes in addition to textual cues while answering these questions. We imagine the scenes as a visual abstraction. Our approach outperforms a strong text-only baseline on these tasks. Our proposed tasks can serve as benchmarks to quantitatively evaluate progress in solving tasks that go "beyond recognition". Our code and datasets are publicly available.}
}
@inproceedings{LinandShohamAAAI-91,
       author = {Lin, Fangzhen and Shoham, Yoav},
        title = {Provably Correct Theories of Action},
    booktitle = {Proceedings AAAI-91},
     location = {Anaheim, California},
 organization = {AAAI},
         year = 1991,
        pages = {349-354},
}
@inproceedings{LinandVitterSTOC-92,
       author = {Lin, Jyh-Han and Vitter, Jeffery Scott},
        title = {$\epsilon$-Approximations with Minimum Packing Constraint Violation},
    booktitle = {Proceedings of the 24th Annual {ACM} Symposium on Theory of Computing},
     location = {Victoria, Canada},
         year = 1992,
}
@article{LindebergIJCV-93,
       author = {Lindeberg, Tony},
        title = {Detecting Salient Blob-Like Image Structures and Their Scales with a Scale-Space Primal Sketch: A Method for Focus-of-Attention},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 11,
        issue = 3,
        pages = {283-318},
         year = 1993,
}
@article{LindebergIJCV-98,
       author = {Tony Lindeberg},
        title = {Feature Detection with Automatic Scale Selection},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
         year = 1998,
       volume = 30,
        pages = {79-116}
}
@book{Lindley80,
       author = {Lindley, D. V.},
        title = {Introduction to Probability and Statistics},
    publisher = {Cambridge University Press},
         year = 1980,
}
@article{LinetalCELL-REPORTS-13,
       author = {Lin, Chih-Yung and Chuang, Chao-Chun and Hua, Tzu-En and Chen, Chun-Chao and Dickson, Barry J. and Greenspan, Ralph J. and Chiang, Ann-Shyn},
        title = {A Comprehensive Wiring Diagram of the Protocerebral Bridge for Visual Information Processing in the <em>Drosophila</em> Brain},
      journal = {Cell Reports},
    publisher = {Elsevier},
       volume = {3},
        issue = {5},
         year = {2013},
        pages = {1739-1753},
     abstract = {How the brain perceives sensory information and generates meaningful behavior depends critically on its underlying circuitry. The protocerebral bridge (PB) is a major part of the insect central complex (CX), a premotor center that may be analogous to the human basal ganglia. Here, by deconstructing hundreds of PB single neurons and reconstructing them into a common three-dimensional framework, we have constructed a comprehensive map of PB circuits with labeled polarity and predicted directions of information flow. Our analysis reveals a highly ordered information processing system that involves directed information flow among CX subunits through 194 distinct PB neuron types. Circuitry properties such as mirroring, convergence, divergence, tiling, reverberation, and parallel signal propagation were observed; their functional and evolutional significance is discussed. This layout of PB neuronal circuitry may provide guidelines for further investigations on transformation of sensory (e.g., visual) input into locomotor commands in fly brains.},
}
@article{LinetalCoRR-18,
       author = {Xi Victoria Lin and Chenglong Wang and Luke Zettlemoyer and Michael D. Ernst},
        title = {{NL2Bash}: {A} Corpus and Semantic Parser for Natural Language Interface to the Linux Operating System},
      journal = {CoRR},
       volume = {arXiv:1802.08979},
         year = {2018},
     abstract = {We present new data and semantic parsing methods for the problem of mapping English sentences to Bash commands (NL2Bash). Our long-term goal is to enable any user to perform operations such as file manipulation, search, and application-specific scripting by simply stating their goals in English. We take a first step in this domain, by providing a new dataset of challenging but commonly used Bash commands and expert-written English descriptions, along with baseline methods to establish performance levels on this task.}
}
@inproceedings{LinetalECCV-18,
       author = {Lin, Di and Ji, Yuanfeng and Lischinski, Dani and Cohen-Or, Daniel and Huang, Hui},
        title = {Multi-Scale Context Intertwining for Semantic Segmentation},
    booktitle = {The European Conference on Computer Vision (ECCV)},
         year = {2018},
        pages = {603-619},
     abstract = {Accurate semantic image segmentation requires the joint consideration of local appearance, semantic information, and global scene context. In today's age of pre-trained deep networks and their powerful convolutional features, state-of-the-art semantic segmentation approaches differ mostly in how they choose to combine together these different kinds of information. In this work, we propose a novel scheme for aggregating features from different scales, which we refer to as Multi-Scale Context Intertwining (MSCI). In contrast to previous approaches, which typically propagate information between scales in a one-directional manner, we merge pairs of feature maps in a bidirectional and recurrent fashion, via connections between two LSTM chains. By training the parameters of the LSTM units on the segmentation task, the above approach learns how to extract powerful and effective features for pixel-level semantic segmentation, which are then combined hierarchically. Furthermore, rather than using fixed information propagation routes, we subdivide images into super-pixels, and use the spatial relationship between them in order to perform image-adapted context aggregation. Our extensive evaluation on public benchmarks indicates that all of the aforementioned components of our approach increase the effectiveness of information propagation throughout the network, and significantly improve its eventual segmentation accuracy.}
}
@inproceedings{LinetalICCV-09,
     author = {Zhe Lin and Zhuolin Jiang and Larry S. Davis},
      title = {Recognizing Actions by Shape-Motion Prototype Trees},
  booktitle = {{IEEE} 12th International Conference on Computer Vision},
      pages = {444-451},
       year = 2009,
}
@article{LinetalNATURE-12,
        title = {Submicrometre geometrically encoded fluorescent barcodes self-assembled from {DNA}},
       author = {Lin, Chenxiang and Jungmann, Ralf and Leifer, Andrew M. and Li, Chao and Levner, Daniel and Church, George M. and Shih, William M. and Yin, Peng},
      journal = {Nature Chemistry},
    publisher = {Nature Publishing Group},
       volume = 4,
        issue = 10,
         year = 2012,
        pages = {832-839},
     abstract = {The identification and differentiation of a large number of distinct molecular species with high temporal and spatial resolution is a major challenge in biomedical science. Fluorescence microscopy is a powerful tool, but its multiplexing ability is limited by the number of spectrally distinguishable fluorophores. Here, we used (deoxy)ribonucleic acid (DNA)-origami technology to construct submicrometre nanorods that act as fluorescent barcodes. We demonstrate that spatial control over the positioning of fluorophores on the surface of a stiff DNA nanorod can produce 216 distinct barcodes that can be decoded unambiguously using epifluorescence or total internal reflection fluorescence microscopy. Barcodes with higher spatial information density were demonstrated via the construction of super-resolution barcodes with features spaced by \~40 nm. One species of the barcodes was used to tag yeast surface receptors, which suggests their potential applications as in situ imaging probes for diverse biomolecular and cellular entities in their native environments.}
}
@article{LinetalNEURON-15,
       author = {I-Chun Lin and Michael Okun and Matteo Carandini and Kenneth D. Harris},
        title = {The Nature of Shared Cortical Variability},
      journal = {Neuron},
       volume = {87},
       number = {3},
         year = {2015},
        pages = {644-656},
     abstract = {Summary Neuronal responses of sensory cortex are highly variable, and this variability is correlated across neurons. To assess how variability reflects factors shared across a neuronal population, we analyzed the activity of many simultaneously recorded neurons in visual cortex. We developed a simple model that comprises two sources of shared variability: a multiplicative gain, which uniformly scales each neuron's sensory drive, and an additive offset, which affects different neurons to different degrees. This model captured the variability of spike counts and reproduced the dependence of pairwise correlations on neuronal tuning and stimulus orientation. The relative contributions of the additive and multiplicative fluctuations could vary over time and had marked impact on population coding. These observations indicate that shared variability of neuronal populations in sensory cortex can be largely explained by two factors that modulate the whole population.}
}
@techreport{LinetalUW-CSE-TR-17,
       author = {Xi Victoria Lin and Chenglong Wang and Deric Pang and Kevin Vu and Luke Zettlemoyer and Michael D. Ernst},
        title = {Program synthesis from natural language using recurrent neural networks},
  institution = {University of Washington Department of Computer Science and Engineering},
       number = {UW-CSE-17-03-01},
      address = {Seattle, WA, USA},
         year = {2017},
     abstract = {Oftentimes, a programmer may have difficulty implementing a desired operation. Even when the programmer can describe her goal in English, it can be difficult to translate into code. Existing resources, such as question-and-answer websites, tabulate specific operations that someone has wanted to perform in the past, but they are not effective in generalizing to new tasks, to compound tasks that require combining previous questions, or sometimes even to variations of listed tasks. Our goal is to make programming easier and more productive by letting programmers use their own words and concepts to express the intended operation, rather than forcing them to accommodate the machine by memorizing its grammar. We have built a system that lets a programmer describe a desired operation in natural language, then automatically translates it to a programming language for review and approval by the programmer. Our system, Tellina, does the translation using recurrent neural networks (RNNs), a state-of-the-art natural language processing technique that we augmented with slot (argument) filling and other enhancements. We evaluated Tellina in the context of shell scripting. We trained Tellina's RNNs on textual descriptions of file system operations and bash one-liners, scraped from the web. Although recovering completely correct commands is challenging, Tellina achieves top-3 accuracy of 80\% for producing the correct command structure. In a controlled study, programmers who had access to Tellina outperformed those who did not, even when Tellina's predictions were not completely correct, to a statistically significant degree.}
}
@article{LinnetalTCAS-14,
       author = {Linn, E. and Siemon, A. and Waser, R. and Menzel, S.},
      journal = {Circuits and Systems I: Regular Papers, IEEE Transactions on},
        title = {Applicability of Well-Established Memristive Models for Simulations of Resistive Switching Devices},
         year = 2014,
       volume = 61,
        issue = 8,
        pages = {2402-2410},
     abstract = {Highly accurate and predictive models of resistive switching devices are needed to enable future memory and logic design. Widely used is the memristive modeling approach considering resistive switches as dynamical systems. Here we introduce three evaluation criteria for memristor models, checking for plausibility of the I-V characteristics, the presence of a sufficiently nonlinearity of the switching kinetics, and the feasibility of predicting the behavior of two antiserially connected devices correctly. We analyzed two classes of models: the first class comprises common linear memristor models and the second class widely used nonlinear memristive models. The linear memristor models are based on Strukov's initial memristor model extended by different window functions, while the nonlinear models include Pickett's physics-based memristor model and models derived thereof. This study reveals lacking predictivity of the first class of models, independent of the applied window function. Only the physics-based model is able to fulfill most of the basic evaluation criteria.},
}
@phdthesis{LippmanTR-86,
       author = {Lippman, Alan F.},
        title = {A Maximum Entropy Method for Expert System Construction},
       school = {Brown University},
         year = 1986,
}
@book{LipsonandKurman2013,
        title = {Fabricated: The New World of {3D} Printing, {1st Edition}},
       author = {Hod Lipson and Melba Kurman},
         year = {2013},
    publisher = {Wiley}
}        
@article{LismanNEURON-15,
        title = {The Challenge of Understanding the Brain: Where We Stand in 2015},
       author = {John Lisman},
      journal = {Neuron},
       volume = 86,
        issue = 4,
        pages = {864-882},
         year = 2015,
     abstract = {Starting with the work of Cajal more than 100 years ago, neuroscience has sought to understand how the cells of the brain give rise to cognitive functions. How far has neuroscience progressed in this endeavor? This Perspective assesses progress in elucidating five basic brain processes: visual recognition, long-term memory, short-term memory, action selection, and motor control. Each of these processes entails several levels of analysis: the behavioral properties, the underlying computational algorithm, and the cellular/network mechanisms that implement that algorithm. At this juncture, while many questions remain unanswered, achievements in several areas of research have made it possible to relate specific properties of brain networks to cognitive functions. What has been learned reveals, at least in rough outline, how cognitive processes can be an emergent property of neurons and their connections.},
}
@article{LismanetalNEURAL-NETWORKS-05,
        title = {Recall of memory sequences by interaction of the dentate and {CA3}: A revised model of the phase precession},
       author = {John E. Lisman and Lucia M. Talamini and Antonino Raffone},
      journal = {Neural Networks},
       volume = {18},
       number = {9},
        pages = {1191-1201},
         year = {2005},
     abstract = {Behavioral and electrophysiological evidence indicates that the hippocampus has a special role in the encoding and recall of memory sequences. Importantly, the hippocampal phase precession, a phenomenon recorded as a rat moves through place fields, can be interpreted as cued recall of the sequence of upcoming places. The phase precession can be recorded in all hippocampal regions, but the role of each region has been unclear. Here, we suggest how the dentate and CA3 regions can work together to learn sequences, recall sequences, and generate the phase precession. Our proposal is constrained by information regarding synaptic plasticity rules, network connectivity, timing delays and theta/gamma oscillations.}
}
@article{LitkeNIMP-98,
       author = {Litke, A. M.},
        title = {The retinal readout system: an application of microstrip detector technology to neurobiology},
      journal = {Nuclear Instruments and Methods in Physics Research Section A},
       volume = 418,
        issue = 1,
         year = 1998,
        pages = {203-209},
     abstract = {We describe a system to study how the retina processes and encodes a visual image. The "Retinal Readout System" will allow the simultaneous detection of signals from hundreds of output cells in live retinal tissue as a visual image is focused on the input cells. The signals will be detected in real-time, with all spatial and temporal correlations recorded. The system will be based, in part, on the technology of silicon microstrip detectors.},
}
@article{LitkeetalTNS-04,
        title = {What Does the Eye Tell the Brain?: Development of a System for the Large-Scale Recording of Retinal Output Activity},
       author = {A. M. Litke and N. Bezayiff and E. J. Chichilnisky and W. Cunningham and W. Dabrowski and A. A. Grillo and M. Grivich and P. Grybos and P. Hottowy and S. Kachiguine and R. S. Kalmar and K. Mathieson and D. Petrusca and M. Rahman and A. Sher},
      journal = {{IEEE} Transactions on Nuclear Science},
       volume = 51,
        issue = 4,
         year = 2004,
        pages = {1434-1440},
}
@article{LitmanandAllenCS-87,
       author = {Diane J. Litman and James F. Allen},
        title = {A Plan Recognition Model for Subdialogues in Conversations},
      journal = {Cognitive Science},
         year = 1987,
       volume = 11,
        issue = 2,
        pages = {163-200},
}
@article{LittleandSommerFNC-13,
       author = {Little, Daniel Ying-Jeh and  Sommer, Friedrich Tobias},
        title = {Learning and exploration in action-perception loops},
      journal = {Frontiers in Neural Circuits},
       volume = 7,
        issue = 37,
         year = 2013,
     abstract = {Discovering the structure underlying observed data is a recurring problem in machine learning with important applications in neuroscience. It is also a primary function of the brain. When data can be actively collected in the context of a closed action-perception loop, behavior becomes a critical determinant of learning efficiency. Psychologists studying exploration and curiosity in humans and animals have long argued that learning itself is a primary motivator of behavior. However, the theoretical basis of learning-driven behavior is not well understood. Previous computational studies of behavior have largely focused on the control problem of maximizing acquisition of rewards and have treated learning the structure of data as a secondary objective. Here, we study exploration in the absence of external reward feedback. Instead, we take the quality of an agent's learned internal model to be the primary objective. In a simple probabilistic framework, we derive a Bayesian estimate for the amount of information about the environment an agent can expect to receive by taking an action, a measure we term the predicted information gain (PIG). We develop exploration strategies that approximately maximize PIG. One strategy based on value-iteration consistently learns faster, across a diverse range of environments, than previously developed reward-free exploration strategies. Psychologists believe the evolutionary advantage of learning-driven exploration lies in the generalized utility of an accurate internal model. Consistent with this hypothesis, we demonstrate that agents which learn more efficiently during exploration are later better able to accomplish a range of goal-directed tasks. We will conclude by discussing how our work elucidates the explorative behaviors of animals and humans, its relationship to other computational models of behavior, and its potential application to experimental design, such as in closed-loop neurophysiology studies.}
}
@phdthesis{LittlestonePhD,
       author = {Littlestone, N.},
        title = {Mistake Bounds and Logarithmic Linear-Threshold Learning Algorithms},
     location = {University of California, Santa Cruz},
         year = 1989,
}
@article{LittlestoneandWarmuth,
       author = {Littlestone, N. and Warmuth, M.},
        title = {The weighted majority algorithm},
      journal = {Information and Computation},
       volume = 108,
        issue = 2,
        pages = {221-261},
         year = 1994,
}
@inproceedings{LittlewoodetalPOTOMAC-15,
       author = {Peter Littlewood},
        title = {Toward a National Brain Observatory},
    booktitle = {America's Next Frontier: Conquering the Mind},
       editor = {Jennifer Buss, Director of the Center for Neurotechnology Studies},
    publisher = {Potomac Institute for Policy Studies Press},
         year = {2015},
        pages = {23-27},
     abstract = {The digital, biological, and nanotechnology disciplines have all provided their own technological leaps for society, and now, neuroscience is at the cusp of sparking its own revolution. The convergence of all of these disciplines will revolutionize the way in which society functions through the enhancement of human intelligence and physical abilities. When neurotechnology begins to reach into every aspect of our lives, the benefits to our health, education, and economy will be limitless. [...] Brain mapping alone does not improve our understanding of the brain. [...] The most informative neuroscientific findings originate from understanding communication between the different scales and functional areas of the brain - not from simply mapping their connections. [...] Neuroscience research should include methodological development of technologies that incorporate the dynamics of neural systems from multiple research perspectives. Just as Google Maps provides salient information about changing traffic patterns overlaid onto a road map, neuroscience research should augment mapping data with improved, high resolution, dynamic information.}
}
@inproceedings{LittmanAAAI-97,
       author = {Littman, Michael L.},
        title = {Probabilistic Propositional Planning: Representations and Complexity},
    booktitle = {Proceedings AAAI-97},
     location = {Providence, Rhode Island},
 organization = {AAAI},
         year = 1997,
}
@phdthesis{LittmanPhD,
       author = {Michael Lederman Littman},
        title = {Algorithms for Sequential Decision Making},
       school = {Department of Computer Science, Brown University},
      comment = {Technical Report CS-96-09},
         year = 1996,
}
@techreport{LittmanTR-94,
       author = {Michael L. Littman},
        title = {The Witness Algorithm for Solving Partially Observable {M}arkov Decision Processes},
         year = 1994,
        issue = {CS-94-40},
  institution = {Brown University},
      address = {Providence, Rhode Island},
}
@inproceedings{LittmanetalUAI-95,
       author = {Littman, Michael and Dean, Thomas and Kaelbling, Leslie},
        title = {On the Complexity of Solving {M}arkov Decision Problems},
    booktitle = {Proceedings of the 11th Conference on Uncertainty in Artificial Intelligence},
     location = {Montreal, Quebec},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
 organization = {AUAI},
        pages = {394-402},
         year = 1995,
}
@book{Liu2014,
        title = {The Three-Body Problem},
       author = {Liu, C. and Liu, K.},
       series = {Remembrance of Earth's Past},
         year = {2014},
    publisher = {Tom Doherty Associates}
}
@book{Liu2015,
        title = {The Dark Forest},
       author = {Liu, C. and Martinsen, J.},
       series = {Remembrance of Earth's Past},
         year = {2015},
    publisher = {Tom Doherty Associates}
}
@book{Liu2016,
        title = {Death's End},
       author = {Liu, C. and Liu, K.},
       series = {Remembrance of Earth's Past},
         year = {2016},
    publisher = {Tom Doherty Associates}
}
@article {LiuandCarterJoN18,
       author = {Liu, Xingchen and Carter, Adam G.},
        title = {Ventral Hippocampal Inputs Preferentially Drive Corticocortical Neurons in the Infralimbic Prefrontal Cortex},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {38},
       number = {33},
         year = {2018},
        pages = {7351-7363},
     abstract = {Inputs from the ventral hippocampus (vHPC) to the prefrontal cortex (PFC) play a key role in working memory and emotional control. However, little is known about how excitatory inputs from the vHPC engage different populations of neurons in the PFC. Here we use optogenetics and whole-cell recordings to study the cell-type specificity of synaptic connections in acute slices from the mouse PFC. We first show that vHPC inputs target pyramidal neurons whose cell bodies are located in layer (L)2/3 and L5 of infralimbic (IL) PFC, but only in L5 of prelimbic (PL) PFC, and not L6 of either IL or PL. We then compare connections onto different classes of projection neurons located in these layers and subregions of PFC. We establish vHPC inputs similarly contact corticocortical (CC) and cortico-amygdala neurons in L2/3 of IL, but preferentially target CC neurons over cortico-pontine neurons in L5 of both IL and PL. Of all these neurons, we determine that vHPC inputs are most effective at driving action potential (AP) firing of CC neurons in L5 of IL. We also show this connection exhibits frequency-dependent facilitation, with repetitive activity enhancing AP firing of IL L5 CC neurons, even in the presence of feedforward inhibition. Our findings reveal how vHPC inputs engage defined populations of projection neurons in the PFC, allowing preferentially activation of the intratelencephalic network. SIGNIFICANCE STATEMENT We examined the impact of connections from the ventral hippocampus (vHPC) onto different projection neurons in the mouse prefrontal cortex (PFC). We found vHPC inputs were strongest at corticocortical neurons in layer 5 of infralimbic PFC, where they robustly evoked action potential firing, including during repetitive activity with intact feedforward inhibition.},
}
@article{LiuandFletcherBJ-06,
       author = {Liu, Allen P. and Fletcher, Daniel A.},
        title = {Actin Polymerization Serves as a Membrane Domain Switch in Model Lipid Bilayers},
      journal = {Biophysical Journal},
         year = 2006,
    publisher = {Biophysical Society},
       volume = 91,
        issue = 11,
        pages = {4064-4070},
     abstract = {The ability of cells to mount localized responses to external or internal stimuli is critically dependent on organization of lipids and proteins in the plasma membrane. Involvement of the actin cytoskeleton in membrane organization has been documented, but an active role for actin networks that directly links internal organization of the cytoskeleton with membrane organization has not yet been identified. Here we show that branched actin networks formed on model lipid membranes enriched with the lipid second messenger PIP(2) trigger both temporal and spatial rearrangement of membrane components. Using giant unilamellar vesicles able to separate into two coexisting liquid phases, we demonstrate that polymerization of dendritic actin networks on the membrane induces phase separation of initially homogenous vesicles. This switch-like behavior depends only on the PIP(2)-N-WASP link between the membrane and actin network, and we find that the presence of a preexisting actin network spatially biases the location of phase separation. These results show that dynamic, membrane-bound actin networks alone can control when and where membrane domains form and may actively contribute to membrane organization during cell signaling.},
}
@article{LiuetalACS-11,
       author = {Liu, Jianfei and Geng, Yanli and Pound, Elisabeth and Gyawali, Shailendra and Ashton, Jeffrey R. and Hickey, John and Woolley, Adam T. and Harb, John N.},
        title = {Metallization of Branched {DNA} Origami for Nanoelectronic Circuit Fabrication},
      journal = {{ACS} Nano},
       volume = 5,
        issue = 3,
         year = 2011,
        pages = {2240-2247},
     abstract = {This work examines the metallization of folded DNA, known as DNA origami, as an enabling step toward the use of such DNA as templates for nanoelectronic circuits. DNA origami, a simple and robust method for creating a wide variety of shapes and patterns, makes possible the increased complexity and flexibility needed for both the design and assembly of useful circuit templates. In addition, selective metallization of the DNA template is essential for circuit fabrication. Metallization of DNA origami presents several challenges over and above those associated with the metallization of other DNA templates such as \lambda-DNA. These challenges include (1) the stability of the origami in the processes used for metallization, (2) the enhanced selectivity required to metallize small origami structures, (3) the increased difficulty of adhering small structures to the surface so that they will not be removed when subject to multiple metallization steps, and (4) the influence of excess staple strands present with the origami. This paper describes our efforts to understand and address these challenges. Specifically, the influence of experimental conditions on template stability and on the selectivity of metal deposition was investigated for small DNA origami templates. These templates were seeded with Ag and then plated with Au via an electroless deposition process. Both staple strand concentration and the concentration of ions in solution were found to have a significant impact. Selective continuous metal deposition was achieved, with an average metallized height as small as 32 nm. The shape of branched origami was also retained after metallization. These results represent important progress toward the realization of DNA-templated nanocircuits.},
}
@article{LiuetalCD-09,
       author = {Liu, D.  and Meltzoff, A. N.  and Wellman, H. M. },
        title = {{{N}eural correlates of belief- and desire-reasoning}},
      journal = {Child Development},
       volume = {80},
       number = {4},
         year = {2009},
        pages = {1163-1171},
     abstract = {Theory of mind requires an understanding of both desires and beliefs. Moreover, children understand desires before beliefs. Little is known about the mechanisms underlying this developmental lag. Additionally, previous neuroimaging and neurophysiological studies have neglected the direct comparison of these developmentally critical mental-state concepts. Event-related brain potentials were recorded as participants (N = 24; mean age = 22 years) reasoned about diverse-desires, diverse-beliefs, and parallel physical situations. A mid-frontal late slow wave (LSW) was associated with desire and belief judgments. A right-posterior LSW was only associated with belief judgments. These findings demonstrate neural overlap and critical differences in reasoning explicitly about desires and beliefs, and they suggest children recruit additional neural processes for belief judgments beyond a common, more general, mentalizing neural system.}
}
@inproceedings{LiuetalCVPR-09,
       author = {Jingen Liu and Jiebo Luo and Mubarak Shah},
        title = {Recognizing Realistic Actions from Videos "in the Wild"},
    booktitle = {{IEEE} Computer Society Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 2009,
}
@inproceedings{LiuetalCVPR-2009,
       author = {Jingen Liu and Jiebo Luo and Mubarak Shah},
        title = {Recognizing Realistic Actions from Videos "in the Wild"},
    booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
         year = 2009,
}
@article{LiuetalCoRR-16,
       author = {Ziwei Liu and Xiaoxiao Li and Ping Luo and Chen Change Loy and Xiaoou Tang},
        title = {Deep Learning Markov Random Field for Semantic Segmentation},
      journal = {CoRR},
       volume = {arXiv:1606.07230},
         year = {2016},
     abstract = {Semantic segmentation tasks can be well modeled by Markov Random Field (MRF). This paper addresses semantic segmentation by incorporating high-order relations and mixture of label contexts into MRF. Unlike previous works that optimized MRFs using iterative algorithm, we solve MRF by proposing a Convolutional Neural Network (CNN), namely Deep Parsing Network (DPN), which enables deterministic end-to-end computation in a single forward pass. Specifically, DPN extends a contemporary CNN to model unary terms and additional layers are devised to approximate the mean field (MF) algorithm for pairwise terms. It has several appealing properties. First, different from the recent works that required many iterations of MF during back-propagation, DPN is able to achieve high performance by approximating one iteration of MF. Second, DPN represents various types of pairwise terms, making many existing models as its special cases. Furthermore, pairwise terms in DPN provide a unified framework to encode rich contextual information in high-dimensional data, such as images and videos. Third, DPN makes MF easier to be parallelized and speeded up, thus enabling efficient inference. DPN is thoroughly evaluated on standard semantic image/video segmentation benchmarks, where a single DPN model yields state-of-the-art segmentation accuracies on PASCAL VOC 2012, Cityscapes dataset and CamVid dataset.}
}
@article{LiuetalCoRR-19,
       author = {Zinan Liu and Arne Hitzmann and Shuhei Ikemoto and Svenja Stark and Jan Peters and Koh Hosoda},
        title = {Local Online Motor Babbling: Learning Motor Abundance of {A} Musculoskeletal Robot Arm},
      journal = {CoRR},
       volume = {arXiv:1906.09013},
         year = {2019},
     abstract = {Motor babbling and goal babbling has been used for sensorimotor learning of highly redundant systems in soft robotics. Recent works in goal babbling has demonstrated successful learning of inverse kinematics (IK) on such systems, and suggests that babbling in the goal space better resolves motor redundancy by learning as few sensorimotor mapping as possible. However, for musculoskeletal robot systems, motor redundancy can be of useful information to explain muscle activation patterns, thus the term motor abundance. In this work, we introduce some simple heuristics to empirically define the unknown goal space, and learn the inverse kinematics of a 10 DoF musculoskeletal robot arm using directed goal babbling. We then further propose local online motor babbling using Covariance Matrix Adaptation Evolution Strategy (CMA-ES), which bootstraps on the collected samples in goal babbling for initialization, such that motor abundance can be queried for any static goal within the defined goal space. The result shows that our motor babbling approach can efficiently explore motor abundance, and gives useful insights in terms of muscle stiffness and synergy.}
} 
@inproceedings{LiuetalIEEE-89,
       author = {Shih, W-K. and Liu, J. W. S. and Chung, J-Y.},
        title = {Fast Algorithms for Scheduling Imprecise Computations},
    booktitle = {Proceedings of the Real-Time Systems Symposium},
     location = {Santa Monica, California},
 organization = {IEEE},
        pages = {12-19},
         year = 1989,
}
@inproceedings{LiuetalIUI-03,
       author = {Liu, Hugo and Lieberman, Henry and Selker, Ted},
        title = {A model of textual affect sensing using real-world knowledge},
    booktitle = {Proceedings of the 8th international conference on Intelligent user interfaces},
    publisher = {ACM},
         year = 2003,
        pages = {125-132},
     abstract = {This paper presents a novel way for assessing the affective qualities of natural language and a scenario for its use. Previous approaches to textual affect sensing have employed keyword spotting, lexical affinity, statistical methods, and hand-crafted models. This paper demonstrates a new approach, using large-scale real-world knowledge about the inherent affective nature of everyday situations (such as "getting into a car accident") to classify sentences into "basic" emotion categories. This commonsense approach has new robustness implications. Open Mind Commonsense was used as a real world corpus of 400,000 facts about the everyday world. Four linguistic models are combined for robustness as a society of commonsense-based affect recognition. These models cooperate and compete to classify the affect of text. Such a system that analyzes affective qualities sentence by sentence is of practical value when people want to evaluate the text they are writing. As such, the system is tested in an email writing application. The results suggest that the approach is robust enough to enable plausible affective text user interfaces.}
}
@article{LiuetalNATURE-15,
       author = {Liu, Yan and Lai, Puxiang and Ma, Cheng and Xu, Xiao and Grabar, Alexander A. and Wang, Lihong V.},
        title = {Optical focusing deep inside dynamic scattering media with near-infrared time-reversed ultrasonically encoded (TRUE) light},
      journal = {Nature Communications},
    publisher = {Nature Publishing Group, a division of Macmillan Publishers Limited. All Rights Reserved.},
       volume = {6},
         year = {2015},
     abstract = {Focusing light deep inside living tissue has not been achieved despite its promise to play a central role in biomedical imaging, optical manipulation and therapy. To address this challenge, internal-guide-star-based wavefront engineering techniques\emdash{}for example, time-reversed ultrasonically encoded (TRUE) optical focusing\emdash{}were developed. The speeds of these techniques, however, were limited to no greater than 1\emdash{}Hz, preventing them from in vivo applications. Here we improve the speed of optical focusing deep inside scattering media by two orders of magnitude, and focus diffuse light inside a dynamic scattering medium having a speckle correlation time as short as 5.6\emdash{}ms, typical of living tissue. By imaging a target, we demonstrate the first focusing of diffuse light inside a dynamic scattering medium containing living tissue. Since the achieved focusing speed approaches the tissue decorrelation rate, this work is an important step towards in vivo deep tissue noninvasive optical imaging, optogenetics and photodynamic therapy.},
}
@article{LiuetalNATURE-NANO-15,
       author = {Liu, Jia and Fu, Tian-Ming and Cheng, Zengguang and Hong, Guosong and Zhou, Tao and Jin, Lihua and Duvvuri, Madhavi and Jiang, Zhe and Kruskal, Peter and Xie, Chong and Suo, Zhigang and Fang, Ying and Lieber, Charles M.},
        title = {Syringe-injectable electronics},
      journal = {Nature Nanotechnology},
       volume = 10, 
         year = 2015,
        pages = {629-636},
    publisher = {Nature Publishing Group},
     abstract = {Seamless and minimally invasive three-dimensional interpenetration of electronics within artificial or natural structures could allow for continuous monitoring and manipulation of their properties. Flexible electronics provide a means for conforming electronics to non-planar surfaces, yet targeted delivery of flexible electronics to internal regions remains difficult. Here, we overcome this challenge by demonstrating the syringe injection (and subsequent unfolding) of sub-micrometre-thick, centimetre-scale macroporous mesh electronics through needles with a diameter as small as 100 Angsrtom. Our results show that electronic components can be injected into man-made and biological cavities, as well as dense gels and tissue, with >90\% device yield. We demonstrate several applications of syringe-injectable electronics as a general approach for interpenetrating flexible electronics with three-dimensional structures, including (1) monitoring internal mechanical strains in polymer cavities, (2) tight integration and low chronic immunoreactivity with several distinct regions of the brain, and (3) in vivo multiplexed neural recording. Moreover, syringe injection enables the delivery of flexible electronics through a rigid shell, the delivery of large-volume flexible electronics that can fill internal cavities, and co-injection of electronics with other materials into host structures, opening up unique applications for flexible electronics.},
}
@article{LiuetalNATURE-NANO-15.supplement,
       author = {Liu, Jia and Fu, Tian-Ming and Cheng, Zengguang and Hong, Guosong and Zhou, Tao and Jin, Lihua and Duvvuri, Madhavi and Jiang, Zhe and Kruskal, Peter and Xie, Chong and Suo, Zhigang and Fang, Ying and Lieber, Charles M.},
        title = {Syringe-injectable electronics: Supplementary Materials},
      journal = {Nature Nanotechnology},
         year = 2015,
}
@incollection{LiuetalNIPS-05,
        title = {An Investigation of Practical Approximate Nearest Neighbor Algorithms},
       author = {Ting Liu and Andrew W. Moore and Ke Yang and Alexander G. Gray},
    booktitle = {Advances in Neural Information Processing Systems 17},
       editor = {L. K. Saul and Y. Weiss and L. Bottou},
    publisher = {MIT Press},
         year = {2005},
        pages = {825-832},
}
@incollection{LiuetalNIPS-06,
        title = {Analysis of Contour Motions},
       author = {Ce Liu and William Freeman and Edward Adelson},
    booktitle = {Advances in Neural Information Processing Systems 19},
       editor = {B. Sch\"{o}lkopf and J. Platt and T. Hofmann},
    publisher = {MIT Press},
      address = {Cambridge, MA},
        pages = {913-920},
         year = 2007
}
@article{LiuetalPNAS-09,
       author = {Liu, Y. S. and Stevens, C. F. and Sharpee, T. O.},
        title = {Predictable irregularities in retinal receptive fields},
      journal = {Proceedings of the National Academy of Science U.S.A.},
         year = {2009},
       volume = {106},
       number = {38},
        pages = {16499-16504},
     abstract = {Understanding how the nervous system achieves reliable performance using unreliable components is important for many disciplines of science and engineering, in part because it can suggest ways to lower the energetic cost of computing. In vision, retinal ganglion cells partition visual space into approximately circular regions termed receptive fields (RFs). Average RF shapes are such that they would provide maximal spatial resolution if they were centered on a perfect lattice. However, individual shapes have fine-scale irregularities. Here, we find that irregular RF shapes increase the spatial resolution in the presence of lattice irregularities from approximately 60\% to approximately 92\% of that possible for a perfect lattice. Optimization of RF boundaries around their fixed center positions reproduced experimental observations neuron-by-neuron. Our results suggest that lattice irregularities determine the shapes of retinal RFs and that similar algorithms can improve the performance of retinal prosthetics where substantial irregularities arise at their interface with neural tissue.}
}
@article{LiuetalPNAS-19,
       author = {Liu, Shari and Brooks, Neon B. and Spelke, Elizabeth S.},
        title = {Origins of the concepts cause, cost, and goal in prereaching infants},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {116},
       number = {36},
         year = {2019},
        pages = {17747-17752},
     abstract = {We view ourselves and others as causal agents who pursue goals and act efficiently to make things happen, but where do these intuitions come from? Five looking-time experiments with 3-month-old infants show that infants interpret actions they cannot yet perform as causally efficacious. When people reach for and cause state changes in objects, young infants interpret these actions as goal-directed and look longer when they are inefficient rather than efficient. In contrast, infants show no consistent responses to similar actions that cause no changes in an object. An early-emerging sensitivity to the causal powers of agents, when they engage in costly, goal-directed actions, may provide one important foundation for the rich causal and social learning that characterizes our species.We investigated the origins and interrelations of causal knowledge and knowledge of agency in 3-month-old infants, who cannot yet effect changes in the world by reaching for, grasping, and picking up objects. Across 5 experiments, n = 152 prereaching infants viewed object-directed reaches that varied in efficiency (following the shortest physically possible path vs. a longer path), goal (lifting an object vs. causing a change in its state), and causal structure (action on contact vs. action at a distance and after a delay). Prereaching infants showed no strong looking preference between a person{\textquoteright}s efficient and inefficient reaches when the person grasped and displaced an object. When the person reached for and caused a change in the state of the object on contact, however, infants looked longer when this action was inefficient than when it was efficient. Three-month-old infants also showed a key signature of adults{\textquoteright} and older infants{\textquoteright} causal inferences: This looking preference was abolished if a short spatial and temporal gap separated the action from its effect. The basic intuition that people are causal agents, who navigate around physical constraints to change the state of the world, may be one important foundation for infants{\textquoteright} ability to plan their own actions and learn from the acts of others.},
} 
@misc{LiuetalUSPATENT-09,
        title = {Building parallel hybrid spill trees to facilitate parallel nearest-neighbor matching operations},
       author = {Liu, T. and Rosenberg, C.J. and Rowley, H.A.},
         year = {2009},
    publisher = {Google Patents},
         note = {US Patent 7,539,657}
}
@inproceedings{LiuetalWACV-07,
       author = {T. Liu and C. Rosenberg and H. A. Rowley},
    booktitle = {Applications of Computer Vision, 2007. WACV '07. IEEE Workshop on},
        title = {Clustering Billions of Images with Large Scale Nearest Neighbor Search},
         year = {2007},
        pages = {28-28},
     abstract = {The proliferation of the Web and digital photography have made large scale image collections containing billions of images a reality. Image collections on this scale make performing even the most common and simple computer vision, image processing, and machine learning tasks nontrivial. An example is nearest neighbor search, which not only serves as a fundamental subproblem in many more sophisticated algorithms, but also has direct applications, such as image retrieval and image clustering. In this paper, we address the nearest neighbor problem as the first step towards scalable image processing. We describe a scalable version of an approximate nearest neighbor search algorithm and discuss how it can be used to find near duplicates among over a billion images},
}
@article{LivnietalCoRR-13,
    author = {Roi Livni and Shai Shalev-Shwartz and Ohad Shamir},
     title = {A Provably Efficient Algorithm for Training Deep Networks},
   journal = {CoRR},
    volume = {arXiv:1304.7045},
      year = {2013},
       URL = {http://arxiv.org/abs/1304.7045},
  abstract = {We consider deep neural networks, in which the output of each node is a quadratic function of its inputs. Similar to other deep architectures, these networks can compactly represent any function on a finite training set. The main goal of this paper is the derivation of an efficient layer-by-layer algorithm for training such networks, which we denote as the Basis Learner. The algorithm is a universal learner in the sense that the training error is guaranteed to decrease at every iteration, and can eventually reach zero under mild conditions. We present practical implementations of this algorithm, as well as preliminary experimental results. We also compare our deep architecture to other shallow architectures for learning polynomials, in particular kernel learning.},
}
@inproceedings{LjunglofWSPD-09,
       author = {Ljungl\"{o}f, Peter},
        title = {Dialogue Management as Interactive Tree Building},
    booktitle = {DiaHolmia 2009, 13th Workshop on the Semantics and Pragmatics of Dialogue},
         year = {2009},
     abstract = {We introduce a new dialogue model and a formalism for limited-domain dialogue systems, which works by interactively building dialogue trees. The model borrows its fundamental ideas from type theoretical grammars and Dynamic Syntax. The resulting dialogue theory is a simple and light-weight formalism, which is still capable of advanced dialogue behaviour.},
}
@incollection{LlinasOLIVOCEREBELLAR-SYSTEM-18,
       author = {Llin\`{a}s, Rudolfo R.},
        title = {Olivocerebellar System},
    booktitle = {Handbook of Brain Microcircuits},
       editor = {Shepherd, G.M. and Grillner, S.},
    publisher = {Oxford University Press},
         year = {2018},
        pages = {301-308}
}
@article{LlinasSCIENCE-88,
       author = {Llin\`{a}s, Rudolfo R.},
        title = {The intrinsic electrophysiological properties of mammalian neurons: insights into central nervous system function},
      journal = {Science},
       volume = {242},
        issue = {4886},
        pages = {1654-1664},
         year = {1988},
     abstract = {This article reviews the electroresponsive properties of single neurons in the mammalian central nervous system (CNS). In some of these cells the ionic conductances responsible for their excitability also endow them with autorhythmic electrical oscillatory properties. Chemical or electrical synaptic contacts between these neurons often result in network oscillations. In such networks, autorhythmic neurons may act as true oscillators (as pacemakers) or as resonators (responding preferentially to certain firing frequencies). Oscillations and resonance in the CNS are proposed to have diverse functional roles, such as (i) determining global functional states (for example, sleep-wakefulness or attention), (ii) timing in motor coordination, and (iii) specifying connectivity during development. Also, oscillation, especially in the thalamo-cortical circuits, may be related to certain neurological and psychiatric disorders. This review proposes that the autorhythmic electrical properties of central neurons and their connectivity form the basis for an intrinsic functional coordinate system that provides internal context to sensory input.},
}
@article{LocascioetalCoRR-16,
       author = {Nicholas Locascio and Karthik Narasimhan and Eduardo DeLeon and Nate Kushman and Regina Barzilay},
        title = {Neural Generation of Regular Expressions from Natural Language with Minimal Domain Knowledge},
      journal = {CoRR},
       volume = {arXiv:1608.03000},
         year = {2016},
     abstract = {This paper explores the task of translating natural language queries into regular expressions which embody their meaning. In contrast to prior work, the proposed neural model does not utilize domain-specific crafting, learning to translate directly from a parallel corpus. To fully explore the potential of neural models, we propose a methodology for collecting a large corpus of regular expression, natural language pairs. Our resulting model achieves a performance gain of 19.6\% over previous state-of-the-art models.}
}
@article{LockerandSchmidPLoS-13,
       author = {Krijnse Locker, Jacomine and Schmid, Sandra L.},
      journal = {{PLoS} Biology},
    publisher = {Public Library of Science},
        title = {Integrated Electron Microscopy: Super-Duper Resolution},
         year = 2013,
       volume = 11,
        issue = 8,
        pages = {e1001639},
     abstract = {New developments in electron microscopy used in combination reveal the elegant architecture of cellular structures at very high resolution.},
}
@article{LockhartetalFiHN-12,
       author = {Lockhart, Samuel N. and Mayda, Adriane B. V. and Roach, Alexandra E. and Fletcher, Evan and Carmichael, Owen  and Maillard, Pauline and Schwarz, Christopher G. and Yonelinas, Andrew P. and Ranganath, Charan and Decarli, Charles},
        title = {Episodic memory function is associated with multiple measures of white matter integrity in cognitive aging},
      journal = {Frontiers in human neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = {6},
         year = {2012},
        pages = {56-56},
     abstract = {Previous neuroimaging research indicates that white matter injury and integrity, measured respectively by white matter hyperintensities (WMH) and fractional anisotropy (FA) obtained from diffusion tensor imaging (DTI), differ with aging and cerebrovascular disease (CVD) and are associated with episodic memory deficits in cognitively normal older adults. However, knowledge about tract-specific relationships between WMH, FA, and episodic memory in aging remains limited. We hypothesized that white matter connections between frontal cortex and subcortical structures as well as connections between frontal and temporo-parietal cortex would be most affected. In the current study, we examined relationships between WMH, FA and episodic memory in 15 young adults, 13 elders with minimal WMH and 15 elders with extensive WMH, using an episodic recognition memory test for object-color associations. Voxel-based statistics were used to identify voxel clusters where white matter measures were specifically associated with variations in episodic memory performance, and white matter tracts intersecting these clusters were analyzed to examine white matter-memory relationships. White matter injury and integrity measures were significantly associated with episodic memory in extensive regions of white matter, located predominantly in frontal, parietal, and subcortical regions. Template based tractography indicated that white matter injury, as measured by WMH, in the uncinate and inferior longitudinal fasciculi were significantly negatively associated with episodic memory performance. Other tracts such as thalamo-frontal projections, superior longitudinal fasciculus, and dorsal cingulum bundle demonstrated strong negative associations as well. The results suggest that white matter injury to multiple pathways, including connections of frontal and temporal cortex and frontal-subcortical white matter tracts, plays a critical role in memory differences seen in older individuals.},
}
@book{Lodge01,
      author = {David Lodge},
       title = {Thinks},
   publisher = {Penguin},
     address = {London},
        year = 2001,
}
@article{LodgeandBischofbergerBBR-19,
       author = {Lodge, M. and Bischofberger, J.},
        title = {Synaptic properties of newly generated granule cells support sparse coding in the adult hippocampus},
      journal = {Behavior and Brain Research},
         year = {2019},
       volume = {372},
        pages = {112036},
     abstract = {In the adult hippocampus new neurons are continuously generated throughout life and integrate into the existing network via the formation of thousands of new synapses. Adult-born granule cells are known to improve learning and memory at about 3-6 weeks post mitosis by enhancing the brains ability to discriminate similar memory items. However, the underlying mechanisms are still controversial. Here we review the distinct functional properties of the newborn young neurons, including enhanced excitability, reduced GABAergic inhibition, NMDA-receptor dependent electrogenesis and enhanced synaptic plasticity. Although these cellular properties provide a competitive advantage for synapse formation, they do not generate 'hyperactivity' of young neurons. By contrast, in vivo evidence from immediate early gene expression and calcium imaging indicates that young neurons show sparse activity during learning. Similarly, in vitro data show a low number of high-impact synapses, leading to activation young cells by distinct subsets of afferent fibers with minimal overlap. Overall, the enhanced excitability of young cells does not generate hyperactivity but rather counterbalance the low number of excitatory input synapses. Finally, sparse coding in young neurons has been shown to be crucial for neurogenesis-dependent improvement of learning behavior. Taken together, converging evidence from cell physiology and behavioral studies suggests a mechanism that can explain the beneficial effects of adult neurogenesis on brain function.}
}
@book{Loewenstein2013,
       author = {Werner R. Loewenstein},
        title = {Physics in Mind: A Quantum View of the Brain},
         year = 2013,
    publisher = {Basic Books},
      address = {New York, NY},
}
@book{Loftus1996eyewitness,
        title = {Eyewitness Testimony},
       author = {Loftus, Elizabeth F.},
         year = {1996},
    publisher = {Harvard University Press}
}
@article{LoftusandPalmerJVB-74,
       author = {Loftus, E. F. and Palmer, J. C.},
        title = {Reconstruction of automobile destruction: An example of the interaction between language and memory},
      journal = {Journal of Verbal Learning \& Verbal Behavior},
       volume = {13},
       number = {5},
         year = {1974},
        pages = {585-589}
}
@article{LoftusandPalmerVLVB-74,
        title = {Reconstruction of automobile destruction: An example of the interaction between language and memory},
      journal = {Journal of Verbal Learning and Verbal Behavior},
       volume = {13},
       number = {5},
        pages = {585-589},
         year = {1974},
       author = {Elizabeth F. Loftus and John C. Palmer},
     abstract = {Two experiments are reported in which subjects viewed films of automobiled accidents and then answered questions about events occurring in the films. The question, "About how fast were the cars going when they smashed into each other?" elicited higher estimates of speed than questions which used the verbs collided, bumped, contacted, or hit in place of smashed. On a retest one week later, those subjects who received the verb smashed were more likely to say "yes" to the question, "Did you see any broken glass?", even though broken glass was not present in the film. These results are consistent with the view that the questions asked subsequent to an event can cause a reconstruction in one's memory of that event.}
}
@article{LolloFRONTIERS-PSYCHOLOGY,
       author = {Di Lollo, Vincent},
        title = {Reentrant processing mediates object substitution masking: comment on Põder (2013)},
      journal = {Frontiers in Psychology},
       volume = {5},
        pages = {819},
         year = {2014},
     abstract = {Object-substitution masking (OSM) occurs when a target stimulus and a surrounding mask are displayed briefly together, and the display then continues with the mask alone. Target identification is accurate when the stimuli co-terminate but is progressively impaired as the duration of the trailing mask is increased. In reentrant accounts, OSM is said to arise from iterative exchanges between brain regions connected by two-way pathways. In an alternative account, OSM is explained on the basis of exclusively feed-forward processes, without recourse to reentry. Here I show that the feed-forward account runs afoul of the extant phenomenological, behavioral, brain-imaging, and electrophysiological evidence. Further, the feed-forward assumption that masking occurs when attention finds a degraded target is shown to be entirely ad hoc. In contrast, the evidence is uniformly consistent with a reentrant-processing account of OSM.}
}
@article{LondonandHausserARN-05,
       author = {Michael London and Michael H\"{a}usser},
        title = {Dendritic Computation},
      journal = {Annual Review of Neuroscience},
       volume = {28},
       number = {1},
         year = {2005},
        pages = {503-532},
     abstract = {One of the central questions in neuroscience is how particular tasks, or computations, are implemented by neural networks to generate behavior. The prevailing view has been that information processing in neural networks results primarily from the properties of synapses and the connectivity of neurons within the network, with the intrinsic excitability of single neurons playing a lesser role. As a consequence, the contribution of single neurons to computation in the brain has long been underestimated. Here we review recent work showing that neuronal dendrites exhibit a range of linear and nonlinear mechanisms that allow them to implement elementary computations. We discuss why these dendritic properties may be essential for the computations performed by the neuron and the network and provide theoretical and experimental examples to support this view.}
}
@article{LondonetalNATURE-10,
       author = {London, M. and Roth, A. and Beeren, L. and Hausser, M. and Latham, P. E.},
        title = {Sensitivity to perturbations in vivo implies high noise and suggests rate coding in cortex},
      journal = {Nature},
         year = {2010},
       volume = {466},
       number = {7302},
        pages = {123-127},
     abstract = {It is well known that neural activity exhibits variability, in the sense that identical sensory stimuli produce different responses, but it has been difficult to determine what this variability means. Is it noise, or does it carry important information-about, for example, the internal state of the organism? Here we address this issue from the bottom up, by asking whether small perturbations to activity in cortical networks are amplified. Based on in vivo whole-cell patch-clamp recordings in rat barrel cortex, we find that a perturbation consisting of a single extra spike in one neuron produces approximately 28 additional spikes in its postsynaptic targets. We also show, using simultaneous intra- and extracellular recordings, that a single spike in a neuron produces a detectable increase in firing rate in the local network. Theoretical analysis indicates that this amplification leads to intrinsic, stimulus-independent variations in membrane potential of the order of +/-2.2-4.5 mV-variations that are pure noise, and so carry no information at all. Therefore, for the brain to perform reliable computations, it must either use a rate code, or generate very large, fast depolarizing events, such as those proposed by the theory of synfire chains. However, in our in vivo recordings, we found that such events were very rare. Our findings are thus consistent with the idea that cortex is likely to use primarily a rate code.}
}
@inproceedings{LongandMartinICSE-16,
       author = {Long, Fan and Rinard, Martin},
        title = {An Analysis of the Search Spaces for Generate and Validate Patch Generation Systems},
    booktitle = {Proceedings of the 38th International Conference on Software Engineering},
     location = {Austin, Texas},
    publisher = {ACM},
      address = {New York, NY, USA},
         year = {2016},
        pages = {702-713},
     abstract = {We present the first systematic analysis of key characteristics of patch search spaces for automatic patch generation systems. We analyze sixteen different configurations of the patch search spaces of SPR and Prophet, two current state-of-the-art patch generation systems. The analysis shows that (i) correct patches are sparse in the search spaces (typically at most one correct patch per search space per defect), (ii)) incorrect patches that nevertheless pass all of the test cases in the validation test suite are typically orders of magnitude more abundant, and (iii) leveraging information other than the test suite is therefore critical for enabling the system to successfully isolate correct patches.}
}
@article{LongetalPNAS-18,
       author = {Long, Bria and Yu, Chen-Ping and Konkle, Talia},
        title = {Mid-level visual features underlie the high-level categorical organization of the ventral stream},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {115},
       number = {38},
        pages = {E9015-E9024},
         year = {2018},
     abstract = {While neural responses to object categories are remarkably systematic across human visual cortex, the nature of these responses has been hotly debated for the past 20 y. In this paper, a class of stimuli (texforms) is used to examine how mid-level features contribute to the large-scale organization of the ventral visual stream. Despite their relatively primitive visual appearance, these unrecognizable texforms elicited the entire large-scale organizations of the ventral stream by animacy and object size. This work demonstrates that much of ventral stream organization can be explained by relatively primitive mid-level features without requiring explicit recognition of the objects themselves. Human object-selective cortex shows a large-scale organization characterized by the high-level properties of both animacy and object size. To what extent are these neural responses explained by primitive perceptual features that distinguish animals from objects and big objects from small objects? To address this question, we used a texture synthesis algorithm to create a class of stimuli called "texforms" that preserve some mid-level texture and form information from objects while rendering them unrecognizable. We found that unrecognizable texforms were sufficient to elicit the large-scale organizations of object-selective cortex along the entire ventral pathway. Further, the structure in the neural patterns elicited by texforms was well predicted by curvature features and by intermediate layers of a deep convolutional neural network, supporting the mid-level nature of the representations. These results provide clear evidence that a substantial portion of ventral stream organization can be accounted for by coarse texture and form information without requiring explicit recognition of intact objects.},
}
@misc{LookSmart03,
        title = {Collaborative Search using {Z}eal and {G}rub},
       author = {LookSmart, Ltd.},
 howpublished = {\url{http://www.looksmart.com/}},
         year = 2003,
}
@article{LoosetalCoRR-17,
       author = {Sarah M. Loos and Geoffrey Irving and Christian Szegedy and Cezary Kaliszyk},
        title = {Deep Network Guided Proof Search},
      journal = {CoRR},
       volume = {arXiv:1701.06972},
         year = {2017},
     abstract = {Deep learning techniques lie at the heart of several significant AI advances in recent years including object recognition and detection, image captioning, machine translation, speech recognition and synthesis, and playing the game of Go. Automated first-order theorem provers can aid in the formalization and verification of mathematical theorems and play a crucial role in program analysis, theory reasoning, security, interpolation, and system verification. Here we suggest deep learning based guidance in the proof search of the theorem prover E. We train and compare several deep neural network models on the traces of existing ATP proofs of Mizar statements and use them to select processed clauses during proof search. We give experimental evidence that with a hybrid, two-phase approach, deep learning based guidance can significantly reduce the average number of proof search steps while increasing the number of theorems proved. Using a few proof guidance strategies that leverage deep neural networks, we have found first-order proofs of 7.36\% of the first-order logic translations of the Mizar Mathematical Library theorems that did not previously have ATP generated proofs. This increases the ratio of statements in the corpus with ATP generated proofs from 56\% to 59\%.},
}
@article{Lopez-RojasaandKreutzNBR-16,
       author = {Jeffrey Lopez-Rojas and Michael R. Kreutz},
        title = {Mature granule cells of the dentate gyrus-Passive bystanders or principal performers in hippocampal function?},
      journal = {Neuroscience and Biobehavioral Reviews},
       volume = {64},
        pages = {167-174},
         year = {2016},
      abstract = {The dentate gyrus is the main entrance of highly processed information to the hippocampus which derives from associative cortices and it is one of the few privileged areas in the brain where adult neurogenesis occurs. This creates the unique situation that neurons of diverse maturation stages are part of one neuronal network at any given point in life. While recently adult-born cells have a low induction threshold for long-term potentiation several studies suggest that following maturation granule cells are poorly excitable and they exhibit reduced Hebbian synaptic plasticity to an extent that it was even suggested that they functionally retire. Here, we review the functional properties of mature granule cells and discuss how plasticity of intrinsic excitability and alterations in excitation-inhibition balance might impact on their role in hippocampal information processing.}
}
@article{LopezandBalaguerFiHN-17,
       author = {L\1{o}pez-Barroso, Diana and de Diego-Balaguer, Ruth},
        title = {Language Learning Variability within the Dorsal and Ventral Streams as a Cue for Compensatory Mechanisms in Aphasia Recovery},
      journal = {Frontiers in Human Neuroscience},
       volume = {11},
         year = {2017},
        pages = {476},
     abstract = {Dorsal and ventral pathways connecting perisylvian language areas have been shown to be functionally and anatomically segregated. Whereas the dorsal pathway integrates the sensory-motor information required for verbal repetition, the ventral pathway has classically been associated with semantic processes. The great individual differences characterizing language learning through life partly correlate with brain structure and function within these dorsal and ventral language networks.  Variability and plasticity within these networks also underlie inter-individual differences in the recovery of linguistic abilities in aphasia. Despite the division of labor of the dorsal and ventral streams, studies in healthy individuals have shown how the interaction of them and the redundancy in the areas they connect allow for compensatory strategies in functions that are usually segregated. In this mini-review we highlight the need to examine compensatory mechanisms between streams in healthy individuals as a helpful guide to choosing the most appropriate rehabilitation strategies, using spared functions and targeting preserved compensatory networks for brain plasticity.}
}
@article{LopreoreetalBJ-08,
       author = {Courtney L. Lopreore and Thomas M. Bartol and Jay S. Coggan and Daniel X. Keller and Gina E. Sosinsky and Mark H. Ellisman and Terrence J. Sejnowski},
        title = {Computational Modeling of Three-Dimensional Electrodiffusion in Biological Systems: {Application to the Node of Ranvier}},
      journal = {Biophysical Journal},
       volume = {95},
        issue = 6,
         year = {2008},
        pages = {2624-2635},
     abstract = {A computational model is presented for the simulation of three-dimensional electrodiffusion of ions. Finite volume techniques were used to solve the Poisson-Nernst-Planck equation, and a dual Delaunay-Voronoi mesh was constructed to evaluate fluxes of ions, as well as resulting electric potentials. The algorithm has been validated and applied to a generalized node of Ranvier, where numerical results for computed action potentials agree well with cable model predictions for large clusters of voltage-gated ion channels. At smaller channel clusters, however, the three-dimensional electrodiffusion predictions diverge from the cable model predictions and show a broadening of the action potential, indicating a significant effect due to each channel's own local electric field. The node of Ranvier complex is an elaborate organization of membrane-bound aqueous compartments, and the model presented here represents what we believe is a significant first step in simulating electrophysiological events with combined realistic structural and physiological data.}
}
@article{LorenzetalPNAS-11,
        title = {How social influence can undermine the wisdom of crowd effect},
       author = {Jan Lorenz and Heiko Rauhut and Frank Schweitzer and Dirk Helbing},
      journal = {Proceedings of the National Academy of Science},
       volume = 108,
        issue = 36,
         year = 2011,
}
@article{LotteretalCoRR-16,
       author = {William Lotter and Gabriel Kreiman and David D. Cox},
        title = {Deep Predictive Coding Networks for Video Prediction and Unsupervised Learning},
      journal = {CoRR},
       volume = {arXiv:1605.08104},
         year = {2016},
     abstract = {While great strides have been made in using deep learning algorithms to solve supervised learning tasks, the problem of unsupervised learning - leveraging un- labeled examples to learn about the structure of a domain - remains a difficult unsolved challenge. Here, we explore prediction of future frames in a video sequence as an unsupervised learning rule for learning about the structure of the visual world. We describe a predictive neural network (PredNet) architecture that is inspired by the concept of - predictive coding - from the neuroscience literature. These networks learn to predict future frames in a video sequence, with each layer in the network making local predictions and only forwarding deviations from those predictions to subsequent network layers. We show that these networks are able to robustly learn to predict the movement of synthetic (rendered) objects, and that in doing so, the networks learn internal representations that are useful for decoding latent object parameters (e.g. pose) that support object recognition with fewer training views. We also show that these networks can scale to complex natural image streams (car-mounted camera videos), capturing key aspects of both egocentric movement and the movement of objects in the visual scene, and the representation learned in this setting is useful for estimating the steering angle. Altogether, these results suggest that prediction represents a powerful framework for unsupervised learning, allowing for implicit learning of object and scene structure.}
}
@article{LotteretalICLR-17,
        title = {Deep predictive coding networks for video prediction and unsupervised learning},
       author = {William Lotter and Gabriel Kreiman and David Cox},
      journal = {Proceedings of the 5th International Conference on Learning Representations},
         year = {2017},
     abstract = {While great strides have been made in using deep learning algorithms to solve supervised learning tasks, the problem of unsupervised learning -- leveraging unlabeled  examples to learn about the structure of a domain -- remains a difficult  unsolved challenge. Here, we explore prediction of future frames in a video sequence  as an unsupervised learning rule for learning about the structure of the  visual world. We describe a predictive neural network ("PredNet") architecture  that is inspired by the concept of "predictive coding" from the neuroscience literature.  These networks learn to predict future frames in a video sequence, with  each layer in the network making local predictions and only forwarding deviations  from those predictions to subsequent network layers. We show that these networks  are able to robustly learn to predict the movement of synthetic (rendered) objects,  and that in doing so, the networks learn internal representations that are useful  for decoding latent object parameters (e.g. pose) that support object recognition  with fewer training views. We also show that these networks can scale to complex  natural image streams (car-mounted camera videos), capturing key aspects  of both egocentric movement and the movement of objects in the visual scene,  and the representation learned in this setting is useful for estimating the steering  angle. Altogether, these results suggest that prediction represents a powerful  framework for unsupervised learning, allowing for implicit learning of object and  scene structure.},
}
@article{LourencoPIAGET_AND_VYGOTSKY-12,
        title = {Piaget and {V}ygotsky: {M}any resemblances, and a crucial difference},
       author = {Orlando Louren\c{c}o},
      journal = {New Ideas in Psychology},
       volume = {30},
       number = {3},
         year = {2012},
        pages = {281-295},
     abstract = {Piaget and Vygotsky are two influential developmental psychologists. One can even say that their contributions to developmental psychology, albeit different, are similarly remarkable and unique. This article is in four parts. In the first part, I refer briefly to a commonly noticed difference between Vygotsky's and Piaget's theories. In the second part, I show that there are many resemblances between Vygotsky and Piaget. In the third part, I argue that in spite of such resemblances, there exists a crucial, and generally unnoticed, difference between Piaget's and Vygotsky's theories, and that this difference underlies the way each author addresses the following issues: 1) the origins of development and the motor of development; 2) the relationships among equal peers vs. those based on authorities, as they are sources of development and learning; 3) the more appropriate methods for studying developmental changes; 4) the importance of the distinction between true vs. necessary knowledge; and 5) the role of transformation and personal reconstruction vs. that of transmission and social influence in the phenomena of development and learning. Finally, I summarize the main ideas and arguments which I elaborate throughout this article, and mention what can be gained when the generally ignored aforementioned difference is noticed.}
}
@inproceedings{LouwerseandCrossleyFLAIRS-06,
      title = {Dialog act classification using N-gram algorithms},
     author = {Louwerse, Max M. and Crossley, Scott A.},
  booktitle = {{FLAIRS} Conference},
     editor = {Sutcliffe, Geoff and Goebel, Randy},
  publisher = {AAAI Press},
       year = 2006,
      pages = {758-763},
}
@article{LovejoyAOR-91,
       author = {Lovejoy, William S.},
        title = {A Survey of Algorithmic Methods for Partially Observed {M}arkov
              Decision Processes},
      journal = {Annals of Operations Research},
       volume = 28,
        pages = {47-66},
         year = 1991,
}
@article{LovejoyOR-91,
       author = {Lovejoy, William S.},
        title = {Computationally Feasible Bounds for Partially Observed {M}arkov Decision Processes},
      journal = {Operations Research},
       volume = 39,
        issue = 1,
         year = 1991,
        pages = {162-175},
}
@article{LoweIJCV-04,
       author = {Lowe, David},
        title = {Distinctive Image Features from Scale-Invariant Keypoints},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 60,
        issue = 2,
        pages = {91-110},
         year = 2004,
}
@article{Lozano-Perez83,
       author = {Tom\'{a}s Lozano-P\'{e}rez},
        title = {Spatial Planning: A Configuration Space Approach},
      journal = {{IEEE} Transactions on Computers},
       volume = 32,
         year = 1983,
        pages = {108-120},
}
@article{Lozano-PerezIEEE-83,
       author = {Lozano-P\'{e}rez, Tom\'{a}s},
        title = {Spatial Planning: A Configuration Space Approach},
      journal = {{IEEE} Transactions on Computers},
       volume = 32,
         year = 1983,
        pages = {108-120},
}
@inproceedings{LozanoandSindhwaniNIPS-10,
       author = {Lozano, Aur\'{e}lie C. and Sindhwani, Vikas},
        title = {Block Variable Selection in Multivariate Regression and High-dimensional Causal Inference},
    booktitle = {Proceedings of the 23rd International Conference on Neural Information Processing Systems},
         year = {2010},
     location = {Vancouver, British Columbia, Canada},
        pages = {1486-1494},
    publisher = {Curran Associates Inc.},
     abstract = {We consider multivariate regression problems involving high-dimensional predictor and response spaces. To efficiently address such problems, we propose a variable selection method, Multivariate Group Orthogonal Matching Pursuit, which extends the standard Orthogonal Matching Pursuit technique. This extension accounts for arbitrary sparsity patterns induced by domain-specific groupings over both input and output variables, while also taking advantage of the correlation that may exist between the multiple outputs. Within this framework, we then formulate the problem of inferring causal relationships over a collection of high-dimensional time series variables. When applied to time-evolving social media content, our models yield a new family of causality-based influence measures that may be seen as an alternative to the classic PageRank algorithm traditionally applied to hyperlink graphs. Theoretical guarantees, extensive simulations and empirical studies confirm the generality and value of our framework.}
}
@article{LubecketalNATURE-METHODS-14,
       author = {Lubeck, Eric and Coskun, Ahmet F. and Zhiyentayev, Timur and Ahmad, Mubhij and Cai, Long},
        title = {Single-cell in situ RNA profiling by sequential hybridization},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = 11,
        issue = 4,
         year = 2014,
        pages = {360-361},
}
@article{LubkeMRT-93,
       author = {Joachim L\"{u}bke},
        title = {Photoconversion of diaminobenzidine with different fluorescent neuronal markers into a light and electron microscopic dense reaction product},
      journal = {Microscopy Research and Technique},
       volume = 24,
         year = 1993,
        pages = {2-14},
     abstract = {This article describes methods for photoconverting diaminobenzidine (DAB) into a stable, light and electron microscopically visible dark reaction product in neurons which contain a fluorescent dye. Photoconversion of DAB has been achieved so far with the following fluorescent dyes: rhodamine labeled latex microspheres (RLM), 4,6-diamidino-2-phenylindole (DAPI), 5,7-di-hydroxytryptamine (5,7-DHT), Fast Blue (FB), Nuclear Yellow (NY), Diamidino Yellow (DY), Evans Blue (EB), acridine orange (AO), ethidium bromide (EBR),1,1′-dioctadecyl-3,3,3′,3′-tetramethylindolcarbocyanine perchlorate, D-282 (DiI), propidium iodide (PI), and intracellularly injected Lucifer Yellow (LY). The dye is introduced into the neurons by tinctorial staining, retrograde transport, or intracellular injection. Photoconversion is conducted by incubating the tissue with the fluorescent substance-containing cells in a DAB solution under simultaneous strong illumination with ultraviolet (UV) light. During the formation of the reaction product, the fluorescence disappears from the cell. In all cases, photoconversion provided a stable, nonfading DAB reaction product for light microscopy. In addition, at the electron microscopic level, it appeared that the photoconversion results in a homogeneously distributed, fine granular, dark, intracellularly located reaction product. With most of the retrograde tracers tested, photoconversion led only to staining of the cell bodies and the proximal portions of primary dendrites. Following photoconversion with intracellularly LY-filled neurons and cells labeled retrogradely with DiI, DiO, and 5,7-DHT, the reaction product was present throughout the cells, extending from the cell bodies into dendrites and dendritic appendices, and into axons. The high selectivity and methodological simplicity of photoconversion of DAB with fluorescent dyes into a stable, light and electron microscopical dense reaction product provide a promising alternative to classical neuroanatomical techniques and a new useful application of fluorescent neuronal tracers to light and electron microscopy.}
}
@article{Lubyetal01,
       author = {M. G. Luby and M. Mitzenmacher and M. A. Shokrollahi and D. A. Spielman},
        title = {Improved Low-Density Parity-Check Codes Using Irregular Graphs},
      journal = {IEEE Transactions on Information Theory},
       volume = {47:2},
        pages = {585-599},
         year = 2001,
}
@article {LubyetalPNAS-12,
       author = {Luby, Joan L. and Barch, Deanna M. and Belden, Andy and Gaffrey, Michael S. and Tillman, Rebecca and Babb, Casey and Nishino, Tomoyuki and Suzuki, Hideo and Botteron, Kelly N.},
        title = {Maternal support in early childhood predicts larger hippocampal volumes at school age},
    publisher = {National Academy of Sciences},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {109},
       number = {8},
         year = {2012},
        pages = {2854-2859},
     abstract = {Early maternal support has been shown to promote specific gene expression, neurogenesis, adaptive stress responses, and larger hippocampal volumes in developing animals. In humans, a relationship between psychosocial factors in early childhood and later amygdala volumes based on prospective data has been demonstrated, providing a key link between early experience and brain development. Although much retrospective data suggests a link between early psychosocial factors and hippocampal volumes in humans, to date there has been no prospective data to inform this potentially important public health issue. In a longitudinal study of depressed and healthy preschool children who underwent neuroimaging at school age, we investigated whether early maternal support predicted later hippocampal volumes. Maternal support observed in early childhood was strongly predictive of hippocampal volume measured at school age. The positive effect of maternal support on hippocampal volumes was greater in nondepressed children. These findings provide prospective evidence in humans of the positive effect of early supportive parenting on healthy hippocampal development, a brain region key to memory and stress modulation.},
}
@inproceedings{LucasandKanadeIJCAI-81,
       author = {Bruce Lucas and Takeo Kanade},
        title = {An Iterative Image Registration Technique with an Application to Stereo Vision},
    booktitle = {Proceedings of the 7th International Joint Conference on Artificial Intelligence},
        pages = {674-679},
         year = 1981,
}
@article{LucicandBaumeisterSCIENCE-05,
       author = {Vladan Lucic and Wolfgang Baumeister},
        title = {Monte Carlo Places Strong Odds on Ectopic Release},
      journal = {Science},
       volume = 309,
        issue = 5733,
         year = 2005,
        pages = {387-388},
}
@article{LucinietalCoRR-19,
       author = {Francesca Arese Lucini and Gino Del Ferraro and Mariano Sigman and Hernan A. Makse},
        title = {How the Brain Transitions from Conscious to Subliminal Perception},
      journal = {CoRR},
       volume = {arXiv:1903.09630},
         year = {2019},
     abstract = {We study the transition in the functional networks that characterize the human brains' conscious-state to an unconscious subliminal state of perception by using k-core percolation. We find that the most inner core (i.e., the most connected kernel) of the conscious-state functional network corresponds to areas which remain functionally active when the brain transitions from the conscious-state to the subliminal-state. That is, the inner core of the conscious network coincides with the subliminal-state. Mathematical modeling allows to interpret the conscious to subliminal transition as driven by k-core percolation, through which the conscious state is lost by the inactivation of the peripheral k-shells of the conscious functional network. Thus, the inner core and most robust component of the conscious brain corresponds to the unconscious subliminal state. This finding imposes constraints to theoretical models of consciousness, in that the location of the core of the functional brain network is in the unconscious part of the brain rather than in the conscious state as previously thought.}
}
@book{Lucky91,
       author = {Silicon Dreams: Information, Man, and Machine},
        title = {Robert W. Lucky},
    publisher = {St. Martin's Press},
      address = {New York},
         year = 1991,
}
@misc{LuddenPSYCHOLOGY-TODAY-16,
        title = {Why You Can't Trust Your Memory (of Anything)},
       author = {David Ludden},
 howpublished = {Psychology Today},
         year = {2016},
}
@article{Ludwig95,
        title = {A Subexponential Randomized Algorithm for the Simple Stochastic Game Problem},
       author = {Walter Ludwig},
      journal = {Information and Computation},
         year = 1995,
}
@book{Luenberger1973,
       author = {Luenberger, D. G.},
        title = {Introduction to Linear and Nonlinear Programming},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1973,
}
@book{Luetal2017medical,
        title = {Deep Learning and Convolutional Neural Networks for Medical Image Computing: Precision Medicine, High Performance and Large-Scale Datasets},
       author = {Lu, L. and Zheng, Y. and Carneiro, G. and Yang, L.},
       series = {Advances in Computer Vision and Pattern Recognition},
         year = {2017},
    publisher = {Springer International Publishing}
}
@article{LuetalCoRR-17,
       author = {Jiasen Lu and Anitha Kannan and Jianwei Yang and Devi Parikh and Dhruv Batra},
        title = {Best of Both Worlds: Transferring Knowledge from Discriminative Learning to a Generative Visual Dialog Model},
      journal = {CoRR},
       volume = {arXiv:1706.01554},
         year = {2017},
          url = {http://arxiv.org/abs/1706.01554},
     abstract = {We present a novel training framework for neural sequence models, particularly for grounded dialog generation. The standard training paradigm for these models is maximum likelihood estimation (MLE), or minimizing the cross-entropy of the human responses. Across a variety of domains, a recurring problem with MLE trained generative neural dialog models (G) is that they tend to produce 'safe' and generic responses ("I don't know", "I can't tell"). In contrast, discriminative dialog models (D) that are trained to rank a list of candidate human responses outperform their generative counterparts; in terms of automatic metrics, diversity, and informativeness of the responses. However, D is not useful in practice since it cannot be deployed to have real conversations with users.  Our work aims to achieve the best of both worlds -- the practical usefulness of G and the strong performance of D -- via knowledge transfer from D to G. Our primary contribution is an end-to-end trainable generative visual dialog model, where G receives gradients from D as a perceptual (not adversarial) loss of the sequence sampled from G. We leverage the recently proposed Gumbel-Softmax (GS) approximation to the discrete distribution -- specifically, an RNN augmented with a sequence of GS samplers, coupled with the straight-through gradient estimator to enable end-to-end differentiability. We also introduce a stronger encoder for visual dialog, and employ a self-attention mechanism for answer encoding along with a metric learning loss to aid D in better capturing semantic similarities in answer responses. Overall, our proposed model outperforms state-of-the-art on the VisDial dataset by a significant margin (2.67\% on recall@10). The source code can be downloaded from this https URL.}
}
@article{LuetalCoRR-21,
        title = {Pretrained Transformers as Universal Computation Engines},
       author = {Kevin Lu and Aditya Grover and Pieter Abbeel and Igor Mordatch},
         year = {2021},
       volume = {arXiv:2103.05247},
     abstract = {We investigate the capability of a transformer pretrained on natural language to generalize to other modalities with minimal finetuning - in particular, without finetuning of the self-attention and feedforward layers of the residual blocks. We consider such a model, which we call a Frozen Pretrained Transformer (FPT), and study finetuning it on a variety of sequence classification tasks spanning numerical computation, vision, and protein fold prediction. In contrast to prior works which investigate finetuning on the same modality as the pretraining dataset, we show that pretraining on natural language improves performance and compute efficiency on non-language downstream tasks. In particular, we find that such pretraining enables FPT to generalize in zero-shot to these modalities, matching the performance of a transformer fully trained on these tasks.}
}
@article{LuetalNATURE-COMMUNICATIONS-18,
       author = {Lu, Ake T. and Xue, Luting and Salfati, Elias L. and Chen, Brian H. and Ferrucci, Luigi and Levy, Daniel and Joehanes, Roby and Murabito, Joanne M. and Kiel, Douglas P. and Tsai, Pei-Chien and Yet, Idil and Bell, Jordana T. and Mangino, Massimo and Tanaka, Toshiko and McRae, Allan F. and Marioni, Riccardo E. and Visscher, Peter M. and Wray, Naomi R. and Deary, Ian J. and Levine, Morgan E. and Quach, Austin and Assimes, Themistocles and Tsao, Philip S. and Absher, Devin and Stewart, James D. and Li, Yun and Reiner, Alex P. and Hou, Lifang and Baccarelli, Andrea A. and Whitsel, Eric A. and Aviv, Abraham and Cardona, Alexia and Day, Felix R. and Wareham, Nicholas J. and Perry, John R. B. and Ong, Ken K. and Raj, Kenneth and Lunetta, Kathryn L. and Horvath, Steve},
        title = {{GWAS} of epigenetic aging rates in blood reveals a critical role for {TERT}},
      journal = {Nature Communications},
    publisher = {Nature Publishing Group},
       volume = {6},
       number = {387},
         year = {2018},
     abstract = {DNA methylation age is an accurate biomarker of chronological age and predicts lifespan, but its underlying molecular mechanisms are unknown. In this genome-wide association study of 9907 individuals, we find gene variants mapping to five loci associated with intrinsic epigenetic age acceleration (IEAA) and gene variants in three loci associated with extrinsic epigenetic age acceleration (EEAA). Mendelian randomization analysis suggests causal influences of menarche and menopause on IEAA and lipoproteins on IEAA and EEAA. Variants associated with longer leukocyte telomere length (LTL) in the telomerase reverse transcriptase gene (TERT) paradoxically confer higher IEAA (P = 6.6 10^{-7}11). Causal modeling indicates TERT-specific and independent effects on LTL and IEAA. Experimental hTERT-expression in primary human fibroblasts engenders a linear increase in DNA methylation age with cell population doubling number. Together, these findings indicate a critical role for hTERT in regulating the epigenetic clock, in addition to its established role of compensating for cell replication-dependent telomere shortening.}
}
@inproceedings{LumSIGMOD-84,
       author = {Lum, V. and Dadam, P. and Erbe, R. and Guenauer, J. and Pistor, P. and Walch, G. and Werner, H. and Woodfill, J.},
        title = {Designing DBMS Support for the Temporal Dimension},
    booktitle = {Proceedings SIGMOD},
 organization = {ACM},
         year = 1984,
        pages = {115-130},
}
@article{Lumelsky,
       author = {Lumelsky, Vladimir J.},
        title = {Effect of Kinematics on Motion Planning for Planar Robot Arms Moving Amidst Unknown Obstacles},
      journal = {{IEEE} Journal of Robotics and Automation},
       volume = 3,
         year = 1987,
        pages = {207-223},
}
@article{LuoFiP-18,
       author = {Luo, Jun},
        title = {The Neural Basis of and a Common Neural Circuitry in Different Types of Pro-social Behavior},
      journal = {Frontiers in psychology},
    publisher = {Frontiers Media S.A.},
       volume = {9},
         year = {2018},
        pages = {859-859},
     abstract = {Pro-social behaviors are voluntary behaviors that benefit other people or society as a whole, such as charitable donations, cooperation, trust, altruistic punishment, and fairness. These behaviors have been widely described through non self-interest decision-making in behavioral experimental studies and are thought to be increased by social preference motives. Importantly, recent studies using a combination of neuroimaging and brain stimulation, designed to reveal the neural mechanisms of pro-social behaviors, have found that a wide range of brain areas, specifically the prefrontal cortex, anterior insula, anterior cingulate cortex, and amygdala, are correlated or causally related with pro-social behaviors. In this review, we summarize the research on the neural basis of various kinds of pro-social behaviors and describe a common shared neural circuitry of these pro-social behaviors. We introduce several general ways in which experimental economics and neuroscience can be combined to develop important contributions to understanding social decision-making and pro-social behaviors. Future research should attempt to explore the neural circuitry between the frontal lobes and deeper brain areas.},
}
@article {LuoetalSTEM-13,
       author = {Luo, Min and Ling, Te and Xie, Wenbing and Sun, He and Zhou, Yonggang and Zhu, Qiaoyun and Shen, Meili and Zong, Le and Lyu, Guoliang and Zhao, Yun and Ye, Tao and Gu, Jun and Tao, Wei and Lu, Zhigang and Grummt, Ingrid},
        title = {{NuRD} Blocks Reprogramming of Mouse Somatic Cells into Pluripotent Stem Cells},
      journal = {Stem Cells},
       volume = 31,
        issue = 7,
    publisher = {Wiley Subscription Services, Inc., A Wiley Company},
         year = 2013,
        pages = {1278-1286},
     abstract = {Reprogramming of somatic cells into induced pluripotent stem cells (iPSCs) by overexpression of a defined set of transcription factors requires epigenetic changes in pluripotency genes. Nuclear reprogramming is an inefficient process and the molecular mechanisms that reset the epigenetic state during iPSC generation are largely unknown. Here, we show that downregulation of the nucleosome remodeling and deacetylation (NuRD) complex is required for efficient reprogramming. Overexpression of Mbd3, a subunit of NuRD, inhibits induction of iPSCs by establishing heterochromatic features and silencing embryonic stem cell-specific marker genes, including Oct4 and Nanog. Depletion of Mbd3, on the other hand, improves reprogramming efficiency and facilitates the formation of pluripotent stem cells that are capable of generating viable chimeric mice, even in the absence of c-Myc or Sox2. The results establish Mbd3/NuRD as an important epigenetic regulator that restricts the expression of key pluripotency genes, suggesting that drug-induced downregulation of Mbd3/NuRD may be a powerful means to improve the efficiency and fidelity of reprogramming.},
}
@inproceedings{LuongetalEMNLP-15,
       author = {Minh{-}Thang Luong and Hieu Pham and Christopher D. Manning},
        title = {Effective Approaches to Attention-based Neural Machine Translation},
    booktitle = {Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing},
        pages = {1412-1421},
         year = {2015},
     abstract = {An attentional mechanism has lately been used to improve neural machine translation (NMT) by selectively focusing on parts of the source sentence during translation. However, there has been little work exploring useful architectures for attention-based NMT. This paper examines two simple and effective classes of attentional mechanism: a global approach which always attends to all source words and a local one that only looks at a subset of source words at a time. We demonstrate the effectiveness of both approaches over the WMT translation tasks between English and German in both directions. With local attention, we achieve a significant gain of 5.0 BLEU points over non-attentional systems which already incorporate known techniques such as dropout. Our ensemble model using different attention architectures has established a new state-of-the-art result in the WMT'15 English to German translation task with 25.9 BLEU points, an improvement of 1.0 BLEU points over the existing best system backed by NMT and an n-gram reranker.}
}
@article{LuqueetalIJNS-11,
       author = {Luque, N. R. and Garrido, J. A. and Carrillo, R. R. and Tolu, S. and Ros, E.},
        title = {{{A}daptive cerebellar spiking model embedded in the control loop: context switching and robustness against noise}},
      journal = {International Journal of Neural Systems},
       volume = {21},
       number = {5},
         year = {2011},
        pages = {385-401},
     abstract = {This work evaluates the capability of a spiking cerebellar model embedded in different loop architectures (recurrent, forward, and forward \& recurrent) to control a robotic arm (three degrees of freedom) using a biologically-inspired approach. The implemented spiking network relies on synaptic plasticity (long-term potentiation and long-term depression) to adapt and cope with perturbations in the manipulation scenario: changes in dynamics and kinematics of the simulated robot. Furthermore, the effect of several degrees of noise in the cerebellar input pathway (mossy fibers) was assessed depending on the employed control architecture. The implemented cerebellar model managed to adapt in the three control architectures to different dynamics and kinematics providing corrective actions for more accurate movements. According to the obtained results, coupling both control architectures (forward \& recurrent) provides benefits of the two of them and leads to a higher robustness against noise.}
}
@book{LuriaHIGHER-CORTICAL_FUNCTION-IN-MAN-66,
        title = {Higher Cortical Function in Man},
       author = {Luria, A.R.},
         year = {1966},
    publisher = {Basic Books}
}
@article{LuuetalPLOS-10,
       author = {Luu, Phan AND Geyer, Alexandra AND Fidopiastis, Cali AND Campbell, Gwendolyn AND Wheeler, Tracey AND Cohn, Joseph AND Tucker, Don M.},
        title = {Reentrant Processing in Intuitive Perception},
      journal = {PLOS ONE},
    publisher = {Public Library of Science},
       volume = {5},
       number = {3},
         year = {2010},
        pages = {1-10},
     abstract = {The process of perception requires not only the brain's receipt of sensory data but also the meaningful organization of that data in relation to the perceptual experience held in memory. Although it typically results in a conscious percept, the process of perception is not fully conscious. Research on the neural substrates of human visual perception has suggested that regions of limbic cortex, including the medial orbital frontal cortex (mOFC), may contribute to intuitive judgments about perceptual events, such as guessing whether an object might be present in a briefly presented fragmented drawing. Examining dense array measures of cortical electrical activity during a modified Waterloo Gestalt Closure Task, results show, as expected, that activity in medial orbital frontal electrical responses (about 250 ms) was associated with intuitive judgments. Activity in the right temporal-parietal-occipital (TPO) region was found to predict mOFC (∼150 ms) activity and, in turn, was subsequently influenced by the mOFC at a later time (∼300 ms). The initial perception of gist or meaning of a visual stimulus in limbic networks may thus yield reentrant input to the visual areas to influence continued development of the percept. Before perception is completed, the initial representation of gist may support intuitive judgments about the ongoing perceptual process.},
}
@inproceedings{LyonICASP-82,
       author = {Lyon, R.F.},
    booktitle = {IEEE International Conference on Acoustics, Speech, and Signal Processing},
        title = {A computational model of filtering, detection, and compression in the cochlea},
         year = 1982,
       volume = 7,
        pages = {1282-1285},
     abstract = {We claim that speech analysis algorithms should be based on computational models of human audition, starting at the ears. While much is known about how hearing works, little of this knowledge has been applied in the speech analysis field. We propose models of the inner ear, or cochlea, which are expressed as time- and place-domain signal processing operations; i.e. the models are computational expressions of the important functions of the cochlea. The main parts of the models concern mechanical filtering effects and the mapping of mechanical vibrations into neural representation. Our model cleanly separates these effects into time-invariant linear filtering based on a simple cascade/parallel filterbank network of second-order sections, plus transduction and compression based on half-wave rectification with a nonlinear coupled automatic gain control network. Compared to other speech analysis techniques, this model does a much better job of preserving important detail in both time and frequency, which is important for robust sound analysis. We discuss the ways in which this model differs from more detailed cochlear models.},
}
@book{Lytton2002,
        title = {From Computer to Brain: Foundations of Computational Neuroscience},
       author = {Lytton, William W.},
         year = 2002,
    publisher = {Springer}
}
@article{MaasandMarkramJoCSS-04,
       author = {Wolfgang Maass and Henry Markram},
        title = {On the computational power of circuits of spiking neurons},
      journal = {Journal of Computer and System Sciences},
       volume = {69},
       number = {4},
         year = {2004},
        pages = {593-616},
     abstract = {Complex real-time computations on multi-modal time-varying input streams are carried out by generic cortical microcircuits. Obstacles for the development of adequate theoretical models that could explain the seemingly universal power of cortical microcircuits for real-time computing are the complexity and diversity of their computational units (neurons and synapses), as well as the traditional emphasis on offline computing in almost all theoretical approaches towards neural computation. In this article, we initiate a rigorous mathematical analysis of the real-time computing capabilities of a new generation of models for neural computation, liquid state machines, that can be implemented with - in fact benefit from - diverse computational units. Hence, realistic models for cortical microcircuits represent special instances of such liquid state machines, without any need to simplify or homogenize their diverse computational units. We present proofs of two theorems about the potential computational power of such models for real-time computing, both on analog input streams and for spike trains as inputs.}
}
@inproceedings{MaasetalCHIME-13,
        title = {Recurrent Neural Network Feature Enhancement: The 2nd {CHiME} Challenge},
       author = {Andrew L. Maas and Tyler M. O'Neil and Awni Y. Hannun and Andrew Y. Ng},
    booktitle = {The 2nd International Workshop on Machine Listening in Multisource Environments},
         year = 2013,
}
@incollection{MaassLIQUID-10,
       author = {W. Maass},
        title = {Liquid State Machines: Motivation, Theory, and Applications},
    booktitle = {Computability in Context: Computation and Logic in the Real World},
    publisher = {Imperial College Press},
       editor = {B. Cooper and A. Sorbi},
         year = {2010},
        pages = {275-296},
     abstract = {The Liquid State Machine (LSM) has emerged as a computational model that is more adequate than the Turing machine for describing computations in biological networks of neurons. Characteristic features of this new model are (i) that it is a model for adaptive computational systems, (ii) that it provides a method for employing randomly connected circuits, or eve "found" physical objects for meaningful computations, (iii) that it provides a theoretical context where heterogeneous, rather than stereotypical, local gates or processors increase the computational power of a circuit, (iv) that it provides a method for multiplexing different computations (on a common input) within the same circuit. This chapter reviews the motivation for this model, its theoretical background, and current work on implementations of this model in innovative artificial computing devices.}
}
@article{MaassetalJPP-04,
       author = {Maass, W. and Natschl\"{a}ger, T. and Markram, H.},
        title = {Fading memory and kernel properties of generic cortical microcircuit models},
      journal = {Journal of Physiology Paris},
         year = {2004},
       volume = {98},
       number = {4-6},
        pages = {315-330},
     abstract = {It is quite difficult to construct circuits of spiking neurons that can carry out complex computational tasks. On the other hand even randomly connected circuits of spiking neurons can in principle be used for complex computational tasks such as time-warp invariant speech recognition. This is possible because such circuits have an inherent tendency to integrate incoming information in such a way that simple linear readouts can be trained to transform the current circuit activity into the target output for a very large number of computational tasks. Consequently we propose to analyze circuits of spiking neurons in terms of their roles as analog fading memory and non-linear kernels, rather than as implementations of specific computational operations and algorithms. This article is a sequel to [W. Maass, T. Natschlger, H. Markram, Real-time computing without stable states: a new framework for neural computation based on perturbations, Neural Comput. 14 (11) (2002) 2531-2560, Online available as #130 from: ], and contains new results about the performance of generic neural microcircuit models for the recognition of speech that is subject to linear and non-linear time-warps, as well as for computations on time-varying firing rates. These computations rely, apart from general properties of generic neural microcircuit models, just on capabilities of simple linear readouts trained by linear regression. This article also provides detailed data on the fading memory property of generic neural microcircuit models, and a quick review of other new results on the computational power of such circuits of spiking neurons.}
}
@article{MaassetalNC-02,
       author = {Maass, W. and Natschl\"{a}ger, T. and Markram, H.},
        title = {Real-time computing without stable states: a new framework for neural computation based on perturbations},
      journal = {Neural Compututation},
       volume = {14},
       number = {11},
         year = {2002},
        pages = {2531-2560},
     abstract = {A key challenge for neural modeling is to explain how a continuous stream of multimodal input from a rapidly changing environment can be processed by stereotypical recurrent circuits of integrate-and-fire neurons in real time. We propose a new computational model for real-time computing on time-varying input that provides an alternative to paradigms based on Turing machines or attractor neural networks. It does not require a task-dependent construction of neural circuits. Instead, it is based on principles of high-dimensional dynamical systems in combination with statistical learning theory and can be implemented on generic evolved or found recurrent circuitry. It is shown that the inherent transient dynamics of the high-dimensional dynamical system formed by a sufficiently large and heterogeneous neural circuit may serve as universal analog fading memory. Readout neurons can learn to extract in real time from the current state of such recurrent neural circuit information about current and past inputs that may be needed for diverse tasks. Stable internal states are not required for giving a stable output, since transient internal states can be transformed by readout neurons into stable target outputs due to the high dimensionality of the dynamical system. Our approach is based on a rigorous computational model, the liquid state machine, that, unlike Turing machines, does not require sequential transitions between well-defined discrete internal states. It is supported, as the Turing machine is, by rigorous mathematical results that predict universal computational power under idealized conditions, but for the biologically more realistic scenario of real-time processing of time-varying inputs. Our approach provides new perspectives for the interpretation of neural coding, the design of experiments and data analysis in neurophysiology, and the solution of problems in robotics and neurotechnology.}
}
@misc{MacCartneyPhD-09,
        title = {Natural Language Inference},
       author = {MacCartney, Bill},
         year = 2009,
 howpublished = {Ph.D. Thesis. Computer Science Department, Stanford University}
}
@inproceedings{MacCartneyandManningACL-07,
       author = {Bill MacCartney and Christopher Manning},
        title = {Natural logic for textual inference},
    booktitle = {Proceedings of {ACL} Workshop on Textual Entailment and Paraphrasing},
         year = 2007
}
@inproceedings{MacCartneyandManningACL-08,
       author = {MacCartney, Bill and Manning, Christopher D.},
        title = {Modeling Semantic Containment and Exclusion in Natural Language Inference},
    booktitle = {Proceedings of the 22nd International Conference on Computational Linguistics - Volume 1},
         year = 2008,
     location = {Manchester, United Kingdom},
        pages = {521-528},
    publisher = {Association for Computational Linguistics},
      address = {Stroudsburg, PA, USA},
}
@inproceedings{MacCartneyandManningIWCS-09,
       author = {MacCartney, Bill and Manning, Christopher D.},
        title = {An Extended Model of Natural Logic},
    booktitle = {Proceedings of the Eighth International Conference on Computational Semantics},
         year = 2009,
     location = {Tilburg, The Netherlands},
        pages = {140-156},
    publisher = {Association for Computational Linguistics},
      address = {Stroudsburg, PA, USA},
}
@article{MacDonaldetalNEURON-11,
       author = {MacDonald, C. J. and Lepage, K. Q. and Eden, U. T. and Eichenbaum, H.},
        title = {Hippocampal 'time cells' bridge the gap in memory for discontiguous events},
      journal = {Neuron},
       volume = {71},
       number = {4},
         year = {2011},
        pages = {737-749},
     abstract = {The hippocampus is critical to remembering the flow of events in distinct experiences and, in doing so, bridges temporal gaps between discontiguous events. Here, we report a robust hippocampal representation of sequence memories, highlighted by "time cells" that encode successive moments during an empty temporal gap between the key events, while also encoding location and ongoing behavior. Furthermore, just as most place cells "remap" when a salient spatial cue is altered, most time cells form qualitatively different representations ("retime") when the main temporal parameter is altered. Hippocampal neurons also differentially encode the key events and disambiguate different event sequences to compose unique, temporally organized representations of specific experiences. These findings suggest that hippocampal neural ensembles segment temporally organized memories much the same as they represent locations of important events in spatially defined environments.}
}
@inproceedings{MacGlashanetalRSS-15,
        title = {Grounding English Commands to Reward Functions},
       author = {MacGlashan, James and Babes-Vroman, Monica and Marie desJardins and Littman, Michael L. and Muresan, Smaranda and Squire, Shawn and Tellex, Stefanie and Arumugam, Dilip and Yang, Lei},
    booktitle = {Robotics: Science and Systems},
         year = {2015},
     abstract = {As intelligent robots become more prevalent, methods to make interaction with the robots more accessible are increasingly important. Communicating the tasks that a person wants the robot to carry out via natural language, and training the robot to ground the natural language through demonstration, are especially appealing approaches for interaction, since they do not require a technical background. However, existing approaches map natural language commands to robot command languages that directly express the sequence of actions the robot should execute. This sequence is often specific to a particular situation and does not generalize to new situations. To address this problem, we present a system that grounds natural language commands into reward functions using demonstrations of different natural language commands being carried out in the environment. Because language is grounded to reward functions, rather than explicit actions that the robot can perform, commands can be high-level, carried out in novel environments autonomously, and even transferred to other robots with different action spaces. We demonstrate that our learned model can be both generalized to novel environments and transferred to a robot with a different action space than the action space used during training}
}
@book{MacKay2003information,
       author = {MacKay, David J. C.},
        title = {Information Theory, Inference \& Learning Algorithms},
         year = {2003},
    publisher = {Cambridge University Press},
      address = {New York, NY, USA},
}
@article{MacKayELECTRONIC-LETTERS-95,
       author = {D. J. C. MacKay},
      journal = {Electronics Letters},
        title = {Free energy minimisation algorithm for decoding and cryptanalysis},
       volume = {31},
       number = {6},
         year = {1995},
        pages = {446-447},
     abstract = {An algorithm is derived for inferring a binary vector s given noisy observations of As module 2, where A is a binary matrix. The binary vector is replaced by a vector of probabilities, optimised by free energy minimisation. Experiments on the inference of the state of a linear feedback shift register indicate that this algorithm supersedes the Meier and Staffelbach polynomial algorithm.},
}
@book{MacLean1990,
        title = {The Triune Brain in Evolution: Role in Paleocerebral Functions},
       author = {Paul D. MacLean},
    publisher = {Springer},
      address = {Berlin},
         year = 1990,
}
@article{MacLeanPNAS-16,
       author = {MacLean, E. L.},
        title = {Unraveling the evolution of uniquely human cognition},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {113},
       number = {23},
         year = {2016},
        pages = {6348-6354},
     abstract = {A satisfactory account of human cognitive evolution will explain not only the psychological mechanisms that make our species unique, but also how, when, and why these traits evolved. To date, researchers have made substantial progress toward defining uniquely human aspects of cognition, but considerably less effort has been devoted to questions about the evolutionary processes through which these traits have arisen. In this article, I aim to link these complementary aims by synthesizing recent advances in our understanding of what makes human cognition unique, with theory and data regarding the processes of cognitive evolution. I review evidence that uniquely human cognition depends on synergism between both representational and motivational factors and is unlikely to be accounted for by changes to any singular cognitive system. I argue that, whereas no nonhuman animal possesses the full constellation of traits that define the human mind, homologies and analogies of critical aspects of human psychology can be found in diverse nonhuman taxa. I suggest that phylogenetic approaches to the study of animal cognition-which can address questions about the selective pressures and proximate mechanisms driving cognitive change-have the potential to yield important insights regarding the processes through which the human cognitive phenotype evolved.}
}
@incollection{MacQueen67,
       author = {MacQueen, J.},
        title = {Some Methods for Classification and Analysis of Multivariate Observations},
     crossref = {LeCamandNeyman67},
        pages = {281-297},
}
@article {MachensetalSCIENCE-05,
       author = {Machens, Christian K. and Romo, Ranulfo and Brody, Carlos D.},
        title = {Flexible Control of Mutual Inhibition: A Neural Model of Two-Interval Discrimination},
      journal = {Science},
       volume = {307},
       number = {5712},
        pages = {1121-1124},
         year = {2005},
    publisher = {American Association for the Advancement of Science},
     abstract = {Networks adapt to environmental demands by switching between distinct dynamical behaviors. The activity of frontal-lobe neurons during two-interval discrimination tasks is an example of these adaptable dynamics. Subjects first perceive a stimulus, then hold it in working memory, and finally make a decision by comparing it with a second stimulus. We present a simple mutual-inhibition network model that captures all three task phases within a single framework. The model integrates both working memory and decision making because its dynamical properties are easily controlled without changing its connectivity. Mutual inhibition between nonlinear units is a useful design motif for networks that must display multiple behaviors.},
}
@article {MachensetalSCIENCE-05.supplement,
       author = {Machens, Christian K. and Romo, Ranulfo and Brody, Carlos D.},
        title = {Flexible Control of Mutual Inhibition: A Neural Model of Two-Interval Discrimination},
      journal = {Science},
       volume = {307},
       number = {5712},
        pages = {1121-1124},
         year = {2005},
}
@inproceedings{MacheryPSA-06,
        title = {Massive Modularity and Brain Evolution},
       author = {Edouard Machery},
    booktitle = {Philosophy of Science Association 20th Biennial Meeting},
     location = {Vancouver},
         year = 2006,
     abstract = {Quartz (2002) argues that some recent findings about the evolution of the brain  Finlay \& Darlington, 1995) are inconsistent with evolutionary psychologists? massive  modularity hypothesis. In substance, Quartz contends that since the volume of the neocortex  evolved in a concerted manner, natural selection did not act on neocortical systems  independently of each other, which is a necessary condition for the massive modularity of our  cognition to be true. I argue however that Quartz's argument fails to undermine the massive  modularity hypothesis.}
}
@article{MachnickaetalBBA-14,
       author = {Beata Machnicka and Aleksander Czogalla and Anita Hryniewicz-Jankowska and Dżamila M. Bogusawska and Renata Grochowalska and Elzbieta Heger and Aleksander F. Sikorski},
        title = {Spectrins: A structural platform for stabilization and activation of membrane channels, receptors and transporters},
      journal = {Biochimica et Biophysica Acta - Special issue on biomembranes},
       volume = 1838,
        issue = 2,
         year = 2014,
        pages = {620-634},
     abstract = {This review focuses on structure and functions of spectrin as a major component of the membrane skeleton. Recent advances on spectrin function as an interface for signal transduction mediation and a number of data concerning interaction of spectrin with membrane channels, adhesion molecules, receptors and transporters draw a picture of multifaceted protein. Here, we attempted to show the current depiction of multitask role of spectrin in cell physiology.},
}
@incollection{MackeetalNIPS-11,
        title = {Empirical models of spiking in neural populations},
       author = {Macke, Jakob H and Buesing, Lars and Cunningham, John P and Yu, Byron M and Shenoy, Krishna V and Sahani, Maneesh},
    booktitle = {Advances in Neural Information Processing Systems 24},
       editor = {J. Shawe-Taylor and R. S. Zemel and P. L. Bartlett and F. Pereira and K. Q. Weinberger},
    publisher = {Curran Associates, Inc.},
         year = {2011},
        pages = {1350-1358},
     abstract = {Neurons in the neocortex code and compute as part of a locally interconnected population. Large-scale multi-electrode recording makes it possible to access these population processes empirically by fitting statistical models to unaveraged data. What statistical structure best describes the concurrent spiking of cells within a local network? We argue that in the cortex, where firing exhibits extensive correlations in both time and space and where a typical sample of neurons still reflects only a very small fraction of the local population, the most appropriate model captures shared variability by a low-dimensional latent process evolving with smooth dynamics, rather than by putative direct coupling. We test this claim by comparing a latent dynamical model with realistic spiking observations to coupled generalised linear spike-response models (GLMs) using cortical recordings. We find that the latent dynamical approach outperforms the GLM in terms of goodness-of-fit, and reproduces the temporal correlations in the data more accurately. We also compare models whose observations models are either derived from a Gaussian or point-process models, finding that the non-Gaussian model provides slightly better goodness-of-fit and more realistic population spike counts.},
}
@book{Macrae1992,
       author = {Norman Macrae},
        title = {John von Neumann},
    publisher = {Pantheon Books},
      address = {New York, NY},
         year = 1992,
}
@book{Macwhinney1987,
       author = {MacWhinney, Brian},
        title = {Mechanisms of Language Acquisition},
       series = {Carnegie Mellon Symposia on Cognition Series},
    publisher = {Lawrence Erlbaum Associates},
      address = {Hillsdale, NJ, US},
         year = {1987},
}
@article{MadoreetalULTRASONICS-09,
        title = {Accelerated Focused Ultrasound Imaging},
       author = {Bruno Madore and P. Jason White and Kai Thomenius and Gregory T. Clement},
      journal = {IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control},
       volume = 56,
        issue = 12,
         year = 2009,
        pages = {2612-2623},
}
@article{MadryetalNEURON-18,
       author = {Madry, Christian and Kyrargyri, Vasiliki and Arancibia-C\'{a}rcamo, I. Lorena and Jolivet, Renaud and Kohsaka, Shinichi and Bryan, Robert M. and Attwell, David},
        title = {Microglial Ramification, Surveillance, and Interleukin-1Beta Release Are Regulated by the Two-Pore Domain K+ Channel THIK-1},
      journal = {Neuron},
         year = {2018},
    publisher = {Elsevier},
       volume = {97},
        issue = {2},
        pages = {299-312},
     abstract = {Microglia exhibit two modes of motility: they constantly extend and retract their processes to survey the brain, but they also send out targeted processes to envelop sites of tissue damage. We now show that these motility modes differ mechanistically. We identify the two-pore domain channel THIK-1 as the main K+ channel expressed in microglia in situ. THIK-1 is tonically active, and its activity is potentiated by P2Y12 receptors. Inhibiting THIK-1 function pharmacologically or by gene knockout depolarizes microglia, which decreases microglial ramification and thus reduces surveillance, whereas blocking P2Y12 receptors does not affect membrane potential, ramification, or surveillance. In contrast, process outgrowth to damaged tissue requires P2Y12 receptor activation but is unaffected by blocking THIK-1. Block of THIK-1 function also inhibits release of the pro-inflammatory cytokine interleukin-1? from activated microglia, consistent with K+ loss being needed for inflammasome assembly. Thus, microglial immune surveillance and cytokine release require THIK-1 channel activity.},
}
@book{Maes1990,
       editor = {Maes, Pattie},
        title = {Designing Autonomous Agents: Theory and Practice from Biology to Engineering and Back},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1990,
}
@article{MaesetalML-09,
       author = {Maes, Francis and Denoyer, Ludovic and Gallinari, Patrick},
        title = {Structured Prediction with Reinforcement Learning},
      journal = {Machine Learning},
       volume = 77,
        issue = {2-3},
         year = 2009,
        pages = {271-301},
      address = {We formalize the problem of Structured Prediction as a Reinforcement Learning task. We first define a Structured Prediction Markov Decision Process (SP-MDP), an instantiation of Markov Decision Processes for Structured Prediction and show that learning an optimal policy for this SP-MDP is equivalent to minimizing the empirical loss. This link between the supervised learning formulation of structured prediction and reinforcement learning (RL) allows us to use approximate RL methods for learning the policy. The proposed model makes weak assumptions both on the nature of the Structured Prediction problem and on the supervision process. It does not make any assumption on the decomposition of loss functions, on data encoding, or on the availability of optimal policies for training. It then allows us to cope with a large range of structured prediction problems. Besides, it scales well and can be used for solving both complex and large-scale real-world problems. We describe two series of experiments. The first one provides an analysis of RL on classical sequence prediction benchmarks and compares our approach with state-of-the-art SP algorithms. The second one introduces a tree transformation problem where most previous models fail. This is a complex instance of the general labeled tree mapping problem. We show that RL exploration is effective and leads to successful results on this challenging task. This is a clear confirmation that RL could be used for large size and complex structured prediction problems.}
}
@article{MaetalNATURE-16,
       author = {Ma, Zhiguo and Stork, Tobias and Bergles, Dwight E. and Freeman, Marc R.},
        title = {Neuromodulators signal through astrocytes to alter neural circuit activity and behaviour},
      journal = {Nature},
    publisher = {Macmillan Publishers Limited},
       volume = {539},
        issue = {7629},
         year = {2016},
        pages = {428-432},
     abstract = {Astrocytes associate with synapses throughout the brain and express receptors for neurotransmitters that can increase intracellular calcium (Ca2+). Astrocytic Ca2+ signalling has been proposed to modulate neural circuit activity, but the pathways that regulate these events are poorly defined and in vivo evidence linking changes in astrocyte Ca2+ levels to alterations in neurotransmission or behaviour is limited. Here we show that Drosophila astrocytes exhibit activity-regulated Ca2+ signalling in vivo. Tyramine and octopamine released from neurons expressing tyrosine decarboxylase 2 (Tdc2) signal directly to astrocytes to stimulate Ca2+ increases through the octopamine/tyramine receptor (Oct-TyrR) and the transient receptor potential (TRP) channel Water witch (Wtrw), and astrocytes in turn modulate downstream dopaminergic neurons. Application of tyramine or octopamine to live preparations silenced dopaminergic neurons and this inhibition required astrocytic Oct-TyrR and Wtrw. Increasing astrocyte Ca2+ signalling was sufficient to silence dopaminergic neuron activity, which was mediated by astrocyte endocytic function and adenosine receptors. Selective disruption of Oct-TyrR or Wtrw expression in astrocytes blocked astrocytic Ca2+ signalling and profoundly altered olfactory-driven chemotaxis and touch-induced startle responses. Our work identifies Oct-TyrR and Wtrw as key components of the astrocytic Ca2+ signalling machinery, provides direct evidence that octopamine- and tyramine-based neuromodulation can be mediated by astrocytes, and demonstrates that astrocytes are essential for multiple sensory-driven behaviours in Drosophila}
}
@article{MaetalNEUROSCIENCE-03,
       author = {Ma, W. and Chabot, J. G. and Powell, K. J. and Jhamandas, K. and Dickerson, I. M. and Quirion, R.},
        title = {Localization and modulation of calcitonin gene-related peptide-receptor component protein-immunoreactive cells in the rat central and peripheral nervous systems},
      journal = {Neuroscience},
       volume = {120},
       number = {3},
         year = {2003},
        pages = {677-694},
     abstract = {Calcitonin gene-related peptide (CGRP) is widely distributed in the central and peripheral nervous system. Its highly diverse biological activities are mediated via the G protein-coupled receptor that uniquely requires two accessory proteins for optimal function. CGRP receptor component protein (RCP) is a coupling protein necessary for CGRP-receptor signaling. In this study, we established the anatomical distribution of RCP in the rat central and peripheral nervous system and its relationship to CGRP immunoreactivity. RCP-immunoreactive (IR) perikarya are widely and selectively distributed in the cerebral cortex, septal nuclei, hippocampus, various hypothalamic nuclei, amygdala, nucleus colliculus, periaqueductal gray, parabrachial nuclei, locus coeruleus, cochlear nuclei, dorsal raphe nuclei, the solitary tractus nucleus and gracile nucleus, cerebellar cortex, various brainstem motor nuclei, the spinal dorsal and ventral horns. A sub-population of neurons in the dorsal root ganglia (DRG) and trigeminal ganglia were strongly RCP-IR. Overall, the localization of RCP-IR closely matched with that of CGRP-IR. We also determined whether RCP in DRG and dorsal horn neurons can be modulated by CGRP receptor blockade and pain-related pathological stimuli. The intrathecal injection of the antagonist CGRP(8-37) markedly increased RCP expression in the lumbar DRG and spinal dorsal horn. Carrageenan-induced plantar inflammation produced a dramatic bilateral increase in RCP expression in the dorsal horn while a partial sciatic nerve ligation reduced RCP expression in the ipsilateral superficial dorsal horn. Our data suggest that the distribution of RCP immunoreactivity is closely matched with CGRP immunoreactivity in most of central and peripheral nervous systems. The co-localization of RCP and CGRP in motoneurons and primary sensory neurons suggests that CGRP has an autocrine or paracrine effect on these neurons. Moreover, our data also suggest that RCP expression in DRG and spinal cord can be modulated during CGRP receptor blockade, inflammation or neuropathic pain and this CGRP receptor-associated protein is dynamically regulated.}
}
@article{MagistrettiandPellerinPTRS-1991,
     title = {Cellular mechanisms of brain energy metabolism and their relevance to functional brain imaging},
    author = {Magistretti, P.J. and Pellerin, L.},
   journal = {Philosophyical Transactions Royal Society London B Biological Science},
    volume = 354,
     issue = 1387,
      year = 1999,
     pages = {1155-1163},
  abstract = {Despite striking advances in functional brain imaging, the cellular and molecular mechanisms that underlie the signals detected by these techniques are still largely unknown. The basic physiological principle of functional imaging is represented by the tight coupling existing between neuronal activity and the associated local increase in both blood flow and energy metabolism. Positron emission tomography (PET) signals detect blood flow, oxygen consumption and glucose use associated with neuronal activity; the degree of blood oxygenation is currently thought to contribute to the signal detected with functional magnetic resonance imaging, while magnetic resonance spectroscopy (MRS) identifies the spatio-temporal pattern of the activity-dependent appearance of certain metabolic intermediates such as glucose or lactate. Recent studies, including those of neurotransmitter-regulated metabolic fluxes in purified preparations and analyses of the cellular localization of enzymes and transporters involved in energy metabolism, as well as in vivo microdialysis and MRS approaches have identified the neurotransmitter glutamate and astrocytes, a specific type of glial cell, as pivotal elements in the coupling of synaptic activity with energy metabolism. Astrocytes are ideally positioned to sense increases in synaptic activity and to couple them with energy metabolism. Indeed they possess specialized processes that cover the surface of intraparenchymal capillaries, suggesting that astrocytes may be a likely site of prevalent glucose uptake. Other astrocyte processes are wrapped around synaptic contacts which possess receptors and reuptake sites for neurotransmitters. Glutamate stimulates glucose uptake into astrocytes. This effect is mediated by specific glutamate transporters present on these cells. The activity of these transporters, which is tightly coupled to the synaptic release of glutamate and operates the clearance of glutamate from the extracellular space, is driven by the electrochemical gradient of Na+. This Na(+)-dependent uptake of glutamate into astrocytes triggers a cascade of molecular events involving the Na+/K(+)-ATPase leading to the glycolytic processing of glucose and the release of lactate by astrocytes. The stoichiometry of this process is such that for one glutamate molecule taken up with three Na+ ions, one glucose molecule enters an astrocyte, two ATP molecules are produced through aerobic glycolysis and two lactate molecules are released. Within the astrocyte, one ATP molecule fuels one 'turn of the pump' while the other provides the energy needed to convert glutamate to glutamine by glutamine synthase. Evidence has been accumulated from structural as well as functional studies indicating that, under aerobic conditions, lactate may be the preferred energy substrate of activated neurons. Indeed, in the presence of oxygen, lactate is converted to pyruvate, which can be processed through the tricarboxylic acid cycle and the associated oxidative phosphorylation, to yield 17 ATP molecules per lactate molecule. These data suggest that during activation the brain may transiently resort to aerobic glycolysis occurring in astrocytes, followed by the oxidation of lactate by neurons. The proposed model provides a direct mechanism to couple synaptic activity with glucose use and is consistent with the notion that the signals detected during physiological activation with 18F-deoxyglucose (DG)-PET may reflect predominantly uptake of the tracer into astrocytes. This conclusion does not question the validity of the 2-DG-based techniques, rather it provides a cellular and molecular basis for these functional brain imaging techniques.}
}
@article{MaglioneandSigristNATURE-NEUROSCIENCE-13,
       author = {Maglione, Marta and Sigrist, Stephan J.},
        title = {Seeing the forest tree by tree: super-resolution light microscopy meets the neurosciences},
      journal = {Nature Neuroscience},
         year = 2013,
    publisher = {Nature Publishing Group},
       volume = 16,
        issue = 7,
        pages = {790-797},
     abstract = {Light microscopy can be applied in vivo and can sample large tissue volumes, features crucial for the study of single neurons and neural circuits. However, light microscopy per se is diffraction-limited in resolution, and the substructure of core signaling compartments of neuronal circuits--axons, presynaptic active zones, postsynaptic densities and dendritic spines--can be only insufficiently characterized by standard light microscopy. Recently, several forms of super-resolution light microscopy breaking the diffraction-imposed resolution limit have started to allow highly resolved, dynamic imaging in the cell-biologically highly relevant 10-100 nanometer range ('mesoscale'). New, sometimes surprising answers concerning how protein mobility and protein architectures shape neuronal communication have already emerged. Here we start by briefly introducing super-resolution microscopy techniques, before we describe their use in the analysis of neuronal compartments. We conclude with long-term prospects for super-resolution light microscopy in the molecular and cellular neurosciences.}
}
@article{MaguireetalHIPPOCAMPUS-00,
       author = {Maguire, Eleanor A. and Mummery, Catherine J. and B\"{u}chel, Christian},
        title = {Patterns of hippocampal-cortical interaction dissociate temporal lobe memory subsystems},
      journal = {Hippocampus},
       volume = {10},
       number = {4},
         year = {2000},
        pages = {475-482},
     abstract = {A distributed network of brain regions supports memory retrieval in humans, but little is known about the functional interactions between areas within this system. During functional magnetic resonance imaging (fMRI), subjects retrieved real-world memories: autobiographical events, public events, autobiographical facts, and general knowledge. A common memory retrieval network was found to support all memory types. However, examination of the correlations (i.e., effective connectivity) between the activity of brain regions within the temporal lobe revealed significant changes dependent on the type of memory being retrieved. Medially, effective connectivity between the parahippocampal cortex and hippocampus increased for recollection of autobiographical events relative to other memory types. Laterally, effective connectivity between the middle temporal gyrus and temporal pole increased during retrieval of general knowledge and public events. The memory types that dissociate the common system into its subsystems correspond to those that typically distinguish between patients at initial phases of Alzheimer's disease or semantic dementia. This approach, therefore, opens the door to new lines of research into memory degeneration, capitalizing on the functional integration of different memory-involved regions. Indeed, the ability to examine interregional interactions may have important diagnostic and prognostic implications.},
}
@article{MaguireetalJEPLMC-12,
       author = {Maguire, Mandy J. and McClelland, M. Michelle and Donovan, Cohn M. and Tihiman, Gail D. and Krawczyk, Daniel C.},
      journal = {Journal of Experimental Psychology. Learning, Memory\& Cognition},
       number = {2},
        pages = {273-281},
        title = {Tracking Cognitive Phases in Analogical Reasoning With Event-Related Potentials.},
       volume = {38},
         year = {2012},
     abstract = {Analogical reasoning consists of multiple phases. Four-term analogies (A:B::C:D) have an encoding period in which the A:B pair is evaluated prior to a mapping phase. The electrophysiological timing associated with analogical reasoning has remained unclear. We used event-related potentials to identify neural timing related to analogical reasoning relative to perceptual and semantic control conditions. Spatiotemporal principal-components analyses revealed differences primarily in left frontal electrodes during encoding and mapping phases of analogies relative to the other conditions. The timing of the activity differed depending upon the phase of the problem. During the encoding of A:B terms, analogies elicited a positive deflection compared to the control conditions between 400 and 1,200 ms, but for the mapping phase analogical processing elicited a negative deflection that occurred earlier and for a shorter time period, between 350 and 625 ms. These results provide neural and behavior},
}
@article{MaguireetalJoEP-12,
       author = {Maguire, M. J. and McClelland, M. M. and Donovan, C. M. and Tillman, G. D. and Krawczyk, D. C.},
        title = {Tracking cognitive phases in analogical reasoning with event-related potentials},
      journal = {Journal of Experimental Psychology: Learning, Memory and Cognition},
       volume = {38},
       number = {2},
         year = {2012},
        pages = {273-281},
     abstract = {Analogical reasoning consists of multiple phases. Four-term analogies (A:B::C:D) have an encoding period in which the A:B pair is evaluated prior to a mapping phase. The electrophysiological timing associated with analogical reasoning has remained unclear. We used event-related potentials to identify neural timing related to analogical reasoning relative to perceptual and semantic control conditions. Spatiotemporal principal-components analyses revealed differences primarily in left frontal electrodes during encoding and mapping phases of analogies relative to the other conditions. The timing of the activity differed depending upon the phase of the problem. During the encoding of A:B terms, analogies elicited a positive deflection compared to the control conditions between 400 and 1,200 ms, but for the mapping phase analogical processing elicited a negative deflection that occurred earlier and for a shorter time period, between 350 and 625 ms. These results provide neural and behavioral evidence that 4-term analogy problems involve a highly active evaluation phase of the A:B pair.}
}
@article{MaguireetalNEUROPSYCHOLOGIA-10,
       author = {Eleanor A. Maguire and Faraneh Vargha-Khadem and Demis Hassabis},
        title = {Imagining fictitious and future experiences: Evidence from developmental amnesia},
      journal = {Neuropsychologia},
       volume = {48},
       number = {11},
         year = {2010},
        pages = {3187-3192},
     abstract = {Patients with bilateral hippocampal damage acquired in adulthood who are amnesic for past events have also been reported to be impaired at imagining fictitious and future experiences. One such patient, P01, however, was found to be unimpaired on these tasks despite dense amnesia and 50\% volume loss in both hippocampi. P01 might be an atypical case, and in order to investigate this we identified another patient with a similar neuropsychological profile. Jon is a well-characterised patient with developmental amnesia and 50\% volume loss in his hippocampi. Interestingly both Jon and P01 retain some recognition memory ability, and show activation of residual hippocampal tissue during fMRI. Jon's ability to construct fictitious and future scenarios was compared with the adult-acquired cases previously reported on this task and control participants. In contrast to the adult-acquired cases, but similar to P01, Jon was able to richly imagine both fictitious and future experiences in a comparable manner to control participants. Moreover, his constructions were spatially coherent. We speculate that the hippocampal activation during fMRI noted previously in P01 and Jon might indicate some residual hippocampal function which is sufficient to support their preserved ability to imagine fictitious and future scenarios.}
}
@incollection{Mahadevan02,
    author = {Sridhar Mahadevan},
     title = {Spatiotemporal Abstraction of Stochastic Sequential Processes},
 booktitle = {Lectures Notes in Artificial Intelligence},
    volume = 2371,
    editor = {Sven Koenig and Robert Holte},
 publisher = {Springer-Verlag},
     pages = {33-50},
      year = 2002,
}
@inproceedings{MahadevanandConnellAAAI-91,
       author = {Mahadevan, Sridhar and Connell, Jonathan},
        title = {Automatic Programming of Behavior-Based Robots Using Reinforcement Learning},
    booktitle = {Proceedings AAAI-91},
     location = {Anaheim, California},
 organization = {AAAI},
         year = 1991,
}
@inproceedings{MahmudandRayNIPS-07,
       author = {Mahmud, M. M. Hassan and Ray, Sylvian R.},
        title = {Transfer Learning Using Kolmogorov Complexity: Basic Theory and Empirical Evaluations},
    booktitle = {Proceedings of the 20th International Conference on Neural Information Processing Systems},
     location = {Vancouver, British Columbia, Canada},
    publisher = {Curran Associates Inc.},
         year = {2007},
        pages = {985-992},
     abstract = {In transfer learning we aim to solve new problems using fewer examples using information gained from solving related problems. Transfer learning has been successful in practice, and extensive PAC analysis of these methods has been developed. However it is not yet clear how to define relatedness between tasks. This is considered as a major problem as it is conceptually troubling and it makes it unclear how much information to transfer and when and how to transfer it. In this paper we propose to measure the amount of information one task contains about another using conditional Kolmogorov complexity between the tasks. We show how existing theory neatly solves the problem of measuring relatedness and transferring the 'right' amount of information in sequential transfer learning in a Bayesian setting. The theory also suggests that, in a very formal and precise sense, no other reasonable transfer method can do much better than our Kolmogorov Complexity theoretic transfer method, and that sequential transfer is always justified. We also develop a practical approximation to the method and use it to transfer information between 8 arbitrarily chosen databases from the UCI ML repository.},
}
@article{MahonandCaramazzaARP-09,
       author = {Mahon, Bradford Z. and Caramazza, Alfonso},
        title = {Concepts and categories: {A} cognitive neuropsychological perspective},
      journal = {Annual review of psychology},
         year = {2009},
       volume = {60},
        pages = {27-51},
     abstract = {One of the most provocative and exciting issues in cognitive science is how neural specificity for semantic categories of common objects arises in the functional architecture of the brain. More than two decades of research on the neuropsychological phenomenon of category-specific semantic deficits has generated detailed claims about the organization and representation of conceptual knowledge. More recently, researchers have sought to test hypotheses developed on the basis of neuropsychological evidence with functional imaging. From those two fields, the empirical generalization emerges that object domain and sensory modality jointly constrain the organization of knowledge in the brain. At the same time, research within the embodied cognition framework has highlighted the need to articulate how information is communicated between the sensory and motor systems, and processes that represent and generalize abstract information. Those developments point toward a new approach for understanding category specificity in terms of the coordinated influences of diverse regions and cognitive systems.},
}
@article {Maier-HeinetalBIORXIV-16,
        title = {Tractography-based connectomes are dominated by false-positive connections},
       author = {Maier-Hein, Klaus and Neher, Peter and Houde, Jean-Christophe and Cote, Marc-Alexandre and Garyfallidis, Eleftherios and Zhong, Jidan and Chamberland, Maxime and Yeh, Fang-Cheng and Lin, Ying Chia and Ji, Qing and Reddick, Wilburn E. and Glass, John O. and Chen, David Qixiang and Feng, Yuanjing and Gao, Chengfeng and Wu, Ye and Ma, Jieyan and Renjie, He and Li, Qiang and Westin, Carl-Fredrik and Deslauriers-Gauthier, Samuel and Gonzalez, J. Omar Ocegueda and Paquette, Michael and St-Jean, Samuel and Girard, Gabriel and Rheault, Francois and Sidhu, Jasmeen and Tax, Chantal M. W. and Guo, Fenghua and Mesri, Hamed Y. and David, Szabolcs and Froeling, Martijn and Heemskerk, Anneriet M. and Leemans, Alexander and Bore, Arnaud and Pinsard, Basile and Bedetti, Christophe and Desrosiers, Matthieu and Brambati, Simona and Doyon, Julien and Sarica, Alessia and Vasta, Roberta and Cerasa, Antonio and Quattrone, Aldo and Yeatman, Jason and Khan, Ali R. and Hodges, Wesley and Alexander, Simon and Romascano, Darcvi and Barakovic, Muhamed and Auria, Anna and Esteban, Oscar and Lemkaddem, Alia and Thiran, Jean-Philippe and Cetingul, H. Ertan and Odry, Benjamin L. and Mailhe, Boris and Nadar, Mariappan and Pizzagalli, Fabrizio and Prasad, Gautam and Villalon-Reina, Julio and Galvis, Justin and Thompson, Paul and Requejo, Francisco and Laguna, Pedro and Lacerda, Luis and Barrett, Rachel and Dell{\textquoteright}Acqua, Flavio and Catani, Marco and Petit, Laurent and Caruyer, Emmanuel and Daducci, Alessandro and Dyrby, Tim and Holland-Letz, Tim and Hilgetag, Claus and Stieltjes, Bram and Descoteaux, Maxime},
      journal = {bioRxiv},
         year = {2016},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {Fiber tractography based on non-invasive diffusion imaging is at the heart of connectivity studies of the human brain. To date, the approach has not been systematically validated in ground truth studies. Based on a simulated human brain dataset with ground truth white matter tracts, we organized an open international tractography challenge, which resulted in 96 distinct submissions from 20 research groups. While most state-of-the-art algorithms reconstructed 90\% of ground truth bundles to at least some extent, on average they produced four times more invalid than valid bundles. About half of the invalid bundles occurred systematically in the majority of submissions. Our results demonstrate fundamental ambiguities inherent to tract reconstruction methods based on diffusion orientation information, with critical consequences for the approach of diffusion tractography in particular and human connectivity studies in general.},
}
@book{MaierandWarren1988,
       author = {Maier, David and Warren, David},
        title = {Computing with Logic: Logic Programming with Prolog},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1988,
}
@article{MainenetalNATURE-16,
       author = {Zachary F. Mainen and Michael H\"{a}̈usser and Alexandre Pouget},
        title = {A better way to crack the brain},
      journal = {Nature},
         year = {2016},
    publisher = {Nature Publishing Group},
       volume = {539},
        issue = {7628},
        pages = {159-161},
     abstract = {At least half a dozen major initiatives to study the mammalian brain have sprung up across the world in the past five years. This wave of national and international projects has arisen in part from the realization that deciphering the principles of brain function will require collaboration on a grand scale. Yet it is unclear whether any of these mega-projects, which include scientists from many subdisciplines, will be effective. Researchers with complementary skill sets often team up on grant proposals. But once funds are awarded, the labs involved often return to work on their parts of the project in relative isolation. We propose an alternative strategy: grass-roots collaborations involving researchers who may be distributed around the globe, but who are already working on the same problems. Such self-motivated groups could start small and expand gradually over time. But they would essentially be built from the ground up, with those involved encouraged to follow their own shared interests rather than responding to the strictures of funding sources or external directives.}
}
@article{MairaletalPAMI-12,
        title = {Task-Driven Dictionary Learning.},
       author = {Mairal, Julien and Bach, Francis and Ponce, Jean},
      journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
        issue = 4,
       volume = 34,
         year = 2012,
        pages = {791-804},
     abstract = {Modeling data with linear combinations of a few elements from a learned dictionary has been the focus of much recent research in machine learning, neuroscience, and signal processing. For signals such as natural images that admit such sparse representations, it is now well established that these models are well suited to restoration tasks. In this context, learning the dictionary amounts to solving a large-scale matrix factorization problem, which can be done efficiently with classical optimization tools. The same approach has also been used for learning features from data for other purposes, e.g., image classification, but tuning the dictionary in a supervised way for these tasks has proven to be more difficult. In this paper, we present a general formulation for supervised dictionary learning adapted to a wide variety of tasks, and present an efficient algorithm for solving the corresponding optimization problem. Experiments on handwritten digit classification, digital art identification, nonlinear inverse image problems, and compressed sensing demonstrate that our approach is effective in large-scale settings, and is well suited to supervised and semi-supervised classification, as well as regression tasks for data that admit sparse representations.},
}
@article{MaisaketalNATURE-13,
       author = {Maisak, M. S. and Haag, J. and Ammer, G. and Serbe, E. and Meier, M. and Leonhardt, A. and Schilling, T. and Bahl, A. and Rubin, G. M. and Nern, A. and Dickson, B. J. and Reiff, D. F. and Hopp, E. and Borst, A.},
        title = {{A} directional tuning map of {D}rosophila elementary motion detectors},
      journal = {Nature},
         year = 2013,
       volume = 500,
        issue = 7461,
        pages = {212-216},
     abstract = {The extraction of directional motion information from changing retinal images is one of the earliest and most important processing steps in any visual system. In the fly optic lobe, two parallel processing streams have been anatomically described, leading from two first-order interneurons, L1 and L2, via T4 and T5 cells onto large, wide-field motion-sensitive interneurons of the lobula plate. Therefore, T4 and T5 cells are thought to have a pivotal role in motion processing; however, owing to their small size, it is difficult to obtain electrical recordings of T4 and T5 cells, leaving their visual response properties largely unknown. We circumvent this problem by means of optical recording from these cells in Drosophila, using the genetically encoded calcium indicator GCaMP5 (ref. 2). Here we find that specific subpopulations of T4 and T5 cells are directionally tuned to one of the four cardinal directions; that is, front-to-back, back-to-front, upwards and downwards. Depending on their preferred direction, T4 and T5 cells terminate in specific sublayers of the lobula plate. T4 and T5 functionally segregate with respect to contrast polarity: whereas T4 cells selectively respond to moving brightness increments (ON edges), T5 cells only respond to moving brightness decrements (OFF edges). When the output from T4 or T5 cells is blocked, the responses of postsynaptic lobula plate neurons to moving ON (T4 block) or OFF edges (T5 block) are selectively compromised. The same effects are seen in turning responses of tethered walking flies. Thus, starting with L1 and L2, the visual input is split into separate ON and OFF pathways, and motion along all four cardinal directions is computed separately within each pathway. The output of these eight different motion detectors is then sorted such that ON (T4) and OFF (T5) motion detectors with the same directional tuning converge in the same layer of the lobula plate, jointly providing the input to downstream circuits and motion-driven behaviours.},
}
@article {MajajetalJoN-15,
       author = {Majaj, Najib J. and Hong, Ha and Solomon, Ethan A. and James J. {DiCarlo}},
        title = {Simple Learned Weighted Sums of Inferior Temporal Neuronal Firing Rates Accurately Predict Human Core Object Recognition Performance},
      journal = {The Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {35},
         year = {2015},
        pages = {13402-13418},
     abstract = {To go beyond qualitative models of the biological substrate of object recognition, we ask: can a single ventral stream neuronal linking hypothesis quantitatively account for core object recognition performance over a broad range of tasks? We measured human performance in 64 object recognition tests using thousands of challenging images that explore shape similarity and identity preserving object variation. We then used multielectrode arrays to measure neuronal population responses to those same images in visual areas V4 and inferior temporal (IT) cortex of monkeys and simulated V1 population responses. We tested leading candidate linking hypotheses and control hypotheses, each postulating how ventral stream neuronal responses underlie object recognition behavior. Specifically, for each hypothesis, we computed the predicted performance on the 64 tests and compared it with the measured pattern of human performance. All tested hypotheses based on low- and mid-level visually evoked activity (pixels, V1, and V4) were very poor predictors of the human behavioral pattern. However, simple learned weighted sums of distributed average IT firing rates exactly predicted the behavioral pattern. More elaborate linking hypotheses relying on IT trial-by-trial correlational structure, finer IT temporal codes, or ones that strictly respect the known spatial substructures of IT ("face patches") did not improve predictive power. Although these results do not reject those more elaborate hypotheses, they suggest a simple, sufficient quantitative model: each object recognition task is learned from the spatially distributed mean firing rates (100 ms) of \~{}60,000 IT neurons and is executed as a simple weighted sum of those firing rates.},
}
@inproceedings{MajietalCVPR-08,
       author = {Subhransu Maji and Alexander C. Berg and Jitendra Malik},
        title = {Classification using Intersection Kernel Support Vector Machines is Efficient},
    booktitle = {IEEE Computer Society Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
      address = {Los Alamitos, CA, USA},
         year = 2008,
}
@inproceedings{MajietalCVPR-11,
       author = {Subhransu Maji and Lubomir Bourdev and Jitendra Malik},
        title = {Action Recognition from a Distributed Representation of Pose and Appearance},
    booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
         year = 2011,
}
@article{MajoretalANN-13,
       author = {Major, G. and Larkum, M. E. and Schiller, J.},
        title = {{A}ctive properties of neocortical pyramidal neuron dendrites},
      journal = {Annual Review Neuroscience},
         year = 2013,
       volume = 36,
        pages = {1-24},
     abstract = {Dendrites are the main recipients of synaptic inputs and are important sites that determine neurons' input-output functions. This review focuses on thin neocortical dendrites, which receive the vast majority of synaptic inputs in cortex but also have specialized electrogenic properties. We present a simplified working-model biophysical scheme of pyramidal neurons that attempts to capture the essence of their dendritic function, including the ability to behave under plausible conditions as dynamic computational subunits. We emphasize the electrogenic capabilities of NMDA receptors (NMDARs) because these transmitter-gated channels seem to provide the major nonlinear depolarizing drive in thin dendrites, even allowing full-blown NMDA spikes. We show how apparent discrepancies in experimental findings can be reconciled and discuss the current status of dendritic spikes in vivo; a dominant NMDAR contribution would indicate that the input-output relations of thin dendrites are dynamically set by network activity and cannot be fully predicted by purely reductionist approaches.},
}
@article{MakhzaniandFreyCoRR-15,
        title = {Winner-Take-All Autoencoders},
       author = {Alireza Makhzani and Brendan Frey},
      journal = {CoRR},
       volume = {arXiv:1409.2752},
         year = 2015,
     abstract = {In this paper, we propose a winner-take-all method for learning hierarchical sparse representations in an unsupervised fashion. We first introduce fully-connected winner-take-all autoencoders which use mini-batch statistics to directly enforce a lifetime sparsity in the activations of the hidden units. We then propose the convolutional winner-take-all autoencoder which combines the benefits of convolutional architectures and autoencoders for learning shift-invariant sparse representations. We describe a way to train convolutional autoencoders layer by layer, where in addition to lifetime sparsity, a spatial sparsity within each feature map is achieved using winner-take-all activation functions. We will show that winner-take-all autoencoders can be used to to learn deep sparse representations from the MNIST, CIFAR-10, ImageNet, Street View House Numbers and Toronto Face datasets, and achieve competitive classification performance.}
}
@inproceedings{MalikIJCAI-83,
       author = {Malik, Jitendra and Binford, Thomas O.},
        title = {Reasoning in Time and Space},
    booktitle = {Proceedings {IJCAI} 8},
     location = {Karlsruhe, West Germany},
 organization = {IJCAII},
         year = 1983,
        pages = {343-345},
}
@article{MalikandPeronaJOSA-90,
       author = {Jitendra Malik and Pietro Perona},
        title = {Preattentive texture discrimination with early vision mechanisms},
      journal = {Journal of the Optical Society of America A},
         year = 1990,
       volume = 7,
        pages = {923-932}
}
@inproceedings{MaliketalICCV-99,
     author = {Jitendra Malik and Serge Belongie and Jianbo Shi and Thomas K. Leung},
      title = {Textons, Contours and Regions: Cue Integration in Image Segmentation},
  booktitle = {Proceedings of the 7th {IEEE} International Conference on Computer Vision},
      pages = {918-925},
       year = 1999,
}
@article{MalinenandJohnstonEAR-13,
       author = {Malinen, S. and Johnston, L.},
        title = {Workplace ageism: discovering hidden bias},
      journal = {Experimental Aging Research},
         year = {2013},
       volume = {39},
       number = {4},
        pages = {445-465},
     abstract = {Research largely shows no performance differences between older and younger employees, or that older workers even outperform younger employees, yet negative attitudes towards older workers can underpin discrimination. Unfortunately, traditional "explicit" techniques for assessing attitudes (i.e., self-report measures) have serious drawbacks. Therefore, using an approach that is novel to organizational contexts, the authors supplemented explicit with implicit (indirect) measures of attitudes towards older workers, and examined the malleability of both. This research consists of two studies. The authors measured self-report (explicit) attitudes towards older and younger workers with a survey, and implicit attitudes with a reaction-time-based measure of implicit associations. In addition, to test whether attitudes were malleable, the authors measured attitudes before and after a mental imagery intervention, where the authors asked participants in the experimental group to imagine respected and valued older workers from their surroundings. Negative, stable implicit attitudes towards older workers emerged in two studies. Conversely, explicit attitudes showed no age bias and were more susceptible to change intervention, such that attitudes became more positive towards older workers following the experimental manipulation. This research demonstrates the unconscious nature of bias against older workers, and highlights the utility of implicit attitude measures in the context of the workplace. In the current era of aging workforce and skill shortages, implicit measures may be necessary to illuminate hidden workplace ageism.}
}
@inproceedings{MalisiewiczetalICCV-11,
        title = {Ensemble of Exemplar-{SVM}s for Object Detection and Beyond},
       author = {Tomasz Malisiewicz and Abhinav Gupta and Alexei A. Efros},
    booktitle = {{IEEE} International Conference on Computer Vision},
    publisher = {IEEE},
         year = 2011
}
@book{Malkiel1996,
       author = {Malkiel, Burton G.},
        title = {A Random Walk Down Wall Street},
    publisher = {Norton},
      address = {New York},
         year = 1996,
}
@article{MallamaciandStoykovaEJN-06,
        title = {Gene networks controlling early cerebral cortex arealization},
       author = {Mallamaci, Antonello and Stoykova, Anastassia},
       volume = 23,
        issue = 4,
      journal = {European Journal of Neuroscience},
    publisher = {Oxford University Press},
      address = {Oxford, UK},
         year = 2006,
        pages = {847-856}
}
@article{MallatandZhangSP-93,
       author = {S. Mallat and Z. Zhang},
        title = {Matching pursuits with time-frequency dictionaries},
      journal = {IEEE Transactions on Signal Processing},
       volume = 41,
        pages = {3397-3415},
         year = 1993,
}
@incollection{MalleLANGUAGE-THEORY-OF-MIND-02,
       author = {Malle, B. F.},
        title = {The relation between language and theory of mind in development and evolution},
       editor = {T. Giv\'{o}n and B. F. Malle},
    booktitle = {The evolution of language out of pre-language},
    publisher = {Benjamins},
      address = {Amsterdam},
         year = {2002},
        pages = {265-284},
     abstract = {There is reason to believe that language and theory of mind have co- evolved, given their close relation in development and their tight connection in social behavior. However, they are clearly not inseparable--neurologically, cognitively, or functionally. So the question becomes, "What is the exact relation between language and theory of mind, in evolution, development, and social behavior?" To answer this question is a daunting task; I will try merely to clear a path toward an answer. I will consider several possible relations between the two faculties, bring conceptual arguments and empirical evidence to bear on them, and end up arguing for an escalation process in which language and theory of mind have fueled each other's evolution.}
}
@inproceedings{MalonandBaiACL-13,
    author = {Christopher Malon and Bing Bai},
     title = {Answer Extraction by Recursive Parse Tree Descent},
 booktitle = {Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics, 
              Proceedings of the Workshop on Continuous Vector Space Models and their Compositionality},
 publisher = {Association for Computational Linguistics},
   address = {Stroudsburg, PA, USA},
      year = 2013,
}
@book{Malone2014,
        title = {The Intel Trinity: How Robert Noyce, Gordon Moore, and Andy Grove Built the World's Most Important Company},
       author = {Michael S. Malone},
         year = 2014,
    publisher = {HarperCollins Publishers},
      address = {New York},
}
@article{MalsburgCOIN-95,
        title = {Binding in models of perception and brain function},
       author = {Christoph von der Malsburg},
      journal = {Current Opinion in Neurobiology},
       volume = 5,
        issue = 4,
         year = 1995,
        pages = {520-526},
     abstract = {The development of the concept of feature binding as fundamental to neural dynamics has made possible recent advances in the modeling of difficult problems of perception and brain function. Major weaknesses of past neural modeling (most prominently its inability to work with natural stimuli and its "learning-time" barrier) have been traced back to improper treatment of the binding issue. Signal synchrony is now seen as playing a major role in binding. Inclusion of temporal binding in neural models has led to recent breakthroughs in solving important perceptual problems. Among them is perceptual segmentation, invariant object recognition and natural language parsing, as well as overcoming the "learning-time" barrier.}
}
@article{MancusoetalEXPER_PHYS-10,
        title = {Optogenetic probing of functional brain circuitry},
       author = {Mancuso, James J. and Kim, Jinsook and Lee, Soojung and Tsuda, Sachiko and Chow, Nicholas B. H. and Augustine, George J.},
      journal = {Experimental Physiology},
       volume = 96,
        issue = 1,
        pages = {26-33},
         year = 2011,
     abstract = {Recently developed optogenetic technologies offer the promise of high-speed mapping of brain circuitry. Genetically targeted light-gated channels and pumps, such as channelrhodopsins and halorhodopsin, allow optical control of neuronal activity with high spatial and temporal resolution. Optogenetic probes of neuronal activity, such as Clomeleon and Mermaid, allow light to be used to monitor the activity of a genetically defined population of neurons. Combining these two complementary sets of optogenetic probes will make it possible to perform all-optical circuit mapping. Owing to the improved efficiency and higher speed of data acquisition, this hybrid approach should enable high-throughput mapping of brain circuitry.},
}
@article{Mandl74,
       author = {Mandl, P.},
        title = {Estimation and control in {M}arkov chains},
      journal = {Advances in Applied Probability},
       volume = 6,
        pages = {40-60},
         year = 1974,
}
@article{ManessietalCoRR-17,
       author = {Franco Manessi and Alessandro Rozza and Mario Manzo},
        title = {Dynamic Graph Convolutional Networks},
      journal = {CoRR},
       volume = {arXiv:1704.06199},
         year = {2017},
         abstract = {Many different classification tasks need to manage structured data, which are usually modeled as graphs. Moreover, these graphs can be dynamic, meaning that the vertices/edges of each graph may change during time. Our goal is to jointly exploit structured data and temporal information through the use of a neural network model. To the best of our knowledge, this task has not been addressed using these kind of architectures. For this reason, we propose two novel approaches, which combine Long Short-Term Memory networks and Graph Convolutional Networks to learn long short-term dependencies together with graph structure. The quality of our methods is confirmed by the promising results achieved.}
}
@techreport{ManfrediandMahadevanTR-05,
       author = {Victoria Manfredi and Sridhar Mahadevan},
        title = {Dynamic Abstraction Networks},
  institution = {University of Massachusetts, Amherst},
        issue = {TR-2005-33},
         year = 2005,
}
@book{MannaandWaldinger83,
       author = {Manna, Zohar and Waldinger, Richard},
        title = {The Logical Basis for Computer Programming, Volume 1: Deductive
                Reasoning},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1983,
}
@incollection{MannaandWaldingerPW-86,
       author = {Manna, Z. and Waldinger, R. J.},
        title = {A Theory of Plans},
     crossref = {GeorgeffandLanskyPW-86},
}
@book{ManningandSchutze99,
       author = {Chris Manning and Hinrich Sch\"{u}tze},
        title = {Foundations of Statistical Natural Language Processing},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1999
}
@book{Manningetal2008,
       author = {Christopher D. Manning and Prabhakar Raghavan and Hinrich Sch\"{u}tze},
        title = {Introduction to Information Retrieval},
    publisher = {Cambridge University Press},
         year = 2008,
}
@article{ManningetalPNAS-20,
       author = {Christopher D. Manning and Kevin Clark and John Hewitt and Urvashi Khandelwal and Omer Levy},
        title = {Emergent linguistic structure in artificial neural networks trained by self-supervision},
      journal = {Proceedings of the National Academy of Sciences}, 
    publisher = {National Academy of Sciences},			   
       volume = {},						   
       number = {},						   
        pages = {},					   
         year = {2020},						   
     abstract = {This paper explores the knowledge of linguistic structure learned by large artificial neural networks, trained via self-supervision, whereby the model simply tries to predict a masked word in a given context. Human language communication is via sequences of words, but language understanding requires constructing rich hierarchical structures that are never observed explicitly. The mechanisms for this have been a prime mystery of human language acquisition, while engineering work has mainly proceeded by supervised learning on treebanks of sentences hand labeled for this latent structure. However, we demonstrate that modern deep contextual language models learn major aspects of this structure, without any explicit supervision. We develop methods for identifying linguistic hierarchical structure emergent in artificial neural networks and demonstrate that components in these models focus on syntactic grammatical relationships and anaphoric coreference. Indeed, we show that a linear transformation of learned embeddings in these models captures parse tree distances to a surprising degree, allowing approximate reconstruction of the sentence tree structures normally assumed by linguists. These results help explain why these models have brought such large improvements across many language-understanding tasks.}
}
@article{MannionandCraigheadBIOPOLYMERS-07,
       author = {Mannion, J. T. and Craighead, H. G.},
        title = {Nanofluidic structures for single biomolecule fluorescent detection},
      journal = {Biopolymers},
       volume = 85,
        issue = 2,
    publisher = {Wiley Subscription Services, Inc., A Wiley Company},
        pages = {131-143},
         year = 2007,
     abstract = {Fluid-filled nanofabricated cavities can be used to increase the spatial resolution of single molecule confocal microscopy based techniques by creating smaller and more uniformly illuminated probe volumes. Such structures may also be used to temporarily stretch single macromolecules, permitting the resolution of molecular details that would otherwise be beyond the capabilities of a diffraction limited system. © 2006 Wiley Periodicals, Inc. Biopolymers 85:131-143, 2007.This article was originally published online as an accepted preprint. The "Published Online" date corresponds to the preprint version. You can request a copy of the preprint by emailing the Biopolymers editorial office at biopolymers@wiley.com},
}
@article{ManoretalNEUROSCIENCE-99,
       author = {Manor, Yair and Nadim, Farzan and Epstein, Steven and Ritt, Jason and Marder, Eve and Kopell, Nancy},
        title = {Network Oscillations Generated by Balancing Graded Asymmetric Reciprocal Inhibition in Passive Neurons},
      journal = {The Journal of Neuroscience},
       volume = 19,
        issue = 7,
         year = 1999,
        pages = {2765-2779},
     abstract = {We describe a novel mechanism by which network oscillations can arise from reciprocal inhibitory connections between two entirely passive neurons. The model was inspired by the activation of the gastric mill rhythm in the crab stomatogastric ganglion by the modulatory commissural ganglion neuron 1 (MCN1), but it is studied here in general terms. One model neuron has a linear current-voltage (I-V) curve with a low (L) resting potential, and the second model neuron has a linear current-voltage curve with a high (H) resting potential. The inhibitory connections between them are graded. There is an extrinsic modulatory excitatory input to the L neuron, and the L neuron presynaptically inhibits the modulatory neuron. Activation of the extrinsic modulatory neuron elicits stable network oscillations in which the L and H neurons are active in alternation. The oscillations arise because the graded reciprocal synapses create the equivalent of a negative-slope conductance region in theI-V curves for the cells. Geometrical methods are used to analyze the properties of and the mechanism underlying these network oscillations.},
}
@article{ManteetalNATURE-13,
       author = {Mante, Valerio and Sussillo, David and Shenoy, Krishna V. and Newsome, William T.},
        title = {Context-dependent computation by recurrent dynamics in prefrontal cortex},
      journal = {Nature},
       volume = 503,
        issue = 7474,
         year = 2013,
        pages = {78-84},
     abstract = {Prefrontal cortex is thought to have a fundamental role in flexible, context-dependent behaviour, but the exact nature of the computations underlying this role remains largely unknown. In particular, individual prefrontal neurons often generate remarkably complex responses that defy deep understanding of their contribution to behaviour. Here we study prefrontal cortex activity in macaque monkeys trained to flexibly select and integrate noisy sensory inputs towards a choice. We find that the observed complexity and functional roles of single neurons are readily understood in the framework of a dynamical process unfolding at the level of the population. The population dynamics can be reproduced by a trained recurrent neural network, which suggests a previously unknown mechanism for selection and integration of task-relevant inputs. This mechanism indicates that selection and integration are two aspects of a single dynamical process unfolding within the same prefrontal circuits, and potentially provides a novel, general framework for understanding context-dependent computations.}
}
@article{ManteetalNATURE-13.supplement,
       author = {Mante, Valerio and Sussillo, David and Shenoy, Krishna V. and Newsome, William T.},
        title = {Context-dependent computation by recurrent dynamics in prefrontal cortex: Supplementary Materials},
      journal = {Nature},
       volume = 503,
        issue = 7474,
         year = 2013,
        pages = {78-84},
}
@article{MantonetalBIORXIV-14,
       author = {Manton, James D. and Ostrovsky, Aaron D. and Goetz, Lea and Costa, Marta and Rohlfing, Torsten and Jefferis, Gregory S. X. E.},
        title = {Combining genome-scale Drosophila 3D neuroanatomical data by bridging template brains},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Labs Journals},
         year = {2014},
     abstract = {The stereotyped structure of mammalian and invertebrate brains is a crucial determinant of their circuit organization. Thus large scale efforts to map circuit organization using 3D image data are underway in a number of model systems, including flies and mice. Many of these studies use registration of sample images to a standard template brain to enable co-visualization and spatial querying. However, studies often use distinct template brains, resulting in large islands of data which cannot be directly compared. To enable this comparison, we have constructed bridging registrations between template brains accounting for the vast majority of Drosophila melanogaster 3D neuroanatomical data. Furthermore, we solve the related problem of mapping data between the left and right brain hemispheres via the construction of mirroring registrations. Finally, we extend our approach across species to demonstrate its potential use in evolutionary studies of neural circuit structure and provide bridging registrations that link a new set of template brains generated for four Drosophila species that are divergent over 40 million years of evolution. We describe our strategy, document the freely available anatomical data and open source com- puter tools that we have generated and provide numerous examples of their use. This effort has unified data from over 30,000 publicly available images, with resources including the 3D atlas embodying the new standard Drosophila anatomical nomenclature and the largest single neuron databank yet available in any species. Over 20,000 registered images have been contributed to the Virtual Fly Brain project and can be viewed online at www.virtualflybrain.org.},
}
@inproceedings{MaoetalICLR-19,
        title = {The Neuro-Symbolic Concept Learner: Interpreting Scenes, Words, and Sentences From Natural Supervision},
       author = {Jiayuan Mao and Chuang Gan and Pushmeet Kohli and Joshua B. Tenenbaum and Jiajun Wu},
    booktitle = {International Conference on Learning Representations},
         year = {2019},
     abstract = {We propose the Neuro-Symbolic Concept Learner (NS-CL), a model that learns visual concepts, words, and semantic parsing of sentences without explicit supervision on any of them; instead, our model learns by simply looking at images and reading paired questions and answers. Our model builds an object-based scene representation and translates sentences into executable, symbolic programs. To bridge the learning of two modules, we use a neuro-symbolic reasoning module that executes these programs on the latent scene representation. Analogical to human concept learning, the perception module learns visual concepts based on the language description of the object being referred to. Meanwhile, the learned visual concepts facilitate learning new words and parsing new sentences. We use curriculum learning to guide the searching over the large compositional space of images and language. Extensive experiments demonstrate the accuracy and efficiency of our model on learning visual concepts, word representations, and semantic parsing of sentences. Further, our method allows easy generalization to new object attributes, compositions, language concepts, scenes and questions, and even new program domains. It also empowers applications including visual question answering and bidirectional image-text retrieval.},
}
@misc{MarblestonePhD-15,
        title = {Designing scalable biological interfaces},
       author = {Adam Marblestone},
         year = 2015,
 howpublished = {Ph.D. Thesis. Biophysics Department, Harvard University}
}
@article{MarblestoneandBoydenNEURON-14,
       author = {Marblestone, A. H. and Boyden, E. S.},
        title = {{D}esigning tools for assumption-proof brain mapping},
      journal = {Neuron},
         year = 2014,
       volume = 83,
        issue = 6,
        pages = {1239-1241},
}
@article{MarblestoneetalARXIV-13,
        title = {Physical Principles for Scalable Neural Recording},
       author = {Adam H. Marblestone and Bradley M. Zamft and Yael G. Maguire and Mikhail G. Shapiro and Thaddeus R. Cybulski and Joshua I. Glaser and Ben Stranges and Reza Kalhor and David A. Dalrymple and Dongjin Seo and Elad Alon and Michel M. Maharbiz and Jose Carmena and Jan Rabaey and Edward S. Boyden and George M. Church and Konrad P. Kording},
         year = 2013,
      journal = {CoRR},
       volume = {arXiv:1306.5709},
          url = {http://arxiv.org/abs/1306.5709},
}
@article {MarblestoneetalBIORXIV-14,
       author = {Marblestone, Adam H and Daugharthy, Evan R and Kalhor, Reza and Peikon, Ian D and Kebschull, Justus M and Shipman, Seth L and Mishchenko, Yuriy and Lee, Je Hyuk and Dalrymple, David A and Zamft, Bradley M and Kording, Konrad P and Boyden, Edward S and Zador, Anthony M and Church, George M},
        title = {Conneconomics: The Economics of Dense, Large-Scale, High-Resolution Neural Connectomics},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Labs Journals},
         year = 2014,
     abstract = {We analyze the scaling and cost-performance characteristics of current and projected connectomics approaches, with reference to the potential implications of recent advances in diverse contributing fields. Three generalized strategies for dense connectivity mapping at the scale of whole mammalian brains are considered: electron microscopic axon tracing, optical imaging of combinatorial molecular markers at synapses, and bulk DNA sequencing of trans-synaptically exchanged nucleic acid barcode pairs. Due to advances in parallel-beam instrumentation, whole mouse brain electron microscopic image acquisition could cost less than \$100 million, with total costs presently limited by image analysis to trace axons through large image stacks. Optical microscopy at 50 to 100 nm isotropic resolution could potentially read combinatorially multiplexed molecular information from individual synapses, which could indicate the identifies of the pre-synaptic and post-synaptic cells without relying on axon tracing. An optical approach to whole mouse brain connectomics may be achievable for less than \$10 million and could be enabled by emerging technologies to sequence nucleic acids in-situ in fixed tissue via fluorescent microscopy. Novel strategies relying on bulk DNA sequencing, which would extract the connectome without direct imaging of the tissue, could produce a whole mouse brain connectome for \$100k to \$1 million or a mouse cortical connectome for \$10k to \$100k. Anticipated further reductions in the cost of DNA sequencing could lead to a \$1000 mouse cortical connectome.},
}
@article{MarblestoneetalCoRR-16a,
        title = {Towards an integration of deep learning and neuroscience},
       author = {Adam H. Marblestone and Wayne, Greg and Kording, Konrad P.},
      journal = {CoRR},
       volume = {arXiv:1606.03813},
         year = 2016,
     abstract = {Neuroscience has focused on the detailed implementation of computation, studying neural codes, dynamics and circuits. In machine learning, however, artificial neural networks tend to eschew precisely designed codes, dynamics or circuits in favor of brute force optimization of a cost function, often using simple and relatively uniform initial architectures. Two recent developments have emerged within machine learning that create an opportunity to connect these seemingly divergent perspectives. First, structured architectures are used, including dedicated systems for attention, recursion and various forms of short- and long-term memory storage. Second, cost functions and training procedures have become more complex and are varied across layers and over time. Here we think about the brain in terms of these ideas. We hypothesize that (1) the brain optimizes cost functions, (2) these cost functions are diverse and differ across brain locations and over development, and (3) optimization operates within a pre-structured architecture matched to the computational problems posed by behavior. Such a heterogeneously optimized system, enabled by a series of interacting cost functions, serves to make learning data-efficient and precisely targeted to the needs of the organism. We suggest directions by which neuroscience could seek to refine and test these hypotheses.}
}
@article{MarblestoneetalCoRR-16b,
        title = {Do brain areas optimize diverse cost functions using specialized structures?},
       author = {Adam H. Marblestone and Wayne, Greg and Kording, Konrad P. and Hayworth, Kenneth J},
         year = 2016,
      comment = {/u/tld/Drive/write/lectures/stanford/stanford_course_archives/spring_2016/calendar_invited_talks/lectures/05/18/protected/MarblestoneetalCoRR-16.05-14-16.pdf},
      journal = {CoRR},
     abstract = {Neuroscience has tended to focus on the detailed implementation of computation, studying representations, circuits and local algorithms. In machine learning, artificial neural networks tend to eschew precisely designed codes or circuits in favor of brute force optimization of a cost function, often using simple and relatively uniform initial architectures. Two recent developments have emerged within machine learning. First, structured architectures are used, including dedicated systems for attention, recursion and various forms of short- and long-term memory storage. Second, cost functions and training procedures have become more complex: they are varied across layers and time, incorporate a wide range of regularization terms, and sometimes use one part of a more complex architecture to generate cost signals for another part. Here we think about the brain in terms of these developments. We hypothesize that (1) the brain optimizes cost functions, (2) these cost functions are diverse and differ across time and brain region, and (3) optimization operates within a pre-structured architecture matched to the computational problems encountered by the brain. We suggest directions by which neuroscience could seek to test and refine this hypothesis.},
}
@article{MarblestoneetalFiCN-16,
       author = {Marblestone, Adam H. and Wayne, Greg and Kording, Konrad P.},
        title = {Toward an Integration of Deep Learning and Neuroscience},
      journal = {Frontiers in Computational Neuroscience},
       volume = {10},
        pages = {94},
         year = {2016},
     abstract = {Neuroscience has focused on the detailed implementation of computation, studying neural codes, dynamics and circuits. In machine learning, however, artificial neural networks tend to eschew precisely designed codes, dynamics or circuits in favor of brute force optimization of a cost function, often using simple and relatively uniform initial architectures. Two recent developments have emerged within machine learning that create an opportunity to connect these seemingly divergent perspectives. First, structured architectures are used, including dedicated systems for attention, recursion and various forms of short- and long-term memory storage. Second, cost functions and training procedures have become more complex and are varied across layers and over time. Here we think about the brain in terms of these ideas. We hypothesize that (1) the brain optimizes cost functions, (2) the cost functions are diverse and differ across brain locations and over development, and (3) optimization operates within a pre-structured architecture matched to the computational problems posed by behavior. In support of these hypotheses, we argue that a range of implementations of credit assignment through multiple layers of neurons are compatible with our current knowledge of neural circuitry, and that the brain's specialized systems can be interpreted as enabling efficient optimization for specific problem classes. Such a heterogeneously optimized system, enabled by a series of interacting cost functions, serves to make learning data-efficient and precisely targeted to the needs of the organism. We suggest directions by which neuroscience could seek to refine and test these hypotheses.}
}
@article{MarblestoneetalROSETTA-14,
       author = {Adam H. Marblestone and Evan R. Daugharthy and Reza Kalhor and Ian D. Peikon and Justus M. Kebschull and Seth L. Shipman and Yuriy Mishchenko and Je Hyuk Lee and Konrad P. Kording and Edward S. Boyden and Anthony M. Zador and George M. Church},
        title = {Rosetta Brains: A Strategy for Molecularly-Annotated Connectomics},
      journal = {CoRR},
       volume = {arXiv:1404.5103},
         year = 2014,
     abstract = {We propose a neural connectomics strategy called Fluorescent In-Situ Sequencing of Barcoded Individual Neuronal Connections (FISSEQ-BOINC), leveraging fluorescent in situ nucleic acid sequencing in fixed tissue (FISSEQ). FISSEQ-BOINC exhibits different properties from BOINC, which relies on bulk nucleic acid sequencing. FISSEQ-BOINC could become a scalable approach for mapping whole-mammalian-brain connectomes with rich molecular annotations.}
}
@article{MarcinandKurachCoRR-16,
       author = {Marcin Andrychowicz and Karol Kurach},
        title = {Learning Efficient Algorithms with Hierarchical Attentive Memory},
      journal = {CoRR},
       volume = {arXiv:1602.03218},
         year = {2016},
     abstract = {In this paper, we propose and investigate a novel memory architecture for neural networks called Hierarchical Attentive Memory (HAM). It is based on a binary tree with leaves corresponding to memory cells. This allows HAM to perform memory access in O(log n) complexity, which is a significant improvement over the standard attention mechanism that requires O(n) operations, where n is the size of the memory.
  
We show that an LSTM network augmented with HAM can learn algorithms for problems like merging, sorting or binary searching from pure input-output examples. In particular, it learns to sort n numbers in time O(n log n) and generalizes well to input sequences much longer than the ones seen during the training. We also show that HAM can be trained to act like classic data structures: a stack, a FIFO queue and a priority queue.}
}
@book{Marcus2004,
        title = {Kluge: The Haphazard Evolution of the Human Mind},
       author = {Marcus, Gary, F.},
         year = 2004,
    publisher = {Houghton Mifflin},
}
@book{Marcus2004birth,
        title = {The Birth of the Mind: How a Tiny Number of Genes Creates The Complexities of Human Thought},
       author = {Marcus, Gary},
         year = {2004},
    publisher = {Basic Books}
}
@article{MarcusCP-98,
        title = {Rethinking Eliminative Connectionism},
       author = {Gary F. Marcus},
      journal = {Cognitive Psychology},
       volume = 37,
        issue = 3,
         year = 1998,
        pages = {243-282},
     abstract = {Humans routinely generalize universal relationships to unfamiliar instances. If we are told "if glork then frum," and "glork," we can infer "frum"; any name that serves as the subject of a sentence can appear as the object of a sentence. These universals are pervasive in language and reasoning. One account of how they are generalized holds that humans possess mechanisms that manipulate symbols and variables; an alternative account holds that symbol-manipulation can be eliminated from scientific theories in favor of descriptions couched in terms of networks of interconnected nodes. Can these "eliminative" connectionist models offer a genuine alternative? This article shows that eliminative connectionist models cannot account for how we extend universals to arbitrary items. The argument runs as follows. First, if these models, as currently conceived, were to extend universals to arbitrary instances, they would have to generalize outside the space of training examples. Next, it is shown that the class of eliminative connectionist models that is currently popular cannot learn to extend universals outside the training space. This limitation might be avoided through the use of an architecture that implements symbol manipulation.}
}
@article{MarcusCoRR-20,
        title = {The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence},
       author = {Gary F. Marcus},
      journal = {CoRR},
         year = {2020},
       volume = {arXiv:2002.06177},
     abstract = {Recent research in artificial intelligence and machine learning has largely emphasized general-purpose learning and ever-larger training sets and more and more compute. In contrast, I propose a hybrid, knowledge-driven, reasoning-based approach, centered around cognitive models, that could provide the substrate for a richer, more robust AI than is currently possible.}
}
@article{MarcusandMarblestoneandDeanSCIENCE-14,
       author = {Gary Marcus and Adam Marblestone and Thomas Dean},
        title = {The atoms of neural computation},
      journal = {Science},
       volume = {346},
        issue = {6289},
         year = {2014},
        pages = {551-552},
} 
@article{MarderNEURON-12,
       author = {Marder, Eve},
        title = {{N}euromodulation of neuronal circuits: back to the future},
      journal = {Neuron},
       volume = 76,
        issue = 1,
         year = 2012,
        pages = {1-11},
}
@article{MarderPLoS-15,
       author = {Marder, Eve},
        title = {Understanding Brains: Details, Intuition, and Big Data},
      journal = {{PLoS} Biology},
         year = 2015,
       volume = 13,
        issue = 5,
        pages = {e1002147},
     abstract = {Understanding how the brain works requires a delicate balance between the appreciation of the importance of a multitude of biological details and the ability to see beyond those details to general principles. As technological innovations vastly increase the amount of data we collect, the importance of intuition into how to analyze and treat these data may, paradoxically, become more important.}
} 
@article{MareketalNEURON-18,
       author = {Scott Marek and Joshua S. Siegel and Evan M. Gordon and Ryan V. Raut and Caterina Gratton and Dillan J. Newbold and Mario Ortega and Timothy O. Laumann and Babatunde Adeyemo and Derek B. Miller and Annie Zheng and Katherine C. Lopez and Jeffrey J. Berg and Rebecca S. Coalson and Annie L. Nguyen and Donna Dierker and Andrew N. Van and Catherine R. Hoyt and Kathleen B. McDermott and Scott A. Norris and Joshua S. Shimony and Abraham Z. Snyder and Steven M. Nelson and Deanna M. Barch and Bradley L. Schlaggar and Marcus E. Raichle and Steven E. Petersen and Deanna J. Greene and Nico U.F. Dosenbach},
        title = {Spatial and Temporal Organization of the Individual Human Cerebellum},
      journal = {Neuron},
         year = {2018},
     abstract = {The cerebellum contains the majority of neurons in the human brain and is unique for its uniform cytoarchitecture, absence of aerobic glycolysis, and role in adaptive plasticity. Despite anatomical and physiological differences between the cerebellum and cerebral cortex, group-average functional connectivity studies have identified networks related to specific functions in both structures. Recently, precision functional mapping of individuals revealed that functional networks in the cerebral cortex exhibit measurable individual specificity. Using the highly sampled Midnight Scan Club (MSC) dataset, we found the cerebellum contains reliable, individual-specific network organization that is significantly more variable than the cerebral cortex. The frontoparietal network, thought to support adaptive control, was the only network overrepresented in the cerebellum compared to the cerebral cortex (2.3-fold). Temporally, all cerebellar resting state signals lagged behind the cerebral cortex (125-380 ms), supporting the hypothesis that the cerebellum engages in a domain-general function in the adaptive control of all cortical processes.}
}
@inproceedings{MarialetalCVPR-08,
       author = {Julien Mairal and Francis Bach and Jean Ponce and Guillermo Sapiro and Andrew Zisserman},
        title = {Discriminative learned dictionaries for local image analysis},
    booktitle = {IEEE Computer Society Conference on Computer Vision and Pattern Recognition},
         year = 2008,
        pages = {1-8},
    publisher = {IEEE Computer Society},
      address = {Los Alamitos, CA, USA},
}
@article{MarienandBeatonCEREBELLUM-ATAXIAS-14,
       author = {Mari{\"{e}}n, Peter and Beaton, Alan},
        title = {The enigmatic linguistic cerebellum: clinical relevance and unanswered questions on nonmotor speech and language deficits in cerebellar disorders},
      journal = {Cerebellum \& ataxias},
         year = {2014},
    publisher = {BioMed Central},
       volume = {1},
        pages = {12-12},
     abstract = {Clinical case descriptions and experimental evidence dating back to the early part of the 19th century from time to time documented a range of nonmotor cognitive and affective impairments following cerebellar pathology. However, a causal relationship between disruption of nonmotor cognitive and affective skills and cerebellar disease was dismissed for several decades and the classical view of the cerebellum as a mere coordinator of autonomic and somatic sensorimotor function prevailed for more than two centuries in behavioural neuroscience. The ignorance of early clinical evidence suggesting a much richer and complex role for the cerebellum than a pure sensorimotor one is remarkable given that in addition: 1) the cerebellum contains more neurons than the rest of the combined cerebral cortex and 2) no other structure has as many connections with other parts of the brain as the cerebellum. During the past decades, the long-standing view of the cerebellum as pure coordinator of sensorimotor function has been substantially modified. From the late 1970s onwards, major advances were made in elucidating the many functional neuroanatomical connections of the cerebellum with the supratentorial association cortices that subserve nonmotor language, cognition and affect. Combined with evidence derived from experimental functional neuroimaging studies in healthy subjects and neurophysiological and neuropsychological research in patients, the role of the cerebellum has been substantially extended to include that of a crucial modulator of cognitive and affective processes. In addition to its long-established role in coordinating motor aspects of speech production, clinical and experimental studies with patients suffering from etiologically different cerebellar disorders have identified involvement of the cerebellum in a variety of nonmotor language functions, including motor speech planning, language dynamics and verbal fluency, phonological and semantic word retrieval, expressive and receptive syntax processing, various aspects of reading and writing and aphasia-like phenomena. Despite considerable efforts currently devoted to further refine typology and anatomoclinical configurations of nonmotor linguistic dysfunctions linked to cerebellar pathology, the exact underlying pathophysiological mechanisms of cerebellar involvement remain to be elucidated.},
}
@incollection{MarinoEAB-10,
       author = {Lori Marino},
        title = {Sentience},
       editor = {Breed, Michael D.  and Moore, Janice},
    booktitle = {Encyclopedia of Animal Behavior},
    publisher = {Academic Press},
         year = {2010},
        pages = {132-138},
     abstract = {Sentience refers to the depth of awareness an individual possesses about himself or herself and others. There appear to be three related, but separable, general domains of sentience. These are self-awareness, metacognition, and theory of mind. To date, evidence shows that these three capacities are found in nonhuman animals, including primates, dolphins, dogs, rodents, and corvids. These findings are evidence of the deep psychological continuity that exists across the animal kingdom.}
}
@article{MarkovSCIENCE-13,
       author = {Markov, Nikola T. and Ercsey-Ravasz, M\'{a}ria and Van Essen, David C. and Knoblauch, Kenneth and Toroczkai, Zolt\'{o}n and Kennedy, Henry},
        title = {Cortical High-Density Counterstream Architectures},
      journal = {Science},
       volume = 342,
        issue = 6158,
         year = 2013,
     abstract = {Small-world networks provide an appealing description of cortical architecture owing to their capacity for integration and segregation combined with an economy of connectivity. Previous reports of low-density interareal graphs and apparent small-world properties are challenged by data that reveal high-density cortical graphs in which economy of connections is achieved by weight heterogeneity and distance-weight correlations. These properties define a model that predicts many binary and weighted features of the cortical network including a core-periphery, a typical feature of self-organizing information processing systems. Feedback and feedforward pathways between areas exhibit a dual counterstream organization, and their integration into local circuits constrains cortical computation. Here, we propose a bow-tie representation of interareal architecture derived from the hierarchical laminar weights of pathways between the high-efficiency dense core and periphery.},
}
@article{MarkramNATURE-06,
     author = {Markram, Henry},
      title = {The blue brain project},
    journal = {Nature Reviews Neuroscience},
     volume = 7, 
      pages = {153-160},
       year = 2006,
}
@article{MarkramandSakmannPNAS-94,
       author = {Markram, Henry and Sakmann, Bert},
        title = {Calcium transients in dendrites of neocortical neurons evoked by single subthreshold excitatory postsynaptic potentials via low-voltage-activated calcium channels.},
      journal = {Proceedings of the National Academy of Science},
         year = {1994},
       volume = {91},
        issue = {11},
        pages = {5207-5211},
     abstract = {Simultaneous recordings of membrane voltage and concentration of intracellular Ca2+ ([Ca2+]i) were made in apical dendrites of layer 5 pyramidal cells of rat neocortex after filling dendrites with the fluorescent Ca2+ indicator Calcium Green-1. Subthreshold excitatory postsynaptic potentials (EP-SPs), mediated by the activation of glutamate receptor channels, caused a brief increase in dendritic [Ca2+]i. This rise in dendritic [Ca2+]i was mediated by the opening of low-voltage-activated Ca2+ channels in the dendritic membrane. The results provide direct evidence that dendrites do not function as passive cables even at low-frequency synaptic activity; rather, a single subthreshold EPSP changes the dendritic membrane conductance by opening Ca2+ channels and generating a [Ca2+]i transient that may propagate towards the soma. The activation of these Ca2+ channels at a low-voltage threshold is likely to influence the way in which dendritic EPSPs contribute to the electrical activity of the neuron.},
}
@article{MarkrametalCELL-15,
       author = {Markram, Henry
and Muller, Eilif
and Ramaswamy, Srikanth
and Reimann, Michael W.
and Abdellah, Marwan
and Sanchez, Carlos Aguado
and Ailamaki, Anastasia
and Alonso-Nanclares, Lidia
and Antille, Nicolas
and Arsever, Selim
and Kahou, Guy Antoine Atenekeng
and Berger, Thomas K.
and Bilgili, Ahmet
and Buncic, Nenad
and Chalimourda, Athanassia
and Chindemi, Giuseppe
and Courcol, Jean-Denis
and Delalondre, Fabien
and Delattre, Vincent
and Druckmann, Shaul
and Dumusc, Raphael
and Dynes, James
and Eilemann, Stefan
and Gal, Eyal
and Gevaert, Michael Emiel
and Ghobril, Jean-Pierre
and Gidon, Albert
and Graham, Joe W.
and Gupta, Anirudh
and Haenel, Valentin
and Hay, Etay
and Heinis, Thomas
and Hernando, Juan B.
and Hines, Michael
and Kanari, Lida
and Keller, Daniel
and Kenyon, John
and Khazen, Georges
and Kim, Yihwa
and King, James G.
and Kisvarday, Zoltan
and Kumbhar, Pramod
and Lasserre, Sebastien
and Le B, Jean-Vincent
and Magalhes, Bruno R. C.
and Merchn-Prez, Angel
and Meystre, Julie
and Morrice, Benjamin Roy
and Muller, Jeffrey
and Muoz-Cspedes, Alberto
and Muralidhar, Shruti
and Muthurasa, Keerthan
and Nachbaur, Daniel
and Newton, Taylor H.
and Nolte, Max
and Ovcharenko, Aleksandr
and Palacios, Juan
and Pastor, Luis
and Perin, Rodrigo
and Ranjan, Rajnish
and Riachi, Imad
and Rodrguez, Jos-Rodrigo
and Riquelme, Juan Luis
and Rssert, Christian
and Sfyrakis, Konstantinos
and Shi, Ying
and Shillcock, Julian C.
and Silberberg, Gilad
and Silva, Ricardo
and Tauheed, Farhan
and Telefont, Martin
and Toledo-Rodriguez, Maria
and Tr\"{a}nkler, Thomas
and Van Geit, Werner
and Daz, Jafet Villafranca
and Walker, Richard
and Wang, Yun
and Zaninetta, Stefano M.
and DeFelipe, Javier
and Hill, Sean L.
and Segev, Idan
and Sch\"{u}rmann, Felix},
        title = {Reconstruction and Simulation of Neocortical Microcircuitry},
      journal = {Cell},
         year = {2015},
    publisher = {Elsevier},
       volume = {163},
        issue = {2},
        pages = {456-492},
     abstract = {We present a first-draft digital reconstruction of the microcircuitry of somatosensory cortex of juvenile rat. The reconstruction uses cellular and synaptic organizing principles to algorithmically reconstruct detailed anatomy and physiology from sparse experimental data. An objective anatomical method defines a neocortical volume of 0.29 plus-or-minus 0.01 mm3 containing 31,000 neurons, and patch-clamp studies identify 55 layer-specific morphological and 207 morpho-electrical neuron subtypes. When digitally reconstructed neurons are positioned in the volume and synapse formation is restricted to biological bouton densities and numbers of synapses per connection, their overlapping arbors form \approx{}8 million connections with \approx{}37 million synapses. Simulations reproduce an array of in vitro and in vivo experiments without parameter tuning. Additionally, we find a spectrum of network states with a sharp transition from synchronous to asynchronous activity, modulated by physiological mechanisms. The spectrum of network states, dynamically reconfigured around this transition, supports diverse information processing strategies.},
}
@misc{MarkrametalHBP-12,
     title = {{The Human Brain Project: A Report to the European Commission}},
    author = {Henry Markram {\it{et al}}},
      year = 2012,
howpublished = {{The HBP-PS Consortium, Lausanne}},
}
@article{MarkrametalNATURE-04,
       author = {Markram, Henry and Toledo-Rodriguez, Maria and Wang, Yun and Gupta, Anirudh and Silberberg, Gilad and Wu, Caizhi},
        title = {Interneurons of the neocortical inhibitory system},
      journal = {Nature Reviews Neuroscience},
         year = 2004,
       volume = 5,
        issue = 10,
        pages = {793-807}
}
@article{MarkrametalNEUROPHARMACOLOGY-98,
       author = {Henry Markram and Dimitri Pikus and Anirudh Gupta and Misha Tsodyks},
        title = {Potential for multiple mechanisms, phenomena and algorithms for synaptic plasticity at single synapses},
      journal = {Neuropharmacology},
       volume = {37},
       number = {4-5},
         year = {1998},
        pages = {489-500},
     abstract = {Recent experimental evidence indicates that in the neocortex, the manner in which each synapse releases neurotransmitter in response to trains of presynaptic action potentials is potentially unique. These unique transmission characteristics arise because of a large heterogeneity in various synaptic properties that determine frequency dependence of transmission such as those governing the rates of synaptic depression and facilitation. A theoretical analysis was therefore undertaken to explore the phenomenologies of changes in the values of these synaptic parameters. The results illustrate how the change in any one of several synaptic parameters produces a distinctive effect on synaptic transmission and how these distinctive effects can point to the most likely biophysical mechanisms. These results could therefore be useful in studies of synaptic plasticity in order to obtain a full characterization of the phenomenologies of synaptic modifications and to isolate potential biophysical mechanisms. Based on this theoretical analysis and experimental data, it is proposed that there exists multiple mechanisms, phenomena and algorithms for synaptic plasticity at single synapses. Finally, it is shown that the impact of changing the values of synaptic parameters depends on the values of the other parameters. This may indicate that the various mechanisms, phenomena and algorithms are interlinked in a 'synaptic plasticity code'.}
}
@article{MarkrametalNLM-98,
       author = {Henry Markram and Anirudh Gupta and Asher Uziel and Yun Wang and Misha Tsodyks},
        title = {Information Processing with Frequency-Dependent Synaptic Connections},
      journal = {Neurobiology of Learning and Memory},
       volume = {70},
       number = {1-2},
         year = {1998},
        pages = {101-112},
     abstract = {The efficacy of synaptic transmission between two neurons changes as a function of the history of previous activations of the synaptic connection. This history dependence can be characterized by examining the dependence of transmission on the frequency of stimulation. In this framework synaptic plasticity can also be examined in terms of changes in the frequency dependence of transmission and not merely in terms of synaptic strength which constitutes only a linear scaling mechanism. Recent work shows that the frequency dependence of transmission determines the content of information transmitted between neurons and that synaptic modifications can change the content of information transmitted. Multipatch-clamp recordings revealed that the frequency dependence of transmission is potentially unique for each synaptic connection made by a single axon and that the class of pre-postsynaptic neuron determines the class of frequency dependence (activity independent), while the unique activity relationship between any two neurons could determine the precise values of the parameters within a specific class (activity dependent). The content of information transmitted between neurons is also formalized to provide synaptic transfer functions which can be used to determine the role of the synaptic connection within a network of neurons. It is proposed that deriving synaptic transfer functions is crucial in order to understand the link between synaptic transmission and information processing within networks of neurons and to understand the link between synaptic plasticity and learning and memory.}
}
@article{MarkrametalPNAS-98,
       author = {Markram, Henry and Wang, Yun and Tsodyks, Misha},
        title = {Differential signaling via the same axon of neocortical pyramidal neurons},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {95},
       number = {9},
        pages = {5323-5328},
         year = {1998},
     abstract = {The nature of information stemming from a single neuron and conveyed simultaneously to several hundred target neurons is not known. Triple and quadruple neuron recordings revealed that each synaptic connection established by neocortical pyramidal neurons is potentially unique. Specifically, synaptic connections onto the same morphological class differed in the numbers and dendritic locations of synaptic contacts, their absolute synaptic strengths, as well as their rates of synaptic depression and recovery from depression. The same axon of a pyramidal neuron innervating another pyramidal neuron and an interneuron mediated frequency-dependent depression and facilitation, respectively, during high frequency discharges of presynaptic action potentials, suggesting that the different natures of the target neurons underlie qualitative differences in synaptic properties. Facilitating-type synaptic connections established by three pyramidal neurons of the same class onto a single interneuron, were all qualitatively similar with a combination of facilitation and depression mechanisms. The time courses of facilitation and depression, however, differed for these convergent connections, suggesting that different pre-postsynaptic interactions underlie quantitative differences in synaptic properties. Mathematical analysis of the transfer functions of frequency-dependent synapses revealed supra-linear, linear, and sub-linear signaling regimes in which mixtures of presynaptic rates, integrals of rates, and derivatives of rates are transferred to targets depending on the precise values of the synaptic parameters and the history of presynaptic action potential activity. Heterogeneity of synaptic transfer functions therefore allows multiple synaptic representations of the same presynaptic action potential train and suggests that these synaptic representations are regulated in a complex manner. It is therefore proposed that differential signaling is a key mechanism in neocortical information processing, which can be regulated by selective synaptic modifications.},
}
@book{Marmarelis2004,
        title = {Nonlinear Dynamic Modeling of Physiological Systems},
       author = {Vasilis Z. Marmarelis},
       series = {{IEEE} Press Series on Biomedical Engineering},
         year = {2004},
    publisher = {John Wiley \& Sons}
}
@article{MarmarelisABME-97,
        title = {Modeling Methodology for Nonlinear Physiological Systems},
       author = {Vasilis Z. Marmarelis},
      journal = {Annals of Biomedical Engineering},
       volume = {25},
        pages = {239-251},
         year = {1997},
     abstract = {A general modeling approach for a broad class of nonlinear systems is presented that uses the concept of principal dynamic modes (PDMs). These PDMs constitute a filter bank whose outputs feed into a multi-input static nonlinearity of multinomial (polynomial) form to yield a general model for the broad class of Volterra systems. Because the practically obtainable models (from stimulus-response data) are of arbitrary order of nonlinearity, this approach is applicable to many nonlinear physiological systems heretofore beyond our methodological means. Two specific methods are proposed for the estimation of these PDMs and the associated nonlinearities from stimulus-response data. Method I uses eigen decomposition of a properly constructed matrix using the first two kernel estimates (obtained by existing methods). Method II uses a particular class of feedforward artificial neural networks with polynomial activation functions. The efficacy of these two methods is demonstrated with computer-simulated examples, and their relative performance is discussed. The advent of this approach promises a practicable solution to the vexing problem of modeling highly nonlinear physiological systems, provided that experimental data be available for reliable estimation of the requisite PDMs.}
}
@book{Marr82,
       author = {Marr, David},
        title = {Vision},
    publisher = {W. H. Freeman and Company},
      address = {New York},
         year = 1982,
} 
@article{MarrJoP-69,
       author = {Marr, David},
        title = {A Theory of Cerebellar Cortex},
      journal = {Journal of Physiology},
       volume = 202,
         year = 1969,
        pages = {437-470},
}
@article{MarrandBrindleyPTRS_B-71,
       author = {D. Marr and Giles Skey Brindley},
        title = {Simple memory: a theory for archicortex},
      journal = {Philosophical Transactions of the Royal Society of London. B, Biological Sciences},
       volume = {262},
       number = {841},
        pages = {23-81},
         year = {1971},
     abstract = {It is proposed that the most important characteristic of archicortex is its ability to perform a simple kind of memorizing task. It is shown that rather general numerical constraints roughly determine the dimensions of memorizing models for the mammalian brain, and from these is derived a general model for archicortex. The addition of further constraints leads to the notion of a simple representation, which is a way of translating a great deal of information into the firing of about 200 out of a population of 105 cells. It is shown that if about 105 simple representations are stored in such a population of cells, very little information about a single learnt event is necessary to provoke its recall. A detailed numerical examination is made of a particular example of this kind of memory, and various general conclusions are drawn from the analysis. The insight gained from these models is used to derive theories for various archicortical areas. A functional interpretation is given of the cells and synapses of the area entorhinalis, the presubiculum, the prosubiculum, the cornu ammonis and the fascia dentata. Many predictions are made, a substantial number of which must be true if the theory is correct. A general functional classification of typical archicortical cells is proposed. }
}
@article{MarreetalNEUROSCIENCE-12,
        title = {Mapping a complete neural population in the retina},
       author = {Marre, O. and Amodei, D. and Deshmukh, N. and Sadeghi, K. and Soo, F. and Holy, T.E. and Berry 2nd, M.J.},
      journal = {Journal Neuroscience},
       volume = 32,
        issue = 43,
         year = 2012,
        pages = {14859-73},
     abstract = {Recording simultaneously from essentially all of the relevant neurons in a local circuit is crucial to understand how they collectively represent information. Here we show that the combination of a large, dense multielectrode array and a novel, mostly automated spike-sorting algorithm allowed us to record simultaneously from a highly overlapping population of >200 ganglion cells in the salamander retina. By combining these methods with labeling and imaging, we showed that up to 95\% of the ganglion cells over the area of the array were recorded. By measuring the coverage of visual space by the receptive fields of the recorded cells, we concluded that our technique captured a neural population that forms an essentially complete representation of a region of visual space. This completeness allowed us to determine the spatial layout of different cell types as well as identify a novel group of ganglion cells that responded reliably to a set of naturalistic and artificial stimuli but had no measurable receptive field. Thus, our method allows unprecedented access to the complete neural representation of visual information, a crucial step for the understanding of population coding in sensory systems.}
}        
@article{MarsacetalMEDICAL_PHYSICS-12,
       author = {Marsac, L. and Chauvet, D. and Larrat, B. and Pernot, M. and Robert, B. and Fink, M. and Boch, A.L. and Aubry, J.F. and Tanter, M.},
        title = {MR-guided adaptive focusing of therapeutic ultrasound beams in the human head},
      journal = {Medical Physics},
       volume = 39,
        issue = 2,
         year = 2012,
        pages = {1141-1149},
     abstract = {PURPOSE: This study aims to demonstrate, using human cadavers the feasibility of energy-based adaptive focusing of ultrasonic waves using magnetic resonance acoustic radiation force imaging (MR-ARFI) in the framework of non-invasive transcranial high intensity focused ultrasound (HIFU) therapy.

METHODS: Energy-based adaptive focusing techniques were recently proposed in order to achieve aberration correction. The authors evaluate this method on a clinical brain HIFU system composed of 512 ultrasonic elements positioned inside a full body 1.5 T clinical magnetic resonance (MR) imaging system. Cadaver heads were mounted onto a clinical Leksell stereotactic frame. The ultrasonic wave intensity at the chosen location was indirectly estimated by the MR system measuring the local tissue displacement induced by the acoustic radiation force of the ultrasound (US) beams. For aberration correction, a set of spatially encoded ultrasonic waves was transmitted from the ultrasonic array and the resulting local displacements were estimated with the MR-ARFI sequence for each emitted beam. A noniterative inversion process was then performed in order to estimate the spatial phase aberrations induced by the cadaver skull. The procedure was first evaluated and optimized in a calf brain using a numerical aberrator mimicking human skull aberrations. The full method was then demonstrated using a fresh human cadaver head.

RESULTS: The corrected beam resulting from the direct inversion process was found to focus at the targeted location with an acoustic intensity 2.2 times higher than the conventional non corrected beam. In addition, this corrected beam was found to give an acoustic intensity 1.5 times higher than the focusing pattern obtained with an aberration correction using transcranial acoustic simulation-based on X-ray computed tomography (CT) scans.

CONCLUSIONS: The proposed technique achieved near optimal focusing in an intact human head for the first time. These findings confirm the strong potential of energy-based adaptive focusing of transcranial ultrasonic beams for clinical applications.}
}
@article{MarshallandSchnitzerACS-13,
       author = {Jesse D. Marshall and Mark J. Schnitzer},
        title = {Optical Strategies for Sensing Neuronal Voltage Using Quantum Dots and Other Semiconductor Nanocrystals},
      journal = {ACS Nano},
    publisher = {American Chemical Society},
       volume = 7,
        issue = 5,
         year = 2013,
        pages = {4601-4609}
}
@article{MarshalletalJN-04,
        title = {Transcranial direct current stimulation during sleep improves declarative memory},
       author = {L. Marshall and M. M\"{o}lle and M. Hallschmid and J. Born},
      journal = {Journal Neuroscience},
       volume = 24,
        issue = 44,
         year = 2004,
        pages = {9985-9992},
}
@article{MarsheletalNEURON-10,
        title = {Targeting single neuronal networks for gene expression and cell labeling in vivo},
       author = {Marshel, J.H. and Mori, T. and Nielsen, K.J. and Callaway, E.M.},
      journal = {Neuron},
       volume = 67,
        issue = 4,
        pages = {562-574},
         year = 2010,
     abstract = {To understand fine-scale structure and function of single mammalian neuronal networks, we developed and validated a strategy to genetically target and trace monosynaptic inputs to a single neuron in vitro and in vivo. The strategy independently targets a neuron and its presynaptic network for specific gene expression and fine-scale labeling, using single-cell electroporation of DNA to target infection and monosynaptic retrograde spread of a genetically modifiable rabies virus. The technique is highly reliable, with transsynaptic labeling occurring in every electroporated neuron infected by the virus. Targeting single neocortical neuronal networks in vivo, we found clusters of both spiny and aspiny neurons surrounding the electroporated neuron in each case, in addition to intricately labeled distal cortical and subcortical inputs. This technique, broadly applicable for probing and manipulating single neuronal networks with single-cell resolution in vivo, may help shed new light on fundamental mechanisms underlying circuit development and information processing by neuronal networks throughout the brain.}
}
@article{MarsheletalNEURON-11,
       author = {Marshel, James H. and Garrett, Marina E. and Nauhaus, Ian and Callaway, Edward M.},
        title = {Functional Specialization of Seven Mouse Visual Cortical Areas},
      journal = {Neuron},
    publisher = {Elsevier},
       volume = {72},
        issue = {6},
         year = {2011},
        pages = {1040-1054},
     abstract = {To establish the mouse as a genetically tractable model for high-order visual processing, we characterized fine-scale retinotopic organization of visual cortex and determined functional specialization of layer 2/3 neuronal populations in seven retinotopically identified areas. Each area contains a distinct visuotopic representation and encodes a unique combination of spatiotemporal features. Areas LM, AL, RL, and AM prefer up to three times faster temporal frequencies and significantly lower spatial frequencies than V1, while V1 and PM prefer high spatial and low temporal frequencies. LI prefers both high spatial and temporal frequencies. All extrastriate areas except LI increase orientation selectivity compared to V1, and three areas are significantly more direction selective (AL, RL, and AM). Specific combinations of spatiotemporal representations further distinguish areas. These results reveal that mouse higher visual areas are functionally distinct, and separate groups of areas may be specialized for motion-related versus pattern-related computations, perhaps forming pathways analogous to dorsal and ventral streams in other species.},
}
@book{Marsonetal86,
       author = {Marsan, M. Ajmone and Balbo, G. and Conte, G.},
        title = {Performance Models of Multiprocessor Systems},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1986,
}
@inproceedings{MarszaleketalCVPR-09,
       author = {Marcin Marsza{\l}ek and Ivan Laptev and Cordelia Schmid},
        title = {Actions in Context},
    booktitle = {{IEEE} Computer Society Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 2009,
}
@article{MartelletalNATURE-BIOTECHNOLOGY-12,
       author = {Martell, Jeffrey D. and Deerinck, Thomas J. and Sancak, Yasemin and Poulos, Thomas L. and Mootha, Vamsi K. and Sosinsky, Gina E. and Ellisman, Mark H. and Ting, Alice Y.},
        title = {Engineered ascorbate peroxidase as a genetically encoded reporter for electron microscopy},
      journal = {Nature Biotechnology},
         year = 2012,
    publisher = {Nature Publishing Group},
       volume = 30,
        issue = 11,
        pages = {1143-1148},
}
@inproceedings{MartensICML-10,
       author = {James Martens},
        title = {Deep learning via Hessian-free optimization},
    booktitle = {Proceedings of the 27th International Conference on Machine Learning},
       editor = {Johannes F{\"{u}}rnkranz and Thorsten Joachims},
         year = {2010},
        pages = {735-742},
         note = {http://www.cs.toronto.edu/~jmartens/docs/HF_talk.pdf},
     abstract = {We develop a 2nd-order optimization method based on the "Hessian-free" approach, and apply it to training deep auto-encoders. Without using pre-training, we obtain results superior to those reported by Hinton \& Salakhutdinov (2006) on the same tasks they considered. Our method is practical, easy to use, scales nicely to very large datasets, and isn't limited in applicability to auto- encoders, or any specific model class. We also discuss the issue of "pathological curvature" as a possible explanation for the difficulty of deep- learning and how 2nd-order optimization, and our method in particular, effectively deals with it.},
}
@book{Martin01,
       author = {Fred Martin},
        title = {Robotic Explorations: A Hands-On Introduction to Engineering},
    publisher = {Prentice Hall},
      address = {Upper Saddle River, NJ},
         year = 2001,
}
@book{Martin66,
       author = {J. J. Martin},
        title = {Bayesian Decision Problems and {M}arkov Chains},
    publisher = {{John Wiley \& Sons, Inc.}},
         year = 1966,
}
@inproceedings{MartinetalICCV-01,
       author = {D. Martin and C. Fowlkes and D. Tal and J. Malik},
        title = {A database of human segmented natural images and its application to evaluating segmentation algorithms and measuring ecological statistics},
    booktitle = {Proceedings of the 8th International Conference on Computer Vision},
     location = {Vancouver, Canada},
       volume = 2,
        pages = {416-423},
         year = 2001,
}
@article{MartinetalJAMS-09,
        title = {Toward an artificial Golgi: redesigning the biological activities of heparan sulfate on a digital microfluidic chip},
       author = {Martin, J.G. and Gupta, M. and Xu, Y. and Akella, S. and Liu, J. and Dordick, J.S. and Linhardt, R.J.},
      journal = {Journal of the American Chemical Society},
       volume = 131,
        issue = 31,
        pages = {11041-8},
         year = 2009,
     abstract = {Using digital microfluidics, recombinant enzyme technology, and magnetic nanoparticles, we have created a functional prototype of an artificial Golgi organelle. Analogous to the natural Golgi, which is responsible for the enzymatic modification of glycosaminoglycans immobilized on proteins, this artificial Golgi enzymatically modifies glycosaminoglycans, specifically heparan sulfate (HS) chains immobilized onto magnetic nanoparticles. Sulfo groups were transferred from adenosine 3'-phosphate 5'-phosphosulfate to the 3-hydroxyl group of the D-glucosamine residue in an immobilized HS chain using D-glucosaminyl 3-O-sulfotransferase. After modification, the nanoparticles with immobilized HS exhibited increased affinity for fluorescently labeled antithrombin III as detected by confocal microscopy. Since the biosynthesis of HS involves an array of specialized glycosyl transferases, epimerase, and sulfotransferases, this approach should mimic the synthesis of HS in vivo. Furthermore, our method demonstrates the feasibility of investigating the effects of multienzyme systems on the structure of final glycan products for HS-based glycomic studies.}
} 
@article{MartinetalPAMI-04,
       author = {D. Martin and C. Fowlkes and J. Malik},
        title = {Learning to Detect Natural Image Boundaries Using Local Brightness, Color and Texture Cues},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
       volume = 26,
        issue = 5,
        pages = {530-549},
         year = 2004,
} 
@inproceedings{MartinsandShapiroIJCAI-83,
       author = {Martins, J. P. and Shapiro, S. C.},
        title = {Reasoning in Multiple Belief Spaces},
    booktitle = {Proceedings {IJCAI} 8},
     location = {Karlsruhe, West Germany},
 organization = {IJCAII},
         year = 1983,
}
@article{MarxNATURE-13,
       author = {Marx, Vivien},
        title = {Neurobiology: Brain mapping in high resolution},
      journal = {Nature},
         year = {2013},
       volume = {503},
        issue = {7474},
        pages = {147-152},
}
@article{MarxNATURE-METHODS-18,
       author = {Marx, Vivien},
        title = {Neuroscience models: {C}hoose your dimension},
      journal = {Nature Methods},
         year = {2018},
     abstract = {New types of 2D and 3D brain models capture more physiology.},
}
@article{MarxetalNATURE-PROTOCOLS-12,
        title = {Improved biocytin labeling and neuronal {3D} reconstruction},
       author = {Marx, Manuel and Gunter, Robert H. and Hucko, Werner and Radnikow, Gabriele and Feldmeyer, Dirk},
      journal = {Nature Protocols},
         year = 2012,
    publisher = {Nature Publishing Group},
       volume = 7,
        issue = 2,
        pages = {394-407},
     abstract = {In this report, we describe a reliable protocol for biocytin labeling of neuronal tissue and diaminobenzidine (DAB)-based processing of brain slices. We describe how to embed tissues in different media and how to subsequently histochemically label the tissues for light or electron microscopic examination. We provide a detailed dehydration and embedding protocol using Eukitt that avoids the common problem of tissue distortion and therefore prevents fading of cytoarchitectural features (in particular, lamination) of brain tissue; as a result, additional labeling methods (such as cytochrome oxidase staining) become unnecessary. In addition, we provide correction factors for tissue shrinkage in all spatial dimensions so that a realistic neuronal morphology can be obtained from slice preparations. Such corrections were hitherto difficult to calculate because embedding in viscous media resulted in highly nonlinear tissue deformation. Fixation, immunocytochemistry and embedding procedures for light microscopy (LM) can be completed within 42-48 h. Subsequent reconstructions and morphological analyses take an additional 24 h or more.}
}
@article{MashouretalNEURON-20,
       author = {George A. Mashour and Pieter Roelfsema and Jean-Pierre Changeux and Stanislas Dehaene},
        title = {Conscious Processing and the Global Neuronal Workspace Hypothesis},
      journal = {Neuron},
       volume = {105},
       number = {5},
        pages = {776-798},
         year = {2020},
     abstract = {We review the central tenets and neuroanatomical basis of the global neuronal workspace (GNW) hypothesis, which attempts to account for the main scientific observations regarding the elementary mechanisms of conscious processing in the human brain. {\it{The GNW hypothesis proposes that, in the conscious state, a non-linear network ignition associated with recurrent processing amplifies and sustains a neural representation, allowing the corresponding information to be globally accessed by local processors.}} We examine this hypothesis in light of recent data that contrast brain activity evoked by either conscious or non-conscious contents, as well as during conscious or non-conscious states, particularly general anesthesia. {\it{We also discuss the relationship between the intertwined concepts of conscious processing, attention, and working memory.}}}
}
@article{MasipetalNATURE-METHODS-16,
       author = {Masip, Martin E. and Huebinger, Jan and Christmann, Jens and Sabet, Ola and Wehner, Frank and Konitsiotis, Antonios and Fuhr, Gunther R. and Bastiaens, Philippe I. H.},
        title = {Reversible cryo-arrest for imaging molecules in living cells at high spatial resolution},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
         year = {2016},
       volume = {13},
        issue = {8},
        pages = {665-672},
     abstract = {The dynamics of molecules in living cells hampers precise imaging of molecular patterns by functional and super-resolution microscopy. We developed a method that circumvents lethal chemical fixation and allows on-stage cryo-arrest for consecutive imaging of molecular patterns within the same living, but arrested, cells. The reversibility of consecutive cryo-arrests was demonstrated by the high survival rate of different cell lines and by intact growth factor signaling that was not perturbed by stress response. Reversible cryo-arrest was applied to study the evolution of ligand-induced receptor tyrosine kinase activation at different scales. The nanoscale clustering of epidermal growth factor receptor (EGFR) in the plasma membrane was assessed by single-molecule localization microscopy, and endosomal microscale activity patterns of ephrin receptor A2 (EphA2) were assessed by fluorescence lifetime imaging microscopy. Reversible cryo-arrest allows the precise determination of molecular patterns while conserving the dynamic capabilities of living cells.},
}
@article{MaslandCURRENT-BIOLOGY-04,
       author = {Masland, Richard},
        title = {Neuronal Cell Types},
      journal = {Current Biology},
       volume = {14},
        issue = {13},
         year = {2004},
}
@article{MasseetalFiN-12,
       author = {Masse, Nicolas Y. and Cachero, Sebastian and Ostrovsky, Aaron D. and Jefferis, Gregory S.X.E.},
        title = {A Mutual Information Approach to Automate Identification of Neuronal Clusters in Drosophila Brain Images},
      journal = {Frontiers in Neuroinformatics},
         year = {2012},
    publisher = {Frontiers Research Foundation},
       volume = {6},
        pages = {21},
     abstract = {Mapping neural circuits can be accomplished by labeling a small number of neural structures per brain, and then combining these structures across multiple brains. This sparse labeling method has been particularly effective in Drosophila melanogaster, where clonally related clusters of neurons derived from the same neural stem cell (neuroblast clones) are functionally related and morphologically highly stereotyped across animals. However identifying these neuroblast clones (approximately 180 per central brain hemisphere) manually remains challenging and time consuming. Here, we take advantage of the stereotyped nature of neural circuits in Drosophila to identify clones automatically, requiring manual annotation of only an initial, smaller set of images. Our procedure depends on registration of all images to a common template in conjunction with an image processing pipeline that accentuates and segments neural projections and cell bodies. We then measure how much information the presence of a cell body or projection at a particular location provides about the presence of each clone. This allows us to select a highly informative set of neuronal features as a template that can be used to detect the presence of clones in novel images. The approach is not limited to a specific labeling strategy and can be used to identify partial (e.g., individual neurons) as well as complete matches. Furthermore this approach could be generalized to studies of neural circuits in other organisms.},
}
@article{MassiminietalEJN-09,
     title = {Slow waves, synaptic plasticity and information processing: insights from transcranial magnetic stimulation and high-density EEG experiments},
    author = {M. Massimini and G. Tononi and R. Huber},
   journal = {European Journal of Neuroscience}, 
    volume = 29,
     issue = 9,
      year = 2009,
     pages = {1761-1770},
}
@article{MasulliandVillaCoRR-15,
       author = {Paolo Masulli, Alessandro E. P. Villa},
        title = {The topology of the directed clique complex as a network invariant},
      journal = {CoRR},
       volume = {arXiv:1510.00660},
         year = {2015},
     abstract = {We introduce new algebro-topological invariants of directed networks, based on the topological construction of the directed clique complex. The shape of the underlying directed graph is encoded in a way that can be studied mathematically to obtain network invariants such as the Euler characteristic and the Betti numbers. Two different cases illustrate the application of these invariants. We investigate how the evolution of a Boolean recurrent artificial neural network is influenced by its topology in a dynamics involving pruning and strengthening of the connections, and to show that the topological features of the directed clique complex influence the dynamical evolution of the network. The second application considers the directed clique complex in a broader framework, to define an invariant of directed networks, the network degree invariant, which is constructed by computing the topological invariant on a sequence of sub-networks filtered by the minimum in- or out-degree of the nodes. The application of the new invariants presented here can be extended to any directed network. These invariants provide a new method for the assessment of specific functional features associated with the network topology.},
}
@techreport{MataricMSc,
       author = {Mataric, Maja J.},
        title = {A Distributed Model for Mobile Robot Environment-Learning and Navigation},
         type = {Technical Report},
        issue = 1128,
  institution = {{MIT} {AI} Laboratory},
         year = 1990,
}
@inproceedings{MatasetalBMVC-02,
       author = {J. Matas and O. Chum and M. Urban and T. Pajdla},
        title = {Robust Wide Baseline Stereo from Maximally Stable Extremal Regions},
    booktitle = {In the Proceedings of the British Machine Vision Conference},
        pages = {384-393},
         year = 2002,
}
@article{MatchinandHickokCEREBRAL-CORTEX-19,
       author = {Matchin, W.  and Hickok, G.},
        title = {{The Cortical Organization of Syntax}},
      journal = {Cerebral Cortex},
         year = {2019},
     abstract = {Syntax, the structure of sentences, enables humans to express an infinite range of meanings through finite means. The neurobiology of syntax has been intensely studied but with little consensus. Two main candidate regions have been identified: the posterior inferior frontal gyrus (pIFG) and the posterior middle temporal gyrus (pMTG). Integrating research in linguistics, psycholinguistics, and neuroscience, we propose a neuroanatomical framework for syntax that attributes distinct syntactic computations to these regions in a unified model. The key theoretical advances are adopting a modern lexicalized view of syntax in which the lexicon and syntactic rules are intertwined, and recognizing a computational asymmetry in the role of syntax during comprehension and production. Our model postulates a hierarchical lexical-syntactic function to the pMTG, which interconnects previously identified speech perception and conceptual-semantic systems in the temporal and inferior parietal lobes, crucial for both sentence production and comprehension. These relational hierarchies are transformed via the pIFG into morpho-syntactic sequences, primarily tied to production. We show how this architecture provides a better account of the full range of data and is consistent with recent proposals regarding the organization of phonological processes in the brain.}
}
@article{MathieuetalCoRR-19,
        title = {Disentangling Disentanglement in Variational Autoencoders},
       author = {Emile Mathieu and Tom Rainforth and N. Siddharth and Yee Whye Teh},
      journal = {CoRR},
         year = {2019},
       volume = {arXiv:1812.02833},
     abstract = {We develop a generalisation of disentanglement in VAEs---decomposition of the latent representation---characterising it as the fulfilment of two factors: a) the latent encodings of the data having an appropriate level of overlap, and b) the aggregate encoding of the data conforming to a desired structure, represented through the prior. Decomposition permits disentanglement, i.e. explicit independence between latents, as a special case, but also allows for a much richer class of properties to be imposed on the learnt representation, such as sparsity, clustering, independent subspaces, or even intricate hierarchical dependency relationships. We show that the \Beta-VAE varies from the standard VAE predominantly in its control of latent overlap and that for the standard choice of an isotropic Gaussian prior, its objective is invariant to rotations of the latent representation. Viewed from the decomposition perspective, breaking this invariance with simple manipulations of the prior can yield better disentanglement with little or no detriment to reconstructions. We further demonstrate how other choices of prior can assist in producing different decompositions and introduce an alternative training objective that allows the control of both decomposition factors in a principled manner.},
} 
@inproceedings{MatikainenetalICCV-09,
        title = {Trajectons: Action recognition through the motion analysis of tracked},
       author = {P. Matikainen and M. Hebert and R. Sukthankar},
    booktitle = {Proceedings of {ICCV} Workshop on Video-oriented Object and Event Classification},
    publisher = {IEEE Computer Society},
         year = 2009,
}
@inproceedings{MatousekSOCG-92,
       author = {J. Matousek and M. Sharir and E. Welzl},
        title = {A subexponential bound for linear programming},
    booktitle = {Proceedings of the 8th Annual Symposium on Computational Geometry},
         year = 1992,
}
@article{MatsuiDGD-17,
       author = {Matsui, Hideaki},
        title = {Dopamine system, cerebellum, and nucleus ruber in fish and mammals},
      journal = {Development, Growth \& Differentiation},
       volume = {59},
       number = {4},
         year = {2017},
        pages = {219-227},
     abstract = {Small teleost fish including zebrafish and medaka have been used as animal models for research because of their small body size, vast amounts of eggs produced, their rapid development, low husbandry costs, and transparency during embryogenesis. Although the body size and appearance seem different, fish and mammals including human still possess anatomical and functional similarities in their brains. This review summarizes the similarities of brain structures and functions between teleost fish and mammalian brains, focusing on the dopamine system, functional regionalization of the cerebellum, and presence of the nucleus ruber.},
}
@article{MatsukawaetalJN-14,
       author = {Matsukawa, Hiroshi and Akiyoshi-Nishimura, Sachiko and Zhang, Qi and Luj\`{a}n, Rafael and Yamaguchi, Kazuhiko and Goto, Hiromichi and Yaguchi, Kunio and Hashikawa, Tsutomu and Sano, Chie and Shigemoto, Ryuichi and Nakashiba, Toshiaki and Itohara, Shigeyoshi},
        title = {Netrin-G/NGL Complexes Encode Functional Synaptic Diversification},
      journal = {The Journal of Neuroscience},
       volume = {34},
       number = {47},
         year = {2014},
        pages = {15779-15792},
     abstract = {Synaptic cell adhesion molecules are increasingly gaining attention for conferring specific properties to individual synapses. Netrin-G1 and netrin-G2 are trans-synaptic adhesion molecules that distribute on distinct axons, and their presence restricts the expression of their cognate receptors, NGL1 and NGL2, respectively, to specific subdendritic segments of target neurons. However, the neural circuits and functional roles of netrin-G isoform complexes remain unclear. Here, we use netrin-G-KO and NGL-KO mice to reveal that netrin-G1/NGL1 and netrin-G2/NGL2 interactions specify excitatory synapses in independent hippocampal pathways. In the hippocampal CA1 area, netrin-G1/NGL1 and netrin-G2/NGL2 were expressed in the temporoammonic and Schaffer collateral pathways, respectively. The lack of presynaptic netrin-Gs led to the dispersion of NGLs from postsynaptic membranes. In accord, netrin-G mutant synapses displayed opposing phenotypes in long-term and short-term plasticity through discrete biochemical pathways. The plasticity phenotypes in netrin-G-KOs were phenocopied in NGL-KOs, with a corresponding loss of netrin-Gs from presynaptic membranes. Our findings show that netrin-G/NGL interactions differentially control synaptic plasticity in distinct circuits via retrograde signaling mechanisms and explain how synaptic inputs are diversified to control neuronal activity.},
}
@article{MatsuzawaDEVELOPMENT-07,
       author = {Tetsuro Matsuzawa},
        title = {Comparative cognitive development},
      journal = {Developmental Science},
       volume = {10},
       number = {1},
         year = {2007},
        pages = {97-103},
     abstract = {This paper aims to compare cognitive development in humans and chimpanzees to illuminate the evolutionary origins of human cognition. Comparison of morphological data and life history strongly highlights the common features of all primate species, including humans. The human mother-infant relationship is characterized by the physical separation of mother and infant, and the stable supine posture of infants, that enables vocal exchange, face-to-face communication, and manual gestures. The cognitive development of chimpanzees was studied using the participation observation method. It revealed that humans and chimpanzees show similar development until 3 months of age. However, chimpanzees have a unique type of social learning that lacks the social reference observed in human children. Moreover, chimpanzees have unique immediate short-term memory capabilities. Taken together, this paper presents a plausible evolutionary scenario for the uniquely human characteristics of cognition.}
}
@incollection{MatsuzawaHUMAN-CONCEPTS-10,
        title = {Cognitive development in chimpanzees: A trade-off between memory and abstraction?},
       author = {Tetsuro Matsuzawa},
    booktitle = {Oxford series in developmental cognitive neuroscience. The making of human concepts},
       editor = {D. Mareschal and P. C. Quinn and S. E. G. Lea},
    publisher = {Oxford University Press},
      address = {New York, NY, US},
         year = {2010},
        pages = {227-244},
     abstract = {This chapter suggests that the strong, near-photographic memory of chimps for number may be one manifestation of a more general characteristic of a representational system that provides extraordinarily detailed records of visual scenes. Such a system may be viewed as adaptive in a cognitive niche in which rapid, categorical decisions need to be made about objects encountered (e.g. ripe vs. unripe food, friend vs. foe). By contrast, the human cognitive niche emphasizes linguistic descriptions of events that capture an abstract gist which can be communicated to others. In this sense, chimps may be likened to humans with autism who display weak central coherence (i.e. an eye for detail, but without the corresponding big-picture idea).}
}
@article{MattarandDawNATURE-NEUROSCIENCE-18,
       author = {Mattar, Marcelo G. and Daw, Nathaniel D.},
        title = {Prioritized memory access explains planning and hippocampal replay},
      journal = {Nature Neuroscience},
       volume = {21},
        issue = {11},
         year = {2018},
        pages = {1609-1617},
     abstract = {To make decisions, animals must evaluate candidate choices by accessing memories of relevant experiences. Yet little is known about which experiences are considered or ignored during deliberation, which ultimately governs choice. We propose a normative theory predicting which memories should be accessed at each moment to optimize future decisions. Using nonlocal "replay" of spatial locations in hippocampus as a window into memory access, we simulate a spatial navigation task in which an agent accesses memories of locations sequentially, ordered by utility: how much extra reward would be earned due to better choices. This prioritization balances two desiderata: the need to evaluate imminent choices versus the gain from propagating newly encountered information to preceding locations. Our theory offers a simple explanation for numerous findings about place cells; unifies seemingly disparate proposed functions of replay including planning, learning, and consolidation; and posits a mechanism whose dysfunction may underlie pathologies like rumination and craving.},
}
@article{MatterandDowNATURE-NEUROSCIENCE-18,
       author = {Mattar, M. G. and Daw, N. D.},
        title = {Prioritized memory access explains planning and hippocampal replay},
      journal = {Nature Neuroscience},
         year = {2018},
       volume = {21},
       number = {11},
        pages = {1609-1617},
     abstract = {To make decisions, animals must evaluate candidate choices by accessing memories of relevant experiences. Yet little is known about which experiences are considered or ignored during deliberation, which ultimately governs choice. We propose a normative theory predicting which memories should be accessed at each moment to optimize future decisions. Using nonlocal 'replay' of spatial locations in hippocampus as a window into memory access, we simulate a spatial navigation task in which an agent accesses memories of locations sequentially, ordered by utility: how much extra reward would be earned due to better choices. This prioritization balances two desiderata: the need to evaluate imminent choices versus the gain from propagating newly encountered information to preceding locations. Our theory offers a simple explanation for numerous findings about place cells; unifies seemingly disparate proposed functions of replay including planning, learning, and consolidation; and posits a mechanism whose dysfunction may underlie pathologies like rumination and craving.}
}
@article{MaussetalCELL-17,
       author = {Mauss, Alex S. and Pankova, Katarina and Arenz, Alexander and Nern, Aljoscha and Rubin, Gerald M. and Borst, Alexander},
        title = {Neural Circuit to Integrate Opposing Motions in the Visual Field},
      journal = {Cell},
    publisher = {Elsevier},
       volume = {162},
        issue = {2},
         year = {2017},
        pages = {351-362},
     abstract = {When navigating in their environment, animals use visual motion cues as feedback signals that are elicited by their own motion. Such signals are provided by wide-field neurons sampling motion directions at multiple image points as the animal maneuvers. Each one of these neurons responds selectively to a specific optic flow-field representing the spatial distribution of motion vectors on the retina. Here, we describe the discovery of a group of local, inhibitory interneurons in the fruit fly Drosophila key for filtering these cues. Using anatomy, molecular characterization, activity manipulation, and physiological recordings, we demonstrate that these interneurons convey direction-selective inhibition to wide-field neurons with opposite preferred direction and provide evidence for how their connectivity enables the computation required for integrating opposing motions. Our results indicate that, rather than sharpening directional selectivity per se, these circuit elements reduce noise by eliminating non-specific responses to complex visual information.},
}
@techreport{MaybeckTR-72,
       author = {Maybeck, Peter S.},
        title = {The {Kalman} Filter: An Introduction to Potential Users},
        issue = {TM-72-3},
  institution = {Air Force Flight Dynamics Laboratory, Wright Patterson AFB, Ohio},
         year = 1972,
}
@article{MaybergBIOLOGICAL-PSYCHIATRY-07,
        title = {Defining the Neural Circuitry of Depression: Toward a New Nosology With Therapeutic Implications},
       author = {Helen S. Mayberg},
      journal = {Biological Psychiatry},
       volume = {61},
       number = {6},
         year = {2007},
        pages = {729-730},
}        
@article{MaybergetalAJP-99,
        title = {Reciprocal limbic-cortical function and negative mood: converging {PET} findings in depression and normal sadness},
       author = {Mayberg, H. S. and Liotti, M. and Brannan, S. K. and McGinnis, S. and Mahurin, R. K. and Jerabek, P. A. and Silva, J. A. and Tekell, J. L. and Martin, C. C. and Lancaster, J. L.},
      journal = {The American Journal of Psychiatry},
    publisher = {American Psychiatric Association},
       volume = 156,
        issue = 5,
         year = 1999,
        pages = {675-82}
}
@book{Mayer-Schonberger2009,
        title = {Delete: The Virtue of Forgetting in the Digital Age},
       author = {Viktor Mayer-Sch\'{o}nberger},
    publisher = {Princeton University Press},
      address = {Princeton, New Jersey},
         year = 2009,
}
@misc{Mayfield_ClinicBASIC-BRAIN-ANATOMY-16,
        title = {Anatomy of the Brain},
       author = {Tonya Hines},
      address = {University of Cincinnati Department of Neurosurgery, Ohio},
         year = 2016,
 howpublished = {Mayfield Clinic Public Service},
          url = {https://www.mayfieldclinic.com/PDF/PE-AnatBrain.pdf},
}
@article{MaynardetalNATURE-12,
       author = {Maynard, Craig L. and Elson, Charles O. and Hatton, Robin D. and Weaver, Casey T.},
        title = {Reciprocal interactions of the intestinal microbiota and immune system},
      journal = {Nature},
    publisher = {Nature Publishing Group, a division of Macmillan Publishers Limited. All Rights Reserved.},
       volume = {489},
        issue = {7415},
         year = {2012},
        pages = {231-241},
     abstract = {The emergence of the adaptive immune system in vertebrates set the stage for evolution of an advanced symbiotic relationship with the intestinal microbiota. The defining features of specificity and memory that characterize adaptive immunity have afforded vertebrates the mechanisms for efficiently tailoring immune responses to diverse types of microbes, whether to promote mutualism or host defence. These same attributes can put the host at risk of immunemediated diseases that are increasingly linked to the intestinal microbiota. Understanding how the adaptive immune system copes with the remarkable number and diversity of microbes that colonize the digestive tract, and how the system integrates with more primitive innate immune mechanisms to maintain immune homeostasis, holds considerable promise for new approaches to modulate immune networks to treat and prevent disease.}
}
@article{MayrazandHintonPAMI-02,
       author = {Guy Mayraz and Geoffrey E. Hinton},
        title = {Recognizing Handwritten Digits Using Hierarchical Products of Experts},
      journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
       volume = 24,
        issue = 2,
         year = 2002,
        pages = {189-197},
}
@article{MazumderetalJoN-09,
       author = {Sonal Mazumder and Rajib Dey and M.K. Mitra and S. Mukherjee and G.C. Das},
        title = {Review: Biofunctionalized Quantum Dots in Biology and Medicine},
      journal = {Journal of Nanomaterials},
       volume = 2009,
         year = 2009,
}
@misc{McAfee,
      author = {McAfee},
       title = {{McAfee ASaP} introduces {R}umor},
howpublished = {\url{http://www.mcafeeasap.com/content/virusscan-asap/rumor.asp}},
        year = 2000,
}
@techreport{McAllesterTR-80,
      author = {McAllester, David A.},
       title = {The Use of Equality in Deduction and Knowledge Representation},
       issue = 550,
 institution = {{MIT} {AI} Laboratory},
        year = 1980,
}
@techreport{McAllesterTR-82,
      author = {McAllester, David A.},
       title = {Reasoning Utility Package User's Manual},
       issue = 667,
 institution = {{MIT} {AI} Laboratory},
        year = 1982,
}
@inproceedings{McAllesterandRosenblittAAAI-91,
      author = {McAllester, David A. and Rosenblitt, David},
       title = {Systematic Nonlinear Planning},
   booktitle = {Proceedings AAAI-91},
    location = {Anaheim, California},
organization = {AAAI},
        year = 1991,
       pages = {634-639},
}
@article{McAvoyetalNEURON-16,
       author = {Kathleen M. McAvoy and Kimberly N. Scobie and Stefan Berger and Craig Russo and Nannan Guo and Pakanat Decharatanachart and Hugo Vega-Ramirez and Sam Miake-Lye and Michael Whalen and Mark Nelson and Matteo Bergami and Dusan Bartsch and Rene Hen and Benedikt Berninger and Amar Sahay},
        title = {Modulating Neuronal Competition Dynamics in the Dentate Gyrus to Rejuvenate Aging Memory Circuits},
      journal = {Neuron},
       volume = {91},
       number = {6},
        pages = {1356-1373},
         year = {2016},
     abstract = {The neural circuit mechanisms underlying the integration and functions of adult-born dentate granule cell (DGCs) are poorly understood. Adult-born DGCs are thought to compete with mature DGCs for inputs to integrate. Transient genetic overexpression of a negative regulator of dendritic spines, Kruppel-like factor 9 (Klf9), in mature DGCs enhanced integration of adult-born DGCs and increased NSC activation. Reversal of Klf9 overexpression in mature DGCs restored spines and activity and reset neuronal competition dynamics and NSC activation, leaving the DG modified by a functionally integrated, expanded cohort of age-matched adult-born DGCs. Spine elimination by inducible deletion of Rac1 in mature DGCs increased survival of adult-born DGCs without affecting proliferation or DGC activity. Enhanced integration of adult-born DGCs transiently reorganized adult-born DGC local afferent connectivity and promoted global remapping in the DG. Rejuvenation of the DG by enhancing integration of adult-born DGCs in adulthood, middle age, and aging enhanced memory precision.}
}
@inproceedings{McCallumandNigamAAAI-98,
      author = {A. McCallum and K. Nigam},
       title = {A comparison of event models for {N}aive {B}ayes text classification},
   booktitle = {Proceedings of the AAAI-98 Workshop on Learning for Text Categorization},
         url = {citeseer.ist.psu.edu/article/mccallum98comparison.html},
        year = 1998,
}
@incollection{McCarthy68,
       author = {McCarthy, John},
        title = {Programs With Common Sense},
     crossref = {Minsky68},
        pages = {403-418},
}
@article{McCarthyAIJ-80,
       author = {McCarthy, John},
        title = {Circumscription - A Form of Nonmonotonic Reasoning},
      journal = AIJ,
       volume = 13,
         year = 1980,
        pages = {295-323},
}
@article{McCarthyAIJ-86,
       author = {McCarthy, John},
        title = {Applications of Circumscription to Formalizing Commonsense Knowledge},
      journal = AIJ,
       volume = 28,
         year = 1986,
        pages = {89-116},
}
@article{McCarthyCACM-60,
       author = {McCarthy, John},
        title = {Recursive Functions of Symbolic Expressions and their Computation by Machine},
      journal = {Communications of the {ACM}},
       volume = 7,
        pages = {184-195},
         year = 1960,
}
@techreport{McCarthyTR-63,
       author = {McCarthy, John},
        title = {Situations, Actions, and Causal Laws},
         type = {Memo No.},
        issue = 2,
  institution = {Stanford University Artificial Intelligence Project},
         year = 1963,
}
@article{McCarthyandHayesMI-69,
       author = {McCarthy, John and Hayes, Patrick J.},
        title = {Some Philosophical Problems from the Standpoint of Artificial Intelligence},
       editor = {Meltzer, B. and Michie, D.},
      journal = {Machine Intelligence},
    publisher = {Edinburgh University Press},
       volume = 4,
        pages = {463-502},
         year = 1969,
}
@article{McClelland79,
       author = {McClelland, J.},
        title = {On the time relations of mental processes: {A}n examination of systems of processes in cascade},
      journal = {Psychological Review},
       volume = 86,
        pages = {287-330},
         year = 1979,
}
@incollection{McClellandandCleeremansCONNECTIONISM-09,
       author = {James L. McClelland and Axel Cleeremans},
        title = {Connectionist Models},
    booktitle = {Oxford Companion to Consciousness},
       editor = {T. Byrne and A. Cleeremans and P. Wilken},
    publisher = {New York: Oxford University Press},
         year = {2009},
}
@article{McClellandandGoddardHIPPOCAMPUS-97,
       author = {McClelland, J. L. and Goddard, N. H.},
        title = {Considerations arising from a complementary learning systems perspective on hippocampus and neocortex},
      journal = {Hippocampus},
         year = {1996},
       volume = {6},
       number = {6},
        pages = {654-665},
     abstract = {We discuss a framework for the organization of learning systems in the mammalian brain, in which the hippocampus and related areas form a memory system complementary to learning mechanisms in neocortex and other areas. The hippocampal system stores new episodes and "replays" them to the neocortical system, interleaved with ongoing experience, allowing generalization as cortical memories form. The data to account for include: (1) neurophysiological findings concerning representations in hippocampal areas, (2) behavioral evidence demonstrating a spatial role for hippocampus, (3) and effects of surgical and pharmacological manipulations on neuronal firing in hippocampal regions in behaving animals. We hypothesize that the hippocampal memory system consists of three major modules: (1) an invertible encoder subsystem supported by the pathways between neocortex and entorhinal cortex, which provides a stable, compressed, invertible encoding in entorhinal cortex (EC) of cortical activity patterns, (2) a memory separation, storage, and retrieval subsystem, supported by pathways between EC, dentate gyrus and area CA3, including the CA3 recurrent collaterals, which facilitates encoding and storage in CA3 of individual EC patterns, and retrieval of those CA3 encodings, in a manner that minimizes interference, and (3) a memory decoding subsystem, supported by the Shaffer collaterals from area CA1 to area CA3 and the bi-directional pathways between EC and CA3, which provides the means by which a retrieved CA3 coding of an EC pattern can reinstate that pattern on EC. This model has shown that (1) there is a trade-off between the need for information-preserving, structure-extracting encoding of cortical traces and the need for effective storage and recall of arbitrary traces, (2) long-term depression of synaptic strength in the pathways subject to long-term potentiation is crucial in preserving information, (3) area CA1 must be able to exploit correlations in EC patterns in the direct perforant path synapses.}
}
@article{McClellandandRumelhartJoEPG-85,
       author = {McClelland, J. L. and Rumelhart, D. E.},
        title = {Distributed memory and the representation of general and specific information},
      journal = {Journal of Experimental Psychology: General},
       volume = {114},
       number = {2},
         year = {1985},
        pages = {159-197},
     abstract = {We describe a distributed model of information processing and memory and apply it to the representation of general and specific information. The model consists of a large number of simple processing elements which send excitatory and inhibitory signals to each other via modifiable connections. Information processing is thought of as the process whereby patterns of activation are formed over the units in the model through their excitatory and inhibitory interactions. The memory trace of a processing event is the change or increment to the strengths of the interconnections that results from the processing event. The traces of separate events are superimposed on each other in the values of the connection strengths that result from the entire set of traces stored in the memory. The model is applied to a number of findings related to the question of whether we store abstract representations or an enumeration of specific experiences in memory. The model simulates the results of a number of important experiments which have been taken as evidence for the enumeration of specific experiences. At the same time, it shows how the functional equivalent of abstract representations--prototypes, logogens, and even rules--can emerge from the superposition of traces of specific experiences, when the conditions are right for this to happen. In essence, the model captures the structure present in a set of input patterns; thus, it behaves as though it had learned prototypes or rules, to the extent that the structure of the environment it has learned about can be captured by describing it in terms of these abstractions.}
}
@article{McClellandandRumelhartPR-88,
       author = {J. McClelland and D. Rumelhart},
        title = {An interactive activation model of context effects in letter perception: {I}. {A}n account of basic findings},
      journal = {Psychological Review},
        pages = {375-407},
       volume = 88,
         year = 1981,
} 
@inproceedings{McClellandetalCOGNITIVE-SCIENCE-20,
        title = {Cognitive consequences of structured education in a connectionist model of analogical reasoning},
       author = {Anonymous, Submission},
    booktitle = {Proceedings of the 2020 Annual Meeting of the Cognitive Science Society},
       volume = {},
       number = {},
        pages = {},
         year = {2020},
     abstract = {Education has a profound impact on human cognition. People who have participated in education are better at solving abstract reasoning tasks, can flexibly transfer knowledge across domains and are better at explaining their solutions. However, the properties of education that are responsible for these cognitive changes are poorly understood. We explore the hypothesis that a structured education consisting of a cumulative, compositional curricular learning regime using culturally constructed concepts and tools can account for many of these observations. In particular, we demonstrate that a connectionist model that learns to solve difficult analogical reasoning problems using a structured education is better at knowledge reuse, while simultaneously providing explanations for solutions. We predict that premature progression through a curriculum, before proficiency in a foundational stage has been established can fundamentally limit the potential for subsequent abstract reasoning performance or knowledge transfer ability.}
} 
@article{McClellandetalCoRR-19,
        title = {Extending Machine Language Models toward Human-Level Language Understanding},
       author = {James L. McClelland and Felix Hill and Maja Rudolph and Jason Baldridge and Hinrich Sch{\"{u}}tze},
      journal = {CoRR},
       volume = {arXiv:1912.05877},
         year = {2019},
     abstract = {Language is central to human intelligence. We review recent breakthroughs in machine language processing and consider what remains to be achieved. Recent approaches rely on domain general principles of learning and representation captured in artificial neural networks. Most current models, however, focus too closely on language itself. In humans, language is part of a larger system for acquiring, representing, and communicating about objects and situations in the physical and social world, and future machine language models should emulate such a system. We describe existing machine models linking language to concrete situations, and point toward extensions to address more abstract cases. Human language processing exploits complementary learning systems, including a deep neural network-like learning system that learns gradually as machine systems do, as well as a fast-learning system that supports learning new information quickly. Adding such a system to machine language models will be an important further step toward truly human-like language understanding.}
}
@article{McClellandetalNIiM-19,
       author = {James L. McClelland and Bruce L. McNaughton and Andrew K. Lampinen},
        title = {Integration of New Information in Memory: New Insights from a Complementary Learning Systems Perspective},
      journal = {Submitted to the Philosophical transactions of the Royal Society of London. Series B, Biological sciences},
    publisher = {The Royal Society},
         year = {2019},
     abstract = {According to complementary learning systems theory, integrating new memories into the neocortex of the brain without interfering with what is already known depends on a gradual learning process, interleaving new items with items already known. However, empirical studies show that information consistent with prior knowledge can be integrated very quickly. We use artificial neural networks with properties like those we attribute to the neocortex to develop a theoretical understanding of the role of consistency with prior knowledge in putatively neocortex-like learning systems, providing new insights into when integration will be fast or slow and how integration might be made more efficient when the items to be learned are hierarchically structured. The work relies on deep linear networks that capture the qualitative aspects of the learning dynamics of the more complex non-linear networks used in previous work. The time course of learning in these networks can be linked to the hierarchical structure in the training data, captured mathematically as a set of dimensions that correspond to the branches in the hierarchy. In this context, a new item to be learned can be characterized as having aspects that project onto previously known dimensions, and others that require adding a new branch/dimension. The projection onto the known dimensions can be learned rapidly without interleaving, but learning the new dimension requires gradual interleaved learning. When a new item only overlaps with items within one branch of a hierarchy, interleaving can focus on the previously-known items within this branch, resulting in faster integration with less interleaving overall. The discussion considers how the brain might exploit these facts to make learning more efficient and highlights predictions about what aspects of new information might be hard or easy to learn.}
}
@article{McClellandetalPR-95,
        title = {Why There Are Complementary Learning Systems in the Hippocampus and Neocortex: {I}nsights From the Successes and Failures of Connectionist Models of Learning and Memory},
       author = {McClelland, James L. and McNaughton, Bruce L. and O'Reilly, Randall C.},
      journal = {Psychological Review},
       volume = 102,
       number = 3,
         year = 1995,
        pages = {419-457},
     abstract = {Damage to the hippocampal system disrupts recent memory but leaves remote memory intact. The account presented here suggests that memories are first stored via synaptic changes in the hippocampal system, that these changes support reinstatement of recent memories in the neocortex, that neocortical synapses change a little on each reinstatement, and that remote memory is based on accumulated neocortical changes. Models that learn via changes to connections help explain this organization. These models discover the structure in ensembles of items if learning of each item is gradual and interleaved with learning about other items. This suggests that the neocortex learns slowly to discover the structure in ensembles of experiences. The hippocampal system permits rapid learning of new items without disrupting this structure, and reinstatement of new memories interleaves them with others to integrate them into structured neocortical memory systems.}
}
@article{McClellandetalPSYCHOLOGICAL-REVIEW-95,
       author = {McClelland, James L. and McNaughton, Bruce L. and O'Reilly, Randall C.},
        title = {Why there are complementary learning systems in the hippocampus and neocortex: Insights from the successes and failures of connectionist models of learning and memory.},
      journal = {Psychological Review},
    publisher = {American Psychological Association},
       volume = {102},
        issue = {3},
         year = {1995},
        pages = {419-457},
     abstract = {Damage to the hippocampal system disrupts recent memory but leaves remote memory intact. The account presented here suggests that memories are first stored via synaptic changes in the hippocampal system, that these changes support reinstatement of recent memories in the neocortex, that neocortical synapses change a little on each reinstatement, and that remote memory is based on accumulated neocortical changes. Models that learn via changes to connections help explain this organization. These models discover the structure in ensembles of items if learning of each item is gradual and interleaved with learning about other items. This suggests that the neocortex learns slowly to discover the structure in ensembles of experiences. The hippocampal system permits rapid learning of new items without disrupting this structure, and reinstatement of new memories interleaves them with others to integrate them into structured neocortical memory systems.}
}
@book{McCorduck79,
       author = {McCorduck, Pamela},
        title = {Machines Who Think},
    publisher = {W. H. Freeman and Company},
      address = {New York},
         year = 1979,
}
@article{McCormicketalNEUROPSYCHOLOGIA-10,
       author = {C. McCormick and M. Moscovitch and A.B. Protzner and C.G. Huber and M.P. McAndrews},
        title = {Hippocampal-neocortical networks differ during encoding and retrieval of relational memory: Functional and effective connectivity analyses},
      journal = {Neuropsychologia},
       volume = {48},
       number = {11},
         year = {2010},
        pages = {3272-3281},
     abstract = {Encoding and retrieval of relational information requires interaction between the hippocampus and various neocortical regions, but it is unknown whether the connectivity of hippocampal-neocortical networks is different at input and output stages. To examine this, we conducted a network analysis of event-related fMRI data collected during a face-recognition, remember/know paradigm. Directed analyses in the medial temporal lobe identified a small region in the left hippocampus that showed differential activation for encoding and retrieval of recollected versus familiar items. Multivariate seed partial least squares (PLS) analysis was used to identify brain regions that were functionally connected to this hippocampal region at encoding and retrieval of 'remembered' items. Anatomically based structural equation modeling (SEM) was then used to test for differences in effective connectivity of network nodes between these two memory stages. The SEM analysis revealed a reversal of directionality between the left hippocampus (LHC) and left inferior parietal cortex (LIPC) at encoding and retrieval. During encoding, activation of the LHC had a positive influence on the LIPC, whereas during retrieval the reverse pattern was found, i.e., the LIPC activation positively influenced LHC activation. These findings emphasize the importance of hippocampal-parietal connections and underscore the complexity of their interactions in initial binding and retrieval/reintegration of relational memory. We also found that, during encoding, the right hippocampus had a positive influence on the right retrospenial cortex, whereas during retrieval this influence was significantly weaker. We submit that examining patterns of connectivity can be important both to elaborate and constrain models of memory involving hippocampal-neocortical interactions.}
}
@article{McCullochandPitts43,
       author = {McCulloch, W. S. and Pitts, W. H.},
        title = {A Logical Calculus of Ideas Immanent in Nervous Activity},
      journal = {Bulletin of Mathematical Biophysics},
       volume = 5,
         year = 1943,
        pages = {115-133},
}
@book{McDermott01,
       author = {Drew McDermott},
        title = {Mind and Mechanism},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2001,
}
@article{McDermottAIJ-82,
       author = {McDermott, John},
        title = {A Rule-Based Configurer of Computer Systems},
      journal = AIJ,
       volume = 19,
        issue = 1,
         year = 1982,
        pages = {39-88},
}
@article{McDermottCS-78,
       author = {McDermott, Drew V.},
        title = {Planning and Acting},
      journal = {Cognitive Science},
       volume = 2,
         year = 1978,
        pages = {71-109},
}
@article{McDermottCS-78b,
       author = {Drew McDermott},
        title = {Tarskian semantics, or no notation without denotation!},
      journal = {Cognitive Science},
       volume = {2},
       number = {3},
         year = {1978},
        pages = {277-282},
}
@article{McDermottCS-82,
       author = {McDermott, Drew V.},
        title = {A temporal logic for reasoning about processes and plans},
      journal = {Cognitive Science},
       volume = 6,
         year = 1982,
        pages = {101-155},
}
@article{McDermottPAMI-83,
       author = {McDermott, Drew V.},
        title = {Contexts and data dependencies: a synthesis},
      journal = PAMI,
       volume = 5,
         year = 1983,
        pages = {237-246},
}
@techreport{McDermottPhD,
       author = {McDermott, Drew V.},
        title = {Flexibility and Efficiency in a Computer Program for Designing Circuits},
        issue = 402,
      comment = {NASL},
  institution = {{MIT} {AI} Laboratory},
         year = 1977,
} 
@techreport{McDermottTR-73,
       author = {McDermott, Drew V. and Sussman, Gerald J.},
        title = {The Conniver Reference Manual},
        issue = 259,
  institution = {{MIT} {AI} Laboratory},
         year = 1973,
} 
@techreport{McDermottTR-85,
       author = {McDermott, Drew V.},
        title = {The DUCK Manual},
        issue = 399,
  institution = {Yale University Computer Science Department},
         year = 1985,
}
@techreport{McDermottTR-89,
       author = {McDermott, Drew V.},
        title = {Regression Planning},
        issue = 752,
      comment = {PEDASTAL},
  institution = {Yale University Computer Science Department},
         year = 1989,
}
@article{McDermottandDavisAIJ-82,
       author = {McDermott, Drew V. and Ernest Davis},
        title = {Planning Routes Through Uncertain Territory},
      journal = AIJ,
       volume = 22,
      comment = {SPAM},
         year = 1982,
        pages = {107-156},
}
@article{McDermottandDoylsAIJ-80,
       author = {McDermott, Drew V. and Doyle, Jon},
        title = {Non-monotonic Logic {I}},
      journal = AIJ,
       volume = 13,
         year = 1980,
        pages = {41-72},
}
@article{McDermottandHendlerAIJ-95,
       author = {McDermott, Drew and Hendler, James},
        title = {Planning: What it is, What it could be, An Introduction to the Special Issue on Planning and Scheduling},
      journal = AIJ,
       volume = 76,
        issue = {1-2},
        pages = {1-16},
         year = 1995,
}
@misc{McEwenAEON-17,
        title = {When is stress good for you?},
       author = {Bruce McEwen},
         year = 2017,
 howpublished = {Aeon Article},
          url = {https://aeon.co/essays/how-stress-works-in-the-human-body-to-make-or-break-us}
} 
@article{McFaddenNoC-20,
       author = {McFadden, Johnjoe},
        title = {Integrating information in the brain's {EM} field: the {CEMI} field theory of consciousness},
      journal = {Neuroscience of Consciousness},
       volume = {2020},
       number = {1},
         year = {2020},
     abstract = {A key aspect of consciousness is that it represents bound or integrated information, prompting an increasing conviction that the physical substrate of consciousness must be capable of encoding integrated information in the brain. However, as Ralph Landauer insisted, "information is physical" so integrated information must be physically integrated. I argue here that nearly all examples of so-called "integrated information", including neuronal information processing and conventional computing, are only temporally integrated in the sense that outputs are correlated with multiple inputs: the information integration is implemented in time, rather than space, and thereby cannot correspond to physically integrated information. I point out that only energy fields are capable of integrating information in space. I describe the conscious electromagnetic information (CEMI) field theory which has proposed that consciousness is physically integrated, and causally active, information encoded in the brain's global electromagnetic (EM) field. I here extend the theory to argue that consciousness implements algorithms in space, rather than time, within the brain's EM field. I describe how the CEMI field theory accounts for most observed features of consciousness and describe recent experimental support for the theory. I also describe several untested predictions of the theory and discuss its implications for the design of artificial consciousness. The cemi field theory proposes a scientific dualism that is rooted in the difference between matter and energy, rather than matter and spirit.},
}
@book{McGinn02,
       author = {Colin McGinn},
        title = {The making of a Philosopher: My Journey through Twentieth-Century Philosophy},
    publisher = {Harper-Collins},
      address = {New York, NY},
         year = 2002,
}
@article{McGinnisetalCELL-84,
     title = {A homologous protein-coding sequence in Drosophila homeotic genes and its conservation in other metazoans},
    author = {McGinnis, W. and Garber, R. L. and Wirz, J. and Kuroiwa, A. and Gehring, W. J.},
   journal = {Cell},
     issue = 2,
     pages = {403-408},
    volume = 37,
      year = 1984
}
@book{McGrayne2011,
        title = {The Theory That Would Not Die: How Bayes' Rule Cracked the Enigma Code, Hunted Down Russian Submarines and Emerged Triumphant From Two Centuries of Controversy},
       author = {Sharon Bertsch McGrayne},
    publisher = {Yale University Press},
      address = {New Haven, CT},
         year = 2011,
} 
@article{McGuireetalSCIENCE-01,
       author = {McGuire, S. E. and Le, P. T. and Davis, R. L.},
        title = {The role of {D}rosophila mushroom body signaling in olfactory memory},
      journal = {Science},
       volume = {293},
       number = {5533},
         year = {2001},
        pages = {1330-1333},
     abstract = {The mushroom bodies of the Drosophila brain are important for olfactory learning and memory. To investigate the requirement for mushroom body signaling during the different phases of memory processing, we transiently inactivated neurotransmission through this region of the brain by expressing a temperature-sensitive allele of the shibire dynamin guanosine triphosphatase, which is required for synaptic transmission. Inactivation of mushroom body signaling through alpha/beta neurons during different phases of memory processing revealed a requirement for mushroom body signaling during memory retrieval, but not during acquisition or consolidation.}
}
@article{McLaughlinandOLeary2005,
        title = {Molecular gradients and development of retinotopic maps},
       author = {Mclaughlin, T. and O'Leary, D. D.},
      journal = {Annual Review of Neuroscience},
       volume = 28,
        issue = 1,
         year = 2005,
        pages = {327-355},
     abstract = {Gradients of axon guidance molecules have long been postulated to control the development of the organization of neural connections into topographic maps. We review progress in identifying molecules required for mapping and the mechanisms by which they act, focusing on the visual system, the predominant model for map development. The Eph family of receptor tyrosine kinases and their ligands, the ephrins, remain the only molecules that meet all criteria for graded topographic guidance molecules, although others fulfill some criteria. Recent reports further define their modes of action and new roles for them, including EphB/ephrin-B control of dorsal-ventral mapping, bidirectional signaling of EphAs/ephrin-As  bifunctional action of ephrins as attractants or repellents in a context-dependent manner, and complex interactions between multiple guidance molecules. In addition, spontaneous patterned neural activity has recently been shown to be required for map refinement during a brief critical period. We speculate on additional activities required for map development and suggest a synthesis of molecular and cellular mechanisms within the context of the complexities of map development.},
}
@article{McNabandKlingbergNATURE-NEUROSCIENCE-07,
       author = {McNab, Fiona and Klingberg, Torkel},
        title = {Prefrontal cortex and basal ganglia control access to working memory},
      journal = {Nature Neuroscience},
         year = {2008},
       volume = {11},
        issue = {1},
        pages = {103-107},
     abstract = {Our capacity to store information in working memory might be determined by the degree to which only relevant information is remembered. The question remains as to how this selection of relevant items to be remembered is accomplished. Here we show that activity in the prefrontal cortex and basal ganglia preceded the filtering of irrelevant information and that activity, particularly in the globus pallidus, predicted the extent to which only relevant information is stored. The preceding frontal and basal ganglia activity were also associated with inter-individual differences in working memory capacity. These findings reveal a mechanism by which frontal and basal ganglia activity exerts attentional control over access to working memory storage in the parietal cortex in humans, and makes an important contribution to inter-individual differences in working memory capacity.},
}
@article{McRoy-KBS-95,
        title = {Misunderstanding and the Negotiation of Meaning Using Abduction},
       author = {Susan W. McRoy},
      journal = {Knowledge-Based Systems},
       volume = 8,
        issue = {2-3},
        pages = {126-134},
         year = 1995,
       author = {Susan W. McRoy},
     abstract = {The approach described in the paper treats misunderstanding and repair as intrinsic to conversants' core language abilities, accountable by the same processing mechanisms that underlie normal speech. A unified theory has been developed by using default reasoning to generate utterances, and using abduction to characterize interpretation and repair. The model combines intentional and social accounts of discourse to capture the expectations that help constrain discourse, without requiring extended inference about participants' motivations when it is not needed. In intentional accounts, speakers use their beliefs, goals, and expectations to decide what to say; when they interpret an utterance, they identify goals and actions that might account for it. In sociological accounts provided by ethnomethodology, discourse interactions and the resolution of misunderstandings are normal activities guided by social conventions. The approach extends intentional accounts by using expectations deriving from social conventions in order to guide interpretation. This paper considers previous work on misunderstanding, provides an overview of the architecture of a new model, and examines how it accounts for an example repair.}
}
@techreport{McRoyandHirstAAAI-93,
        title = {Misunderstanding and the Negotiation of Meaning},
       author = {Susan McRoy and Graeme Hirst},
         year = 1993,
  institution = {AAAI Technical Report FS-93-05}
}
@inproceedings{McgregorandKyungtaeACL-18,
       author = {Mcgregor, Stephen and Lim, Kyungtae},
        title = {Affordances in Grounded Language Learning},
    booktitle = {ACL 2018 Workshop on Cognitive Aspects of Computational Language Learning and Processing},
         year = {2018},
     abstract = {We present a novel methodology involving mappings between different modes of semantic representations. We propose dis-tributional semantic models as a mechanism for representing the kind of world knowledge inherent in the system of abstract symbols characteristic of a sophisticated community of language users. Then, motivated by insight from ecological psychology , we describe a model approximating affordances, by which we mean a language learner's direct perception of opportunities for action in an environment. We present a preliminary experiment involving mapping between these two rep-resentational modalities, and propose that our methodology can become the basis for a cognitively inspired model of grounded language learning.}
}
@book{Mead1989,
       author = {Mead, Carver},
        title = {Analog VLSI and Neural Systems},
    publisher = {Addison-Wesley Longman Publishing Co., Inc.},
      address = {Boston, MA, USA},
         year = {1989},
}
@article{MeadCIT-87,
       author = {Carver Mead},
        title = {Neural Hardware for Vision},
      journal = {Engineering \& Science},
    publisher = {California Institute of Technlogy},
       volume = {1},
         year = {1987},
        pages = {2-7},
}
@article{MeadENGINEERING-SCIENCE-87,
       author = {Carver Mead},
        title = {Neural Hardware for Vision},
      journal = {Engineering \& Science},
    publisher = {California Institute of Technlogy},
       volume = {1},
         year = {1987},
        pages = {2-7},
}
@inproceedings{MeadIEEE-90,
       author = {Carver Mead},
        title = {Neuromorphic electronic systems},
    booktitle = {Proceedings of the {IEEE}},
         year = {1990},
     abstract = {Biological information-processing systems operate on completely different principles from those with which most engineers are familiar. For many problems, particularly those in which the input data are ill-conditioned and the computation can be specified in a relative manner, biological solutions are many orders of magnitude more effective than those we have been able to implement using digital methods. This advantage can be attributed principally to the use of elementary physical phenomena as computational primitives, and to the representation of information by the relative values of analog signals, rather than by the absolute values of digital signals. This approach requires adaptive techniques to mitigate the effects of component differences. This kind of adaptation leads naturally to systems that learn about their environment. Large-scale adaptive analog systems are more robust to component degredation and failure than are more conventional}
}
@book{MeadandConway80,
       author = {Mead, C. A. and Conway, M. A.},
        title = {Introduction to {VLSI} Systems},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1980,
}
@incollection{MeadandIsmail1989,
       author = {Mead, Carver},
       editor = {Mead, Carver and Ismail, Mohammed},
        title = {Adaptive Retina},
    booktitle = {Analog VLSI Implementation of Neural Systems},
         year = {1989},
    publisher = {Springer US},
      address = {Boston, MA},
        pages = {239-246},
     abstract = {Mahowald describes a silicon model of the computation performed by the first layer of visual processing, located in the outer plexiform layer of the retina. The lateral spread of information at the outer plexiform layer is mediated by a two-dimensional resistive network. The voltage at every point in the network represents a spatially weighted average of the photoreceptor inputs. The farther away an input is from a point in the network, the less weight it is given. The weighting function decreases in a generally exponential manner with distance.},
}
@article{MeansetalBJ-06,
       author = {Means, Shawn and Smith, Alexander J. and Shepherd, Jason and Shadid, John and Fowler, John and Wojcikiewicz, Richard J. H. and Mazel, Tomas and Smith, Gregory D. and Wilson, Bridget S.},
        title = {Reaction Diffusion Modeling of Calcium Dynamics with Realistic {ER} Geometry},
      journal = {Biophysics Journal},
    publisher = {Biophysical Society},
       volume = {91},
        issue = {2},
         year = {2006},
        pages = {537-557},
     abstract = {We describe a finite-element model of mast cell calcium dynamics that incorporates the endoplasmic reticulum's complex geometry. The model is built upon a three-dimensional reconstruction of the endoplasmic reticulum (ER) from an electron tomographic tilt series. Tetrahedral meshes provide volumetric representations of the ER lumen, ER membrane, cytoplasm, and plasma membrane. The reaction-diffusion model simultaneously tracks changes in cytoplasmic and ER intraluminal calcium concentrations and includes luminal and cytoplasmic protein buffers. Transport fluxes via PMCA, SERCA, ER leakage, and Type II IP(3) receptors are also represented. Unique features of the model include stochastic behavior of IP(3) receptor calcium channels and comparisons of channel open times when diffusely distributed or aggregated in clusters on the ER surface. Simulations show that IP(3)R channels in close proximity modulate activity of their neighbors through local Ca(2+) feedback effects. Cytoplasmic calcium levels rise higher, and ER luminal calcium concentrations drop lower, after IP(3)-mediated release from receptors in the diffuse configuration. Simulation results also suggest that the buffering capacity of the ER, and not restricted diffusion, is the predominant factor influencing average luminal calcium concentrations.},
}
@article{Medinaetal00,
       author = {A. Medina and I. Matta and J. Byers},
        title = {On the origin of power laws in {I}nternet topologies},
      journal = {Computer Communication Review},
       volume = {30:2},
        pages = {18-28},
         year = 2000,
}
@inproceedings{MedinandRips2015,
       author = {Douglas L. Medin and Lance J. Rips},
        title = {Memory, Meaning, and Metaphysics},
    booktitle = {Concepts: Core Readings},
       editor = {E. Margolis and S. Laurence},
    publisher = {MIT Press},
         year = {2013}
}
@inproceedings{MehrotraetalICIP-96,
       author = {S. Mehrotra and N. Chaddha and R.M. Gray},
        title = {Predictive Hierarchical Table-Lookup Vector Quantization with Quadtree Encoding},
    booktitle = {International Conference on Image Processing},
         year = 1996,
}
@article{MehtaandSchwabCoRR-14,
       author = {Pankaj Mehta, David J. Schwab},
        title = {An exact mapping between the Variational Renormalization Group and Deep Learning},
      journal = {CoRR},
       volume = {arXiv:1410.3831},
         year = {2014},
     abstract = {Deep learning is a broad set of techniques that uses multiple layers of representation to automatically learn relevant features directly from structured data. Recently, such techniques have yielded record-breaking results on a diverse set of difficult machine learning tasks in computer vision, speech recognition, and natural language processing. Despite the enormous success of deep learning, relatively little is understood theoretically about why these techniques are so successful at feature learning and compression. Here, we show that deep learning is intimately related to one of the most important and successful techniques in theoretical physics, the renormalization group (RG). RG is an iterative coarse-graining scheme that allows for the extraction of relevant features (i.e. operators) as a physical system is examined at different length scales. We construct an exact mapping from the variational renormalization group, first introduced by Kadanoff, and deep learning architectures based on Restricted Boltzmann Machines (RBMs). We illustrate these ideas using the nearest-neighbor Ising Model in one and two-dimensions. Our results suggests that deep learning algorithms may be employing a generalized RG-like scheme to learn relevant features from data.}
}
@article{MehtaandWilson00,
       author = {Mehta, M.R. and Wilson, M.A.},
        title = {From Hippocampus to {V1}: Effect of {LTP} on Spatio-Temporal Dynamics of ReceptiveFields},
      journal = {Neurocomputing},
       volume = 32,
        pages = {905-911},
         year = 2000,
}
@article{Mehtaetal00,
       author = {Mehta, M.R. and M.C. Quirk and M.A. Wilson},
        title = {Experience-dependent asymmetric shape of hippocampal receptive fields},
      journal = {Neuron},
       volume = 25,
        issue = 3,
        pages = {707-715},
         year = 2000,
}
@article{Mehtaetal02,
       author = {Mehta, M.R. and A.K. Lee and M.A. Wilson},
        title = {Role of experience and oscillations in transforming a rate code into a temporal code},
      journal = {Nature},
       volume = 417,
        issue = 6890,
        pages = {741-746},
         year = 2002,
}
@article{Mehtaetal97,
       author = {Mehta, M.R. and C.A. Barnes and B.L. McNaughton},
        title = {Experience-dependent, asymmetric expansion of hippocampal place fields},
      journal = {Proceedings National Academy of Science {(USA)}},
       volume = 94,
        issue = 16,
        pages = {8918-8921},
         year = 1997,
}
@article{MehtaetalNAS-97,
       author = {Mehta, M. and Barnes, C.A. and McNaughton, B.L.},
        title = {Experience-dependent, asymmetric expansion of hippocampal place fields},
      journal = {Proceedings of the National Academy of Science},
       volume = 94,
        issue = 16,
        pages = {8918-8921},
         year = 1997,
}
@techreport{MeilaandJaakkolaTR-00,
       author = {Marina Meil\u{a} and Tommi Jaakkola},
        title = {Tractable bayesian learning of tree belief networks},
        issue = {Technical report CMU-RI-TR-00-15},
  institution = {Carnegie Mellon University},
          url = {citeseer.ist.psu.edu/article/meila00tractable.html},
         year = 2000,
}
@article{MeilaandJordanJMLR-00,
       author = {Marina Meil\u{a} and Michael Jordan},
        title = {Learning with Mixtures of Trees},
      journal = {Journal of Machine Learning Research},
       volume = 1,
        pages = {1-48},
         year = 2000,
          url = {citeseer.ist.psu.edu/meila-predoviciu00learning.html}
}        
@techreport{MelekopoglouTR-90,
        title = {On the Complexity of the Policy Iteration Algorithm for Stochastic Games},
       author = {Mary Melekopoglou and Anne Condon},
  institution = {Computer Sciences Department, University of Wisconsin Madison},
      comment = {To appear in the ORSA Journal on Computing},
        issue = {CS-TR-90-941},
         year = 1990,
}
@article{MelroyetalSENSORS-17,
       author = {Melroy, Samantha and Bauer, Christopher and McHugh, Matthew and Carden, Garret and Stolin, Alexander and Majewski, Stan and Brefczynski-Lewis, Julie and Wuest, Thorsten},
        title = {Development and Design of Next-Generation Head-Mounted Ambulatory Microdose Positron-Emission Tomography (AM-PET) System},
      journal = {Sensors},
       volume = {17},
         year = {2017},
       number = {5},
       volume = {1164},
     abstract = {Several applications exist for a whole brain positron-emission tomography (PET) brain imager designed as a portable unit that can be worn on a patient's head. Enabled by improvements in detector technology, a lightweight, high performance device would allow PET brain imaging in different environments and during behavioral tasks. Such a wearable system that allows the subjects to move their heads and walk--the Ambulatory Microdose PET (AM-PET)--is currently under development. This imager will be helpful for testing subjects performing selected activities such as gestures, virtual reality activities and walking. The need for this type of lightweight mobile device has led to the construction of a proof of concept portable head-worn unit that uses twelve silicon photomultiplier (SiPM) PET module sensors built into a small ring which fits around the head. This paper is focused on the engineering design of mechanical support aspects of the AM-PET project, both of the current device as well as of the coming next-generation devices. The goal of this work is to optimize design of the scanner and its mechanics to improve comfort for the subject by reducing the effect of weight, and to enable diversification of its applications amongst different research activities.},
}
@book{Menand01,
       author = {Louis Menand},
        title = {The Metaphysical Club: A Story of Ideas in America},
    publisher = {Farrar, Straus \& Giroux},
      address = {New York, NY},
         year = 2001,
}
@techreport{Menczer01,
      author = {F. Menczer},
       title = {Links tell us about lexical and semantic {W}eb content},
       issue = {Computer Science Abstract CS.IR/0108004},
 institution = {University of Iowa, Department of Management Science},
     address = {Iowa City, IA},
       month = {August},
        year = 2001,
}
@article{MenczerNAS-02,
   author = {Menczer, F.},
    title = {Growing and navigating the {S}mall {W}orld {W}eb by local content},
  journal = {Proceedings of the National Academy of Science},
   volume = 99,
    pages = {14014-14019},
      url = {http://www.pnas.org/cgi/content/abstract/99/22/14014},
     year = 2002,
}
@book{Mendelson79,
   author =    {Mendelson, Elliot},
   title =     {Introduction to Mathematical Logic},
   publisher = {D. Van Nostrand},
   address =   {New York},
   year =      1979,
}
@book{Mendenhall68,
   author =    {Mendenhall, W.},
   title =     {Introduction to Linear Models and the Design and Analysis of
                Experiments},
   publisher = {Wadsworth},
   address =   {Belmont, California},
   year =      1968,
}
@article{MenegasetalELIFE-17,
        title = {Opposite initialization to novel cues in dopamine signaling in ventral and posterior striatum in mice},
       author = {Menegas, William and Babayan, Benedicte M and Uchida, Naoshige and Watabe-Uchida, Mitsuko},
      journal = {Elife},
       volume = {6},
        pages = {e21886},
         year = {2017},
    publisher = {eLife Sciences Publications Limited}
}
@article{MenendezetalCELL-13,
    author = {Menendez, J. A. and Joven, J. and Cufi, S. and Corominas-Faja, B. and Oliveras-Ferraros, C. and Cuyas, E. and Martin-Castillo, B. and Lopez-Bonet, E. and Alarcon, T. and Vazquez-Martin, A.},
     title = {{T}he {W}arburg effect version 2.0: metabolic reprogramming of cancer stem cells},
   journal = {Cell Cycle},
      year = 2013,
    volume = 12,
     issue = 8,
     pages = {1166-1179},
   article = {When fighting cancer, knowledge on metabolism has always been important. Today, it matters more than ever. The restricted cataloging of cancer genomes is quite unlikely to achieve the task of curing cancer, unless it is integrated into metabolic networks that respond to and influence the constantly evolving cancer stem cell (CSC) cellular states. Once the genomic era of carcinogenesis had pushed the 1920s Otto Warburg's metabolic cancer hypothesis into obscurity for decades, the most recent studies begin to support a new developing paradigm, in which the molecular logic behind the conversion of non-CSCs into CSCs can be better understood in terms of the "metabolic facilitators" and "metabolic impediments" that operate as proximate openings and roadblocks, respectively, for the transcriptional events and signal transduction programs that ultimately orchestrate the intrinsic and/or microenvironmental paths to CSC cellular states. Here we propose that a profound understanding of how human carcinomas install a proper "Warburg effect version 2.0" allowing them to "run" the CSCs' "software" programs should guide a new era of metabolo-genomic-personalized cancer medicine. By viewing metabolic reprogramming of CSCs as an essential characteristic that allows dynamic, multidimensional and evolving cancer populations to compete successfully for their expansion on the organism, we now argue that CSCs bioenergetics might be another cancer hallmark. A definitive understanding of metabolic reprogramming in CSCs may complement or to some extent replace, the 30-year-old paradigm of targeting oncogenes to treat human carcinomas, because it can be possible to metabolically create non-permissive or "hostile" metabotypes to prevent the occurrence of CSC cellular states with tumor- and metastasis-initiating capacity.}
}
@article{MenonPBR-16,
       author = {Menon, Vinod},
        title = {Memory and cognitive control circuits in mathematical cognition and learning},
      journal = {Progress in Brain Research},
       volume = {227},
         year = {2016},
        pages = {159-186},
     abstract = {Numerical cognition relies on interactions within and between multiple functional brain systems, including those subserving quantity processing, working memory, declarative memory, and cognitive control. This chapter describes recent advances in our understanding of memory and control circuits in mathematical cognition and learning. The working memory system involves multiple parietal-frontal circuits which create short-term representations that allow manipulation of discrete quantities over several seconds. In contrast, hippocampal-frontal circuits underlying the declarative memory system play an important role in formation of associative memories and binding of new and old information, leading to the formation of long-term memories that allow generalization beyond individual problem attributes. The flow of information across these systems is regulated by flexible cognitive control systems which facilitate the integration and manipulation of quantity and mnemonic information. The implications of recent research for formulating a more comprehensive systems neuroscience view of the neural basis of mathematical learning and knowledge acquisition in both children and adults are discussed.},
}
@article{MenonandKimTCS-99,
    author = {Ravi S. Menon and Seong-Gi Kim},
     title = {Spatial and temporal limits in cognitive neuroimaging with {fMRI}},
   journal = {Trends in Cognitive Sciences},
    volume = 3,
     issue = 6,
     pages = {207-216},
      year = 1999,
}
@article{MereletalCoRR-19,
        title = {Deep neuroethology of a virtual rodent},
       author = {Josh Merel and Diego Aldarondo and Jesse Marshall and Yuval Tassa and Greg Wayne and Bence {\"{O}}lveczky},
         year = {2019},
      journal = {CoRR},
       volume = {arXiv:1911.09451},
         year = {2019},
     abstract = {Parallel developments in neuroscience and deep learning have led to mutually productive exchanges, pushing our understanding of real and artificial neural networks in sensory and cognitive systems. However, this interaction between fields is less developed in the study of motor control. In this work, we develop a virtual rodent as a platform for the grounded study of motor activity in artificial models of embodied control. We then use this platform to study motor activity across contexts by training a model to solve four complex tasks. Using methods familiar to neuroscientists, we describe the behavioral representations and algorithms employed by different layers of the network using a neuroethological approach to characterize motor activity relative to the rodent's behavior and goals. We find that the model uses two classes of representations which respectively encode the task-specific behavioral strategies and task-invariant behavioral kinematics. These representations are reflected in the sequential activity and population dynamics of neural subpopulations. Overall, the virtual rodent facilitates grounded collaborations between deep reinforcement learning and motor neuroscience.}
}        
@inproceedings{MereletalICLR-19,
        title = {Neural Probabilistic Motor Primitives for Humanoid Control},
       author = {Josh Merel and Leonard Hasenclever and Alexandre Galashov and Arun Ahuja and Vu Pham and Greg Wayne and Yee Whye Teh and Nicolas Heess},
    booktitle = {International Conference on Learning Representations},
         year = {2019},
     abstract = {We focus on the problem of learning a single motor module that can flexibly express a range of behaviors for the control of high-dimensional physically simulated humanoids. To do this, we propose a motor architecture that has the general structure of an inverse model with a latent-variable bottleneck. We show that it is possible to train this model entirely offline to compress thousands of expert policies and learn a motor primitive embedding space. The trained neural probabilistic motor primitive system can perform one-shot imitation of whole-body humanoid behaviors, robustly mimicking unseen trajectories. Additionally, we demonstrate that it is also straightforward to train controllers to reuse the learned motor primitive space to solve tasks, and the resulting movements are relatively naturalistic.}
}
@article{MereletalJNM-16,
       author = {Merel, J. and Shababo, B. and Naka, A. and Adesnik, H. and Paninski, L.},
        title = {Bayesian methods for event analysis of intracellular currents},
      journal = {Journal Neuroscience Methods},
         year = {2016},
     abstract = {Investigation of neural circuit functioning often requires statistical interpretation of events in subthreshold electrophysiological recordings. This problem is non-trivial because recordings may have moderate levels of structured noise and events may have distinct kinetics. In addition, novel experimental designs that combine optical and electrophysiological methods will depend upon statistical tools that combine multimodal data.\\ We present a Bayesian approach for inferring the timing, strength, and kinetics of post-synaptic currents (PSCs) from voltage-clamp electrophysiological recordings on a per event basis. The simple generative model for a single voltage-clamp recording flexibly extends to include additional structure to enable experiments designed to probe synaptic connectivity.\\ We validate the approach on simulated and real data. We also demonstrate that extensions of the basic PSC detection algorithm can handle recordings contaminated with optically evoked currents, and we simulate a scenario in which calcium imaging observations, available for a subset of neurons, can be fused with electrophysiological data to achieve higher temporal resolution.\\ We apply this approach to simulated and real ground truth data to demonstrate its higher sensitivity in detecting small signal-to-noise events and its increased robustness to noise compared to standard methods for detecting PSCs.\\ The new Bayesian event analysis approach for electrophysiological recordings should allow for better estimation of physiological parameters under more variable conditions and help support new experimental designs for circuit mapping.}
}        
@article{MereletalNC-19,
       author = {Merel, Josh and Botvinick, Matt and Wayne, Greg},
        title = {{Hierarchical motor control in mammals and machines}},
      journal = {Nature Communications},
       volume = {10},
       number = {1},
         year = {2019},
        pages = {5489},
     abstract = {Advances in artificial intelligence are stimulating interest in neuroscience. However, most attention is given to discrete tasks with simple action spaces, such as board games and classic video games. Less discussed in neuroscience are parallel advances in "synthetic motor control". While motor neuroscience has recently focused on optimization of single, simple movements, AI has progressed to the generation of rich, diverse motor behaviors across multiple tasks, at humanoid scale. It is becoming clear that specific, well-motivated hierarchical design elements repeatedly arise when engineering these flexible control systems. We review these core principles of hierarchical control, relate them to hierarchy in the nervous system, and highlight research themes that we anticipate will be critical in solving challenges at this disciplinary intersection.}
}
@article{MerityetalCoRR-16,
       author = {Stephen Merity and Caiming Xiong and James Bradbury and Richard Socher},
        title = {Pointer Sentinel Mixture Models},
      journal = {CoRR},
       volume = {arXiv:1609.07843},
         year = {2016},
     abstract = {Recent neural network sequence models with softmax classifiers have achieved their best language modeling performance only with very large hidden states and large vocabularies. Even then they struggle to predict rare or unseen words even if the context makes the prediction unambiguous. We introduce the pointer sentinel mixture architecture for neural sequence models which has the ability to either reproduce a word from the recent context or produce a word from a standard softmax classifier. Our pointer sentinel-LSTM model achieves state of the art language modeling performance on the Penn Treebank (70.9 perplexity) while using far fewer parameters than a standard softmax LSTM. In order to evaluate how well language models can exploit longer contexts and deal with more realistic vocabularies and larger corpora we also introduce the freely available WikiText corpus.},
}
@article{MerkerBBS-07,
       author = {Merker, Bjorn},
        title = {Consciousness without a cerebral cortex: A challenge for neuroscience and medicine},
      journal = {Behavioral and Brain Sciences},
    publisher = {Cambridge University Press},
       volume = {30},
       number = {1},
         year = {2007},
        pages = {63-81},
     abstract = {A broad range of evidence regarding the functional organization of the vertebrate brain - spanning from comparative neurology to experimental psychology and neurophysiology to clinical data - is reviewed for its bearing on conceptions of the neural organization of consciousness. A novel principle relating target selection, action selection, and motivation to one another, as a means to optimize integration for action in real time, is introduced. With its help, the principal macrosystems of the vertebrate brain can be seen to form a centralized functional design in which an upper brain stem system organized for conscious function performs a penultimate step in action control. This upper brain stem system retained a key role throughout the evolutionary process by which an expanding forebrain - culminating in the cerebral cortex of mammals - came to serve as a medium for the elaboration of conscious contents. This highly conserved upper brainstem system, which extends from the roof of the midbrain to the basal diencephalon, integrates the massively parallel and distributed information capacity of the cerebral hemispheres into the limited-capacity, sequential mode of operation required for coherent behavior. It maintains special connective relations with cortical territories implicated in attentional and conscious functions, but is not rendered nonfunctional in the absence of cortical input. This helps explain the purposive, goal-directed behavior exhibited by mammals after experimental decortication, as well as the evidence that children born without a cortex are conscious. Taken together these circumstances suggest that brainstem mechanisms are integral to the constitution of the conscious state, and that an adequate account of neural mechanisms of conscious function cannot be confined to the thalamocortical complex alone.}
}
@article{MerklandMiddendorfEC-02,
       author = {Merkle, Daniel and Middendorf, Martin},
        title = {Modeling the Dynamics of Ant Colony Optimization},
      journal = {Evolutionary Computation},
    publisher = {MIT Press},
       volume = {10},
        issue = {3},
         year = {2002},
        pages = {235-262},
     abstract = {The dynamics of Ant Colony Optimization (ACO) algorithms is studied using a deterministic model that assumes an average expected behavior of the algorithms. The ACO optimization metaheuristic is an iterative approach, where in every iteration, artificial ants construct solutions randomly but guided by pheromone information stemming from former ants that found good solutions. The behavior of ACO algorithms and the ACO model are analyzed for certain types of permutation problems. It is shown analytically that the decisions of an ant are influenced in an intriguing way by the use of the pheromone information and the properties of the pheromone matrix. This explains why ACO algorithms can show a complex dynamic behavior even when there is only one ant per iteration and no competition occurs. The ACO model is used to describe the algorithm behavior as a combination of situations with different degrees of competition between the ants. This helps to better understand the dynamics of the algorithm when there are several ants per iteration as is always the case when using ACO algorithms for optimization. Simulations are done to compare the behavior of the ACO model with the ACO algorithm. Results show that the deterministic model describes essential features of the dynamics of ACO algorithms quite accurately, while other aspects of the algorithms behavior cannot be found in the model.},
}
@article{MerkleNANOTECHNOLOGY-93,
       author = {Ralph C. Merkle},
        title = {Two types of mechanical reversible logic},
      journal = {Nanotechnology},
       volume = 4,
        issue = 2,
         year = 1993,
        pages = {114-131},
     abstract = {The author describes two types of mechanical reversible logic which eliminate sliding contact. In the first type, contact between parts, when it does occur, involves only pressure. In the second, all contact between parts is eliminated. The entire computation could in principle be performed by a single block of complexly shaped oscillating material. The state of the computation is stored in the elastic deformations of this single block. If the material is perfectly elastic, the resulting computer will dissipate no energy. Real material moving sufficiently slowly can approximate a perfectly elastic material as closely as desired. In real materials, energy dissipation occurs during successive cycles of compression and decompression of the material much as energy dissipation occurs when a gas is compressed by a piston. However, at sufficiently low temperatures, this mechanism of energy dissipation might no longer be operative. It is not impossible that energy dissipation can be made to fall exponentially as both temperature and speed of operation are reduced.}
}
@article{MerleretalTMM-12,
     title = {Semantic Model Vectors for Complex Video Event Recognition},
    author = {Michele Merler and Bert Huang and Lexing Xie and Gang Hua and Apostol Natsev},
   journal = {{IEEE} Transactions on Multimedia},
    volume = 14,
     issue = 1,
      year = 2012,
     pages = {88-101},
}
@article{MerollaetalSCIENCE-14,
       author = {Merolla, Paul A. and Arthur, John V. and Alvarez-Icaza, Rodrigo and Cassidy, Andrew S. and Sawada, Jun and Akopyan, Filipp and Jackson, Bryan L. and Imam, Nabil and Guo, Chen and Nakamura, Yutaka and Brezzo, Bernard and Vo, Ivan and Esser, Steven K. and Appuswamy, Rathinakumar and Taba, Brian and Amir, Arnon and Flickner, Myron D. and Risk, William P. and Manohar, Rajit and Modha, Dharmendra S.},
        title = {A million spiking-neuron integrated circuit with a scalable communication network and interface},
      journal = {Science},
       volume = 345,
        issue = 6197,
        pages = {668-673},
         year = 2014,
     abstract = {Inspired by the brain's structure, we have developed an efficient, scalable, and flexible non-von Neumann architecture that leverages contemporary silicon technology. To demonstrate, we built a 5.4-billion-transistor chip with 4096 neurosynaptic cores interconnected via an intrachip network that integrates 1 million programmable spiking neurons and 256 million configurable synapses. Chips can be tiled in two dimensions via an interchip communication interface, seamlessly scaling the architecture to a cortexlike sheet of arbitrary size. The architecture is well suited to many applications that use complex neural networks in real time, for example, multiobject detection and classification. With 400-pixel-by-240-pixel video input at 30 frames per second, the chip consumes 63 milliwatts.},
}
@inproceedings{MesniletalINTERSPEECH-13,
     author = {Mesnil, Gr{\'{e}}goire and He, Xiaodong and Deng, Li and Bengio, Yoshua},
      title = {Investigation of Recurrent-Neural-Network Architectures and Learning Methods for Spoken Language Understanding},
  booktitle = {Interspeech},
       year = 2013,
   abstract = {One of the key problems in spoken language understanding (SLU) is the task of slot filling. In light of the recent success of applying deep neural network technologies in domain detection and intent identification, we carried out an in-depth investigation on the use of recurrent neural networks for the more difficult task of slot filling involving sequence discrimination. In this work, we implemented and compared several important recurrent-neural-network architectures, including the Elman-type and Jordan-type recurrent networks and their variants. To make the results easy to reproduce and compare, we implemented these networks on the common Theano neural network toolkit, and evaluated them on the ATIS benchmark. We also compared our results to a conditional random fields (CRF) baseline. Our results show that on this task, both types of recurrent networks outperform the CRF baseline substantially, and a bi-directional Jordan-type network that takes into account both past and future dependencies among slots works best, outperforming a CRF-based baseline by 14\% in relative error reduction.}
}
@inproceedings{MessingetalICCV-09,
      title = {Activity recognition using the velocity histories of tracked keypoints},
     author = {R. Messing and C. Pal and H. Kautz},
  booktitle = {Proceedings of the International Conference on Computer Vision}, 
  publisher = {IEEE Computer Society},
       year = 2009, 
}
@article{MessomandBarczakPDCAT-08,
     author = {Chris Messom and Andre Barczak},
      title = {Stream Processing of Integral Images for Real-Time Object Detection},
    journal = {International Conference on Parallel and Distributed Computing Applications and Technologies},
      pages = {405-412},
  publisher = {IEEE Computer Society},
       year = 2008,
}
@article{MetinandFrostPNAS-89,
    author = {M\'{e}tin, C. and Frost, D.O.}, 
     title = {Visual responses of neurons in somatosensory cortex of hamsters with experimentally induced retinal projections to somatosensory thalamus}, 
   journal = {Proceedings of the National Academy of Sciences},
    volume = 86, 
     issue = 1, 
      year = 1989, 
     pages = {357-361}, 
  abstract = {These experiments investigate the capacity of thalamic and
  cortical structures in a sensory system to process information of a
  modality normally associated with another system. Retinal ganglion
  cells in newborn Syrian hamsters were made to project permanently to
  the main thalamic somatosensory (ventrobasal) nucleus. When the
  animals were adults, single unit recordings were made in the
  somatosensory cortices, the principal targets of the ventrobasal
  nucleus. The somatosensory neurons responded to visual stimulation
  of distinct receptive fields, and their response properties
  resembled, in several characteristic features, those of normal
  visual cortical neurons. In the visual cortex of normal animals and
  the somatosensory cortex of operated animals, the same functional
  categories of neurons occurred in similar proportions, and the
  neurons' selectivity for the orientation or direction of movement of
  visual stimuli was comparable. These results suggest that thalamic
  nuclei or cortical areas at corresponding levels in the visual and
  somatosensory pathways perform similar transformations on their inputs.}, 
}
@article{Metropolisetal53,
       author = {Nicholas Metropolis and Arianna W. Rosenbluth and Marshall N. Rosenbluth and Augusta H. Teller and Edward Teller},
        title = {Equation of State Calculations by Fast Computing Machines},
      journal = {The Journal of Chemical Physics},
       volume = 21,
        issue = 6,
         year = 1953,
        pages = {1087-1092},
}
@article{Metzler-BaddeleyetalJoN-11,
       author = {Metzler-Baddeley, Claudia and Jones, Derek K. and Belaroussi, Boubakeur and Aggleton, John P. and O{\textquoteright}Sullivan, Michael J.},
        title = {Frontotemporal Connections in Episodic Memory and Aging: A Diffusion MRI Tractography Study},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {31},
       number = {37},
        pages = {13236-13245},
         year = {2011},
     abstract = {Human episodic memory is supported by networks of white matter tracts that connect frontal, temporal, and parietal regions. Degradation of white matter microstructure is increasingly recognized as a general mechanism of cognitive deterioration with aging. However, atrophy of gray matter regions also occurs and, to date, the potential role of specific white matter connections has been largely ignored. Changes to frontotemporal tracts may be important for the decline of episodic memory; while frontotemporal cooperation is known to be critical, the precise pathways of interaction are unknown. Diffusion-weighted MRI tractography was used to reconstruct three candidate fasciculi known to link components of memory networks: the fornix, the parahippocampal cingulum, and the uncinate fasciculus. Age-related changes in the microstructure of these tracts were investigated in 40 healthy older adults between the ages of 53 and 93 years. The relationships between aging, microstructure, and episodic memory were assessed for each individual tract. Age-related reductions of mean fractional anisotropy and/or increased mean diffusivity were found in all three tracts. However, age-related decline in recall was specifically associated with degradation of fornix microstructure, consistent with the view that this tract is important for episodic memory. In contrast, a decline in uncinate fasciculus microstructure was linked to impaired error monitoring in a visual object{\textendash}location association task, echoing the effects of uncinate transection in monkeys. These results suggest that degradation of microstructure in the fornix and the uncinate fasciculus make critical but differential contributions to the mechanisms underlying age-related cognitive decline and subserve distinct components of memory.},
}
@inproceedings{Meuleau99,
       author = {Nicolas Meuleau},
        title = {The Importance of Impossible Trajectories in the {VAPS} Algorithm},
    booktitle = {Preparation},
         year = 1999,
}
@inproceedings{MeuleauetalAAAI-98,
       author = {Nicolas Meuleau and Craig Boutilier and Milos Hauskrecht and Leslie Kaelbling and Kee-Eung Kim and Leonid Peshkin and Thomas Dean},
        title = {Solving very large weakly coupled {M}arkov decision processes},
    booktitle = {Proceedings {AAAI}-98},
     location = {Madison, Wisconsin},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
 organization = {AAAI},
        pages = {165-172},
         year = 1998,
}
@inproceedings{MeuleauetalCUAI-99a,
       author = {Nicolas Meuleau and Kee-Eung Kim and Leslie Kaelbling and Anthony R. Cassandra},
        title = {Solving {POMDP}s by searching in the space of finite policies},
    booktitle = {Proceedings of the Fifteenth Conference on Uncertainty in Artificial Intelligence},
     location = {Stockholm, Sweden},
         year = 1999,
}
@inproceedings{MeuleauetalCUAI-99b,
       author = {Nicolas Meuleau and Leonid Peshkin and Kee-Eung Kim and Leslie Kaelbling},
        title = {Learning Finite-State Controllers for Partially Observable Environments},
    booktitle = {Proceedings of the Fifteenth Conference on Uncertainty in Artificial Intelligence},
     location = {Stockholm, Sweden},
         year = 1999,
}
@article{MeyersetalCoRR-19,
       author = {Richard Meyes and Melanie Lu and Constantin Waubert de Puiseau and Tobias Meisen},
        title = {Ablation Studies in Artificial Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1901.08644},
         year = {2019},
     abstract = {Ablation studies have been widely used in the field of neuroscience to tackle complex biological systems such as the extensively studied Drosophila central nervous system, the vertebrate brain and more interestingly and most delicately, the human brain. In the past, these kinds of studies were utilized to uncover structure and organization in the brain, i.e. a mapping of features inherent to external stimuli onto different areas of the neocortex. considering the growth in size and complexity of state-of-the-art artificial neural networks (ANNs) and the corresponding growth in complexity of the tasks that are tackled by these networks, the question arises whether ablation studies may be used to investigate these networks for a similar organization of their inner representations. In this paper, we address this question and performed two ablation studies in two fundamentally different ANNs to investigate their inner representations of two well-known benchmark datasets from the computer vision domain. We found that features distinct to the local and global structure of the data are selectively represented in specific parts of the network. Furthermore, some of these representations are redundant, awarding the network a certain robustness to structural damages. We further determined the importance of specific parts of the network for the classification task solely based on the weight structure of single units. Finally, we examined the ability of damaged networks to recover from the consequences of ablations by means of recovery training. We argue that ablations studies are a feasible method to investigate knowledge representations in ANNs and are especially helpful to examine a networks robustness to structural damages, a feature of ANNs that will become increasingly important for future safety-critical applications.}
}
@inproceedings{MeystelIC-85,
       author = {Meystel, A.},
        title = {Intelligent Control: Issues and Perspectives},
    booktitle = {Proceedings of the {IEEE} Workshop on Intelligent Control},
     location = {Troy, New York},
 organization = {IEEE},
        pages = {1-15},
         year = 1985,
}
@book{MichalskietalML-83,
       editor = {Michalski, R. S. and Carbonell, J. G. and Mitchell, T. M.},
        title = {Machine Learning: an Artificial Intelligence Approach},
    booktitle = {Machine Learning: an Artificial Intelligence Approach},
    publisher = {Tioga},
         year = 1983,
}
@article{MicheletalSCIENCE-10,
       author = {Michel, Jean-Baptiste and Shen, Yuan Kui and Aiden, Aviva P. and Veres, Adrian and Gray, Matthew K. and The Google Books Team and Pickett, Joseph P. and  Hoiberg, Dale and Clancy, Dan and Norvig, Peter and  Orwant, Jon and Pinker, Steven and Nowak, Martin A. and Aiden, Erez Lieberman},
        title = {Quantitative Analysis of Culture Using Millions of Digitized Books},
      journal = {Science},
         year = 2010,
     abstract = {We constructed a corpus of digitized texts containing about 4\% of all books ever printed. Analysis of this corpus enables us to investigate cultural trends quantitatively. We survey the vast terrain of "culturomics", focusing on linguistic and cultural phenomena that were reflected in the English language between 1800 and 2000. We show how this approach can provide insights about fields as diverse as lexicography, the evolution of grammar, collective memory, the adoption of technology, the pursuit of fame, censorship, and historical epidemiology. "Culturomics" extends the boundaries of rigorous quantitative inquiry to a wide array of new phenomena spanning the social sciences and the humanities.},
}
@article{MichevaandSmithNEURON-07,
      title = {Array Tomography: A New Tool for Imaging the Molecular Architecture and Ultrastructure of Neural Circuits},
     author = {Micheva, Kristina D. and Smith, Stephen J.},
    journal = {Neuron},
  publisher = {Cell Press},
     volume = 55,
      issue = 1,
       year = 2007,
      pages = {25-36}
}
@article{MichevaetalNEURON-10,
     title = {Single-synapse analysis of a diverse synapse population: Proteomic imaging methods and markers},
    author = {Micheva, K.D. and Busse, B.L. and Weiler, N.C. and O'Rourke, N. and  Smith, S.J.},
   journal = {Neuron},
    volume = 68,
     issue = 4,
      year = 2010,
     pages = {639-653},
}
@incollection{MichieetalKBSIC-90,
       author = {D. Michie and M. Bain and J. Hayes-Michie},
       editor = {M. Grimble and S. McGhee and P. Mowforth},
        title = {Cognitive models from subcognitive skills},
    booktitle = {Knowledge-based Systems for Industrial Control},
       series = {Control, Robotics \& Sensors},
    publisher = {Institution of Engineering and Technology},
        pages = {71-99},
         year = {1990},
} 
@inproceedings{MihailFOCS-89,
       author = {Mihail, Milena},
        title = {Conductance and Convergence of {M}arkov Chains: A Combinatorial Treatment of Expanders},
    booktitle = {Proceedings of the Thirty First Annual Symposium on Foundations of Computer Science},
        pages = {526-531},
         year = 1989,
}
@article{MikolajczykandSchmidIJCV-04,
     author = {K. Mikolajczyk and C. Schmid},
      title = {Scale and Affine invariant interest point detectors},
    journal = {International Journal of Computer Vision},
       year = 2004,
     volume = 60,
      issue = 2,
      pages = {63-86},
}
@article{MikolajczykandSchmidPAMI-05,
     author = {Krystian Mikolajczyk and Cordelia Schmid},
      title = {A performance evaluation of local descriptors},
    journal = {IEEE Transactions on Pattern Analysis \& Machine Intelligence},
      issue = 10,
     volume = 27,
      pages = {1615-1630},
       year = 2005,
}
@inproceedings{MikolovandZweigSLT-12, 
     author = {Mikolov, T. and Zweig, G.}, 
  booktitle = {Spoken Language Technology Workshop (SLT), 2012 IEEE}, 
      title = {Context dependent recurrent neural network language model}, 
       year = 2012, 
      pages = {234-239}, 
   abstract = {Recurrent neural network language models (RNNLMs) have recently demonstrated state-of-the-art performance across a variety of tasks. In this paper, we improve their performance by providing a contextual real-valued input vector in association with each word. This vector is used to convey contextual information about the sentence being modeled. By performing Latent Dirichlet Allocation using a block of preceding text, we achieve a topic-conditioned RNNLM. This approach has the key advantage of avoiding the data fragmentation associated with building multiple topic models on different data subsets. We report perplexity results on the Penn Treebank data, where we achieve a new state-of-the-art. We further apply the model to the Wall Street Journal speech recognition task, where we observe improvements in word-error-rate.}, 
}
@inproceedings{MikolovetalASRU-11,
       author = {Tom\`{a}s Mikolov and Stefan Kombrink and Anoop Deoras and Lukas Burget and Jan Cernocky},
        title = {{RNNLM} - Recurrent Neural Network Language Modeling Toolkit},
    booktitle = {Proceedings of ASRU 2011},
    publisher = {{IEEE} Signal Processing Society},
         year = {2011},
        pages = {1-4},
}
@article{MikolovetalCoRR-13a,
     title = {Efficient Estimation of Word Representations in Vector Space},
    author = {Mikolov, Tom\`{a}s and Chen, Kai and Corrado, Greg and Dean, Jeffrey},
   journal = {CoRR},
    volume = {arXiv:1301.3781},
      year = 2013,
       URL = {http://arxiv.org/abs/1301.3781},
}
@article{MikolovetalCoRR-13b,
    author = {Tom\`{a}s Mikolov and Ilya Sutskever and Kai Chen and Greg Corrado and Jeffrey Dean},
     title = {Distributed Representations of Words and Phrases and their Compositionality},
   journal = {CoRR},
    volume = {arXiv:1310.4546},
      year = 2013,
       URL = {http://arxiv.org/abs/1310.4546},
}
@article{MikolovetalCoRR-13c,
       author = {Tom\`{a}s Mikolov and Quoc V. Le and Ilya Sutskever},
        title = {Exploiting Similarities among Languages for Machine Translation},
      journal = {CoRR},
       volume = {arXiv:1309.4168},
         year = 2013,
}
@article{MikolovetalCoRR-15,
       author = {Tomas Mikolov and Armand Joulin and Marco Baroni},
        title = {A Roadmap towards Machine Intelligence},
      journal = {CoRR},
       volume = {abs/1511.08130},
         year = {2015},
}
@inproceedings{MikolovetalHLT-13,
       author = {Tom\`{a}s Mikolov and Wen-tau Yih and Geoffrey Zweig},
        title = {Linguistic Regularities in Continuous Space Word Representations},
    booktitle = {Proceedings of the 2013 Conference of the North American Chapter of the
                Association for Computational Linguistics: Human Language Technologies},
    publisher = {Association for Computational Linguistics},
         year = 2013,
     abstract = {Continuous space language models have recently demonstrated outstanding results across a variety of tasks. In this paper, we examine the vector-space word representations that are implicitly learned by the input-layer weights. We find that these representations are surprisingly good at capturing syntactic and semantic regularities in language, and that each relationship is characterized by a relation-specific vector offset. This allows vector-oriented reasoning based on the offsets between words. For example, the male/female relationship is automatically learned, and with the induced vector representations, "King - Man + Woman" results in a vector very close to "Queen." We demonstrate that the word vectors capture syntactic regularities by means of syntactic analogy questions (provided with this paper), and are able to correctly answer almost 40{\%} of the questions. We demonstrate that the word vectors capture semantic regularities by using the vector offset method to answer SemEval-2012 Task 2 questions. Remarkably, this method outperforms the best previous systems.},
}
@inproceedings{MikolovetalNIPS-13,
    author = {Tom\`{a}s Mikolov and Ilya Sutskever and Kai Chen and Greg Corrado and Jeffrey Dean},
     title = {Distributed Representations of Words and Phrases and their Compositionality},
 booktitle = {Advances in Neural Information Processing Systems 26},
      year = 2013,
     pages = {3111-3119},
}
@techreport{MikolovetalTR-12,
       author = {Tom\'{a}s Mikolov and Ilya Sutskever and Anoop Deoras and Hai-Son Le and Stefan Kombrink and Jan Cernocky},
        title = {Subword Language Modeling With Neural Networks},
  institution = {Brno University of Technology, Brno, Czech Republic},
         year = 2012,
          url = {http://www.fit.vutbr.cz/~imikolov/rnnlm/char.pdf},
}
@article{MikulaandDenkNATURE-METHODS-15,
       author = {Mikula, Shawn and Denk, Winfried},
        title = {High-resolution whole-brain staining for electron microscopic circuit reconstruction},
      journal = {Nature Methods},
         year = 2015,
    publisher = {Nature Publishing Group},
     abstract = {Currently only electron microscopy provides the resolution necessary to reconstruct neuronal circuits completely and with single-synapse resolution. Because almost all behaviors rely on neural computations widely distributed throughout the brain, a reconstruction of brain-wide circuits[mdash]and, ultimately, the entire brain[mdash]is highly desirable. However, these reconstructions require the undivided brain to be prepared for electron microscopic observation. Here we describe a preparation, BROPA (brain-wide reduced-osmium staining with pyrogallol-mediated amplification), that results in the preservation and staining of ultrastructural details throughout the brain at a resolution necessary for tracing neuronal processes and identifying synaptic contacts between them. Using serial block-face electron microscopy (SBEM), we tested human annotator ability to follow neural "wires" reliably and over long distances as well as the ability to detect synaptic contacts. Our results suggest that the BROPA method can produce a preparation suitable for the reconstruction of neural circuits spanning an entire mouse brain.},
}
@article{MikulaetalNATURE-12,
        title = {Staining and embedding the whole mouse brain for electron microscopy},
       author = {Mikula, Shawn and Binding, Jonas and Denk, Winfried},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = 9,
         year = 2012,
        pages = {1198-1201},
}
@book{Mill79,
       author = {John Stuart Mill},
        title = {Utilitarianism},
    publisher = {Reprinted 1979, Hackett Publishing},
      address = {Indianapolis, IN},
         year = 1863,
}
@article{MillerCOIN-16,
       author = {Paul Miller},
        title = {Itinerancy between attractor states in neural systems},
      journal = {Current Opinion in Neurobiology},
       volume = {40},
         year = {2016},
        pages = {14-22},
     abstract = {Converging evidence from neural, perceptual and simulated data suggests that discrete attractor states form within neural circuits through learning and development. External stimuli may bias neural activity to one attractor state or cause activity to transition between several discrete states. Evidence for such transitions, whose timing can vary across trials, is best accrued through analyses that avoid any trial-averaging of data. One such method, hidden Markov modeling, has been effective in this context, revealing state transitions in many neural circuits during many tasks. Concurrently, modeling efforts have revealed computational benefits of stimulus processing via transitions between attractor states. This review describes the current state of the field, with comments on how its perceived limitations have been addressed.}
}
@article{MillerPSYCHOLOGICAL-REVIEW-56,
       author = {George A. Miller},
        title = {The Magical Number Seven, Plus or Minus Two: {S}ome Limits on Our Capacity for Processing Information},
      journal = {Psychological Review},
       volume = {63},
       number = {2},
         year = {1956},
        pages = {81-97},
}
@techreport{MillerPhD,
       author = {Miller, David P.},
        title = {Planning by Search Through Simulations},
        issue = 423,
  institution = {Yale University Computer Science Department},
         year = 1985,
}
@inproceedings{MillerRA-84,
       author = {Miller, David P.},
        title = {A Spatial Representation System for Mobile Robots},
    booktitle = {Proceedings of the {IEEE} International Conference on Robotics and Automation},
     location = {Washington, DC},
         year = 1984,
        pages = {122-128},
}
@techreport{MillerTR-83,
       author = {Miller, David P.},
        title = {Scheduling Heuristics for Problem Solvers},
        issue = 264,
  institution = {Yale University Computer Science Department},
         year = 1983,
}
@article{MillerandSahayNATURE-NEUROSCIENCE-19,
       author = {Miller, Samara M. and Sahay, Amar},
        title = {Functions of adult-born neurons in hippocampal memory interference and indexing},
      journal = {Nature Neuroscience},
         year = {2019},
     abstract = {The dentate gyrus-CA3 circuit of the hippocampus is continuously modified by the integration of adult-born dentate granule cells (abDGCs). All abDGCs undergo a prolonged period of maturation, during which they exhibit heightened synaptic plasticity and refinement of electrophysiological properties and connectivity. Consistent with theoretical models and the known functions of the dentate gyrus-CA3 circuit, acute or chronic manipulations of abDGCs support a role for abDGCs in the regulation of memory interference. In this Review, we integrate insights from studies that examine the maturation of abDGCs and their integration into the circuit with network mechanisms that support memory discrimination, consolidation and clearance. We propose that adult hippocampal neurogenesis enables the generation of a library of experiences, each registered in mature abDGC physiology and connectivity. Mature abDGCs recruit inhibitory microcircuits to support pattern separation and memory indexing.},
}
@article{MillerandZuckerNC-92,
       author = {Miller, Douglas A. and Zucker, Steven W.},
        title = {Efficient Simplex-like Methods for Equilibria of Nonsymmetric Analog Networks},
      journal = {Neural Computing},
       volume = 4,
        issue = 2,
         year = 1992,
        pages = {167-190},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
}
@article{MillerandZuckerNC-99,
       author = {Miller, Douglas A. and Zucker, Steven W.},
        title = {Computing with Self-excitatory Cliques: A Model and an Application to Hyperacuity-scale Computation in Visual Cortex},
      journal = {Neural Computing},
       volume = 11,
        issue = 1,
         year = 1999,
        pages = {21-66},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
abstract = {We present a model of visual computation based on tightly inter-connected cliques of pyramidal cells. It leads to a formal theory of cell assemblies, a specific relationship between correlated firing patterns and abstract functionality, and a direct calculation relating estimates of cortical cell counts to orientation hyperacuity. Our network architecture is unique in that (1) it supports a mode of computation that is both reliable and efficent; (2) the current-spike relations are modeled as an analog dynamical system in which the requisite computations can take place on the time scale required for an early stage of visual processing; and (3) the dynamics are triggered by the spatiotemporal response of cortical cells. This final point could explain why moving stimuli improve vernier sensitivity.}
}
@article{MilleretalCoRR-16,
       author = {Alexander H. Miller and Adam Fisch and Jesse Dodge and Amir{-}Hossein Karimi and Antoine Bordes and Jason Weston},
        title = {Key-Value Memory Networks for Directly Reading Documents},
      journal = {CoRR},
       volume = {arXiv:1606.03126},
         year = {2016},
     abstract = {Directly reading documents and being able to answer questions from them is an unsolved challenge. To avoid its inherent difficulty, question answering (QA) has been directed towards using Knowledge Bases (KBs) instead, which has proven effective. Unfortunately KBs often suffer from being too restrictive, as the schema cannot support certain types of answers, and too sparse, e.g. Wikipedia contains much more information than Freebase. In this work we introduce a new method, Key-Value Memory Networks, that makes reading documents more viable by utilizing different encodings in the addressing and output stages of the memory read operation. To compare using KBs, information extraction or Wikipedia documents directly in a single framework we construct an analysis tool, WikiMovies, a QA dataset that contains raw text alongside a preprocessed KB, in the domain of movies. Our method reduces the gap between all three settings. It also achieves state-of-the-art results on the existing WikiQA benchmark.}
}
@inproceedings{MilleretalEMNLP-16,
        title = {Key-Value Memory Networks for Directly Reading Documents},
       author = {Miller, Alexander  and Fisch, Adam  and Dodge, Jesse and Karimi, Amir-Hossein  and Bordes, Antoine  and Weston, Jason},
    booktitle = {Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing},
    publisher = {Association for Computational Linguistics},
         year = {2016},
        pages = {1400-1409},
}
@inproceedings{MilleretalIJCAI-85,
       author = {Miller, David P. and Firby, James R. and Dean, Thomas L.},
        title = {Deadlines, Travel Time, and Robot Problem Solving},
    booktitle = {Proceedings {IJCAI}-85},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
     location = {Los Angeles, California},
        pages = {1052-1054},
 organization = {IJCAI},
         year = 1985,
}
@article{MilleretalPNAS-12,
       author = {Miller, E. W. and Lin, J. Y. and Frady, E. P. and Steinbach, P. A. and Kristan, W. B. and Tsien, R. Y.},
        title = {{O}ptically monitoring voltage in neurons by photo-induced electron transfer through molecular wires},
      journal = {Proceedings of the National Academy of Sciences},
       volume = 109,
        issue = 6,
         year = 2012,
        pages = {2114-2119},
}
@article{MilleretalPNAS-14,
       author = {Miller, J. E. and Ayzenshtat, I. and Carrillo-Reid, L. and Yuste, R.},
        title = {Visual stimuli recruit intrinsically generated cortical ensembles},
      journal = {Proceedings of the National Academy of Science U.S.A.},
         year = {2014},
       volume = {111},
       number = {38},
        pages = {4053-4061},
     abstract = {The cortical microcircuit is built with recurrent excitatory connections, and it has long been suggested that the purpose of this design is to enable intrinsically driven reverberating activity. To understand the dynamics of neocortical intrinsic activity better, we performed two-photon calcium imaging of populations of neurons from the primary visual cortex of awake mice during visual stimulation and spontaneous activity. In both conditions, cortical activity is dominated by coactive groups of neurons, forming ensembles whose activation cannot be explained by the independent firing properties of their contributing neurons, considered in isolation. Moreover, individual neurons flexibly join multiple ensembles, vastly expanding the encoding potential of the circuit. Intriguingly, the same coactive ensembles can repeat spontaneously and in response to visual stimuli, indicating that stimulus-evoked responses arise from activating these intrinsic building blocks. Although the spatial properties of stimulus-driven and spontaneous ensembles are similar, spontaneous ensembles are active at random intervals, whereas visually evoked ensembles are time-locked to stimuli. We conclude that neuronal ensembles, built by the coactivation of flexible groups of neurons, are emergent functional units of cortical activity and propose that visual stimuli recruit intrinsically generated ensembles to represent visual attributes.}
}
@article{MilleretalRR-87,
       author = {Miller, W. Thomas and Glanz, Filson H. and Kraft, Gordon L.},
        title = {Application of a General Learning Algorithm to the Control of Robotics Manipulators},
      journal = {The International Journal of Robotics Research},
       volume = 6,
         year = 1987,
        pages = {84-98},
}
@article{MilnerEBR-17,
       author = {Milner, A. D.},
        title = {How do the two visual streams interact with each other?},
      journal = {Experimental Brain Research},
       volume = {235},
       number = {5},
         year = {2017},
        pages = {1297-1308},
     abstract = {The current consensus divides primate cortical visual processing into two broad networks or "streams" composed of highly interconnected areas (Milner and Goodale 2006, 2008; Goodale 2014). The ventral stream, passing from primary visual cortex (V1) through to inferior parts of the temporal lobe, is considered to mediate the transformation of the contents of the visual signal into the mental furniture that guides memory, recognition and conscious perception. In contrast the dorsal stream, passing from V1 through to various areas in the posterior parietal lobe, is generally considered to mediate the visual guidance of action, primarily in real time. The brain, however, does not work through mutually insulated subsystems, and indeed there are well-documented interconnections between the two streams. Evidence for contributions from ventral stream systems to the dorsal stream comes from human neuropsychological and neuroimaging research, and indicates a crucial role in mediating complex and flexible visuomotor skills. Complementary evidence points to a role for posterior dorsal-stream visual analysis in certain aspects of 3-D perceptual function in the ventral stream. A series of studies of a patient with visual form agnosia has been instrumental in shaping our knowledge of what each stream can achieve in isolation; but it has also helped us to tease apart the relative dependence of parietal visuomotor systems on direct bottom-up visual inputs versus inputs redirected via perceptual systems within the ventral stream.}
}
@book{MilnerandGoodale95,
       author = {Milner, D. A. and Goodale, M.},
        title = {The Visual Brain in Action},
    publisher = {Oxford University Press},
      address = {New York, NY},
         year = 1995,
}
@article{MiloetalSCIENCE-02,
       author = {Milo, R. and Shen-Orr, S. and Itzkovitz, S. and Kashtan, N. and Chklovskii, D. and Alon, U.},
        title = {Network motifs: simple building blocks of complex networks},
      journal = {Science},
         year = {2002},
       volume = {298},
       number = {5594},
        pages = {824-827},
     abstract = {Complex networks are studied across many fields of science. To uncover their structural design principles, we defined "network motifs," patterns of interconnections occurring in complex networks at numbers that are significantly higher than those in randomized networks. We found such motifs in networks from biochemistry, neurobiology, ecology, and engineering. The motifs shared by ecological food webs were distinct from the motifs shared by the genetic networks of Escherichia coli and Saccharomyces cerevisiae or from those found in the World Wide Web. Similar motifs were found in networks that perform information processing, even though they describe elements as different as biomolecules within a cell and synaptic connections between neurons in Caenorhabditis elegans. Motifs may thus define universal classes of networks. This approach may uncover the basic building blocks of most networks.}
}
@article{MiltneretalNATURE-09,
       author = {Miltner, W. H. and Braun, C. and Arnold, M. and Witte, H. and Taub, E.},
        title = {Coherence of gamma-band {EEG} activity as a basis for associative learning},
      journal = {Nature},
         year = 1999,
       volume = 397,
        issue = 6718,
        pages = {434-436},
     abstract = {Different regions of the brain must communicate with each other to provide the basis for the integration of sensory information, sensory-motor coordination and many other functions that are critical for learning, memory, information processing, perception and the behaviour of organisms. Hebb suggested that this is accomplished by the formation of assemblies of cells whose synaptic linkages are strengthened whenever the cells are activated or 'ignited' synchronously. Hebb's seminal concept has intrigued investigators since its formulation, but the technology to demonstrate its existence had been lacking until the past decade. Previous studies have shown that very fast electroencephalographic activity in the gamma band (20-70 Hz) increases during, and may be involved in, the formation of percepts and memory, linguistic processing, and other behavioural and perceptual functions. We show here that increased gamma-band activity is also involved in associative learning. In addition, we find that another measure, gamma-band coherence, increases between regions of the brain that receive the two classes of stimuli involved in an associative-learning procedure in humans. An increase in coherence could fulfil the criteria required for the formation of hebbian cell assemblies, binding together parts of the brain that must communicate with one another in order for associative learning to take place. In this way, coherence may be a signature for this and other types of learning.}
} 
@inproceedings{MinkaUAI-01,
    author = {Thomas Minka},
     title = {Expectation Propagation for approximate Bayesian inference},
 booktitle = {Proceedings of the 17th Conference in Uncertainty in Artificial Intelligence},
     pages = {362-369},
 publisher = {Morgan Kaufmann Publishers},
   address = {San Francisco, CA},
      year = 2001,
}
@article{MinkeretalCSL-99,
      title = {Stochastically-based semantic analysis for machine translation.},
     author = {Minker, Wolfgang and Gavald`{a}, Marsal and Waibel, Alex},
    journal = {Computer Speech \& Language},
      issue = 2,
     volume = 13,
       year = 1999,
      pages = {177-194},
}
@book{Minsky68,
       editor = {Minsky, Marvin},
        title = {Semantic Information Processing},
    booktitle = {Semantic Information Processing},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1968,
} 
@incollection{Minsky75,
       author = {Minsky, Marvin},
        title = {A Framework for Representing Knowledge},
       editor = {Winston, Patrick},
    booktitle = {The Psychology of Computer Vision},
    publisher = {McGraw-Hill},
        pages = {211-277},
         year = 1975,
}
@incollection{Minsky81,
       author = {Minsky, Marvin},
        title = {A Framework for Representing Knowledge},
       editor = {Haugeland, John},
    booktitle = {Mind Design},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1981,
}
qq@article{LiuetalICLR-17,
       author = {Chang Liu and Xin Wang and Richard Shin and Joseph E. Gonzalez and Dawn Song},
        title = {Neural Code Completion},
      journal = {Submitted to {ICLR} 2017},
         year = {2017},
     abstract = {Code completion, an essential part of modern software development, yet can bechallenging for dynamically typed programming languages.  In this paper we ex-plore the use of neural network techniques to automatically learn code completionfrom  a  large  corpus  of  dynamically  typed  JavaScript  code.   We  show  differentneural networks that leverage not only token level information but also structuralinformation,  and  evaluate  their  performance  on  different  prediction  tasks.   Wedemonstrate that our models can outperform the state-of-the-art approach, whichis based on decision tree techniques, on both next non-terminal and next terminalprediction tasks by 3.8 points and 0.5 points respectively.  We believe that neuralnetwork techniques can play a transformative role in helping software developersmanage the growing complexity of software systems, and we see this work as afirst step in that direction.},
}
@book{Minsky87,
       author = {Marvin Minsky},
        title = {The Society of Mind},
    publisher = {Simon \& Schuster},
      address = {New York, NY},
         year = 1987,
}
@article{MinskyAoM-61,
       author = {Marvin L. Minsky},
        title = {{Recursive Unsolvability of Post's Problem of "Tag" and other Topics in Theory of Turing Machines}},
      journal = {Annals of Mathematics},
    publisher = {Annals of Mathematics},
         year = {1961},
       number = {3},
       volume = {74},
        pages = {437-455},
}
@book{MinskyCOMPUTATION-67,
       author = {Minsky, Marvin L.},
        title = {Computation: Finite and Infinite Machines},
         year = {1967},
    publisher = {Prentice-Hall, Inc.},
      address = {Upper Saddle River, NJ, USA},
     abstract = {Man has within a single generation found himself sharing the world with a strange new species: the computers and computer-like machines. Neither history, nor philosophy, nor common sense will tell us how these machines will affect us, for they do not do "work" as did machines of the Industrial Revolution. Instead of dealing with materials or energy, we are told that they handle "control" and "information" and even "intellectual processes." There are very few individuals today who doubt that the computer and its relatives are developing rapidly in capability and complexity, and that these machines are destined to play important (though not as yet fully understood) roles in society's future. Though only some of us deal directly with computers, all of us are falling under the shadow of their ever-growing sphere of influence, and thus we all need to understand their capabilities and their limitations.}
}
@book{MinskyandPapert69,
       author = {Minsky, Marvin and Papert, Seymour},
        title = {Perceptrons},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1969,
}
@book{Minton92,
        title = {Machine Learning Methods for Planning and Scheduling},
       editor = {Minton, Steve},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1992,
}
@article{MintonAIJ-89,
       author = {Minton, Steven and Carbonell, Jaime G. and Knoblock, Craig A. and Kuokka, Daniel R. and Etzioni, Oren and Gil, Yolanda},
        title = {Explanation-Based Learning: A Problem Solving Perspective},
      journal = AIJ,
       volume = 40,
         year = 1989,
        pages = {63-118},
}
@book{MintonPhD,
       author = {Minton, Steven},
        title = {Learning Search Control Knowledge: An Explanation-Based Approach},
    publisher = {Kluwer},
      address = {Boston, Massachusetts},
         year = 1988,
}
@article{MirietalNATURE-NEUROSCIENCE-11,
       author = {Miri, Andrew and Daie, Kayvon and Arrenberg, Aristides B. and Baier, Herwig and Aksay, Emre and Tank, David W.},
        title = {Spatial gradients and multidimensional dynamics in a neural integrator circuit},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {14},
        issue = {9},
         year = {2011},
        pages = {1150-1159},
     abstract = {In a neural integrator, the variability and topographical organization of neuronal firing-rate persistence can provide information about the circuit's functional architecture. We used optical recording to measure the time constant of decay of persistent firing (persistence time) across a population of neurons comprising the larval zebrafish oculomotor velocity-to-position neural integrator. We found extensive persistence time variation (tenfold; coefficients of variation = 0.58-1.20) across cells in individual larvae. We also found that the similarity in firing between two neurons decreased as the distance between them increased and that a gradient in persistence time was mapped along the rostrocaudal and dorsoventral axes. This topography is consistent with the emergence of persistence time heterogeneity from a circuit architecture in which nearby neurons are more strongly interconnected than distant ones. Integrator circuit models characterized by multiple dimensions of slow firing-rate dynamics can account for our results.},
}
@book{Mischel2014,
        title = {The Marshmallow Test: Understanding Self-Control and How to Master It},
       author = {Mischel, W.},
         year = {2014},
    publisher = {Transworld Publishers Limited}
}
@article{MishchenckoetalAAS-11,
        title = {A Bayesian approach for inferring neuronal connectivity from calcium fluorescent imaging data},
       author = {Mishchenko, Yuriy and Vogelstein, Joshua T. and Paninski, Liam},
      journal = {The Annals of Applied Statistics},
    publisher = {The Institute of Mathematical Statistics},
       volume = 5,
        issue = {2B},
         year = 2011,
        pages = {1229-1261},
     abstract = {Deducing the structure of neural circuits is one of the central problems of modern neuroscience. Recently-introduced calcium fluorescent imaging methods permit experimentalists to observe network activity in large populations of neurons, but these techniques provide only indirect observations of neural spike trains, with limited time resolution and signal quality. In this work, we present a Bayesian approach for inferring neural circuitry given this type of imaging data. We model the network activity in terms of a collection of coupled hidden Markov chains, with each chain corresponding to a single neuron in the network and the coupling between the chains reflecting the net- work's connectivity matrix. We derive a Monte Carlo Expectation-Maximization algorithm for fitting the model parameters; to obtain the sufficient statistics in a computationally-efficient manner, we introduce a specialized blockwise-Gibbs algorithm for sampling from the joint activity of all observed neurons given the observed fluorescence data. We perform large-scale simulations of randomly connected neuronal networks with biophysically realistic parameters and find that the proposed methods can accurately infer the connectivity in these networks given reasonable experimental and computational constraints. In addition, the estimation accuracy may be improved significantly by incorporating prior knowledge about the sparseness of connectivity in the network, via standard $L_1$ penalization methods.},
}
@article{MishchenkoJNM-09,
title = {Automation of {3D} reconstruction of neural tissue from large volume of conventional serial section transmission electron micrographs},
    author = {Mishchenko, Yuriy},
   journal = {Journal Neuroscience Methods},
    volume = 176,
     issue = 2,
     pages = {276-89},
      year = 2009,
  abstract = {We describe an approach for automation of the process of reconstruction of neural tissue from serial section transmission electron micrographs. Such reconstructions require {3D} segmentation of individual neuronal processes (axons and dendrites) performed in densely packed neuropil. We first detect neuronal cell profiles in each image in a stack of serial micrographs with multi-scale ridge detector. Short breaks in detected boundaries are interpolated using anisotropic contour completion formulated in fuzzy-logic framework. Detected profiles from adjacent sections are linked together based on cues such as shape similarity and image texture. Thus obtained {3D} segmentation is validated by human operators in computer-guided proofreading process. Our approach makes possible reconstructions of neural tissue at final rate of about 5 microm3/manh, as determined primarily by the speed of proofreading. To date we have applied this approach to reconstruct few blocks of neural tissue from different regions of rat brain totaling over 1000microm3, and used these to evaluate reconstruction speed, quality, error rates, and presence of ambiguous locations in neuropil ssTEM imaging data.}
}
@article{MishchenkoJNM-11,
        title = {Reconstruction of complete connectivity matrix for connectomics by sampling neural connectivity with fluorescent synaptic markers},
       author = {Mishchenko, Yuriy},
      journal = {Journal Neuroscience Methods},
       volume = 196,
        issue = 2,
        pages = {289-302},
         year = 2011,
     abstract = {Physical organization of the nervous system is a topic of perpetual interest in neuroscience. Despite significant achievements here in the past, many details of the nervous system organization and its role in animals' behavior remain obscure, while the problem of complete connectivity reconstructions has recently re-emerged as one of the major directions in neuroscience research (ie connectomics). We describe a novel paradigm for connectomics reconstructions that can yield connectivity maps with high resolution, high speed of imaging and data analysis, and significant robustness to errors. In essence, we propose that physical connectivity in a neural circuit can be sampled using anatomical fluorescent synaptic markers localized to different parts of the neural circuit with a technique for randomized genetic targeting, and that high-resolution connectivity maps can be extracted from such datasets. We describe how such an approach can be implemented and how neural connectivity matrix can be reconstructed statistically using the methods of Compressive Sensing. Use of Compressive Sensing is the key to allow accurate neural connectivity reconstructions with orders-of-magnitude smaller volumes of experimental data. We test described approach on simulations of neural connectivity reconstruction experiments in C. elegans, where real neural wiring diagram is available from past electron microscopy studies. We show that such wiring diagram can be in principle re-obtained using described approach in 1-7 days of imaging and data analysis. Alternative approaches would require currently at least 1-2 years to produce a single comparable reconstruction. We discuss possible applications of described approach in larger organisms such as Drosophila.},
}
@article{MisicandSpornsCOIN-16,
       author = {Bratislav Misic and Olaf Sporns},
        title = {From regions to connections and networks: new bridges between brain and behavior},
      journal = {Current Opinion in Neurobiology},
       volume = {40},
         year = {2016},
        pages = {1-7},
     abstract = {Connections and interactions among distributed brain areas are increasingly recognized as the basis for cognitive operations and a diverse repertoire of behaviors. Analytic advances have allowed for brain connectivity to be represented and quantified at multiple levels: from single connections to communities and networks. This review traces the trajectory of network neuroscience, focusing on how connectivity patterns can be related to cognition and behavior. As recent initiatives for open science provide access to imaging and phenotypic data with great detail and depth, we argue that approaches capable of directly modeling multivariate relationships between brain and behavior will become increasingly important in the field.}
}
@book{Mitchell96,
       author = {Melanie Mitchell},
        title = {An Introduction to Genetic Algorithms},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1996,
}
@book{Mitchell97,
       author = {Tom Mitchell},
        title = {Machine Learning},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1997,
}
@article{MitchellAIJ-82,
       author = {Mitchell, Thomas M.},
        title = {Generalization as Search},
      journal = AIJ,
       volume = 18,
        pages = {203-226},
         year = 1982,
}
@incollection{MitchellML-83,
       author = {Mitchell, Thomas M. and Utgoff, P. and Banerji, R.},
        title = {Learning by Experimentation: Formulating and Generalizing Plans from Past Experience},
     crossref = {MichalskietalML-83},
}
@techreport{MitchellPhD,
       author = {Mitchell, Thomas M.},
        title = {Version Spaces: An Approach to Concept Learning},
         type = {{Ph.D.} Thesis},
  institution = {Stanford University},
         year = 1977,
}
@article{MitchellREVIEW_FIGMENTS_of_REALITY-REVIEW-97,
        title = {Review of Figments of Reality by Ian Stewart and Jack Cohen Cambridge University Press, 1997},
       author = {Melanie Mitchell},
      journal = {New Scientist},
         year = {1997},
          day = {11},
        month = {August}
}
@techreport{MitchellTR-80,
       author = {Tom M. Mitchell},
        title = {The Need for Biases in Learning Generalizations},
        issue = {CBM-TR-117},
  institution = {Carnegie Mellon University},
         year = {1980}
}
@article{MitchellandLapataCOGSCI-10,
       author = {Mitchell, Jeff and Lapata, Mirella},
        title = {Composition in Distributional Models of Semantics},
      journal = {Cognitive Science},
       volume = 34,
        issue = 8,
    publisher = {Blackwell Publishing Ltd},
        pages = {1388-1429},
         year = 2010,
     abstract = {Vector-based models of word meaning have become increasingly popular in cognitive science. The appeal of these models lies in their ability to represent meaning simply by using distributional information under the assumption that words occurring within similar contexts are semantically similar. Despite their widespread use, vector-based models are typically directed at representing words in isolation, and methods for constructing representations for phrases or sentences have received little attention in the literature. This is in marked contrast to experimental evidence (e.g., in sentential priming) suggesting that semantic similarity is more complex than simply a relation between isolated words. This article proposes a framework for representing the meaning of word combinations in vector space. Central to our approach is vector composition, which we operationalize in terms of additive and multiplicative functions. Under this framework, we introduce a wide range of composition models that we evaluate empirically on a phrase similarity task.},
}
@inproceedings{MitchelletalAAAI-92,
       author = {Mitchell, David and Selman, Bart and Levesque, Hector},
        title = {Hard and Easy Distributions of {SAT} Problems},
    booktitle = {Proceedings AAAI-92},
     location = {San Jose, California},
 organization = {AAAI},
        pages = {459-465},
         year = 1992,
}
@article{MitchelletalCACM-18,
       author = {T. Mitchell and W. Cohen and E. Hruschka and P. Talukdar and B. Yang and J. Betteridge and A. Carlson and B. Dalvi and M. Gardner and B. Kisiel and J. Krishnamurthy and N. Lao and K. Mazaitis and T. Mohamed and N. Nakashole and E. Platanios and A. Ritter and M. Samadi and B. Settles and R. Wang and D. Wijaya and A. Gupta and X. Chen and A. Saparov and M. Greaves and J. Welling},
        title = {Never-Ending Learning},
      journal = {Communications of the {ACM}},
       volume = {61},
       number = {5},
         year = {2018},
        pages = {103-115},
     abstract = {Whereas people learn many different types of knowledge from diverse experiences over many years, and become better learners over time, most current machine learning systems are much more narrow, learning just a single function or data model based on statistical analysis of a single data set. We suggest that people learn better than computers precisely because of this difference, and we suggest a key direction for machine learning research is to develop software architectures that enable intelligent agents to also learn many types of knowledge, continuously over many years, and to become better learners over time. In this paper we define more precisely this never-ending learning paradigm for machine learning, and we present one case study: the Never-Ending Language Learner (NELL), which achieves a number of the desired properties of a never-ending learner. NELL has been learning to read the Web 24hrs/day since January 2010, and so far has acquired a knowledge base with 120mn diverse, confidence-weighted beliefs (e.g., servedWith(tea,biscuits)), while learning thousands of interrelated functions that continually improve its reading competence over time. NELL has also learned to reason over its knowledge base to infer new beliefs it has not yet read from those it has, and NELL is inventing new relational predicates to extend the ontology it uses to represent beliefs. We describe the design of NELL, experimental results illustrating its behavior, and discuss both its successes and shortcomings as a case study in never-ending learning. NELL can be tracked online at http://rtw.ml.cmu.edu, and followed on Twitter at @CMUNELL.}
}
@article{MitchelletalML-04,
       author = {T.M. Mitchell and R. Hutchinson and R.S. Niculescu and F. Pereira and X. Wang and M. Just and S. Newman},
        title = {Learning to Decode Cognitive States from Brain Images},
      journal = {Machine Learning},
       volume = 57,
        issue = {1-2},
        pages = {145-175},
         year = 2004,
}
@article{MitchelletalSCIENCE-08,
   author = {T. M. Mitchell and S. V. Shinkareva and A. Carlson and K.M. Chang and V. L. Malave and R. A. Mason and M. A. Just},
    title = {Predicting Human Brain Activity Associated with the Meanings of Nouns}, 
  journal = {Science},
   volume = 320,
    issue = 5880,
    pages = {1191-1195},
     year = 2008,
}
@article{MitkoandFischerCRPI-20,
       author = {Mitko, Alex and Fischer, Jason},
        title = {When it all falls down: the relationship between intuitive physics and spatial cognition},
      journal = {Cognitive Research: Principles and Implications},
         year = {2020},
       volume = {5},
        issue = {1},
        pages = {24},
     abstract = {Our intuitive understanding of physical dynamics is crucial in daily life. When we fill a coffee cup, stack items in a refrigerator, or navigate around a slippery patch of ice, we draw on our intuitions about how physical interactions will unfold. What mental machinery underlies our ability to form such inferences? Numerous aspects of cognition must contribute - for example, spatial thinking, temporal prediction, and working memory, to name a few. Is intuitive physics merely the sum of its parts - a collection of these and other related abilities that we apply to physical scenarios as we would to other tasks? Or does physical reasoning rest on something extra - a devoted set of mental resources that takes information from other cognitive systems as inputs? Here, we take a key step in addressing this question by relating individual differences on a physical prediction task to performance on spatial tasks, which may be most likely to account for intuitive physics abilities given the fundamentally spatial nature of physical interactions. To what degree can physical prediction performance be disentangled from spatial thinking? We tested 100 online participants in an 'Unstable Towers' task and measures of spatial cognition and working memory. We found a positive relationship between intuitive physics and spatial skills, but there were substantial, reliable individual differences in physical prediction ability that could not be accounted for by spatial measures or working memory. Our findings point toward the separability of intuitive physics from spatial cognition.},
}
@article{MitraAB-03,
    author = {Mitra, R.D. and Shendure, J. and Olejnik, J. and Olejnik, E-K. and Church G.M.},
     title = {Fluorescent in situ sequencing on polymerase colonies},
   journal = {Analytical Biochemistry},
    volume = 320,
     issue = 1,
     pages = {55-65},
      year = 2003,
  abstract = {Integration of DNA isolation, amplification, and sequencing can be achieved by the use of polymerase colonies (polonies) and cycles of fluorescent dNTP incorporation. In this paper, we present four advances that bring us closer to sequencing genomes cost-effectively using the polony technology. First, a polymerase trapping technique enables efficient nucleotide extension by DNA polymerase in a polyacrylamide matrix and eliminates loss of enzyme during sequencing cycles. Next, we present two novel types of reversibly dye-labeled nucleotide analogues, show that DNA polymerase can incorporate these analogues, and demonstrate that the dyes can be removed by thiol reduction or light exposure. Using these nucleotides, we have sequenced multiple polonies in parallel. In addition, we have found that a high density of polonies can be achieved with minimal overlap between adjacent polonies by limiting the concentration of free primer in the polony amplification reactions. Finally, we have developed software for automated image alignment and sequence calling.}
}
@misc{MitzenmacherandUpfal01,
       author = {M. Mitzenmacher and E. Upfal},
        title = {Modeling deletions in the {W}eb graph},
 howpublished = {Draft},
         year = 2001,
}
@article{MiyashitaSCI-04,
       author = {Yasushi Miyashita},
        title = {Cognitive Memory: Cellular and Network Machineries and their Top-Down Control},
      journal = {Science},
       volume = 306,
        issue = 5695,
         year = 2004,
        pages = {435-440},
}
@inproceedings{MnihandHintonICML-07,
    author = {Andriy Mnih and Geoffrey E. Hinton},
     title = {Three new graphical models for statistical language modelling},
 booktitle = {Proceedings of the 24th International Conference on Machine Learning},
    editor = {Zoubin Ghahramani},
      year = 2007,
     pages = {641-648},
}
@inproceedings{MnihandHintonNIPS-08,
     title = {A Scalable Hierarchical Distributed Language Model},
    author = {Andriy Mnih and Geoffrey E. Hinton},
 booktitle = {Advances in Neural Information Processing Systems 21},
    editor = {D. Koller and D. Schuurmans and Y. Bengio and L. Bottou},
      year = 2008,
     pages = {1081-1088},
}
@inproceedings{MnihandTehICML-12,
    author = {Andriy Mnih and Yee Whye Teh},
     title = {A fast and simple algorithm for training neural probabilistic language models},
 booktitle = {Proceedings of the 29th International Conference on Machine Learning},
    editor = {John Langford and Joelle Pineau},
      year = 2012,
}
@article{MnihetalCoRR-13,
       author = {Volodymyr Mnih and Koray Kavukcuoglu and David Silver and Alex Graves and Ioannis Antonoglou and Daan Wierstra and Martin Riedmiller},
        title = {Playing {Atari} with Deep Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1312.5602},
         year = 2013,
     abstract = {We present the first deep learning model to successfully learn control policies directly from high-dimensional sensory input using reinforcement learning. The model is a convolutional neural network, trained with a variant of Q-learning, whose input is raw pixels and whose output is a value function estimating future rewards. We apply our method to seven Atari 2600 games from the Arcade Learning Environment, with no adjustment of the architecture or learning algorithm. We find that it outperforms all previous approaches on six of the games and surpasses a human expert on three of them.}
}
@article{MnihetalCoRR-14,
       author = {Volodymyr Mnih and Nicolas Heess and Alex Graves and Koray Kavukcuoglu},
        title = {Recurrent Models of Visual Attention},
      journal = {CoRR},
       volume = {arXiv:1406.6247},
         year = 2014,
          url = {http://arxiv.org/abs/1406.6247},
     abstract = {Applying convolutional neural networks to large images is computationally expensive because the amount of computation scales linearly with the number of image pixels. We present a novel recurrent neural network model that is capable of extracting information from an image or video by adaptively selecting a sequence of regions or locations and only processing the selected regions at high resolution. Like convolutional neural networks, the proposed model has a degree of translation invariance built-in, but the amount of computation it performs can be controlled independently of the input image size. While the model is non-differentiable, it can be trained using reinforcement learning methods to learn task-specific policies. We evaluate our model on several image classification tasks, where it significantly outperforms a convolutional neural network baseline on cluttered images, and on a dynamic visual control problem, where it learns to track a simple object without an explicit training signal for doing so.}
}
@article{MnihetalCoRR-16,
       author = {Volodymyr Mnih and Adri{\`{a}} Puigdom{\`{e}}nech Badia and Mehdi Mirza and Alex Graves and Timothy P. Lillicrap and Tim Harley and David Silver and Koray Kavukcuoglu},
        title = {Asynchronous Methods for Deep Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1602.01783},
         year = {2016},
      comment = {This is often referred to as the A3C  (asynchronous advantage actor-critic) model.},
     abstract = {We propose a conceptually simple and lightweight framework for deep reinforcement learning that uses asynchronous gradient descent for optimization of deep neural network controllers. We present asynchronous variants of four standard reinforcement learning algorithms and show that parallel actor-learners have a stabilizing effect on training allowing all four methods to successfully train neural network controllers. The best performing method, an asynchronous variant of actor-critic, surpasses the current state-of-the-art on the Atari domain while training for half the time on a single multi-core CPU instead of a GPU. Furthermore, we show that asynchronous actor-critic succeeds on a wide variety of continuous motor control problems as well as on a new task of navigating random 3D mazes using a visual input.}
}
@article{MnihetalNATURE-15,
       author = {Mnih, Volodymyr and Kavukcuoglu, Koray and Silver, David and Rusu, Andrei A. and Veness, Joel and Bellemare, Marc G. and Graves, Alex and Riedmiller, Martin and Fidjeland, Andreas K. and Ostrovski, Georg and Petersen, Stig and Beattie, Charles and Sadik, Amir and Antonoglou, Ioannis and King, Helen and Kumaran, Dharshan and Wierstra, Daan and Legg, Shane and Hassabis, Demis},
        title = {Human-level control through deep reinforcement learning},
      journal = {Nature},
       number = {7540},
        pages = {529-533},
       volume = {518},
         year = {2015},
     abstract = {The theory of reinforcement learning provides a normative account, deeply rooted in psychological and neuroscientific perspectives on animal behaviour, of how agents may optimize their control of an environment. To use reinforcementlearning successfully insituations approaching real-world complexity, however, agents are confronted with a difficult task:they must derive efficient representations of the environment from high-dimensional sensory inputs, and use these to generalize past experience to new situations. Remarkably, humans and other animalsseem to solve this problem through a harmonious combination of reinforcement learning and hierarchical sensory processing systems, the former evidenced by a wealth of neural data revealing not able parallels between the phasic signals emitted by dopaminergic neurons and temporal difference reinforcement learning algorithms. While reinforcement learning agents have achieved some successes in a variety of domains, their applicability has previously been limited to domains in which useful features can be handcrafted, or to domains with fully observed, low-dimensional state spaces. Here we use recent advances in training deep neural networks to develop a novel artificial agent, termed a deep Q-network, that can learn successful policies directly from high-dimensional sensory inputs using end-to-end reinforcement learning. We tested this agent on the challenging domain of classic Atari 2600 games. We demonstrate that the deep Q-network agent, receiving only the pixels and the game score as inputs, was able to surpass the performance of all previous algorithms and achieve a level comparable to that of a professional human games tester across a set of 49 games,using the same algorithm, network architecture and hyperparameters. This work bridges the divide between high-dimensional sensory inputs and actions,resulting in the first artificial agent that is capable of learning to excel at a diverse array of challenging tasks.},
}
@article{MobahiCoRR-16a,
       author = {Hossein Mobahi},
        title = {Training Recurrent Neural Networks by Diffusion},
      journal = {CoRR},
       volume = {arXiv:1601.04114},
         year = {2016},
     abstract = {This work presents a new algorithm for training recurrent neural networks (although ideas are applicable to feedforward networks as well). The algorithm is derived from a theory in nonconvex optimization related to the diffusion equation. The contributions made in this work are two fold. First, we show how some seemingly disconnected mechanisms used in deep learning such as smart initialization, annealed learning rate, layerwise pretraining, and noise injection (as done in dropout and SGD) arise naturally and automatically from this framework, without manually crafting them into the algorithms. Second, we present some preliminary results on comparing the proposed method against SGD. It turns out that the new algorithm can achieve similar level of generalization accuracy of SGD in much fewer number of epochs.},
}
@article{MobahiCoRR-16b,
       author = {Hossein Mobahi},
        title = {A Theory of Local Matching: {SIFT} and Beyond},
      journal = {CoRR},
       volume = {arXiv:1601.05116},
         year = {2016},
     abstract = {Why has SIFT been so successful? Why its extension, DSP-SIFT, can further improve SIFT? Is there a theory that can explain both? How can such theory benefit real applications? Can it suggest new algorithms with reduced computational complexity or new descriptors with better accuracy for matching? We construct a general theory of local descriptors for visual matching. Our theory relies on concepts in energy minimiza- tion and heat diffusion. We show that SIFT and DSP-SIFT approximate the solution the theory suggests. In particular, DSP-SIFT gives a bet- ter approximation to the theoretical solution; justifying why DSP-SIFT outperforms SIFT. Using the developed theory, we derive new descriptors that have fewer parameters and are potentially better in handling affine deformations.}
}
@article {ModhaandSinghPNAS-10,
       author = {Modha, Dharmendra S. and Singh, Raghavendra},
        title = {Network architecture of the long-distance pathways in the macaque brain},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {107},
       number = {30},
        pages = {13485-13490},
         year = {2010},
    publisher = {National Academy of Sciences},
     abstract = {Understanding the network structure of white matter communication pathways is essential for unraveling the mysteries of the brain's function, organization, and evolution. To this end, we derive a unique network incorporating 410 anatomical tracing studies of the macaque brain from the Collation of Connectivity data on the Macaque brain (CoCoMac) neuroinformatic database. Our network consists of 383 hierarchically organized regions spanning cortex, thalamus, and basal ganglia; models the presence of 6,602 directed long-distance connections; is three times larger than any previously derived brain network; and contains subnetworks corresponding to classic corticocortical, corticosubcortical, and subcortico-subcortical fiber systems. We found that the empirical degree distribution of the network is consistent with the hypothesis of the maximum entropy exponential distribution and discovered two remarkable bridges between the brain's structure and function via network-theoretical analysis. First, prefrontal cortex contains a disproportionate share of topologically central regions. Second, there exists a tightly integrated core circuit, spanning parts of premotor cortex, prefrontal cortex, temporal lobe, parietal lobe, thalamus, basal ganglia, cingulate cortex, insula, and visual cortex, that includes much of the task-positive and task-negative networks and might play a special role in higher cognition and consciousness.},
}
@article {ModhaandSinghPNAS-10.Supplement,
       author = {Modha, Dharmendra S. and Singh, Raghavendra},
        title = {Supplement to Network architecture of the long-distance pathways in the macaque brain},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {107},
       number = {30},
         year = {2010},
        pages = {1-57},
}
@article{MoffaertandNoweJMLR-14,
       author = {Kristof Van Moffaert and Ann Now\'{e}},
        title = {Multi-Objective Reinforcement Learning using Sets of Pareto Dominating Policies},
      journal = {Journal of Machine Learning Research},
         year = {2014},
       volume = {15},
        pages = {3663-3692},
     abstract = {Many real-world problems involve the optimization of multiple, possibly conflicting objectives. Multi-objective reinforcement learning (MORL) is a generalization of standard reinforcement learning where the scalar reward signal is extended to multiple feedback signals, in essence, one for each objective. MORL is the process of learning policies that optimize multiple criteria simultaneously. In this paper, we present a novel temporal difference learning algorithm that integrates the Pareto dominance relation into a reinforcement learning approach. This algorithm is a multi-policy algorithm that learns a set of Pareto dominating policies in a single run. We name this algorithm Pareto Q-learning and it is applicable in episodic environments with deterministic as well as stochastic transition functions. A crucial aspect of Pareto Q-learning is the updating mechanism that bootstraps sets of Q-vectors. One of our main contributions in this paper is a mechanism that separates the expected immediate reward vector from the set of expected future discounted reward vectors. This decomposition allows us to update the sets and to exploit the learned policies consistently throughout the state space. To balance exploration and exploitation during learning, we also propose three set evaluation mechanisms. These three mechanisms evaluate the sets of vectors to accommodate for standard action selection strategies, such as ϵ-greedy. More precisely, these mechanisms use multi-objective evaluation principles such as the hypervolume measure, the cardinality indicator and the Pareto dominance relation to select the most promising actions. We experimentally validate the algorithm on multiple environments with two and three objectives and we demonstrate that Pareto Q-learning outperforms current state-of-the-art MORL algorithms with respect to the hypervolume of the obtained policies. We note that (1) Pareto Q-learning is able to learn the entire Pareto front under the usual assumption that each state-action pair is sufficiently sampled, while (2) not being biased by the shape of the Pareto front. Furthermore, (3) the set evaluation mechanisms provide indicative measures for local action selection and (4) the learned policies can be retrieved throughout the state and action space.}
}
@inproceedings{MohajerietalIEEE-09,
       author = {K. {Mohajeri} and G. {Pishehvar} and M. {Seifi}},
    booktitle = {2009 IEEE International Symposium on Computational Intelligence in Robotics and Automation - (CIRA)},
        title = {{CMAC} neural networks structures},
         year = {2009},
        pages = {39-45},
     abstract = {Cerebellar Model Articulation Controller (CMAC) NN is a computational model of cerebellum introduced as an alternative to backpropagated multilayer networks to control robot arms. From then it has seen many improvements and has been applied in many other areas as a general NN. These improvements have been in the context of generalization, learning techniques, differentiability, memory size, fuzzification and hardware implementation. This paper is a systematic review of CMAC's different structures and applications.},
}
@article{MokeichevetalNEURON-07,
    author = {Mokeichev, A. and Okun, M. and Barak, O. and Katz, Y. and Ben-Shahar, O. and Lampl, I.},
     title = {Stochastic emergence of repeating cortical motifs in spontaneous membrane potential fluctuations in vivo},
   journal = {Neuron},
    volume = 53,
     issue = 3,
      year = 2007,
     pages = {413-425},
  abstract = {It was recently discovered that subthreshold membrane potential fluctuations of cortical neurons can precisely repeat during spontaneous activity, seconds to minutes apart, both in brain slices and in anesthetized animals. These repeats, also called cortical motifs, were suggested to reflect a replay of sequential neuronal firing patterns. We searched for motifs in spontaneous activity, recorded from the rat barrel cortex and from the cat striate cortex of anesthetized animals, and found numerous repeating patterns of high similarity and repetition rates. To test their significance, various statistics were compared between physiological data and three different types of stochastic surrogate data that preserve dynamical characteristics of the recorded data. We found no evidence for the existence of deterministically generated cortical motifs. Rather, the stochastic properties of cortical motifs suggest that they appear by chance, as a result of the constraints imposed by the coarse dynamics of subthreshold ongoing activity.}
}
@misc{Molecular_Probes_Handbook-10,
        title = {Molecular Probes Handbook},
       author = {Life Technologies},
         year = 2010,
          url = {file:///u/tld/Drive/write/articles/MOLECULAR_PROBES_HANDBOOK/Molecular_Probes_Handbook-10.pdf},
 howpublished = {Life Technologies Technical Report},
}
@incollection{Molnaretal2010,
        title = {Subplate and the Formation of the Earliest Cerebral Cortical Circuits},
       author = {Zolt\'{a}n Moln\'{a}r and  Wei Zhi Wang and  Maria Carmen Pi\~{n}on and  Daniel Blakey and  Shinichi Kondo and  Franziska Oeschger and Anna Hoerder-Suabedissen},
       editor = {Feldmeyer D. and L\"{u}bke J.},
    publisher = {Springer, Boston, MA},
    booktitle = {New Aspects of Axonal Structure and Function},
      address = {Springer, Boston, MA},
         year = {2010},
        pages = {19-31}
}
@article{MolnaretalCILS-11,
       author = {Ferenc Molnar Jr. and Ferenc Izsak and Robert Miszaros and Istvan Lagzi},
        title = {Simulation of reaction-diffusion processes in three dimensions using {CUDA}},
      journal = {Chemometrics and Intelligent Laboratory Systems},
       volume = {108},
        issue = {1},
         year = {2011},
        pages = {76-85},
     abstract = {Numerical solution of reaction-diffusion equations in three dimensions is one of the most challenging applied mathematical problems. Since these simulations are very time consuming, any ideas and strategies aiming at the reduction of CPU time are important topics of research. A general and robust idea is the parallelization of source codes/programs. Recently, the technological development of graphics hardware created a possibility to use desktop video cards to solve numerically intensive problems. We present a powerful parallel computing framework for solving reaction-diffusion equations numerically using the Graphics Processing Units (GPUs) with CUDA. Four different reaction-diffusion problems, (i) diffusion of chemically inert compound, (ii) Turing pattern formation, (iii) phase separation in the wake of a moving diffusion front and (iv) air pollution dispersion were solved, and additionally both the Shared method and the Moving Tiles method were tested. Our results show that parallel implementation achieves typical acceleration values in the order of 5-40 times compared to CPU using a single-threaded implementation on a 2.8 GHz desktop computer.},
}
@article{MolnaretalJoA-19,
       author = {Moln\'{a}r, Zolt\'{a}n and Clowry, Gavin J. and Sestan, Nenad and Alzu'bi, Ayman and Bakken, Trygve and Hevner, Robert F. and H\"{u}ppi, Petra S. and Kostovi\'{c}, Ivica and Rakic, Pasko and Anton, E. S. and Edwards, David and Garcez, Patricia and Hoerder-Suabedissen, Anna and Kriegstein, Arnold},
        title = {New insights into the development of the human cerebral cortex},
      journal = {Journal of Anatomy},
       volume = {235},
       number = {3},
         year = {2019},
        pages = {432-451},
     abstract = {Abstract The cerebral cortex constitutes more than half the volume of the human brain and is presumed to be responsible for the neuronal computations underlying complex phenomena, such as perception, thought, language, attention, episodic memory and voluntary movement. Rodent models are extremely valuable for the investigation of brain development, but cannot provide insight into aspects that are unique or highly derived in humans. Many human psychiatric and neurological conditions have developmental origins but cannot be studied adequately in animal models. The human cerebral cortex has some unique genetic, molecular, cellular and anatomical features, which need to be further explored. The Anatomical Society devoted its summer meeting to the topic of Human Brain Development in June 2018 to tackle these important issues. The meeting was organized by Gavin Clowry (Newcastle University) and Zolt\'{a}n Moln\'{a}r (University of Oxford), and held at St John's College, Oxford. The participants provided a broad overview of the structure of the human brain in the context of scaling relationships across the brains of mammals, conserved principles and recent changes in the human lineage. Speakers considered how neuronal progenitors diversified in human to generate an increasing variety of cortical neurons. The formation of the earliest cortical circuits of the earliest generated neurons in the subplate was discussed together with their involvement in neurodevelopmental pathologies. Gene expression networks and susceptibility genes associated to neurodevelopmental diseases were discussed and compared with the networks that can be identified in organoids developed from induced pluripotent stem cells that recapitulate some aspects of in vivo development. New views were discussed on the specification of glutamatergic pyramidal and γ-aminobutyric acid (GABA)ergic interneurons. With the advancement of various in vivo imaging methods, the histopathological observations can be now linked to in vivo normal conditions and to various diseases. Our review gives a general evaluation of the exciting new developments in these areas. The human cortex has a much enlarged association cortex with greater interconnectivity of cortical areas with each other and with an expanded thalamus. The human cortex has relative enlargement of the upper layers, enhanced diversity and function of inhibitory interneurons and a highly expanded transient subplate layer during development. Here we highlight recent studies that address how these differences emerge during development focusing on diverse facets of our evolution.},
}
@article{MomennejadCOiBS-20,
        title = {Learning Structures: Predictive Representations, Replay, and Generalization},
       author = {Ida Momennejad},
      journal = {Current Opinion in Behavioral Sciences},
       volume = {32},
        pages = {155-166},
         year = {2020},
     abstract = {Memory and planning rely on learning the structure of relationships among experiences. Compact representations of these structures guide flexible behavior in humans and animals. A century after 'latent learning' experiments summarized by Tolman, the larger puzzle of cognitive maps remains elusive: how does the brain learn and generalize relational structures? This review focuses on a reinforcement learning (RL) approach to learning compact representations of the structure of states. We review evidence showing that capturing structures as predictive representations updated via replay offers a neurally plausible account of human behavior and the neural representations of predictive cognitive maps. We highlight multi-scale successor representations, prioritized replay, and policy-dependence. These advances call for new directions in studying the entanglement of learning and memory with prediction and planning.}
}
@article{MonahanMS-82,
       author = {Monahan, George E.},
        title = {A Survey of Partially Observable {M}arkov Decision Processes: Theory, Models, and Algorithms},
      journal = {Management Science},
       volume = 28,
        issue = 1,
         year = 1982,
        pages = {1-16},
}
@article{MoncadaetalNP-15,
       author = {Moncada, Diego and Ballarini, Fabricio and Viola, Hayd{\'{e}}e},
        title = {Behavioral Tagging: {A} Translation of the Synaptic Tagging and Capture Hypothesis},
      journal = {Neural Plasticity},
       volume = {215},
        issue = {650780},
         year = {2015},
        pages = {21},
     abstract = {Similar molecular machinery is activated in neurons following an electrical stimulus that induces synaptic changes and after learning sessions that trigger memory formation. Then, to achieve perdurability of these processes protein synthesis is required for the reinforcement of the changes induced in the network. The synaptic tagging and capture theory provided a strong framework to explain synaptic specificity and persistence of electrophysiological induced plastic changes. Ten years later, the behavioral tagging hypothesis (BT) made use of the same argument, applying it to learning and memory models. The hypothesis postulates that the formation of lasting memories relies on at least two processes: the setting of a learning tag and the synthesis of plasticity related proteins, which once captured at tagged sites allow memory consolidation. BT explains how weak events, only capable of inducing transient forms of memories, can result in lasting memories when occurring close in time with other behaviorally relevant experiences that provide proteins. In this review, we detail the findings supporting the existence of BT process in rodents, leading to the consolidation, persistence, and interference of a memory. We focus on the molecular machinery taking place in these processes and describe the experimental data supporting the BT in humans.}
}
@article{MonperrusACM-17,
       author = {Monperrus, Martin},
        title = {Automatic Software Repair: A Bibliography},
      journal = {{ACM} Computing Surveys},
       volume = {51},
       number = {1},
         year = {2018},
        pages = {17:1-17:24},
    publisher = {ACM},
      address = {New York, NY, USA},
     abstract = {This article presents a survey on automatic software repair. Automatic software repair consists of automatically finding a solution to software bugs without human intervention. This article considers all kinds of repairs. First, it discusses behavioral repair where test suites, contracts, models, and crashing inputs are taken as oracle. Second, it discusses state repair, also known as runtime repair or runtime recovery, with techniques such as checkpoint and restart, reconfiguration, and invariant restoration. The uniqueness of this article is that it spans the research communities that contribute to this body of knowledge: software engineering, dependability, operating systems, programming languages, and security. It provides a novel and structured overview of the diversity of bug oracles and repair operators used in the literature.}
}
@article{MontgomeryetalBMCB-10,
     title = {Reconstructing the ups and downs of primate brain evolution: implications for adaptive hypotheses and {\it{Homo floresiensis}}},
    author = {Stephen H. Montgomery and Isabella Capellini and Robert A. Barton and Nicholas I. Mundy},
   journal = {BMC Biology},
    volume = 8,
     issue = 1,
      year = 2010,
     pages = 9,
  abstract = {BACKGROUND: Brain size is a key adaptive trait. It is often assumed that increasing brain size was a general evolutionary trend in primates, yet recent fossil discoveries have documented brain size decreases in some lineages, raising the question of how general a trend there was for brains to increase in mass over evolutionary time. We present the first systematic phylogenetic analysis designed to answer this question. RESULTS: We performed ancestral state reconstructions of three traits (absolute brain mass  absolute body mass, relative brain mass) using 37 extant and 23 extinct primate species and three approaches to ancestral state reconstruction: parsimony, maximum likelihood and Bayesian Markov-chain Monte Carlo. Both absolute and relative brain mass generally increased over evolutionary time, but body mass did not. Nevertheless both absolute and relative brain mass decreased along several branches. Applying these results to the contentious case of Homo floresiensis, we find a number of scenarios under which the proposed evolution of Homo floresiensis' small brain appears to be consistent with patterns observed along other lineages, dependent on body mass and phylogenetic position. CONCLUSIONS: Our results confirm that brain expansion began early in primate evolution and show that increases occurred in all major clades. Only in terms of an increase in absolute mass does the human lineage appear particularly striking, with both the rate of proportional change in mass and relative brain size having episodes of greater expansion elsewhere on the primate phylogeny. However, decreases in brain mass also occurred along branches in all major clades, and we conclude that, while selection has acted to enlarge primate brains, in some lineages this trend has been reversed. Further analyses of the phylogenetic position of Homo floresiensis and better body mass estimates are required to confirm the plausibility of the evolution of its small brain mass. We find that for our dataset the Bayesian analysis for ancestral state reconstruction is least affected by inclusion of fossil data suggesting that this approach might be preferable for future studies on other taxa with a poor fossil record.},
}
@article{MontgomeryetalMBE-11,
     title = {Adaptive Evolution of Four Microcephaly Genes and the Evolution of Brain Size in Anthropoid Primates},
    author = {Stephen H. Montgomery and Isabella Capellini and Chris Venditti and Robert A. Barton and Nicholas I. Mundy},
   journal = {Molecular Biological Evolution},
    volume = 28,
     issue = 1,
      year = 2011,
     pages = {625-638},
}
@article{MontgomeryetalNATURE-15,
       author = {Montgomery, Kate L. and Yeh, Alexander J. and Ho, John S. and Tsao, Vivien and Mohan Iyer, Shrivats and Grosenick, Logan and Ferenczi, Emily A. and Tanabe, Yuji and Deisseroth, Karl and Delp, Scott L. and Poon, Ada S. Y.},
        title = {Wirelessly powered, fully internal optogenetics for brain, spinal and peripheral circuits in mice},
      journal = {Nature Methods},
         year = {2015},
    publisher = {Nature Publishing Group},
       volume = {12},
        issue = {10},
        pages = {969-974},
     abstract = {To enable sophisticated optogenetic manipulation of neural circuits throughout the nervous system with limited disruption of animal behavior, light-delivery systems beyond fiber optic tethering and large, head-mounted wireless receivers are desirable. We report the development of an easy-to-construct, implantable wireless optogenetic device. Our smallest version (20 mg, 10 mm3) is two orders of magnitude smaller than previously reported wireless optogenetic systems, allowing the entire device to be implanted subcutaneously. With a radio-frequency (RF) power source and controller, this implant produces sufficient light power for optogenetic stimulation with minimal tissue heating (<1 [deg]C). We show how three adaptations of the implant allow for untethered optogenetic control throughout the nervous system (brain, spinal cord and peripheral nerve endings) of behaving mice. This technology opens the door for optogenetic experiments in which animals are able to behave naturally with optogenetic manipulation of both central and peripheral targets.},
}
@incollection{MoodyNIPS-89,
       author = {Moody, John E.},
        title = {Fast Learning in Multi-Resolution Hierarchies},
       editor = {Touretsky, David},
    booktitle = {Advances in Neural Information Processing 1},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1989,
}
@techreport{MoodyandDarkenTR-89,
       author = {Moody, John E. and Darken, Christian},
        title = {Fast Learning in Networks of Locally-Tuned Processing Units},
        issue = {654},
  institution = {Yale University Computer Science Department},
         year = 1989,
}
@article{MoonetalNATURE-12,
       author = {Moon, Tae Seok and Lou, Chunbo and Tamsir, Alvin and Stanton, Brynne C. and Voigt, Christopher A.},
        title = {Genetic programs constructed from layered logic gates in single cells},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = 491,
        issue = 7423,
         year = 2012,
        pages = {249-253},
     abstract = {Genetic programs function to integrate environmental sensors, implement signal processing algorithms and control expression dynamics. These programs consist of integrated genetic circuits that individually implement operations ranging from digital logic to dynamic circuits, and they have been used in various cellular engineering applications, including the implementation of process control in metabolic networks and the coordination of spatial differentiation in artificial tissues. A key limitation is that the circuits are based on biochemical interactions occurring in the confined volume of the cell, so the size of programs has been limited to a few circuits1, 7. Here we apply part mining and directed evolution to build a set of transcriptional AND gates in Escherichia coli. Each AND gate integrates two promoter inputs and controls one promoter output. This allows the gates to be layered by having the output promoter of an upstream circuit serve as the input promoter for a downstream circuit. Each gate consists of a transcription factor that requires a second chaperone protein to activate the output promoter. Multiple activator-chaperone pairs are identified from type III secretion pathways in different strains of bacteria. Directed evolution is applied to increase the dynamic range and orthogonality of the circuits. These gates are connected in different permutations to form programs, the largest of which is a 4-input AND gate that consists of 3 circuits that integrate 4 inducible systems, thus requiring 11 regulatory proteins. Measuring the performance of individual gates is sufficient to capture the behaviour of the complete program. Errors in the output due to delays (faults), a common problem for layered circuits, are not observed. This work demonstrates the successful layering of orthogonal logic gates, a design strategy that could enable the construction of large, integrated circuits in single cells.}
}
@article{MooneyCJP-57,
       author = {Craig M. Mooney},
        title = {Age in the development of closure ability in children},
      journal = {Canadian Journal of Psychology},
       volume = {11},
        issue = {4},
         year = {1957},
        pages = {219-226},
}
@incollection{Moore56,
       author = {Moore, Edward F.},
        title = {Gedanken experiments on Sequential Machines},
    booktitle = {Automata Studies},
    publisher = {Princeton University Press},
      address = {Princeton, New Jersey},
        pages = {129-153},
         year = 1956,
 comment      = {Originally published as Annals of Mathematical Studies, No. 34},
}
@article{MooreELECTRONICS-65,
       author = {Gordon Moore},
        title = {Cramming more components onto integrated circuits},
      journal = {Electronics Magazine},
       volume = {38},
        issue = {8},
         year = {1965},
}
@article{MooreLISTENER-37,
       author = {Henry Moore},
        title = {The Sculptor Speaks},
      journal = {Listener},
       volume = {18},
         year = {1937},
        pages = {338-340}
}
@techreport{MooreMSc,
       author = {Moore, Robert C.},
        title = {Reasoning from Incomplete Knowledge in a Procedural Deduction System},
        issue = {AI-TR-347},
  institution = {{MIT} {AI} Laboratory},
         year = 1975,
}
@incollection{MooreNIPS-93,
       author = {Moore, Andrew W.},
        title = {The Parti-game Algorithm for Variable Resolution Reinforcement Learning in Multidimensional State Spaces},
       editor = {Hanson, S. J. and Cowan, J. D. and Giles, C. L.},
    booktitle = {Advances in Neural Information Processing 5},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1993,
         page = {711-718},
}
@techreport{MoorePhD,
       author = {Moore, Robert C.},
        title = {Reasoning About Knowledge and Action},
        issue = {Technical Note 191},
  institution = {SRI International},
         year = 1980,
}
@article{MooreandAtkesonML-93,
       author = {Moore, Andrew W. and Atkeson, Christopher G.},
        title = {Prioritized Sweeping: Reinforcement Learning with Less Data and Less Time},
      journal = {Machine Learning},
       volume = 13,
        issue = 1,
        pages = {103-130},
         year = 1993,
}
@article{Mora-BermudezetalELIFE-16,
       author = {Mora-Berm{\'{u}}dez, Felipe and Badsha, Farhath  and Kanton, Sabina  and Camp, J. Gray  and Vernot, Benjamin  and K{\"{o}}hler, Kathrin  and Voigt, Birger  and Okita, Keisuke  and Maricic, Tomislav  and He, Zhisong  and Lachmann, Robert  and Paabo, Svante  and Treutlein, Barbara  and Huttner, Wieland B.},
        title = {Differences and similarities between human and chimpanzee neural progenitors during cerebral cortex development},
      journal = {eLife},
         year = {2016},
    publisher = {eLife Sciences Publications, Ltd},
       volume = {5},
        pages = {e18683},
     abstract = {Human neocortex expansion likely contributed to the remarkable cognitive abilities of humans. This expansion is thought to primarily reflect differences in proliferation versus differentiation of neural progenitors during cortical development. Here, we have searched for such differences by analysing cerebral organoids from human and chimpanzees using immunohistofluorescence, live imaging, and single-cell transcriptomics. We find that the cytoarchitecture, cell type composition, and neurogenic gene expression programs of humans and chimpanzees are remarkably similar. Notably, however, live imaging of apical progenitor mitosis uncovered a lengthening of prometaphase-metaphase in humans compared to chimpanzees that is specific to proliferating progenitors and not observed in non-neural cells. Consistent with this, the small set of genes more highly expressed in human apical progenitors points to increased proliferative capacity, and the proportion of neurogenic basal progenitors is lower in humans. These subtle differences in cortical progenitors between humans and chimpanzees may have consequences for human neocortex evolution.},
}
@incollection{MoranteandDesplanCSHL-11,
        title = {Dissecting and Staining Drosophila Optic Lobes},
       author = {Javier Morante and Claude Desplan},
    booktitle = {Drosophila Neurobiology: A Laboratory Manual},
       editor = {Bing Zhang and Marc R. Freeman and Scott Waddell},
    publisher = {{CSHL Press}},
      address = {Cold Spring Harbor, New York},
       volume = 2011,
        issue = 6,
         year = 2011,
        pages = {652-656},
     abstract = {The Drosophila visual system is composed of the retina and the optic lobes, which are the ganglia where photoreceptors project and initial processing of visual inputs occur. This protocol outlines procedures for dissecting the optic lobes from Drosophila larvae, pupae, and adults. It also describes methods for visualizing the anatomy of brain neural circuits by staining with fluorescent secondary antibodies and primary antibodies specific for various neuronal populations and architectural features.}
}
@article{MoranteandDesplanCURRENT-BIOLOGY-08,
       author = {Morante, J. and Desplan, C. },
    title = {{{T}he color-vision circuit in the medulla of {D}rosophila}},
   journal = {Current Biology},
    volume = {18},
    number = {8},
      year = {2008},
     pages = {553--565},
   abstract = {Color vision requires comparison between photoreceptors that are sensitive to different wavelengths of light. In Drosophila, this is achieved by the inner photoreceptors (R7 and R8) that contain different rhodopsins. Two types of comparisons can occur in fly color vision: between the R7 (UV sensitive) and R8 (blue- or green sensitive) photoreceptor cells within one ommatidium (unit eye) or between different ommatidia that contain spectrally distinct inner photoreceptors. Photoreceptors project to the optic lobes: R1-R6, which are involved in motion detection, project to the lamina, whereas R7 and R8 reach deeper in the medulla. This paper analyzes the neural network underlying color vision into the medulla. We reconstruct the neural network in the medulla, focusing on neurons likely to be involved in processing color vision. We identify the full complement of neurons in the medulla, including second-order neurons that contact both R7 and R8 from a single ommatidium, or contact R7 and/or R8 from different ommatidia. We also examine third-order neurons and local neurons that likely modulate information from second-order neurons. Finally, we present highly specific tools that will allow us to functionally manipulate the network and test both activity and behavior. This precise characterization of the medulla circuitry will allow us to understand how color vision is processed in the optic lobe of Drosophila, providing a paradigm for more complex systems in vertebrates.}
}
@article{MoranteandDesplanSCDB-04,
       author = {Javier Morante and Claude Desplan},
        title = {Building a projection map for photoreceptor neurons in the Drosophila optic lobes},
      journal = {Seminars in Cell \& Developmental Biology},
       volume = {15},
       number = {1},
        pages = {137-143},
         year = {2004},
     abstract = {The sensory tasks performed by the eye are diverse and complex. In Drosophila, the eye performs motion detection for navigation as well as detection of the quality of light (color and polarized light). Both types of inputs are processed separately, as different photoreceptors are specialized in these tasks and contact different target cell layers in the optic lobe. However, their respective outputs are likely to be integrated in higher brain centers. Here, we discuss the cell diversity and potential role of the several ganglia that form the fly optic lobe. We also discuss the power of modern genetic tools to provide the potential to trace the visual neural networks.}
}
@book{Moravec88,
       author = {Hans P. Moravec},
        title = {Mind Children: The Future of Robot and Human Intelligence},
    publisher = {Harvard University Press},
      address = {Cambridge, MA},
         year = 1988,
} 
@book{Moravec99,
       author = {Hans P. Moravec},
        title = {Robot: Mere Machine to Transcendent Mind},
    publisher = {Oxford University Press},
      address = {Oxford},
         year = 1999,
}
@inproceedings{MoravecandElfesRA-85,
       author = {Moravec, H. P. and A. Elfes},
        title = {High Resolution Maps from Wide Angle Sonar},
    booktitle = {{IEEE} International Conference on Robotics and Automation},
     location = {St. Louis, Missouri},
         year = 1985,
        pages = {138-145},
}
@incollection{MordatchetalNIPS-15,
        title = {Interactive Control of Diverse Complex Characters with Neural Networks},
       author = {Mordatch, Igor and Lowrey, Kendall and Andrew, Galen and Popovic, Zoran and Todorov, Emanuel V.},
    booktitle = {Advances in Neural Information Processing Systems 28},
       editor = {C. Cortes and N. D. Lawrence and D. D. Lee and M. Sugiyama and R. Garnett},
    publisher = {Curran Associates, Inc.},
         year = {2015},
        pages = {3132-3140},
     abstract = {We present a method for training recurrent neural networks to act as near-optimal feedback controllers. It is able to generate stable and realistic behaviors for a range of dynamical systems and tasks -- swimming, flying, biped and quadruped walking with different body morphologies. It does not require motion capture or task-specific features or state machines. The controller is a neural network, having a large number of feed-forward units that learn elaborate state-action mappings, and a small number of recurrent units that implement memory states beyond the physical system state. The action generated by the network is defined as velocity. Thus the network is not learning a control policy, but rather the dynamics under an implicit policy. Essential features of the method include interleaving supervised learning with trajectory optimization, injecting noise during training, training for unexpected changes in the task specification, and using the trajectory optimizer to obtain optimal feedback gains in addition to optimal actions.}
}
@book{MordfinASTM-02,
    editor = {Leonard Mordfin},
     title = {Handbook of Reference Data for Nondestructive Testing},
 publisher = {American Society for Testing \& Materials},
   address = {West Conshohocken, PA},
      year = 2002,
}
@article{Moreno-JimenezetalNATURE-MEDICINE-19,
       author = {Moreno-Jim{\'{e}}nez, Elena P. and Flor-Garc{\'{i}}a, Miguel and Terreros-Roncal, Julia and R{\'{a}}bano, Alberto and Cafini, Fabio and Pallas-Bazarra, Noem{\'{i}} and {\'{A}}vila, Jes{\'{u}}s and Llorens-Mart{\'{i}}n, Mar{\'{i}}n},
        title = {Adult hippocampal neurogenesis is abundant in neurologically healthy subjects and drops sharply in patients with Alzheimer's disease},
      journal = {Nature Medicine},
         year = {2019},
       volume = {25},
        issue = {4},
        pages = {554-560},
     abstract = {The hippocampus is one of the most affected areas in Alzheimer's disease (AD). Moreover, this structure hosts one of the most unique phenomena of the adult mammalian brain, namely, the addition of new neurons throughout life. This process, called adult hippocampal neurogenesis (AHN), confers an unparalleled degree of plasticity to the entire hippocampal circuitry. Nonetheless, direct evidence of AHN in humans has remained elusive. Thus, determining whether new neurons are continuously incorporated into the human dentate gyrus (DG) during physiological and pathological aging is a crucial question with outstanding therapeutic potential. By combining human brain samples obtained under tightly controlled conditions and state-of-the-art tissue processing methods, we identified thousands of immature neurons in the DG of neurologically healthy human subjects up to the ninth decade of life. These neurons exhibited variable degrees of maturation along differentiation stages of AHN. In sharp contrast, the number and maturation of these neurons progressively declined as AD advanced. These results demonstrate the persistence of AHN during both physiological and pathological aging in humans and provide evidence for impaired neurogenesis as a potentially relevant mechanism underlying memory deficits in AD that might be amenable to novel therapeutic strategies.},
}
@article{MorganSA-93,
        title = {Risk Analysis and Management},
       author = {Morgan, M. Granger},
      journal = {Scientific American},
       volume = {July},
        pages = {32-41},
         year = 1993,
}
@article{MorganetalCELL-16,
       author = {Morgan, Josh Lyskowski and Berger, Daniel Raimund and Wetzel, Arthur Willis and Lichtman, Jeff William},
        title = {The Fuzzy Logic of Network Connectivity in Mouse Visual Thalamus},
      journal = {Cell},
    publisher = {Elsevier},
       volume = {165},
        issue = {1},
         year = {2016},
        pages = {192-206},
     abstract = {In an attempt to chart parallel sensory streams passing through the visual thalamus, we acquired a 100-trillion-voxel electron microscopy (EM) dataset and identified cohorts of retinal ganglion cell axons (RGCs) that innervated each of a diverse group of postsynaptic thalamocortical neurons (TCs). Tracing branches of these axons revealed the set of TCs innervated by each RGC cohort. Instead of finding separate sensory pathways, we found a single large network that could not be easily subdivided because individual RGCs innervated different kinds of TCs and different kinds of RGCs co-innervated individual TCs. We did find conspicuous network subdivisions organized on the basis of dendritic rather than neuronal properties. This work argues that, in the thalamus, neural circuits are not based on a canonical set of connections between intrinsically different neuronal types but, rather, may arise by experience-based mixing of different kinds of inputs onto individual postsynaptic cells.},
}
@inproceedings{MorgensternAAAI-88,
       author = {Morgenstern, Leora and Stein, Lynn Andrea},
        title = {Why Things Go Wrong: A Formal Theory of Causal Reasoning},
    booktitle = {Proceedings AAAI-88},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
         year = 1988,
        pages = {518-523},
}
@inproceedings{MorgensternDARPA-87,
       author = {Morgenstern, Leora},
        title = {Replanning},
    booktitle = {Proceedings of the DARPA Knowledge-Based Planning Workshop},
     location = {Austin, Texas},
 organization = {DARPA},
         year = 1988,
}
@phdthesis{MorgensternPhD,
       author = {Morgenstern, Leora},
        title = {Foundations of a Logic of Knowledge, Action, and Communication},
       school = {New York University},
         year = 1988,
}
@inproceedings{MorgensternTARK-86,
       author = {Morgenstern, Leora},
        title = {A First Order Theory of Planning, Knowledge, and Action},
     crossref = {HalpernTARK-86a},
        pages = {83-98},
}
@article{MorietalPAMI-05,
       author = {Greg Mori and Serge Belongie and Jitendra Malik},
        title = {Efficient Shape Matching Using Shape Contexts},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
       volume = 27,
        issue = 11,
        pages = {1832-1837},
         year = 2005,
}        
@inproceedings{MorinandBengioAISTATS-05,
       author = {Frederic Morin and Yoshua Bengio},
        title = {Hierarchical probabilistic neural network language model},
    booktitle = {{AI} \& Statistics},
         year = 2005,
        pages = {246-252},
}
@article{MormannandKoch07,
       author = {Mormann, F. and Koch, C.},
        title = {{N}eural correlates of consciousness},
         year = {2007},
      journal = {Scholarpedia},
       volume = {2},
       number = {12},
        pages = {1740},
     abstract = {In contrast to diffuse cortical damage, relatively discrete bilateral injuries to midline (paramedian) subcortical structures can also cause a complete loss of consciousness. These structures are therefore part of the enabling factors that control the level of brain arousal (as determined by metabolic or electrical activity) and that are needed for any form of consciousness to occur. One such example is the heterogeneous collection of more than two dozen (on each side) of nuclei in the upper brainstem (pons, midbrain and in the posterior hypothalamus) collectively referred to as the reticular activating system (RAS). These nuclei - three-dimensional collections of neurons with their own cytoarchitecture and neurochemical identity - release distinct neuromodulators such as acetylcholine, noradrenaline/norepinephrine, serotonin, histamine and orexin/hypocretin. Their axons project widely throughout the brain (Figure 3). These neuromodulators control the excitability of thalamus and forebrain and mediate the alternation between wakefulness and sleep as well as the general level of both behavioral and brain arousal. Acute lesions of nuclei in the RAS can result in loss of consciousness and coma. However, eventually the excitability of the thalamus and forebrain can recover and consciousness can return (Villablanca 2004). Another enabling factor for consciousness are the five or more intralaminar nuclei (ILN) of the thalamus. These receive input from many brainstem nuclei and project strongly to the basal ganglia and, in a more distributed manner, into layer I of much of neocortex. Comparatively small (1 cm3 or less), bilateral lesions in the thalamic ILN completely knock out all awareness (Bogen 1995).}
}
@article{MorrisetalNEURON-06,
       author = {Richard G.M. Morris and Jennifer Inglis and James A. Ainge and Henry J. Olverman and Jane Tulloch and Yadin Dudai and Paul A.T. Kelly},
        title = {Memory Reconsolidation: Sensitivity of Spatial Memory to Inhibition of Protein Synthesis in Dorsal Hippocampus during Encoding and Retrieval},
      journal = {Neuron},
       volume = {50},
       number = {3},
        pages = {479-489},
         year = {2006},
     abstract = {Reconsolidation is a putative neuronal process in which the retrieval of a previously consolidated memory returns it to a labile state that is once again subject to stabilization. This study explored the idea that reconsolidation occurs in spatial memory when animals retrieve memory under circumstances in which new memory encoding is likely to occur. Control studies confirmed that intrahippocampal infusions of anisomycin inhibited protein synthesis locally and that the spatial training protocols we used are subject to overnight protein synthesis-dependent consolidation. We then compared the impact of anisomycin in two conditions: when memory retrieval occurred in a reference memory task after performance had reached asymptote over several days; and after a comparable extent of training of a delayed matching-to-place task in which new memory encoding was required each day. Sensitivity to intrahippocampal anisomycin was observed only in the protocol involving new memory encoding at the time of retrieval.}
}
@article{MorrisetalNLM-2013,
        title = {Microglia: A new frontier for synaptic plasticity, learning and memory, and neurodegenerative disease research},
       author = {Gary P. Morris and Ian A. Clark and Raphael Zinn and Bryce Vissel},
      journal = {Neurobiology of Learning and Memory},
       volume = {105},
         year = {2013},
        pages = {40-53},
     abstract = {We focus on emerging roles for microglia in synaptic plasticity, cognition and disease. We outline evidence that ramified microglia, traditionally thought to be functionally "resting" (i.e. quiescent) in the normal brain, in fact are highly dynamic and plastic. Ramified microglia continually and rapidly extend processes, contact synapses in an activity and experience dependent manner, and play a functionally dynamic role in synaptic plasticity, possibly through release of cytokines and growth factors. Ramified microglial also contribute to structural plasticity through the elimination of synapses via phagocytic mechanisms, which is necessary for normal cognition. Microglia have numerous mechanisms to monitor neuronal activity and numerous mechanisms also exist to prevent them transitioning to an activated state, which involves retraction of their surveying processes. Based on the evidence, we suggest that maintaining the ramified state of microglia is essential for normal synaptic and structural plasticity that supports cognition. Further, we propose that change of their ramified morphology and function, as occurs in inflammation associated with numerous neurological disorders such as Alzheimer's and Parkinson's disease, disrupts their intricate and essential synaptic functions. In turn altered microglia function could cause synaptic dysfunction and excess synapse loss early in disease, initiating a range of pathologies that follow. We conclude that the future of learning and memory research depends on an understanding of the role of non-neuronal cells and that this should include using sophisticated molecular, cellular, physiological and behavioural approaches combined with imaging to causally link the role of microglia to brain function and disease including Alzheimer's and Parkinson's disease and other neuropsychiatric disorders.}
}
@article{MorrisonetalNC-06,
        title = {Advancing the Boundaries of High-Connectivity Network Simulation with Distributed Computing},
       author = {Abigail Morrison and Carsten Mehring and Theo Geisel and Ad Aertsen and Markus Diesmann},
      journal = {Neural Computation},
       volume = 17,
        issue = 8,
        pages = {1776-1801},
         year = 2005,
}
@article{MoscovitchetalARP-16,
       author = {Moscovitch, M. and Cabeza, R. and Winocur, G. and Nadel, L.},
        title = {Episodic Memory and Beyond: {T}he Hippocampus and Neocortex in Transformation},
      journal = {Annual Review of Psychology},
         year = {2016},
       volume = {67},
        pages = {105-134},
     abstract = {The last decade has seen dramatic technological and conceptual changes in research on episodic memory and the brain. New technologies, and increased use of more naturalistic observations, have enabled investigators to delve deeply into the structures that mediate episodic memory, particularly the hippocampus, and to track functional and structural interactions among brain regions that support it. Conceptually, episodic memory is increasingly being viewed as subject to lifelong transformations that are reflected in the neural substrates that mediate it. In keeping with this dynamic perspective, research on episodic memory (and the hippocampus) has infiltrated domains, from perception to language and from empathy to problem solving, that were once considered outside its boundaries. Using the component process model as a framework, and focusing on the hippocampus, its subfields, and specialization along its longitudinal axis, along with its interaction with other brain regions, we consider these new developments and their implications for the organization of episodic memory and its contribution to functions in other domains.}
}
@article{MoseretalNEUROSCIENCE-08,
       author = {Moser, E. I. and Kropff, E. and Moser, M. B.},
        title = {Place cells, grid cells, and the brain's spatial representation system},
      journal = {Annual Reviews Neuroscience},
         year = {2008},
       volume = {31},
        pages = {69-89},
     abstract = {More than three decades of research have demonstrated a role for hippocampal place cells in representation of the spatial environment in the brain. New studies have shown that place cells are part of a broader circuit for dynamic representation of self-location. A key component of this network is the entorhinal grid cells, which, by virtue of their tessellating firing fields, may provide the elements of a path integration-based neural map. Here we review how place cells and grid cells may form the basis for quantitative spatiotemporal representation of places, routes, and associated experiences during behavior and in memory. Because these cell types have some of the most conspicuous behavioral correlates among neurons in nonsensory cortical systems, and because their spatial firing structure reflects computations internally in the system, studies of entorhinal-hippocampal representations may offer considerable insight into general principles of cortical network dynamics.}
}
@article{MouetalCoRR-14,
       author = {Lili Mou and Ge Li and Yuxuan Liu and Hao Peng and Zhi Jin and Yan Xu and Lu Zhang},
        title = {Building Program Vector Representations for Deep Learning},
      journal = {CoRR},
       volume = {arXiv:1409.3358},
         year = {2014},
     abstract = {Deep learning has made significant breakthroughs in various fields of artificial intelligence. Advantages of deep learning include the ability to capture highly complicated features, weak involvement of human engineering, etc. However, it is still virtually impossible to use deep learning to analyze programs since deep architectures cannot be trained effectively with pure back propagation. In this pioneering paper, we propose the "coding criterion" to build program vector representations, which are the premise of deep learning for program analysis. Our representation learning approach directly makes deep learning a reality in this new field. We evaluate the learned vector representations both qualitatively and quantitatively. We conclude, based on the experiments, the coding criterion is successful in building program representations. To evaluate whether deep learning is beneficial for program analysis, we feed the representations to deep neural networks, and achieve higher accuracy in the program classification task than "shallow" methods, such as logistic regression and the support vector machine. This result confirms the feasibility of deep learning to analyze programs. It also gives primary evidence of its success in this new field. We believe deep learning will become an outstanding technique for program analysis in the near future.}
} 
@article{MouhotandVillaniCoRR-09,
       author = {Mouhot, Cl\`{e}ment and Villani, C\`{e}dric},
        title = {On {L}andau damping},
      journal = {CoRR},
       volume = {arXiv:0904.2760},
         year = 2009,
     abstract = {Going beyond the linearized study has been a longstanding problem in the theory of Landau damping. In this paper we establish exponential Landau damping in analytic regularity. The damping phenomenon is reinterpreted in terms of transfer of regularity between kinetic and spatial variables, rather than exchanges of energy; phase mixing is the driving mechanism. The analysis involves new families of analytic norms, measuring regularity by comparison with solutions of the free transport equation; new functional inequalities; a control of nonlinear echoes; sharp scattering estimates; and a Newton approximation scheme. Our results hold for any potential no more singular than Coulomb or Newton interaction; the limit cases are included with specific technical effort. As a side result, the stability of homogeneous equilibria of the nonlinear Vlasov equation is established under sharp assumptions. We point out the strong analogy with the KAM theory, and discuss physical implications.},
} 
@incollection{Mountcastle1978,
        title = {An organizing principle for cerebral function: the unit model and the distributed system},
       author = {Mountcastle, Vernon},
    booktitle = {The Mindful Brain},
       editor = {Gerald Edelman and Vernon Mountcastle},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1978,
        pages = {7-50},
}
@book{Mountcastle98,
       author = {Vernon B. Mountcastle},
        title = {Perceptual Neuroscience: The Cerebral Cortex},
      address = {Cambridge, MA},
    publisher = {Harvard University Press},
         year = 1998,
} 
@article{MountcastleBRAIN-97,
       author = {Vernon B. Mountcastle},
        title = {The columnar organization of the neocortex},
      journal = {Brain},
    publisher = {Oxford University Press},
       volume = 120,
        issue = 4,
        pages = {701-722},
         year = 1997,
}
@article{MountcastleCC-03,
       author = {Vernon B. Mountcastle},
        title = {Introduction to the Special Issue on Computation in Cortical Columns},
      journal = {Cerebral Cortex},
         year = 2003,
       volume = 13,
        issue = 1,
        pages = {2-4},
}
@article{MovshonandNewsomeJoN-96,
       author = {Movshon, J. Anthony and Newsome, William T.},
        title = {Visual Response Properties of Striate Cortical Neurons Projecting to Area MT in Macaque Monkeys},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {16},
       number = {23},
         year = {1996},
        pages = {7733-7741},
     abstract = {We have previously shown that some neurons in extrastriate area MT are capable of signaling the global motion of complex patterns; neurons randomly sampled from V1, on the other hand, respond only to the motion of individual oriented components. Because only a small fraction of V1 neurons projects to MT, we wished to establish the processing hierarchy more precisely by studying the properties of those neurons projecting to MT, identified by antidromic responses to electrical stimulation of MT. The neurons that project from V1 to MT were directionally selective and, like other V1 neurons, responded only to the motion of the components of complex patterns. The projection neurons were predominantly "special complex", responsive to a broad range of spatial and temporal frequencies, and sensitive to very low stimulus contrasts. The projection neurons thus comprise a homogeneous and highly specialized subset of V1 neurons, consistent with the notion that V1 acts as clearing house of basic visual measurements, distributing information appropriately to higher cortical areas for specialized analysis.},
}
@article{MrksicetalCoRR-15,
       author = {Nikola Mrksic and Diarmuid {\'{O}} S{\'{e}}aghdha and Blaise Thomson and Milica Gasic and Pei{-}hao Su and David Vandyke and Tsung{-}Hsien Wen and Steve J. Young},
        title = {Multi-domain Dialog State Tracking using Recurrent Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1506.07190},
         year = {2015},
     abstract = {Dialog state tracking is a key component of many modern dialog systems, most of which are designed with a single, well-defined domain in mind. This paper shows that dialog data drawn from different dialog domains can be used to train a general belief tracking model which can operate across all of these domains, exhibiting superior performance to each of the domain-specific models. We propose a training procedure which uses out-of-domain data to initialise belief tracking models for entirely new domains. This procedure leads to improvements in belief tracking performance regardless of the amount of in-domain data available for training the model.}
}
@article{MrowcaetalCoRR-18,
       author = {Damian Mrowca and Chengxu Zhuang and Elias Wang and Nick Haber and Li Fei{-}Fei and Joshua B. Tenenbaum and Daniel L. K. Yamins},
        title = {Flexible Neural Representation for Physics Prediction},
      journal = {CoRR},
       volume = {arXiv:1806.08047},
         year = {2018},
     abstract = {Humans have a remarkable capacity to understand the physical dynamics of objects in their environment, flexibly capturing complex structures and interactions at multiple levels of detail. Inspired by this ability, we propose a hierarchical particle-based object representation that covers a wide variety of types of three-dimensional objects, including both arbitrary rigid geometrical shapes and deformable materials. We then describe the Hierarchical Relation Network (HRN), an end-to-end differentiable neural network based on hierarchical graph convolution, that learns to predict physical dynamics in this representation. Compared to other neural network baselines, the HRN accurately handles complex collisions and nonrigid deformations, generating plausible dynamics predictions at long time scales in novel settings, and scaling to large scene configurations. These results demonstrate an architecture with the potential to form the basis of next-generation physics predictors for use in computer vision, robotics, and quantitative cognitive science.}
}
@article{MuellerFiN-12,
       author = {Mueller, Thomas},
        title = {What is the Thalamus in Zebrafish?},
      journal = {Frontiers in Neuroscience},
    publisher = {Frontiers Research Foundation},
       volume = {6},
        issue = {64},
         year = {2012},
     abstract = {Current research on the thalamus and related structures in the zebrafish diencephalon identifies an increasing number of both neurological structures and ontogenetic processes as evolutionary conserved between teleosts and mammals. The patterning processes, for example, which during the embryonic development of zebrafish form the thalamus proper appear largely conserved. Yet also striking differences between zebrafish and other vertebrates have been observed, particularly when we look at mature and histologically differentiated brains. A case in point is the migrated preglomerular complex of zebrafish which evolved only within the lineage of ray-finned fish and has no counterpart in mammals or tetrapod vertebrates. Based on its function as a sensory relay station with projections to pallial zones, the preglomerular complex has been compared to specific thalamic nuclei in mammals. However, no thalamic projections to the zebrafish dorsal pallium, which corresponds topologically to the mammalian isocortex, have been identified. Merely one teleostean thalamic nucleus proper, the auditory nucleus, projects to a part of the dorsal telencephalon, the pallial amygdala. Studies on patterning mechanisms identify a rostral and caudal domain in the embryonic thalamus proper. In both, teleosts and mammals, the rostral domain gives rise to GABAergic neurons, whereas glutamatergic neurons originate in the caudal domain of the zebrafish thalamus. The distribution of GABAergic derivatives in the adult zebrafish brain, furthermore, revealed previously overlooked thalamic nuclei and redefined already established ones. These findings require some reconsideration regarding the topological origin of these adult structures. In what follows, I discuss how evolutionary conserved and newly acquired features of the developing and adult zebrafish thalamus can be compared to the mammalian situation.}
}
@book{MuellerandWullimann2015,
        title = {Atlas of Early Zebrafish Brain Development: A Tool for Molecular Neurogenetics},
       author = {Mueller, T. and Wullimann, M.},
    publisher = {Elsevier Science},
          url = {https://books.google.com/books?id=5dScBAAAQBAJ},
         year = {2015},
}
@article{MuhlethaleretalEJN-93,
       author = {M. M\"{u}hlethaler, M. de Curtis, K. Walton and R. Llin\`{a}s},
        title = {The Isolated and Perfused Brain of the Guinea-pig in Vitro},
      journal = {European Journal of Neuroscience},
       volume = 5,
         year = 1993,
        pages = {915-926},
}
@article {MuhlfriedeletalEJN-07,
       author = {M\"{u}hlfriedel, Sven and Kirsch, Friederike and Gruss, Peter and Chowdhury, Kamal and Stoykova, Anastassia},
        title = {Novel genes differentially expressed in cortical regions during late neurogenesis},
      journal = {European Journal of Neuroscience},
       volume = 26,
        issue = 1,
    publisher = {Blackwell Publishing Ltd},
        pages = {33-50},
         year = 2007,
     abstract = {Differential gene expression across the embryonic cerebral cortex is assumed to play a role in the subdivision of the cortex into distinct areas with specific morphology, physiology and function   In a search for genes that may be involved in the cortical regionalization during late neurogenesis in mouse, we performed an extensive in-situ expression analysis at embryonic day E16 and E18 The examined candidate genes were selected beforehand by microarray  screen by virtue of their preferential expression in the anlagen of the motor, somatosensory, visual and cingulate cortices or hippocampus. We present new information about graded or regionally enriched expression of 25 genes (nine of which are novel genes  across the mouse embryonic cortex, in progenitor cells as well as in the cortical plate. The established differential expression of most of these genes is persistent at both stages studied, suggesting that their expression is regulated by an intrinsic programme. For some of the genes, the concept of intrinsic regulation is further substantiated by the high similarity of the reported expression patterns at E16 and E18 and published data from earlier stages. Few genes with robust expression in the E16 caudal cortex showed a more restricted pattern at E18, possibly because of their response to extrinsic cues. In addition, several genes appeared to be suitable novel markers for amygdalar and diencephalic nuclei. Taken together   our findings reveal novel molecular partitions of the late mouse cortex that are in accordance with the model of a leading role of intrinsic mechanisms in cortical arealization.},
}
@book{Mukherjee2016,
       author = {Siddhartha Mukherjee},
        title = {{The Gene: An Intimate History}},
    publisher = {Simon and Schuster Incorporated},
         year = {2016},
}
@article{MukherjiandSloviterJB-87,
    author = {B. Mukherji and H. A. Sloviter},
     title = {Metabolism of an isolated brain perfused with perfluoro blood substitute},
   journal = {Journal of Bioscience},
    volume = 11,
     issue = {1-4},
      year = 1987,
     pages = {23-33},
} 
@article{Mukhopadhyay,
   author =  {Mukhopadhyay, Uttam and Stephens, Larry and Huhns, Michael and
              Bonnell, Ronald},
   title =   {An Intelligent System for Document Retrieval in Distributed
              Office Environments},
   journal = {Journal of the American Society for Information Science},
   volume =  37,
   year =    1986,
   pages =   {123-135},
}
@article{MuldoonetalNSR-16,
       author = {Muldoon, Sarah Feldt and Bridgeford, Eric W. and Bassett, Danielle S.},
        title = {Small-World Propensity and Weighted Brain Networks},
      journal = {Scientific Reports},
         year = {2016},
       volume = {6},
        pages = {22057},
     abstract = {Quantitative descriptions of network structure can provide fundamental insights into the function of interconnected complex systems. Small-world structure, diagnosed by high local clustering yet short average path length between any two nodes, promotes information flow in coupled systems, a key function that can differ across conditions or between groups. However, current techniques to quantify small-worldness are density dependent and neglect important features such as the strength of network connections, limiting their application in real-world systems. Here, we address both limitations with a novel metric called the Small-World Propensity (SWP). In its binary instantiation, the SWP provides an unbiased assessment of small-world structure in networks of varying densities. We extend this concept to the case of weighted brain networks by developing (i) a standardized procedure for generating weighted small-world networks, (ii) a weighted extension of the SWP, and (iii) a method for mapping observed brain network data onto the theoretical model. In applying these techniques to compare real-world brain networks, we uncover the surprising fact that the canonical biological small-world network, the C. elegans neuronal network, has strikingly low SWP. These metrics, models, and maps form a coherent toolbox for the assessment and comparison of architectural properties in brain networks.}
}
@book{Muller2010,
      title = {Information Retrieval for Music and Motion},
     author = {Meinard M\"{u}ller}, 
  publisher = {Springer-Verlag},
       year = 2010,
}
@article{MulleretalCHEMPHYSCHEM-12,
       author = {Muller, T. and Schumann, C. and Kraegeloh, A.},
        title = {{STED} microscopy and its applications: new insights into cellular processes on the nanoscale},
      journal = {Chemphyschem},
       volume = {13},
       number = {8},
         year = {2012},
        pages = {1986-2000},
     abstract = {For about a decade, superresolution fluorescence microscopy has been advancing steadily, maturing from the proof-of-principle stage to routine application. Of the various techniques, STED (stimulated emission depletion) microscopy was the first to break the diffraction barrier. Today, it is a prominent and versatile form of superresolution light microscopy. STED microscopy has shed a sharper light on numerous topics in cell biology, but also in material sciences. Both disciplines extend into the nanometer range, making detailed studies of structural and functional relationships difficult or even impossible to achieve using diffraction-limited microscopy. With recent advancements like spectral multiplexing or live-cell imaging, STED microscopy makes nanoscale materials and components of the cell accessible for fluorescence-based investigations. With multicolor superresolution imaging, even the interactions between biological and engineered nanostructures can be studied in detail. This review gives an introduction into the working principle of STED microscopy, provides a detailed overview of recent advancements and new techniques implemented for use with STED microscopy and shows how these have been applied in the life sciences and nanotechnologies.}
}
@article{MullisSCIAM-90,
    author = {Mullis, K. B.},
     title = {The unusual origin of the polymerase chain reaction},
   journal = {Scientific American},
      year = 1990,
    volume = 262,
     issue = 4,
     pages = {56-61},
}
@article{MulveyandVladimirouMS-92,
   author =  {Mulvey, John M. and Vladimirou, Hercules},
   title =   {Stochastic Network Programming for Financial Planning Problems},
   journal = {Management Science},
   year =    1992,
}
@incollection{Mumford94,
   author = {Mumford, David},
    title = {Neuronal architectures for pattern-theoretic problems},
booktitle = {Large Scale Neuronal Theories of the Brain},
  editors = {Koch, C. and Davis, J.L.},
publisher = {MIT Press},
    pages = {125-152},
  comment = {Mumford94.pdf},
     year = 1994,
}
@article{MumfordBC-91,
  author = {Mumford, David},
   title = {On the computational architecture of the neocortex {I}:
            The role of the thalamo-cortical loop},
 journal = {Biological Cybernetics},
  volume = 65,
   pages = {135-145},
    year = 1991,
}
@article{MumfordBC-92,
  author = {Mumford, David},
   title = {On the computational architecture of the neocortex {II}:
            The role of cortico-cortical loops},
 journal = {Biological Cybernetics},
  volume = 66,
   pages = {241-251},
    year = 1992,
}
@inproceedings{MumfordICM-02,
    author = {Mumford, David},
     title = {Pattern Theory: the Mathematics of Perception},
 booktitle = {Proceedings of the International Congress of Mathematicians: Volume {III}},
  location = {Beijing, China},
   comment = {MumfordICM-02.pdf},
      year = 2002,
}
@inproceedings{MuralietalFSE-17,
       author = {Murali, Vijayaraghavan and Chaudhuri, Swarat and Jermaine, Chris},
        title = {Bayesian Specification Learning for Finding API Usage Errors},
    booktitle = {Proceedings of the 2017 11th Joint Meeting on Foundations of Software Engineering},
     location = {Paderborn, Germany},
         year = {2017},
        pages = {151-162},
}
@inproceedings{MuralietalICLR-18,
        title = {Neural Sketch Learning for Conditional Program Generation},
       author = {Vijayaraghavan Murali and Letao Qi and Swarat Chaudhuri and Chris Jermaine},
    booktitle = {International Conference on Learning Representations},
         year = {2018},
     abstract = {We study the problem of generating source code in a strongly typed, Java-like programming language, given a label (for example a set of API calls or types) carrying a small amount of information about the code that is desired. The generated programs are expected to respect a `"realistic" relationship between programs and labels, as exemplified by a corpus of labeled programs available during training. Two challenges in such *conditional program generation* are that the generated programs must satisfy a rich set of syntactic and semantic constraints, and that source code contains many low-level features that impede learning.  We address these problems by training a neural generator not on code but on *program sketches*, or models of program syntax that abstract out names and operations that do not generalize across programs. During generation, we infer a posterior distribution over sketches, then concretize samples from this distribution into type-safe programs using combinatorial techniques.  We implement our ideas in a system for generating API-heavy Java code, and show that it can often predict the entire body of a method given just a few API calls or data types that appear in the method.}
}
@book{Murphy02,
   author = {Gregory L. Murphy},
    title = {The Big Book of Concepts},
publisher = {MIT Press},
  address = {Cambridge, MA},
     year = 2002,
}
@article{MurphyCSS-01,
  author = {Kevin Murphy},
   title = {The {B}ayes {N}et {T}oolbox for {M}atlab},
 journal = {Computing Science and Statistics},
  volume = {33},
    year = 2001,
}
@incollection{MurphyECS-02,
     author = {Kevin Murphy},
      title = {Learning {M}arkov Processes},
  booktitle = {The Encyclopedia of Cognitive Science},
     editor = {L. Nadel, {\emph{et al}}},
  publisher = {Macmillan Company},
       year = 2002,
}
@phdthesis{MurphyPhD,
  author = {Kevin P. Murphy},
   title = {Dynamic Bayesian Networks: Representation, Inference and Learning},
  school = {University of California, Berkeley, Computer Science Division},
    year = 2002,
}
@misc{MurphyUBC-TR-95,
       author = {Kevin P. Murphy},
        title = {Learning Finite Automata},
         year = {1995},
 howpublished = {UBC Computer Science Technical Report},
}
@incollection{MurphyandPaskinNIPS-01,
    author = {K. Murphy and M. Paskin},
     title = {Linear time inference in hierarchical {HMM}s},
 booktitle = {Advances in Neural Information Processing Systems},
    volume = 13,
    editor = {Leen, Todd K. and Dietterich, Thomas G. and Tresp, Volker},
 publisher = {MIT Press},
   address = {Cambridge, MA},
      year = 2001,
}
@incollection{MurphyetalNIPS-03,
     author = {K. Murphy and A. Torralba and W. Freeman},
      title = {Using the forest to see the tree: a graphical model relating features, objects and the scenes},
  booktitle = {Advances in Neural Information Processing Systems 16},
  publisher = {MIT Press},
     editor = {Sebastian Thrun and Lawrence Saul},
       year = 2004,
}
@inproceedings{MurphyetalUAI-00,
    author = {Kevin Murphy and Yair Weiss and Michael Jordan},
     title = {Loopy-belief Propagation for Approximate Inference: An Empirical Study},
 booktitle = {Proceedings of the 16th Conference on Uncertainty in
              Artificial Intelligence},
 publisher = {Morgan Kaufmann},
     pages = {467-475},
      year = 2000,
}
@article{MurrayetalCC-03,
   title = {Processing Shape, Motion and Three-dimensional Shape-from-motion in the Human Cortex},
  author = {Scott O. Murray and Bruno A. Olshausen and David L. Woods},
 journal = {Cerebral Cortex},
  volume = 13, 
   issue = 5, 
     url = {http://cercor.oxfordjournals.org/cgi/reprint/13/5/508},
   pages = {508-516},
    year = 2003,
} 
@inproceedings{MuscettolaandSmithIJCAI-87,
       author = {Muscettola, Nicola and Smith, Steve},
        title = {A Probabilistic Framework for Resource-Constrained Multiagent Planning},
    booktitle = {Proceedings {IJCAI} 10},
     location = {Milan, Italy},
 organization = {IJCAII},
         year = 1987,
} 
@article{MuslineretalIEEE-95,
       author = {Musliner, D. and Hendler, J. and Agrawala, A. and Durfee, E. and Strosnider, J.},
        title = {The Challenges of Real-Time {AI}},
      journal = {{IEEE} Computer},
       volume = 28,
        issue = 1,
        pages = {58-67},
         year = 1995,
}
@inproceedings{MutchandLoweCVPR-06,
       author = {Mutch, Jim and Lowe, David G.},
        title = {Multiclass Object Recognition with Sparse, Localized Features},
    booktitle = {Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
      address = {Washington, DC, USA},
        pages = {11-18},
         year = 2006,
}
@article{MutchandLoweIJCV-08,
     author = {Mutch, Jim and Lowe, David G.},
      title = {Object Class Recognition and Localization Using Sparse Features with Limited Receptive Fields},
    journal = {International Journal on Computer Vision},
     volume = {80},
      issue = {1},
       year = {2008},
      pages = {45-57},
  publisher = {Kluwer Academic Publishers},
    address = {Hingham, MA, USA},
}
@techreport{MutchetalTR-10,
     author = {Jim Mutch and Ulf Knoblich and Tomaso Poggio},
      title = {{CNS}: a {GPU}-based framework for simulating cortically-organized networks},
      issue = {MIT-CSAIL-TR-2010-013 / CBCL-286},
institution = {Massachusetts Institute of Technology},
    address = {Cambridge, MA},
      month = {February},
       year = 2010
}
@article{NadeauetalEMBS-06,
     title = {Quantum dot-FRET systems for imaging of neuronal action potentials},
    author = {Nadeau, J.L. and Clarke, S.J. and Hollmann, C.A. and Bahcheli, D.M.},
   journal = {{IEEE} Engineering in Medicine \& Biology Society Conference},
 publisher = {{IEEE} Computer Society},
    volume = 1,
      year = 2006,
  abstract = {Fluorescent semiconductor quantum dots (QDs) can act as energy donors or acceptors with a wide variety of environmentally-sensitive molecules. Conjugation of a single QD to a select number of the selected molecule can optimize the range of sensitivity for a given application, and the relatively large size of the QDs allows them to be tracked individually in cells. Using QDs as FRET acceptors, we have created first-generation sensors for membrane potential which shows good signal to noise and time resolution, but prohibitive toxicity. The challenges of delivery, calibration, and toxicity and plans for improvement of the sensors are presented, in the context of the eventual aim of monitoring membrane potential in a cultured motor neuron model of amyotrophic lateral sclerosis.}
}
@article{NagaoandKitazawaBRAIN-NERVE-08,
       author = {Nagao, S. and Kitazawa, H.},
        title = {Role of the cerebellum in the acquisition and consolidation of motor memory},
      journal = {Brain Nerve},
       volume = {60},
       number = {7},
         year = {2008},
        pages = {783-790},
     abstract = {Most of our motor skills are acquired through learning. Although the cerebellum is established as the critical site for motor learning, the precise mechanism underlying acquisition and consolidation of motor memory is a topic of on-going debate. We examined the role of the cerebellum in acquisition and storage of motor memory by using adaptations of horizontal vestibulo-ocular reflex (VOR) and optokinetic response (OKR) eye movements in mice and monkeys. A training of the mice to 1h rapid screen oscillation induced an increase in the gains of OKR that declined within 24h. However, when the mice were subjected to 1h of such training successively for a week, a long-term increase occurred in the gains of OKR for 2-3 continuous weeks. We examined the role of the cerebellum in these short (day-long) and long (week-long) -term adaptations. Our pharmacological and electrophysiological experiments revealed that the memory trace of day-long adaptation is located within the cerebellar cortex, while that of week-long adaptation is maintained in the vestibular nuclei where the cerebellar cortex issues its outputs. We obtained the same results in the experiments conducted in monkeys using of adaptations of VOR gains. We further revealed that long-term depression (LTD) of parallel fiber-Purkinje cell synapses plays a critical role in both the day- and week-long adaptations. These results suggest that the trace of motor memory is initially acquired in the cerebellar cortex, and later transferred to the cerebellar/vestibular nuclei for consolidation. LTD plays an essential role in both the acquisition and consolidation of motor memory.}
}
@book{Nagel1989godel,
        title = {G\"{o}del's Proof},
       author = {Nagel, Ernest and Newman, James R.},
         year = {1989},
    publisher = {Routledge}
}
@book{NagelandNewman1958godelproof,
        title = {G\"{o}del's Proof},
       author = {Nagel, Ernest and Newman, James R.},
         year = {1958},
    publisher = {New York University Press}
}
@article{NagumoetalIRE-62,
       author = {Nagumo J. and Arimoto S. and Yoshizawa S.},
        title = {An active pulse transmission line simulating nerve axon},
      journal = {Proceedings of the Intitute of Radio Engineers (IRE)},
       volume = {50},
         year = {1962},
        pages = {2061-2070},
}
@book{Nahin2002,
        title = {Oliver Heaviside: The Life, Work, and Times of an Electrical Genius of the Victorian Age},
       author = {Nahin, P.J.},
         year = {2002},
    publisher = {Johns Hopkins University Press},
     abstract = {"He was a man who often was incapable of conducting himself properly in the most elementary social interactions. His only continuing contacts with women were limited to his mother, nieces, and housekeepers. He was a man who knew the power of money and desired it, but refused to work for it, preferring to live off the sweat of his family and long-suffering friends, whom he often insulted even as they paid his bills." \emdash{} from the book. This, then, was Oliver Heaviside, a pioneer of modern electrical theory. Born into a low social class of Victorian England, Heaviside made advances in mathematics by introducing the operational calculus; in physics, where he formulated the modern-day expressions of Maxwell's Laws of electromagnetism; and in electrical engineering, through his duplex equations. Now available in paperback with a new preface by the author, this acclaimed biography will appeal to historians of technology and science, as well as to scientists and engineers who wish to learn more about this remarkable man.}
}
@inproceedings{NairandHintonICML-10,
     title = {Rectified Linear Units Improve Restricted Boltzmann Machines},
    author = {Vinod Nair and Geoffrey Hinton},
 booktitle = {Proceedings of International Conference on Machine learning},
      year = 2010,
}
@article{NairetalCoRR-15,
        title = {Massively Parallel Methods for Deep Reinforcement Learning},
       author = {Arun Naira and Praveen Srinivasana and Sam Blackwella and Cagdas Alciceka and Rory Fearona and Alessandro De Mariaa and Vedavyas Panneershelvama and Mustafa Suleymana and Charles Beattiea and Stig Petersena and Shane Legga and Volodymyr Mniha and Koray Kavukcuoglua and David Silver},
      journal = {CoRR},
       volume = {arXiv:1507.04296},
         year = {2015},
     abstract = {We present the first massively distributed architecture for deep reinforcement learning. This architecture uses four main components: parallel actors that generate new behaviour; parallel learners that are trained from stored experience; a distributed neural network to represent the value function or behaviour policy; and a distributed store of experience. We used our architecture to implement the Deep Q-Network algorithm (DQN). Our distributed algorithm was applied to 49 games from Atari 2600 games from the Arcade Learning Environment, using identical hyperparameters. Our performance surpassed non-distributed DQN in 41 of the 49 games and also reduced the wall-time required to achieve these results by an order of magnitude on most games.}
}
@inproceedings{NajorkandWiener01,
     author = {Marc Najork and Janet L. Wiener},
      title = {Breadth-first crawling yields high-quality pages},
  booktitle = {Proceedings of the 10th International Conference on World Wide Web},
      pages = {114-118},
  publisher = {ACM Press},
       year = 2001,
}
@article{NakajimaandHalassaCOIN-17,
       author = {Nakajima, M. and Halassa, M. M.},
        title = {Thalamic control of functional cortical connectivity},
      journal = {Current Opinion in Neurobiology},
       volume = {44},
         year = {2017},
        pages = {127-131},
     abstract = {The thalamus is an evolutionarily conserved structure with extensive reciprocal connections to cortical regions. While its role in transmitting sensory signals is well-studied, its broader engagement in cognition is unclear. In this review, we discuss evidence that the thalamus regulates functional connectivity within and between cortical regions, determining how a cognitive process is implemented across distributed cortical microcircuits. Within this framework, thalamic circuits do not necessarily determine the categorical content of a cognitive process (e.g., sensory details in feature-based attention), but rather provide a route by which task-relevant cortical representations are sustained and coordinated. Additionally, thalamic control of cortical connectivity bridges general arousal to the specific processing of categorical content, providing an intermediate level of cognitive and circuit description that will facilitate mapping neural computations onto thought and behavior.}
}
@article{NakataniandVanLeeuwenBC-06,
       author = {Nakatani, H. and van Leeuwen, C.},
        title = {Transient synchrony of distant brain areas and perceptual switching in ambiguous figures},
      journal = {Biological Cybernetics},
         year = 2006,
       volume = 94,
        issue = 6,
        pages = {445-457},
     abstract = {We studied the relationship between perceptual switching in the Necker cube and long-distance transient phase synchronization in EEG. Transient periods of response related synchrony between parietal and frontal areas were observed. They start 800-600, ms prior to the switch response and occur in pairs. Four types of pairs could be distinguished, two of which are accompanied by transient alpha band activity in the occipital area. The results indicate that perceptual switching processes involve parietal and frontal areas; these are the ones that are normally associated with various cognitive processes. Sensory information in the visual areas is involved in some, but not in all, of switching processes. The intrinsic variability, as well as the participating areas, points to the role of strategic cognitive processes in perceptual switching.}
}
@incollection{Nakayamaetal95,
     author = {Nakayama, K. and He, Z. J. and Shimojo, S.},
      title = {Visual surface representation: a critical link between lower-level and higher level vision}, 
     editor = {Kosslyn, S.M. and Osherson, D.N.},
  publisher = {MIT Press},
    address = {Cambridge, MA},
  booktitle = {Vision: In Invitation to Cognitive Science},
      pages = {1-70}, 
       year = 1995
}
@article{NakazawaNEURON-17,
       author = {Kazu Nakazawa},
        title = {Dentate Mossy Cell and Pattern Separation},
      journal = {Neuron},
       volume = {93},
        issue = {3},
         year = {2017},
        pages = {465-467},
}
@book{Nalwa93,
       author = {Nalwa, Vishvjit S.},
        title = {A Guided Tour of Computer Vision},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1993,
}
@article {NamboodirietalBIORXIV-20,
       author = {Namboodiri, Vijay Mohan K and Hobbs, Taylor and Pisanty, Ivan Trujillo and Simon, Rhiana C and Stuber, Garret D},
        title = {Meta-reinforcement learning in a thalamo-orbitofrontal circuit},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Laboratory},
         year = {2020},
     abstract = {Learning to predict rewards is essential for the survival of animals. Contemporary views suggest that such learning is driven by a reward prediction error{\textemdash}the difference between received and predicted rewards. Here we show using two-photon calcium imaging and optogenetics in mice that a different class of reward learning signals exists within the orbitofrontal cortex (OFC). Specifically, the reward responses of many OFC neurons exhibit plasticity consistent with filtering out rewards that are less salient for learning (such as predicted rewards, or, unpredicted rewards available in a context containing highly salient aversive stimuli). We show using quasi-simultaneous imaging and optogenetics that this reward response plasticity is sculpted by medial thalamic inputs to OFC. These results provide a biological substrate for emerging theoretical views of meta-reinforcement learning in prefrontal cortex.Competing Interest StatementThe authors have declared no competing interest.},
}
@article{NambuetalNEUROSCIENCE-RESEARCH-02,
        title = {Functional significance of the cortico-subthalamo-pallidal 'hyperdirect' pathway},
       author = {Nambu, Atsushi and Tokuno, Hironobu and Takada, Masahiko},
      journal = {Neuroscience {R}esearch},
       volume = {43},
         year = {2002},
        pages = {111-117},
     abstract = {How the motor-related cortical areas modulate the activity of the output nuclei of the basal ganglia is an important issue for understanding the mechanisms of motor control by the basal ganglia. The cortico-subthalamo-pallidal 'hyperdirect' pathway conveys powerful excitatory effects from the motor-related cortical areas to the globus pallidus, bypassing the striatum, with shorter conduction time than effects conveyed through the striatum. We emphasize the functional significance of the 'hyperdirect' pathway and propose a dynamic 'center-surround model' of basal ganglia function in the control of voluntary limb movements. When a voluntary movement is about to be initiated by cortical mechanisms, a corollary signal conveyed through the cortico-subthalamo-pallidal 'hyperdirect' pathway first inhibits large areas of the thalamus and cerebral cortex that are related to both the selected motor program and other competing programs. Then, another corollary signal through the cortico-striato-pallidal 'direct' pathway disinhibits their targets and releases only the selected motor program. Finally, the third corollary signal possibly through the cortico-striato-external pallido-subthalamo-internal pallidal 'indirect' pathway inhibits their targets extensively. Through this sequential information processing, only the selected motor program is initiated, executed and terminated at the selected timing, whereas other competing programs are canceled.}
}
@article{NarasimhanetalJAIR-18,
        title = {Grounding Language for Transfer in Deep Reinforcement Learning},
       author = {Karthik Narasimhan and Regina Barzilay and Tommi Jaakkola},
       volume = {63},
        pages = {849-874},
      journal = {Journal of Artificial Intelligence Research},
         year = {2018},
     abstract = {In this paper, we explore the utilization of natural language to drive transfer for reinforcement learning (RL). Despite the wide-spread application of deep RL techniques, learning generalized policy representations that work across domains remains a challenging problem. We demonstrate that textual descriptions of environments provide a compact intermediate channel to facilitate effective policy transfer. Specifically, by learning to ground the meaning of text to the dynamics of the environment such as transitions and rewards, an autonomous agent can effectively bootstrap policy learning on a new domain given its description. We employ a model-based RL approach consisting of a differentiable planning module, a model-free component and a factorized state representation to effectively use entity descriptions. Our model outperforms prior work on both transfer and multi-task scenarios in a variety of different environments. For instance, we achieve up to 14\% and 11.5\% absolute improvement over previously existing models in terms of average and initial rewards, respectively.}
}
@article{NarayananCEREBRAL-CORTEX-15,
       author = {Narayanan, Rajeevan T. and Egger, Robert and Johnson, Andrew S. and Mansvelder, Huibert D. and Sakmann, Bert and de Kock, Christiaan P.J. and Oberlaender, Marcel},
        title = {Beyond Columnar Organization: Cell Type- and Target Layer-Specific Principles of Horizontal Axon Projection Patterns in Rat Vibrissal Cortex},
      journal = {Cerebral Cortex},
         year = 2015,
     abstract = {Vertical thalamocortical afferents give rise to the elementary functional units of sensory cortex, cortical columns. Principles that underlie communication between columns remain however unknown. Here we unravel these by reconstructing in vivo-labeled neurons from all excitatory cell types in the vibrissal part of rat primary somatosensory cortex (vS1). Integrating the morphologies into an exact {3D} model of vS1 revealed that the majority of intracortical (IC) axons project far beyond the borders of the principal column. We defined the corresponding innervation volume as the IC-unit. Deconstructing this structural cortical unit into its cell type-specific components, we found asymmetric projections that innervate columns of either the same whisker row or arc, and which subdivide vS1 into 2 orthogonal [supra-]granular and infragranular strata. We show that such organization could be most effective for encoding multi whisker inputs. Communication between columns is thus organized by multiple highly specific horizontal projection patterns, rendering IC-units as the primary structural entities for processing complex sensory stimuli.},
}
@article{NarayananNEUROCOMP-03,
       author = {Narayanan, Srini},
        title = {The role of cortico-basal-thalamic loops in cognition: a computational model and preliminary results},
      journal = {Neurocomputing},
         year = 2003,
       volume = 5254,
        pages = {605-614},
     abstract = {Clinical and experimental research over the last decade has implicated neuroanatomic loops connecting the frontal cortex to the basal ganglia and thalamus in various aspects of planning and memory. We report on computational model whose central aspects are (1) a model of cortical-striatal-thalamic loops in planning and executive control, and (2) a fine-grained model of basal-ganglia function that exploits specific component connectivity and dynamics. The model is biologically plausible given current literature on the neurophysiology and disease pathology of the relevant brain regions. Specifically, our model has implications for subjects with diseases affecting the relevant brain regions (Parkinson's disease (PD) and Huntington's disease (HD)).}
}
@article{NarayananandLaubachJoN-09,
       author = {Narayanan, Nandakumar S. and and Laubach, Mark},
        title = {Delay activity in rodent frontal cortex during a simple reaction time task},
      journal = {Journal of neurophysiology},
    publisher = {American Physiological Society},
       volume = {101},
        issue = {6},
         year = {2009},
        pages = {2859-2871},
     abstract = {To understand how different parts of the frontal cortex control the timing of action, we characterized the firing patterns of single neurons in two areas of rodent frontal cortex-dorsomedial prefrontal cortex (dmPFC) and motor cortex-during a simple reaction time task. Principal component analysis was used to identify major patterns of delay-related activity in frontal cortex: ramping activity and sustained delay activity. These patterns were similar in dmPFC and motor cortex and did not change as animals learned to respond at novel delays. Many neurons in both areas were modulated early in the delay period. Other neurons were modulated in a persistent manner over the duration of the delay period. Delay-related modulations started earlier in motor cortex than in dmPFC and terminated around different task events (at the time of release in dmPFC, just before release of the lever in motor cortex). A subpopulation of neurons was found in dmPFC, but not motor cortex, that fired in response to the trigger stimulus. These results suggest that populations of neurons in rodent frontal cortex are coordinated during delay periods to enable proactive inhibitory control of action.},
}
@incollection{NarayananetalABA-16,
        title = {{Chapter 9 - Neuronal Cell Types in the Neocortex}},
       author = {Rajeevan T. Narayanan and Robert Egger and Christiaan P.J. de Kock and Marcel Oberlaender},
    booktitle = {Axons and Brain Architecture},
       editor = {Kathleen S. Rockland},
    publisher = {Academic Press},
      address = {San Diego},
         year = {2016},
        pages = {183-202},
     abstract = {One of the major challenges in neuroscience is to provide an answer to the question: Is it possible to define a set of neuronal cell types that is sufficient to describe the organizational principles of cortex and relate structure to function? In this chapter, we will provide an overview of some of the most common current cell type classification schemes and discuss the relationships between them and their present limits. Then, we will review some of our own work that aims to elucidate the relationships between soma location, dendrite morphology, axonal projections, and in vivo spiking of individual excitatory cortical neurons. Using these classification criteria, we introduce a set of 10 input-response-output excitatory cell types in the vibrissal part of rat primary somatosensory cortex, and discuss whether these cell types represent a "minimal set" that is sufficient to describe structural and functional organizational principles of whisker-evoked information processing.}
}
@book{Narendra89,
       author = {Narendra, Kumpati S. and Thathachar, Mandayam A. L.},
        title = {Learning Automata: An Introduction},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1989,
}
@article{NaselarisetalNEURON-11,
        title = {Bayesian reconstruction of natural images from human brain activity},
       author = {Naselaris, Thomas and Prenger, Ryan J. and Kay, Kendrick N. and Oliver, Michael and Gallant, Jack L.},
      journal = {Neuron},
       volume = 63,
        issue = 6,
         year = 2009,
        pages = {902-915},
     abstract = {Recent studies have used fMRI signals from early visual areas to reconstruct simple geometric patterns. Here, we demonstrate a new Bayesian decoder that uses fMRI signals from early and anterior visual areas to reconstruct complex natural images. Our decoder combines three elements: a structural encoding model that characterizes responses in early visual areas, a semantic encoding model that characterizes responses in anterior visual areas, and prior information about the structure and semantic content of natural images. By combining all these elements, the decoder produces reconstructions that accurately reflect both the spatial structure and semantic category of the objects contained in the observed natural image. Our results show that prior information has a substantial effect on the quality of natural image reconstructions. We also demonstrate that much of the variance in the responses of anterior visual areas to complex natural images is explained by the semantic category of the image alone.},
}
@article{NasretalSCIENCE-ADVANCES-19,
       author = {Nasr, Khaled and Viswanathan, Pooja and Nieder, Andreas},
        title = {Number detectors spontaneously emerge in a deep neural network designed for visual object recognition},
      journal = {Science Advances},
    publisher = {American Association for the Advancement of Science},
       volume = {5},
       number = {5},
         year = {2019},
     abstract = {Humans and animals have a "number sense", an innate capability to intuitively assess the number of visual items in a set, its numerosity. This capability implies that mechanisms to extract numerosity indwell the brain's visual system, which is primarily concerned with visual object recognition. Here, we show that network units tuned to abstract numerosity, and therefore reminiscent of real number neurons, spontaneously emerge in a biologically inspired deep neural network that was merely trained on visual object recognition. These numerosity-tuned units underlay the network's number discrimination performance that showed all the characteristics of human and animal number discriminations as predicted by the Weber-Fechner law. These findings explain the spontaneous emergence of the number sense based on mechanisms inherent to the visual system.},
}
@article{NauetalJAIR-03,
       author = {Dana Nau and Tsz-Chiu Au and Okhtay Ilghami and Ugur Kuter and J. William Murdock and Dan Wu and Fusun Yaman},
        title = {{SHOP2}: An {HTN} Planning System},
      journal = {Journal of Artificial Intelligence Research},
       volume = 20,
         year = 2003,
        pages = {379-404}
}
@article{NaumannetalCELL-16,
       author = {Naumann, Eva A. and Fitzgerald, James E. and Dunn, Timothy W. and Rihel, Jason and Sompolinsky, Haim and Engert, Florian},
        title = {From Whole-Brain Data to Functional Circuit Models: {T}he Zebrafish Optomotor Response},
      journal = {Cell},
         year = {2016},
    publisher = {Elsevier},
       volume = {167},
        issue = {4},
        pages = {947-960},
     abstract = {Detailed descriptions of brain-scale sensorimotor circuits underlying vertebrate behavior remain elusive. Recent advances in zebrafish neuroscience offer new opportunities to dissect such circuits via whole-brain imaging, behavioral analysis, functional perturbations, and network modeling. Here, we harness these tools to generate a brain-scale circuit model of the optomotor response, an orienting behavior evoked by visual motion. We show that such motion is processed by diverse neural response types distributed across multiple brain regions. To transform sensory input into action, these regions sequentially integrate eye- and direction-specific sensory streams, refine representations via interhemispheric inhibition, and demix locomotor instructions to independently drive turning and forward swimming. While experiments revealed many neural response types throughout the brain, modeling identified the dimensions of functional connectivity most critical for the behavior. We thus reveal how distributed neurons collaborate to generate behavior and illustrate a paradigm for distilling functional circuit models from whole-brain data.},
}
@article{NaumannetalJCN-12,
       author = {Naumann, R. K. and Anjum, F. and Roth-Alpermann, C. and Brecht, M.},
        title = {{C}ytoarchitecture, areas, and neuron numbers of the {E}truscan shrew cortex},
      journal = {Journal of Comparative Neurolology},
       volume = {520},
       number = {11},
         year = {2012},
        pages = {2512-2530},
     abstract = {The Etruscan shrew, Suncus etruscus, is one of the smallest mammals. Etruscan shrews can recognize prey shape with amazing speed and accuracy, based on whisker-mediated tactile cues. Because of its small size, quantitative analysis of the Etruscan shrew cortex is more tractable than in other animals. To quantitatively assess the anatomy of the Etruscan shrew's brain, we sectioned brains and applied Nissl staining and NeuN (neuronal nuclei) antibody staining. On the basis of these stains, we estimated the number of neurons of 10 cortical hemispheres by using Stereoinvestigator and Neurolucida (MBF Bioscience) software. On average, the neuron number per hemisphere was found to be ~1 million. We also measured cortical surface area and found an average of 11.1 mm² (n = 7) and an average volume of 5.3 mm³ (n = 10) per hemisphere. We identified 13 cortical regions by cytoarchitectonic boundaries in coronal, sagittal, and tangential sections processed for Nissl substance, myelin, cytochrome oxidase, ionic zinc, neurofilaments, and vesicular glutamate transporter 2 (VGluT2). The Etruscan shrew is a highly tactile animal with a large somatosensory cortex, which contains a barrel field, but the barrels are much less clearly defined than in rodents. The anatomically derived cortical partitioning scheme roughly corresponds to physiologically derived maps of neocortical sensory areas.}
}
@article{NavonetalCoRR-20,
        title = {Auxiliary Learning by Implicit Differentiation},
       author = {Aviv Navon and Idan Achituve and Haggai Maron and Gal Chechik and Ethan Fetaya},
         year = {2020},
      journal = {CoRR},
       volume = {arXiv:2007.02693},
     abstract = {Training with multiple auxiliary tasks is a common practice used in deep learning for improving the performance on the main task of interest. Two main challenges arise in this multi-task learning setting: (i) Designing useful auxiliary tasks; and (ii) Combining auxiliary tasks into a single coherent loss. We propose a novel framework, AuxiLearn, that targets both challenges, based on implicit differentiation. First, when useful auxiliaries are known, we propose learning a network that combines all losses into a single coherent objective function. This network can learn non-linear interactions between auxiliary tasks. Second, when no useful auxiliary task is known, we describe how to learn a network that generates a meaningful, novel auxiliary task. We evaluate AuxiLearn in a series of tasks and domains, including image segmentation and learning with attributes. We find that AuxiLearn consistently improves accuracy compared with competing methods.},
}
@inproceedings{NayakandMukerjeeCOLING-12,
        title = {Grounded Language Acquisition: A Minimal Commitment Approach},
       author = {Nayak, Sushobhan  and Mukerjee, Amitabha},
    booktitle = {Proceedings of {COLING} 2012},
      address = {Mumbai, India},
         year = {2012},
        pages = {2059-2076},
     abstract = {We take up the challenge of learning a grounded model of language when our agent has a body of machine learning algorithms and no prior knowledge of either the physical domain or language, in the sense of "least commitment". Based on a 2D video and co-occurring raw text, we demonstrate how this cognitively inspired model segments the world to obtain a meaning space, and combines words into hierarchical patterns for a linguistic pattern space. By associating these two spaces under temporal co-occurrence constraints, we demonstrate the acquisition of term-meaning pairs for names, actions and relations. We next map physical arguments for actions and relations to syntactical constructions resembling a cognitive grammar framework. Thus the system is able to bootstrap a rudimentary lexicon and syntax. While experiments are primarily in English, we present partial results for Hindi obtained without any change in the methods, to indicate its potential application to other languages.}
}
@techreport{NealPhD,
     author = {R. Neal},
      title = {Probabilistic Inference Using {M}arkov Chain {M}onte {C}arlo Methods,},
institution = {Dept. of Computer Science, University of Toronto},
       year = 1993,
      issue = {CRG-TR-93-1},
}
@book{NealandEichenbaum1993,
       author = {Cohen, Neal J. and Eichenbaum, Howard},
        title = {Memory, amnesia, and the hippocampal system.},
         year = {1993},
    publisher = {The MIT Press},
      address = {Cambridge,  MA,  US},
     abstract = {Neal J. Cohen and Howard Eichenbaum bring together converging findings from neuropsychology, neuroscience, and cognitive science that provide the critical clues and constraints for developing a more comprehensive understanding of memory. Specifically, they offer a cognitive neuroscience theory of memory that accounts for the nature of memory impairment exhibited in human amnesia and animal models of amnesia, that specifies the functional role played by the hippocampal system in memory, and that provides further understanding of the componential structure of memory.},
}
@incollection{NealandHinton93,
       author = {Radford M. Neal and Geoffrey E. Hinton},
        title = {A New View of the EM Algorithm that Justifies Incremental and Other Variants},
    booktitle = {Learning in Graphical Models},
         year = {1993},
        pages = {355-368},
    publisher = {Kluwer Academic Publishers},
     abstract = {We present a new view of the EM algorithm for maximum likelihood estimation
in situations with unobserved variables. In this view, both the E and the M steps
of the algorithm are seen as maximizing a joint function of the model parameters
and of the distribution over unobserved variables. From this perspective, it is
easy to justify an incremental variant of the algorithm in which the distribution
for only one of the unobserved variables is recalculated in each E step. This
variant is shown empirically to give faster convergence in a mixture estimation
problem. A wide range of other variant algorithms are also seen to be possible}
}
@incollection{NealandHinton98,
       author = {R.M. Neal and Hinton, G.E.},
        title = {A view of the {EM} algorithm that justifies incremental, sparse, and other variants},
       editor = {Jordan, M.I.},
    booktitle = {Learning in Graphical Models},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1998,
}
@incollection{NealandHintonGEM-98,
       author = {R. M. Neal and G. E. Hinton},
        title = {A view of the {EM} algorithm that justies incremental, sparse, and other variants},
    booktitle = {Learning in Graphical Models},
       editor = {M. I. Jordan},
    publisher = {Kluwer Academic Publishers},
      address = {Norwell, Massachusetts},
         year = 1998,
        pages = {355-368},
}
@book{Neapolitan90,
       author = {Neapolitan, Richard E.},
        title = {Probabilistic Reasoning in Expert Systems: Theory and Algorithms},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1990,
}
@misc{Nebel,
       author = {Nebel, Bernhard and B\"{a}ckstr\"{o}m, Christer},
        title = {On the Computational Complexity of Temporal Projection and some Related Problems},
        month = {October},
         year = 1991,
}
@article{NeeandJonidesPNAS-08,
       author = {Nee, Derek Evan and Jonides, John},
        title = {Neural correlates of access to short-term memory},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {105},
       number = {37},
         year = {2008},
        pages = {14228-14233},
     abstract = {Behavioral research has led to the view that items in short-term memory can be parsed into two categories: a single item in the focus of attention that is available for immediate cognitive processing and a small set of other items that are in a heightened state of activation but require retrieval for further use. We examined this distinction by using an item-recognition task. The results show that the item in the focus of attention is represented by increased activation in inferior temporal representational cortices relative to other information in short-term memory. Functional connectivity analyses suggest that activation of these inferior temporal regions is maintained via frontal- and posterior-parietal contributions. By contrast, other items in short-term memory demand retrieval mechanisms that are represented by increased activation in the medial temporal lobe and left mid-ventrolateral prefrontal cortex. These results show that there are two distinctly different sorts of access to information in short-term memory, and that access by retrieval operations makes use of neural machinery similar to that used in long-term memory retrieval.},
}
@article{NeelakantanetalCoRR-15,
       author = {Arvind Neelakantan and Quoc V. Le and Ilya Sutskever},
        title = {Neural Programmer: Inducing Latent Programs with Gradient Descent},
      journal = {CoRR},
       volume = {arXiv:1511.04834},
         year = {2015},
     abstract = {Deep neural networks have achieved impressive supervised classification performance in many tasks including image recognition, speech recognition, and sequence to sequence learning. However, this success has not been translated to applications like question answering that may involve complex arithmetic and logic reasoning. A major limitation of these models is in their inability to learn even simple arithmetic and logic operations. For example, it has been shown that neural networks fail to learn to add two binary numbers reliably. In this work, we propose Neural Programmer, an end-to-end differentiable neural network augmented with a small set of basic arithmetic and logic operations. Neural Programmer can call these augmented operations over several steps, thereby inducing compositional programs that are more complex than the built-in operations. The model learns from a weak supervision signal which is the result of execution of the correct program, hence it does not require expensive annotation of the correct program itself. The decisions of what operations to call, and what data segments to apply to are inferred by Neural Programmer. Such decisions, during training, are done in a differentiable fashion so that the entire network can be trained jointly by gradient descent. We find that training the model is difficult, but it can be greatly improved by adding random noise to the gradient. On a fairly complex synthetic table-comprehension dataset, traditional recurrent networks and attentional models perform poorly while Neural Programmer typically obtains nearly perfect accuracy.}
}
@inproceedings{NeelakantanetalICLR-17,
       author = {Arvind Neelakantan and Quoc V. Le and Mart\'{i}n Abadi and Andrew McCallum and Dario Amodei},
        title = {Learning a Natural Language Interface with Neural Programmer},
    booktitle = {International Conference on Learning Representations},
       volume = {arXiv:1611.08945},
         year = {2017},
     abstract = {Learning a natural language interface for database tables is a challenging task that involves deep language understanding and multi-step reasoning. The task is often approached by mapping natural language queries to logical forms or programs that provide the desired response when executed on the database. To our knowledge, this paper presents the first weakly supervised, end-to-end neural network model to induce such programs on a real-world dataset. We enhance the objective function of Neural Programmer, a neural network with built-in discrete operations, and apply it on WikiTableQuestions, a natural language question-answering dataset. The model is trained end-to-end with weak supervision of question-answer pairs, and does not require domain-specific grammars, rules, or annotations that are key elements in previous approaches to program induction. The main experimental result in this paper is that a single Neural Programmer model achieves 34.2\% accuracy using only 10,000 examples with weak supervision. An ensemble of 15 models, with a trivial combination technique, achieves 37.7\% accuracy, which is competitive to the current state-of-the-art accuracy of 37.1\% obtained by a traditional natural language semantic parser.}
}
@book{Nelson1989narratives,
        title = {Narratives from the Crib},
       author = {Nelson, Katherine},
         year = {1989},
    publisher = {Harvard University Press}
}
@book{Nelson1998language,
        title = {Language in Cognitive Development: The Emergence of the Mediated Mind},
       author = {Nelson, Katherine},
         year = {1998},
    publisher = {Cambridge University Press}
}
@incollection{NelsonDAD-2011,
        title = {Ageism: The Strange Case of Prejudice Against the Older You},
       author = {Nelson, T.D.},
    booktitle = {Disability and Aging Discrimination},
       editor = {Wiener, R. and Willborn, S.},
    publisher = {Springer},
      address = {New York, NY},
         year = {2011},
     abstract = {In a field that has only been in existence for about 36 years, research on age prejudice has revealed many robust findings, a few counterintuitive results, and it has left open a number of unanswered questions. Today I would like to highlight the major findings in ageism research, the unexpected discoveries, and to point future researchers to issues that still require investigation. The field is growing, as more and more researchers are sensing the theoretical and practical need for empirical attention devoted to a problem that affects millions of older adults. As we begin to shed light on the causes and consequences of age prejudice, we can start to devise ways to reduce age stereotyping, thereby enhancing the quality of life for all older adults. My remarks today will explore the state of the literature on ageism and address the following questions: How does ageism manifest itself, who is most likely to be ageist, and how can we reduce ageism?}
}
@incollection{NelsonNARRATIVE-03,
        title = {Narrative and the Emergence of a Conscious Self},
    booktitle = {Narrative and Consciousness: Literature, Psychology and the Brain},
    publisher = {Oxford University Press},
       editor = {Gary D. Fireman and Ted E. McVay and Owen J. Flanagan},
       author = {Katherine Nelson},
         year = {2003},
        pages = {17-36},
}
@article{NelsonandFivushPSYCHOLOGICAL-REVIEW-04,
       author = {Nelson, K. and Fivush, R.},
        title = {{{T}he emergence of autobiographical memory: a social cultural developmental theory}},
      journal = {Psychological Review},
         year = {2004},
       volume = {111},
       number = {2},
        pages = {486-511},
     abstract = {The authors present a multicomponent dynamic developmental theory of human autobiographical memory that emerges gradually across the preschool years. The components that contribute to the process of emergence include basic memory abilities, language and narrative, adult memory talk, temporal understanding, and understanding of self and others. The authors review the empirical developmental evidence within each of these components to show how each contributes to the timing, quantity, and quality of personal memories from the early years of life. The authors then consider the relevance of the theory to explanations of childhood amnesia and how the theory accounts for and predicts the complex findings on adults' earliest memories, including individual, gender, and cultural differences.}
}
@book{Nemeczek1995vangogh,
        title = {Van Gogh in Arles},
       author = {Nemeczek, A. and van Gogh, V.},
         year = {1995},
    publisher = {Prestel}
}
@incollection{NeriecandDesplanCTDB-16,
        title = {Chapter Fourteen - From the Eye to the Brain: Development of the Drosophila Visual System},
       editor = {Paul M. Wassarman},
    booktitle = {Essays on Developmental Biology, Part A},
    publisher = {Academic Press},
       volume = {116},
       series = {Current Topics in Developmental Biology},
       author = {Nathalie N\`{e}riec and Claude Desplan},
         year = {2016},
        pages = {247-271},
     abstract = {How stem cells produce the huge diversity of neurons that form the visual system, and how these cells are assembled in neural circuits are a critical question in developmental neurobiology. Investigations in Drosophila have led to the discovery of several basic principles of neural patterning. In this chapter, we provide an overview of the field by describing the development of the Drosophila visual system, from the embryo to the adult and from the gross anatomy to the cellular level. We then explore the general molecular mechanisms identified that might apply to other neural structures in flies or in vertebrates. Finally, we discuss the major challenges that remain to be addressed in the field.}
}
@article{NeriecandDesplanCURRENT-TOPICS-DEVELOPMENTAL-BIOLOGY-16,
       author = {Neriec, N.  and Desplan, C. },
        title = {{{F}rom the {E}ye to the {B}rain: {D}evelopment of the {D}rosophila {V}isual {S}ystem}},
      journal = {Current Topics in Developmental Biololgy},
       volume = {116},
         year = {2016},
        pages = {247-271},
     abstract = {How stem cells produce the huge diversity of neurons that form the visual system, and how these cells are assembled in neural circuits are a critical question in developmental neurobiology. Investigations in Drosophila have led to the discovery of several basic principles of neural patterning. In this chapter, we provide an overview of the field by describing the development of the Drosophila visual system, from the embryo to the adult and from the gross anatomy to the cellular level. We then explore the general molecular mechanisms identified that might apply to other neural structures in flies or in vertebrates. Finally, we discuss the major challenges that remain to be addressed in the field.}
}
@book{Nesse2012sick,
        title = {Why We Get Sick: The New Science of Darwinian Medicine},
       author = {Nesse, Randolph M. and Williams, George C.},
         year = {2012},
    publisher = {Knopf Doubleday Publishing Group}
}
@book{Nesse2019good,
        title = {Good Reasons for Bad Feelings: Insights from the Frontier of Evolutionary Psychiatry},
       author = {Nesse, Randolph M.},
         year = {2019},
          url = {https://www.wsj.com/articles/good-reasons-for-bad-feelings-review-sufferings-silver-lining-11550100691},
    publisher = {Penguin Publishing Group}
}
@techreport{Neumann1945,
       author = {Neumann, John von},
        title = {First Draft of a Report on the {EDVAC}},
         year = {1945},
  institution = {Institute for Advanced Study}
}
@incollection{Neumann1956,
       author = {John von Neumann},
        title = {Probabilistic logics and the synthesis of reliable organisms from unreliable components},
    booktitle = {Automata Studies},
       editor = {Claude E. Shannon and John McCarthy},
    publisher = {Princeton University Press},
      address = {Princeton, NJ},
         year = 1956,
        pages = {329-378},
}
@book{Neumann1966,
       author = {John von Neumann},
        title = {Theory of Self-Reproducing Automata},
    publisher = {University of Illinois Press},
      address = {Urbana, IL},
         year = 1966,
}
@book{Neumann1966completeTheory_of_Self_Reproducing_Automata,
       author = {John von Neumann},
        title = {Theory of Self-Reproducing Automata},
    publisher = {University of Illinois Press},
      address = {Urbana, IL},
         year = 1966,
}
@article{NeunuebelandKnierimNEURON-14,
       author = {Joshua P. Neunuebel and James J. Knierim},
        title = {{CA3} Retrieves Coherent Representations from Degraded Input: Direct Evidence for {CA3} Pattern Completion and Dentate Gyrus Pattern Separation},
      journal = {Neuron},
       volume = {81},
       number = {2},
        pages = {416-427},
         year = {2014},
     abstract = {Theories of associative memory suggest that successful memory storage and recall depend on a balance between two complementary processes: pattern separation (to minimize interference) and pattern completion (to retrieve a memory when presented with partial or degraded input cues). Putative attractor circuitry in the hippocampal CA3 region is thought to be the final arbiter between these two processes. Here we present direct, quantitative evidence that CA3 produces an output pattern closer to the originally stored representation than its degraded input patterns from the dentate gyrus (DG). We simultaneously recorded activity from CA3 and DG of behaving rats when local and global reference frames were placed in conflict. CA3 showed a coherent population response to the conflict (pattern completion), even though its DG inputs were severely disrupted (pattern separation). The results thus confirm the hallmark predictions of a longstanding computational model of hippocampal memory processing.}
}
@article{NevelnCTB-90,
       author = {Neveln, B.},
        title = {Comma-free and synchronizable codes},
      journal = {Journal of Theoretical Biology},
         year = {1990},
       volume = {144},
       number = {2},
        pages = {209-212},
     abstract = {Shepherd (1981. Proc. natn. Acad. Sci. U.S.A. 78, 1596-1600), Clarke (1982. J. theor. Biol. 99, 397-403), Crick et al. (1976. Origins of Life 7, 389) have put forward the possibility that there was a code, ancestral to the existing genetic code, which was non-degenerate and comma-free. The point of a comma-free code, as first proposed in Crick et al. (1957), is that it provides immunity from frameshift reading errors. In this paper it is argued that if frameshift error immunity is to be used as a selection factor in guiding the search for a hypothetical ancestral code, then the appropriate mathematical formulation of this characteristic is not comma-freeness but synchronizability. The calling card of the comma-free notion is that it results in a limit of 20 on the number of sense codons when applied to the genetic coding system. But so does synchronizability, which includes comma-freeness as a special case. Further, using synchronizability as a constraint instead of comma-freeness yields roughly 3000 times as many potential models for an ancestral code. In particular, the RNY model becomes more compatible with immunity from frameshift error as an evolutionary constraint.},
}
@phdthesis{Nevill-ManningPhD,
       author = {Craig Nevill-Manning},
        title = {Inferring Sequential Structure},
  institution = {University of Waikato, New Zealand},
         year = 1996,
}
@article{Nevill-ManningandWittenJAIR-94,
       author = {Craig Nevill-Manning and Ian Witten},
        title = {Identifying Hierarchical Structure in Sequences: A Linear-Time Algorithm},
      journal = JAIR,
       volume = 7,
        pages = {67-82},
         year = 1997,
}
@book{Newell90,
       author = {Allen Newell},
        title = {Unified Theories of Cognition},
    publisher = {Harvard University Press},
      address = {Cambridge, MA},
         year = 1990,
}
@incollection{NewelletalCATEGORIES-11,
        title = {Systems of Category Learning: Fact or Fantasy?},
       author = {Ben R. Newell and John C. Dunn and Michael Kalish},
    booktitle = {The Psychology of Learning and Motivation},
       editor = {Brian Ross},
       volume = {54},
    publisher = {Academic Press}, 
      address = {New York},
         year = {2011}, 
        pages = {167-215},
     abstract = {Psychology abounds with vigorous debates about the need for one or more underlying mental processes or systems to explain empirical observations. The field of category learning provides an excellent exemplar. We present a critical examination of this field focusing on empirical, methodological, and mathematical modeling considerations. We review what is often presented as the "best evidence" for multiple systems of category learning and critique the evidence by considering three questions: (1) Are multiple-systems accounts the only viable explanations for reported effects? (2) Are the inferences sound logically and methodologically? (3) Are the mathematical models that can account for behavior sufficiently constrained, and are alternative (single-system) models applicable? We conclude that the evidence for multiple-systems accounts of category learning does not withstand such scrutiny. We end by discussing the varieties of explanation that can be offered for psychological phenomena and highlight why multiple-systems accounts often provide an illusory sense of scientific progress.}
}
@article{Newman-TokerSTROKE-13,
       author = {Newman-Toker, David E. and Saber Tehrani, Ali S. and Mantokoudis, Georgios and Pula, John H. and Guede, Cynthia I. and Kerber, Kevin A. and Blitz, Ari and Ying, Sarah H. and Hsieh, Yu-Hsiang and Rothman, Richard E. and Hanley, Daniel F. and Zee, David S. and Kattah, Jorge C.},
        title = {Quantitative Video-Oculography to Help Diagnose Stroke in Acute Vertigo and Dizziness: Toward an ECG for the Eyes},
      journal = {Stroke},
       volume = 44,
        issue = 4,
         year = 2013,
        pages = {1158-1161},
     abstract = {Background and Purpose--Strokes can be distinguished from benign peripheral causes of acute vestibular syndrome using bedside oculomotor tests (head impulse test, nystagmus, test-of-skew). Using head impulse test, nystagmus, test-of-skew is more sensitive and less costly than early magnetic resonance imaging for stroke diagnosis in acute vestibular syndrome but requires expertise not routinely available in emergency departments. We sought to begin standardizing the head impulse test, nystagmus, test-of-skew diagnostic approach for eventual emergency department use through the novel application of a portable video-oculography device measuring vestibular physiology in real time. This approach is conceptually similar to ECG to diagnose acute cardiac ischemia.Methods--Proof-of-concept study (August 2011 to June 2012). We recruited adult emergency department patients with acute vestibular syndrome defined as new, persistent vertigo/dizziness, nystagmus, and (1) nausea/vomiting, (2) head motion intolerance, or (3) new gait unsteadiness. We recorded eye movements, including quantitative horizontal head impulse testing of vestibulo-ocular-reflex function. Two masked vestibular experts rated vestibular findings, which were compared with final radiographic gold-standard diagnoses. Masked neuroimaging raters determined stroke or no stroke using magnetic resonance imaging of the brain with diffusion-weighted imaging obtained 48 hours to 7 days after symptom onset.Results--We enrolled 12 consecutive patients who underwent confirmatory magnetic resonance imaging. Mean age was 61 years (range 30-73), and 10 were men. Expert-rated video-oculography-based head impulse test, nystagmus, test-of-skew examination was 100\% accurate (6 strokes, 6 peripheral vestibular).Conclusions--Device-based physiological diagnosis of vertebrobasilar stroke in acute vestibular syndrome should soon be possible. If confirmed in a larger sample, this bedside eye ECG approach could eventually help fulfill a critical need for timely, accurate, efficient diagnosis in emergency department patients with vertigo or dizziness who are at high risk for stroke.},
}
@article{NewmanetalNAS-02,
       author = {M. Newman and D. Watts and S. Strogatz},
        title = {Random graph models of social networks},
      journal = {Proceedings of the National Academy of Science},
       volume = 99,
        pages = {2566-2572},
         year = 2002,
}
@article{NewportLANGUAGE_SCIENCES-88,
        title = {Constraints on learning and their role in language acquisition: Studies of the acquisition of American sign language},
       author = {Elissa L. Newport},
      journal = {Language Sciences},
       volume = {10},
       number = {1},
         year = {1988},
        pages = {147-172},
     abstract = {The general question raised here is why the young child is superior to older children and adults at language acquisition, while at the same time inferior to them in many other cognitive tasks. As an example of the general problem, the paper reviews our own work on the acquisition of complex verbs of American Sign Language (ASL). It begins with an outline of the structure of verbs of motion in ASL, along with possible inductive generalizations a language learner might make concerning this structure. Three lines of research on ASL acquisition are then presented. The first line of research demonstrates that young children, exposed to ASL as a native language, acquire ASL verbs in terms of morphological components, piece by piece, as do children learning spoken language. Moreover, they do so despite alternative generalizations which seem potentially simpler. The second line of research compares native learners of ASL with learners exposed to ASL later in life. This research shows that, while native learners make the morphological componential generalization described, later learners in fact do make alternative generalizations. The third line of research investigates native learners of ASL whose parental input models are late learners. Again it appears that natives perform a morphological analysis, despite the fact that their input is not well organized for such analyses. Taken together, the research shows a striking tendency for children-and only children-to acquire language in a particular fashion. The paper concludes with a discussion of possible explanations for these findings. One possibility is that children have a special set of skills for language acquisition which declines with age. A second possibility is that the cognitive limitations of the child provide the basis on which the child's componential learning occurs, and that the expansion of these cognitive abilities with age is in part responsible for the decline in this type of learning.}
}
@incollection{NewtonandSur2004,
       author = {Newton, Jessica R. and Sur, Mriganka},
    booktitle = {Plasticity of the central auditory system and processing of complex acoustic signals},
       editor = {Syka, J. and Merzenich, M. M.},
    publisher = {Springer Verlag},
        title = {Rewiring Cortex: Functional Plasticity of the Auditory Cortex During Development},
          url = {http://web.mit.edu/msur/www/publications.html},
         year = 2004,
        pages = {127-138},
     abstract = {How does the primary auditory cortex (A1) derive its function? Is it primarily determined by intrinsic factors expressed within cortex (i.e. molecular gradients, patterns of gene expression) or extrinsic factors that originate outside cortex (i.e. sensory experience)?  This is a pivotal issue for understanding the development and function of A1, yet one that is difficult to address using traditional experimental methods because these techniques are unable to separate out the relative contributions of intrinsic and extrinsic factors. However  cross-modal experiments where the inputs of one sensory modality are re-directed to a different modality, in effect "rewiring" the brain, are able to distinguish the two and provide insight   Rewiring experiments also reveal how input activity modifies  cortical organization as well as the limitations imposed on this  plasticity by the underlying cortical substrate. In this chapter  we review rewiring experiments in which visual inputs are induced  to innervate the auditory pathway, and describe its physiological and behavioral consequences as well as the implications for normal cortical processing and human cross-modal plasticity.},
}
@article{NgBRAIN-SCIENCE-18,
       author = {Ng, Betsy},
        title = {The Neuroscience of Growth Mindset and Intrinsic Motivation},
      journal = {Brain Science},
         year = {2018},
       volume = {8},
       number = {2},
     abstract = {Our actions can be triggered by intentions, incentives or intrinsic values. Recent neuroscientific research has yielded some results about the growth mindset and intrinsic motivation. With the advances in neuroscience and motivational studies, there is a global need to utilize this information to inform educational practice and research. Yet, little is known about the neuroscientific interplay between growth mindset and intrinsic motivation. This paper attempts to draw on the theories of growth mindset and intrinsic motivation, together with contemporary ideas in neuroscience, outline the potential for neuroscientific research in education. It aims to shed light on the relationship between growth mindset and intrinsic motivation in terms of supporting a growth mindset to facilitate intrinsic motivation through neural responses. Recent empirical research from the educational neuroscience perspective that provides insights into the interplay between growth mindset and intrinsic motivation will also be discussed.}
}
@inproceedings{NgandJordanNIPS-02,
       author = {Andrew Ng and Michael Jordan},
        title = {On Discriminative versus Generative Classifiers: A comparison of logistic regression and na\"{i}ve {B}ayes},
    booktitle = {Advances in Neural Information Processing Systems 14},
       editor = {T. G. Dietterich and S. Becker and Z. Ghahramani},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2002,
}
@inproceedings{NgetalNIPS-02,
       author = {A. Ng and M. Jordan and Y. Weiss},
        title = {On spectral clustering: Analysis and an algorithm},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 14,
      address = {Cambridge, MA},
    publisher = {MIT Press},
          url = {http://citeseer.ist.psu.edu/ng01spectral.html},
         year = 2002,
}
@article{NgoetalDEVELOPMENTAL-COGNITIVE_NEUROSCIENCE-17,
       author = {Chi T. Ngo and Kylie H. Alm and Athanasia Metoki and William Hampton and Tracy Riggins and Nora S. Newcombe and Ingrid R. Olson},
        title = {White matter structural connectivity and episodic memory in early childhood},
      journal = {Developmental Cognitive Neuroscience},
       volume = {28},
         year = {2017},
        pages = {41-53},
     abstract = {Episodic memory undergoes dramatic improvement in early childhood; the reason for this is poorly understood. In adults, episodic memory relies on a distributed neural network. Key brain regions that supporting these processes include the hippocampus, portions of the parietal cortex, and portions of prefrontal cortex, each of which shows different developmental profiles. Here we asked whether developmental differences in the axonal pathways connecting these regions may account for the robust gains in episodic memory in young children. Using diffusion weighted imaging, we examined whether white matter connectivity between brain regions implicated in episodic memory differed with age, and were associated with memory performance differences in 4- and 6-year-old children. Results revealed that white matter connecting the hippocampus to the inferior parietal lobule significantly predicted children's performance on episodic memory tasks. In contrast, variation in the white matter connecting the hippocampus to the medial prefrontal cortex did not relate to memory performance. These findings suggest that structural connectivity between the hippocampus and lateral parietal regions is relevant to the development of episodic memory.}
}
@misc{NguyenBOOSTED-HINDSIGHT-REPLAY-19,
        title = {Four Ways to Boost Experience Replay: {M}aking agents remember the important things},
       author = {Austin Nguyen},
 howpublished = {\url{https://towardsdatascience.com/4-ways-to-boost-experience-replay-999d9f17f7b6}},
         year = {2019},
}
@article{NguyenetalBIORXIV-17,
       author = {Jeffrey P. Nguyen and Ashley N. Linder and George S. Plummer and Joshua and W. Shaevitz and Andrew M. Leifer},
        title = {Automatically tracking neurons in a moving and deforming brain},
      journal = {bioRxiv},
         year = {2017},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {Advances in optical neuroimaging techniques now allow neural activity to be recorded with cellular resolution in awake and behaving animals. Brain motion in these recordings pose a unique challenge. The location of individual neurons must be tracked in 3D over time to accurately extract single neuron activity traces. Recordings from small invertebrates like C. elegans are especially challenging because they undergo very large brain motion and deformation during animal movement. Here we present an automated computer vision pipeline to reliably track populations of neurons with single neuron resolution in the brain of a freely moving C. elegans undergoing large motion and deformation. 3D volumetric fluorescent images of the animal's brain are straightened, aligned and registered, and the locations of neurons in the images are found via segmentation. Each neuron is then assigned an identity using a new time-independent machine-learning approach we call Neuron Registration Vector Encoding. In this approach, non-rigid point-set registration is used to match each segmented neuron in each volume with a set of reference volumes taken from throughout the recording. The way each neuron matches with the references defines a feature vector which is clustered to assign an identity to each neuron in each volume. Finally, thin-plate spline interpolation is used to correct errors in segmentation and check consistency of assigned identities. The Neuron Registration Vector Encoding approach proposed here is uniquely well suited for tracking neurons in brains undergoing large deformations. When applied to whole-brain calcium imaging recordings in freely moving C. elegans, this analysis pipeline located 150 neurons for the duration of an 8 minute recording and consistently found more neurons more quickly than manual or semi-automated approaches.},
}
@article{NguyenetalBIORXIV-17_IEEE_Dataport_Supplement,
       author = {Jeffrey P. Nguyen and Ashley N. Linder and George S. Plummer and Joshua and W. Shaevitz and Andrew M. Leifer},
        title = {Automatically tracking neurons in a moving and deforming brain},
      journal = {IEEE Dataport},
    publisher = {IEEE Press},
      address = {Piscataway, NJ},
         year = {2017},
          url = {https://ieee-dataport.org/open-access/tracking-neurons-moving-and-deforming-brain-dataset},
     abstract = {Advances in optical neuroimaging techniques now allow neural activity to be recorded with cellular resolution in awake and behaving animals.  Brain motion in these recordings pose a unique challenge. The location of individual neurons must  be tracked in 3D over time to accurately extract single neuron activity traces. Recordings from small invertebrates like C. elegans are especially challenging because they undergo very large brain motion and deformation during animal movement. Here we present a dataset to accompany an automated computer vision pipeline we've developed to reliably track populations of neurons with single neuron resolution in the brain of a freely moving C. elegans undergoing large motion and deformation. Two calcium imaging recordings from freely moving C. elegans are provided, worm 1 and worm 2, corresponding to recordings discussed in our corresponding publication Ngueyn et al, "Automatically tracking neurons in a moving and deforming brain," PLoS Comp Bio, in press, 2017.  Worm 1 contains hand annotated tracking of individual neurons, while Worm 2 does not.  When our analysis pipeline is applied to the 8 minute whole-brain calcium imaging recording of Worm 2, we identify 156 neurons for the duration of the recording. Our algorithm consistently finds more neurons more quickly than manual or semi-automated approaches.}
}
@article{NguyenetalPNAS-15,
        title = {Whole-brain calcium imaging with cellular resolution in freely behaving {\it{Caenorhabditis elegans}}},
       author = {Jeffrey P. Nguyen and Frederick B. Shipley and Ashley N. Linder and George S. Plummer and Mochi Liu and Sagar U. Setru and Joshua W. Shaevitz and Andrew M. Leifer},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {113},
        issue = {11},
         year = {2015},
        pages = {E1074-E1081},
     abstract = {The ability to acquire large-scale recordings of neuronal activity in awake and unrestrained animals is needed to provide new insights into how populations of neurons generate animal behavior. We present an instrument capable of recording intracellular calcium transients from the majority of neurons in the head of a freely behaving Caenorhabditis elegans with cellular resolution while simultaneously recording the animal's position, posture, and locomotion. This instrument provides whole-brain imaging with cellular resolution in an unrestrained and behaving animal. We use spinning-disk confocal microscopy to capture 3D volumetric fluorescent images of neurons expressing the calcium indicator GCaMP6s at 6 head-volumes/s. A suite of three cameras monitor neuronal fluorescence and the animal's position and orientation. Custom software tracks the 3D position of the animal's head in real time and two feedback loops adjust a motorized stage and objective to keep the animal's head within the field of view as the animal roams freely. We observe calcium transients from up to 77 neurons for over 4 min and correlate this activity with the animal's behavior. We characterize noise in the system due to animal motion and show that, across worms, multiple neurons show significant correlations with modes of behavior corresponding to forward, backward, and turning locomotion.}
}
@article{NguyenetalPNAS-15.supplement,
        title = {Whole-brain calcium imaging with cellular resolution in freely behaving {\it{Caenorhabditis elegans}}: Supplementary Materials},
       author = {Jeffrey P. Nguyen and Frederick B. Shipley and Ashley N. Linder and George S. Plummer and Mochi Liu and Sagar U. Setru and Joshua W. Shaevitz and Andrew M. Leifer},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {113},
        issue = {11},
         year = {2015},
        pages = {E1074-E1081},
}
@article{NicholsetalJoVE-12,
       author = {Nichols, C. D.  and Becnel, J.  and Pandey, U. B.},
        title = {{{M}ethods to assay {D}rosophila behavior}},
      journal = {Journal of Visualized Experience},
         year = {2012},
       volume = {61},
     abstract = {Drosophila melanogaster, the fruit fly, has been used to study molecular mechanisms of a wide range of human diseases such as cancer, cardiovascular disease and various neurological diseases(1). We have optimized simple and robust behavioral assays for determining larval locomotion, adult climbing ability (RING assay), and courtship behaviors of Drosophila. These behavioral assays are widely applicable for studying the role of genetic and environmental factors on fly behavior. Larval crawling ability can be reliably used for determining early stage changes in the crawling abilities of Drosophila larvae and also for examining effect of drugs or human disease genes (in transgenic flies) on their locomotion. The larval crawling assay becomes more applicable if expression or abolition of a gene causes lethality in pupal or adult stages, as these flies do not survive to adulthood where they otherwise could be assessed. This basic assay can also be used in conjunction with bright light or stress to examine additional behavioral responses in Drosophila larvae. Courtship behavior has been widely used to investigate genetic basis of sexual behavior, and can also be used to examine activity and coordination, as well as learning and memory. Drosophila courtship behavior involves the exchange of various sensory stimuli including visual, auditory, and chemosensory signals between males and females that lead to a complex series of well characterized motor behaviors culminating in successful copulation. Traditional adult climbing assays (negative geotaxis) are tedious, labor intensive, and time consuming, with significant variation between different trials(2-4). The rapid iterative negative geotaxis (RING) assay(5) has many advantages over more widely employed protocols, providing a reproducible, sensitive, and high throughput approach to quantify adult locomotor and negative geotaxis behaviors. In the RING assay, several genotypes or drug treatments can be tested simultaneously using large number of animals, with the high-throughput approach making it more amenable for screening experiments.},
}
@inproceedings{NicholsonECAI-92,
       author = {A. E. Nicholson and J. M. Brady},
        title = {The Data Association Problem when Monitoring Robot Vehicles using Dynamic Belief Networks},
    booktitle = {Proceedings of the 10th European Conference on Artificial Intelligence (ECAI-92)},
         year = 1992,
}
@phdthesis{NicholsonPhD,
       author = {A. E. Nicholson},
        title = {Monitoring Discrete Environments using Dynamic Belief Networks},
       school = {Dept. of Engineering Sciences, Oxford University},
         year = 1992,
}
@inproceedings{NicholsonandKaelblingSSS-94,
       author = {Nicholson, Ann and Kaelbling, Leslie Pack},
        title = {Toward approximate planning in very large stochastic domains},
    booktitle = {Proceedings of the {AAAI} Spring Symposium on Decision Theoretic Planning},
     location = {Stanford, California},
         year = 1994,
}
@incollection{NickelandVolkerTENSOR-13,
    author = {Nickel, Maximilian and Tresp, Volker},
     title = {Tensor Factorization for Multi-relational Learning},
    volume = 8190,
    series = {Lecture Notes in Computer Science},
 booktitle = {Machine Learning and Knowledge Discovery in Databases},
 publisher = {Springer Berlin Heidelberg},
    editor = {Blockeel, Hendrik and Kersting, Kristian and Nijssen, Siegfried and Aelezna, Filip},
      year = 2013,
     pages = {617-621}
}
@book{Nicolelis2011,
     title = {Beyond Boundaries: The New Neuroscience of Connecting Brains with Machines --- and How It Will Change Our Lives},
    author = {Miguel Nicolelis},
 publisher = {Henry Holt and Company},
      year = 2011,
}
@book{Nicolelis2011beyond,
        title = {Beyond Boundaries: The New Neuroscience of Connecting Brains with Machines and How It Will Change Our Lives},
       author = {Nicolelis, Miguel},
         year = {2011},
    publisher = {Henry Holt and Company}
} 
@article{NicolelisandFanselowRS-02,
     title = {Dynamic shifting in thalamocortical processing during different behavioural states.},
    author = {Nicolelis, M.A. and Fanselow, E.E.},
   journal = {Philosophical Transactions of the Royal Society London B Biological Science},
    volume = 357,
     issue = 1428,
      year = 2002,
     pages = {1753-1758},
  abstract = {Recent experiments in our laboratory have indicated that as rats shift the behavioural strategy employed to explore their surrounding environment, there is a parallel change in the physiological properties of the neuronal ensembles that define the main thalamocortical loop of the trigeminal somatosensory system. Based on experimental evidence from several laboratories   we propose that this concurrent shift in behavioural strategy and thalamocortical physiological properties provides rats with an efficient way to optimize either the detection or analysis of complex tactile stimuli.},
}
@inproceedings{NicolescuandMataricMAS-03,
       author = {Monica Nicolescu and Maja J. Matari\'{c}},
        title = {Natural Methods for Robot Task Learning: Instructive Demonstration, Generalization and Practice},
    booktitle = {International Joint Conference on Autonomous Agents and Multiagent Systems},
      address = {Melbourne, Australia},
          url = {http://cres.usc.edu/cgi-bin/print_pub_details.pl?pubid=344},
         year = 2003,
}
@article{NieblesetalJCV-08,
       author = {J.C. Niebles and H. Wang and L. Fei-Fei},
        title = {Unsupervised learning of human action categories using spatial-temporal words},
      journal = {International Journal of Computer Vision},
       volume = 79,
        issue = 3,
         year = 2008,
        pages = {299-318},
}
@article{NielsenNLM-15,
       author = {Nielsen, T. and O'Reilly, C. and Carr, M. and Dumel, G. and Godin, I. and Solomonova, E. and Lara-Carrasco, J. and Blanchette-Carriere, C. and Paquette, T.},
        title = {Overnight improvements in two {R}{E}{M} sleep-sensitive tasks are associated with both {REM} and {NREM} sleep changes, sleep spindle features, and awakenings for dream recall},
      journal = {Neurobiology Learning and Memory},
         year = {2015},
       volume = {122},
        pages = {88-97},
     abstract = {Memory consolidation is associated with sleep physiology but the contribution of specific sleep stages remains controversial. To clarify the contribution of REM sleep, participants were administered two REM sleep-sensitive tasks to determine if associated changes occurred only in REM sleep. Twenty-two participants (7 men) were administered the Corsi Block Tapping and Tower of Hanoi tasks prior to and again after a night of sleep. Task improvers and non-improvers were compared for sleep structure, sleep spindles, and dream recall. Control participants (N = 15) completed the tasks twice during the day without intervening sleep. Overnight Corsi Block improvement was associated with more REM sleep whereas Tower of Hanoi improvement was associated with more N2 sleep. Corsi Block improvement correlated positively with \%REM sleep and Tower of Hanoi improvement with \%N2 sleep. Post-hoc analyses suggest Tower of Hanoi effects-but not Corsi Block effects-are due to trait differences. Sleep spindle density was associated with Tower of Hanoi improvement whereas spindle amplitude correlated with Corsi Block improvement. Number of REM awakenings for dream reporting (but not dream recall per se) was associated with Corsi Block, but not Tower of Hanoi, improvement but was confounded with REM sleep time. This non-replication of one of 2 REM-sensitive task effects challenges both 'dual-process' and 'sequential' or 'sleep organization' models of sleep-dependent learning and points rather to capacity limitations on REM sleep. Experimental awakenings for sampling dream mentation may not perturb sleep-dependent learning effects; they may even enhance them.}
}
@article{NiepertetalCoRR-16,
       author = {Mathias Niepert and Mohamed Ahmed and Konstantin Kutzkov},
        title = {Learning Convolutional Neural Networks for Graphs},
      journal = {CoRR},
       volume = {arXiv:1605.05273},
         year = {2016},
     abstract = {Numerous important problems can be framed as learning from graph data. We propose a framework for learning convolutional neural networks for arbitrary graphs. These graphs may be undirected, directed, and with both discrete and continuous node and edge attributes. Analogous to image-based convolutional networks that operate on locally connected regions of the input, we present a general approach to extracting locally connected regions from graphs. Using established benchmark data sets, we demonstrate that the learned feature representations are competitive with state of the art graph kernels and that their computation is highly efficient.},
}
@article{NiessingandFriedrichNATURE-10,
       author = {Niessing, Jorn and Friedrich, Rainer W.},
        title = {Olfactory pattern classification by discrete neuronal network states},
      journal = {Nature},
         year = {2010},
    publisher = {Macmillan Publishers Limited},
       volume = {465},
        issue = {7294},
        pages = {47-52},
     abstract = {The categorial nature of sensory, cognitive and behavioural acts indicates that the brain classifies neuronal activity patterns into discrete representations. Pattern classification may be achieved by abrupt switching between discrete activity states of neuronal circuits, but few experimental studies have directly tested this. We gradually varied the concentration or molecular identity of odours and optically measured responses across output neurons of the olfactory bulb in zebrafish. Whereas population activity patterns were largely insensitive to changes in odour concentration, morphing of one odour into another resulted in abrupt transitions between odour representations. These transitions were mediated by coordinated response changes among small neuronal ensembles rather than by shifts in the global network state. The olfactory bulb therefore classifies odour-evoked input patterns into many discrete and defined output patterns, as proposed by attractor models. This computation is consistent with perceptual phenomena and may represent a general information processing strategy in the brain.}
}
@book{Nietzsche,
      author = {Friedrich Nietzsche},
       title = {Beyond Good and Evil: Prelude to a Philosophy of the Future},
   publisher = {Reprinted 1966, Vintage Books},
     comment = {Translated by Walter Kaufmann},
     address = {New York, NY},
        year = 1886,
}
@article{NigametalJML-00,
     author = {Kamal Nigam and Andrew K. McCallum and Sebastian Thrun and Tom M. Mitchell},
      title = {Text Classification from Labeled and Unlabeled Documents using {EM}},
    journal = {Machine Learning},
     volume = 39,
      issue = {2-3},
      pages = {103-134},
        url = {citeseer.ist.psu.edu/nigam99text.html},
       year = 2000,
}
@article{NightingaleMEDICAL_IMAGING-11,
    author = {Kathy Nightingale},
     title = {Acoustic Radiation Force Impulse ({ARFI}) Imaging: a Review},
   journal = {Current Medical Imaging Reviews},
    volume = 7,
     issue = 4,
      year = 2011,
     pages = {328-339},
}
@article{NijhawanandWuPTRS-B-09,
       author = {Romi Nijhawan  and Si Wu},
        title = {Compensating time delays with neural predictions: are predictions sensory or motor?},
      journal = {Philosophical Transactions of the Royal Society A: Mathematical, Physical and Engineering Sciences},
       volume = {367},
       number = {1891},
         year = {2009},
        pages = {1063-1078},
     abstract = {Neural delays are a general property of computations carried out by neural circuits. Delays are a natural consequence of temporal summation and coding used by the nervous system to integrate information from multiple resources. For adaptive behaviour, however, these delays must be compensated. In order to sense and interact with moving objects, for example, the visual system must predict the future position of the object to compensate for delays. In this paper, we address two critical questions concerning the implementation of the compensation mechanisms in the brain, namely, where does compensation occur and how is it realized. We present evidence showing that compensation can happen in both the motor and sensory systems, and that compensation using 'diagonal neural pathways' is a suitable strategy for implementing compensation in the visual system. In this strategy, neural signals in the early stage of information processing are sent to the future cortical positions that correspond to the distance the object will travel in the period of transmission delay. We propose a computational model to elucidate this using the retinal visual information pathway.}
}
@article{NikolenkoetalNATURE-07,
     title = {Two-photon photostimulation and imaging of neural circuits},
    author = {Nikolenko, V. and Poskanzer, K.E. and Yuste, R.},
   journal = {Nature Methods},
    volume = 4,
     issue = 11,
     pages = {943-50},
      year = 2007,
  abstract = {We introduce an optical method to stimulate individual neurons in brain slices in any arbitrary spatiotemporal pattern, using two-photon uncaging of MNI-glutamate with beam multiplexing. This method has single-cell and three-dimensional precision. By sequentially stimulating up to a thousand potential presynaptic neurons, we generated detailed functional maps of inputs to a cell. We combined this approach with two-photon calcium imaging in an all-optical method to image and manipulate circuit activity.}
}
@book{Nilsson65,
       author = {Nilsson, Nils},
        title = {Learning Machines},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1965,
}
@book{Nilsson80,
       author = {Nilsson, Nils J.},
        title = {Principles of Artificial Intelligence},
    publisher = {Tioga Publishing Company},
      address = {Palo Alto, California},
         year = 1980,
}        
@article{NilssonAIJ-86,
       author = {Nilsson, Nils},
        title = {Probabilistic Logic},
      journal = AIJ,
       volume = 28,
         year = 1986,
        pages = {71-88},
}
@inproceedings{NilssonRPW-89,
       author = {Nilsson, Nils J.},
        title = {Action Networks},
       editor = {Tenenberg, Josh and Weber, Jay and Allen, James},
    booktitle = {Proceedings from the Rochester Planning Workshop: From Formal Systems to Practical Systems},
         year = 1989,
        pages = {36-68},
}
@article{NimchinskyetalPNAS-99,
        title = {A neuronal morphologic type unique to humans and great apes},
       author = {Nimchinsky, Esther A. and Gilissen, Emmanuel and Allman, John M. and Perl, Daniel P. and Erwin, Joseph M. and Hof, Patrick R.},
      journal = {Proceedings of the National Academy of Sciences},
       volume = 96,
        issue = 9,
        pages = {5268-5273},
         year = 1999,
     abstract = {We report the existence and distribution of an unusual type of projection neuron, a large, spindle-shaped cell, in layer Vb of the anterior cingulate cortex of pongids and hominids. These spindle cells were not observed in any other primate species or any other mammalian taxa, and their volume was correlated with brain volume residuals, a measure of encephalization in higher primates. These observations are of particular interest when considering primate neocortical evolution, as they reveal possible adaptive changes and functional modifications over the last 15--20 million years in the anterior cingulate cortex, a region that plays a major role in the regulation of many aspects of autonomic function and of certain cognitive processes. That in humans these unique neurons have been shown previously to be severely affected in the degenerative process of Alzheimer's disease suggests that some of the differential neuronal susceptibility that occurs in the human brain in the course of age-related dementing illnesses may have appeared only recently during primate evolution.},
}
@article{NimmerjahnetalSCIENCE-05,
       author = {Nimmerjahn, A.  and Kirchhoff, F.  and Helmchen, F. },
        title = {{{R}esting microglial cells are highly dynamic surveillants of brain parenchyma in vivo}},
      journal = {Science},
         year = {2005},
       volume = {308},
       number = {5726},
        pages = {1314-1318},
     abstract = {Microglial cells represent the immune system of the mammalian brain and therefore are critically involved in various injuries and diseases. Little is known about their role in the healthy brain and their immediate reaction to brain damage. By using in vivo two-photon imaging in neocortex, we found that microglial cells are highly active in their presumed resting state, continually surveying their microenvironment with extremely motile processes and protrusions. Furthermore, blood-brain barrier disruption provoked immediate and focal activation of microglia, switching their behavior from patroling to shielding of the injured site. Microglia thus are busy and vigilant housekeepers in the adult brain.}
}
@article{Nishihara84,
       author = {Nishihara, H. K.},
        title = {Practical Real-Time Imaging Stereo Matcher},
      journal = {Optical Engineering},
       volume = 23,
        issue = 5,
         year = 1984,
        pages = {536-545},
}
@article{NishimotoetalCB-11,
        title = {Reconstructing Visual Experiences from Brain Activity Evoked by Natural Movies},
       author = {Shinji Nishimoto and An T. Vu and Thomas Naselaris and Yuval Benjamini and Bin Yu and Jack L. Gallant},
      journal = {Current Biology},
       volume = 21,
        issue = 19,
        pages = {1641-1646},
         year = 2011,
}
@inproceedings{NishimuraetalAPSIPA-12,
       author = {K. Nishimura and H. Kawanami and H. Saruwatari and K. Shikano},
    booktitle = {Asia-Pacific Signal Information Processing Association Annual Summit and Conference},
        title = {Response generation based on statistical machine translation for speech-oriented guidance system},
         year = {2012},
        pages = {1-4},
     abstract = {An example-based response generation is a robust and practical approach for a real-environment information guidance system. However, this framework cannot reflect differences in nuance, because the set of answer sentences are fixed beforehand. To overcome this issue, we have proposed response generation using a statistical machine translation technique. In this paper, we make use of N-best speech recognition candidates instead of manual transcription used in our previous study. As a result, the generation rate of appropriate response sentences was improved by using multiple recognition hypothesis.},
}
@inproceedings{NishitanietalDCR-14,
       author = {Nishitani, Y. and Kaneko, Y. and Ueda, M.},
    booktitle = {Proceedings of the 72nd Annual Device Research Conference},
        title = {Artificial synapses using ferroelectric memristors embedded with CMOS Circuit for image recognition},
         year = {2014},
        pages = {297-298},
     abstract = {Memristors have attracted attention as devices for brain-inspired computing hardware, such as artificial neural networks. Typical neural networks comprise multiple neurons interconnected via synapses. A synapse modulates the signal transmission strength or "weight" between two neurons. Weight controllability is essential to neural network adaptability. Therefore, it is necessary to establish an artificial synapse that can modulate its own electric conductance, which represents the weights. Some researchers have used two-terminal memristors as synapses. However, when using conventional memristors, pulses with complex shapes corresponding to what is learned must be prepared and applied to both terminals simultaneously because of their two-terminal structures. Previously, we showed that a programmable synapse function could be implemented on a three-terminal ferroelectric memristor (3T-FeMEM) fabricated on a single crystal oxide substrate, which enabled simple learning schemes. In this work, synapse chips were fabricated by integrating 3T-FeMEMs on CMOS circuits. We then demonstrated on-chip associative memory function using a neural network circuit with these chips.},
}
@inproceedings{NisterandSteweniusCVPR-06,
    author = {David N\'{i}ster and Henrik Stew\'{}nius},
     title = {Scalable Recognition with a Vocabulary Tree},
 booktitle = {Proceedings of the 2006 IEEE Computer Society 
              Conference on Computer Vision and Pattern Recognition},
 publisher = {IEEE Computer Society},
   address = {Washington, DC, USA},
     pages = {2161-2168},
      year = 2006,
}
@inproceedings{NiuetalISLPED-10,
       author = {Dimin Niu and Yiran Chen and Yuan Xie},
    booktitle = {Proceedings of the International Symposium on Low-Power Electronics and Design},
        title = {Low-power dual-element memristor based memory design},
         year = 2010,
        pages = {25-30},
     abstract = {Recently, the emerging memristor device technology has attracted significant research interests due to its distinctive hysteresis characteristic, which potentially can enable novel circuit designs for future VLSI circuits. In particular, characteristics such as non-volatility, non-linearity, low power consumption, and good scalability make memristor one of the most promising emerging memory technologies. Some important design parameter, however, such as speed, energy consumption, and distingushiablility, are mainly determined by the memristor's physical characteristics. In this paper, a key observation of memristor's asymmetric energy consumption is made by the detailed analysis of the transient power consumption. Based on this observation, we propose a dual-element memory structure in which each memory cell consists of two memristors. By constantly writing complement bits into the two elements within a cell, the dual-element is flexible to satisfy design constraints which are usually difficult to be satisfied with one-cell memory structure. Design space of the dual-element memory cell is studied and it shows that the trade-offs among the energy, speed, and distingushiablility should be explored for different design objectives. In particular, we show that under the energy-driven optimization, the proposed dual-element memory achieves the same programming speed and distinguishability as conventional single-element memory but the energy consumption can be reduced by up to 80\%.},
}
@article{NiziolekSoN-13,
       author = {Niziolek, Caroline A. and Nagarajan, Srikantan S. and Houde, John F.},
        title = {What does motor efference copy represent?: Evidence from speech production},
      journal = {The Journal of neuroscience : the official journal of the Society for Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {33},
        issue = {41},
         year = {2013},
        pages = {16110-16116},
     abstract = {How precisely does the brain predict the sensory consequences of our actions? Efference copy is thought to reflect the predicted sensation of self-produced motor acts, such as the auditory feedback heard while speaking. Here, we use magnetoencephalographic imaging (MEG-I) in human speakers to demonstrate that efference copy prediction does not track movement variability across repetitions of the same motor task. Specifically, spoken vowels were less accurately predicted when they were less similar to a speaker's median production, even though the prediction is thought to be based on the very motor commands that generate each vowel. Auditory cortical responses to less prototypical speech productions were less suppressed, resembling responses to speech errors, and were correlated with later corrective movement, suggesting that the suppression may be functionally significant for error correction. The failure of the motor system to accurately predict less prototypical speech productions suggests that the efferent-driven suppression does not reflect a sensory prediction, but a sensory goal.},
}
@article{NobaueretalNATURE-METHODS-17,
       author = {Nobauer, Tobias and Skocek, Oliver and Pernia-Andrade, Alejandro J. and Weilguny, Lukas and Traub, Francisca Martinez and Molodtsov, Maxim I. and Vaziri, Alipasha},
        title = {Video rate volumetric {Ca2+} imaging across cortex using seeded iterative demixing ({SID}) microscopy},
      journal = {Nature Methods},
         year = {2017},
    publisher = {Nature Publishing Group},
       volume = {advance online publication},
     abstract = {Light-field microscopy (LFM) is a scalable approach for volumetric Ca2+ imaging with high volumetric acquisition rates (up to 100 Hz). Although the technology has enabled whole-brain Ca2+ imaging in semi-transparent specimens, tissue scattering has limited its application in the rodent brain. We introduce seeded iterative demixing (SID), a computational source-extraction technique that extends LFM to the mammalian cortex. SID can capture neuronal dynamics in vivo within a volume of 900 [times] 900 [times] 260 [mu]m located as deep as 380 [mu]m in the mouse cortex or hippocampus at a 30-Hz volume rate while discriminating signals from neurons as close as 20 [mu]m apart, at a computational cost three orders of magnitude less than that of frame-by-frame image reconstruction. We expect that the simplicity and scalability of LFM, coupled with the performance of SID, will open up a range of applications including closed-loop experiments.},
}
@article{NoklandCoRR-16,
        title = {Direct Feedback Alignment Provides Learning in Deep Neural Networks},
       author = {Arild N\o kland},
      journal = {CoRR},
       volume = {arxiv:arXiv:1609.01596},
         year = {2016},
     abstract = {Artificial neural networks are most commonly trained with the back-propagation algorithm, where the gradient for learning is provided by back-propagating the error, layer by layer, from the output layer to the hidden layers. A recently discovered method called feedback-alignment shows that the weights used for propagating the error backward don't have to be symmetric with the weights used for propagation the activation forward. In fact, random feedback weights work evenly well, because the network learns how to make the feedback useful. In this work, the feedback alignment principle is used for training hidden layers more independently from the rest of the network, and from a zero initial condition. The error is propagated through fixed random feedback connections directly from the output layer to each hidden layer. This simple method is able to achieve zero training error even in convolutional networks and very deep networks, completely without error back-propagation. The method is a step towards biologically plausible machine learning because the error signal is almost local, and no symmetric or reciprocal weights are required. Experiments show that the test performance on MNIST and CIFAR is almost as good as those obtained with back-propagation for fully connected networks. If combined with dropout, the method achieves 1.45\% error on the permutation invariant MNIST task.}
}
@article{NoklandEidnesCoRR-19,
        title = {Training Neural Networks with Local Error Signals},
       author = {Arild N{\o}kland and Lars Hiller Eidnes},
      journal = {ArXiv},
         year = {2019},
       volume = {arXiv:1901.06656},
     abstract = {Supervised training of neural networks for classification is typically performed with a global loss function. The loss function provides a gradient for the output layer, and this gradient is back-propagated to hidden layers to dictate an update direction for the weights. An alternative approach is to train the network with layer-wise loss functions. In this paper we demonstrate, for the first time, that layer-wise training can approach the state-of-the-art on a variety of image datasets. We use single-layer sub-networks and two different supervised loss functions to generate local error signals for the hidden layers, and we show that the combination of these losses help with optimization in the context of local learning. Using local errors could be a step towards more biologically plausible deep learning because the global error does not have to be transported back to hidden layers. A completely backprop free variant outperforms previously reported results among methods aiming for higher biological plausibility.},
}
@book{Nolte2009,
       author = {John Nolte},
        title = {The Human Brain: An Introduction to its Functional Anatomy, 6th Edition},
    publisher = {Elsevier},
         year = 2009,
}
@article{NormanHIPPOCAMPUS-10,
       author = {Norman, K. A.},
        title = {How hippocampus and cortex contribute to recognition memory: {R}evisiting the complementary learning systems model},
      journal = {Hippocampus},
       volume = {20},
       number = {11},
         year = {2010},
        pages = {1217-1227},
     abstract = {We describe how the Complementary Learning Systems neural network model of recognition memory (Norman and O'Reilly (2003) Psychol Rev 104:611-646) can shed light on current debates regarding hippocampal and cortical contributions to recognition memory. We review simulation results illustrating three critical differences in how (according to the model) hippocampus and cortex contribute to recognition memory, all of which derive from the hippocampus' use of pattern separated representations. Pattern separation makes the hippocampus especially well-suited for discriminating between studied items and related lures; it makes the hippocampus especially poorly suited for computing global match; and it imbues the hippocampal ROC curve with a Y-intercept > 0. We also describe a key boundary condition on these differences: When the average level of similarity between items in an experiment is very high, hippocampal pattern separation can fail, at which point the hippocampal model will start to behave like the cortical model. We describe the implications of these simulation results for extant debates over how to describe hippocampal versus cortical contributions and how to measure these contributions.}
}
@article{NormanandOReillyPR-03,
       author = {Norman, Kenneth and O'Reilly, Randall},
        title = {Modeling Hippocampal and Neocortical Contributions to Recognition Memory: A Complementary Learning Systems Approach},
      journal = {Psychological review},
       volume = {110},
         year = {2003},
        pages = {611-46},
     abstract = {The authors present a computational neural-network model of how the hippocampus and medial temporal lobe cortex (MTLC) contribute to recognition memory. The hippocampal component contributes by recalling studied details. The MTLC component cannot support recall, but one can extract a scalar familiarity signal from MTLC that tracks how well a test item matches studied items. The authors present simulations that establish key differences in the operating characteristics of the hippocampal-recall and MTLC-familiarity signals and identify several manipulations (e.g., target-lure similarity, interference) that differentially affect the 2 signals. They also use the model to address the stochastic relationship between recall and familiarity and the effects of partial versus complete hippocampal lesions on recognition.},
}
@book{Norvig91,
     author = {Peter Norvig},
      title = {Paradigms of Artificial Intelligence Programming},
  publisher = {Morgan Kaufmann},
    address = {San Francisco, CA},
       year = 1991,
}
@article{NottebohmPLoS-05,
       author = {Nottebohm, Fernando},
      journal = {PLOS Biology},
    publisher = {Public Library of Science},
        title = {The Neural Basis of Birdsong},
         year = {2005},
       volume = {3},
     abstract = {There is a tradition in biology of using specific animal models to study generalizable basic properties of a system. For example, the giant axon of squid was used for the pioneering work on nerve transmission; the fruit fly (Drosophila) has played a key role in researchers discovering the role of homeobox genes in embryogenesis; the sea slug (Aplysia) is used to study the molecular biology of learning; and the round worm (Caenorhabditis elegans) is used to study programmed cell death. Basic insights gained from these four systems apply widely to other multicellular animals. Here, I will review basic discoveries made by studying birdsong that have helped answer more general questions in vertebrate neuroscience.},
}
@article{NowaketalNATURE-10,
     title = {The evolution of eusociality},
    author = {Nowak, Martin A. and Tarnita, Corina E. and Wilson, Edward O.},
   journal = {Nature},
 publisher = {Nature Publishing Group},
    volume = 466,
     issue = 7310,
      year = 2010,
     pages = {1057-1062},
  abstract = {Eusociality, in which some individuals reduce their own
    lifetime reproductive potential to raise the offspring of others,
    underlies the most advanced forms of social organization and the
    ecologically dominant role of social insects and humans. For the
    past four decades kin selection theory, based on the concept of
    inclusive fitness, has been the major theoretical attempt to
    explain the evolution of eusociality. Here we show the limitations
    of this approach. We argue that standard natural selection theory
    in the context of precise models of population structure
    represents a simpler and superior approach, allows the evaluation
    of multiple competing hypotheses, and provides an exact framework
    for interpreting empirical observations.},
}
@techreport{NunezMSc,
       author = {Linda Mensinger Nunez},
        title = {On the Relationship Between Temporal {B}ayes Networks and {M}arkov Chains},
         type = {{M.Sc.} Thesis},
  institution = {Brown University},
         year = 1989,
}
@article{NusbaumandBeenhakkerNATURE-01,
       author = {Michael P. Nusbaum and Mark P. Beenhakker},
        title = {A small-systems approach to motor pattern generation},
      journal = {Nature},
       volume = 417,
         year = 2001,
        pages = {343-350},
     abstract = {How neuronal networks enable animals, humans included, to make coordinated movements is a continuing goal of neuroscience research. The stomatogastric nervous system of decapod crustaceans, which contains a set of distinct but interacting motor circuits, has contributed significantly to the general principles guiding our present understanding of how rhythmic motor circuits operate at the cellular level. This results from a detailed documentation of the circuit dynamics underlying motor pattern generation in this system as well as its modulation by individual transmitters and neurons.}
}
@article{Nusslein-VolhardandWieschausNATURE-80,
    author = {Nusslein-Volhard, Christiane and Wieschaus, Eric},
     title = {Mutations affecting segment number and polarity in Drosophila},
   journal = {Nature},
    volume = 287,
     issue = 5785,
      year = 1980,
     pages = {795-801},
}
@article{Nyquist28,
       author = {Nyquist, H.},
        title = {Certain Topics in Telegraph Transmission Theory},
      journal = {Transactions of the AIEE},
       volume = 47,
         year = 1928,
        pages = {617-644},
}
@book{OKeefeandNadelHIPPOCAMPUS-78,
        title = {The hippocampus as a cognitive map},
       author = {O'Keefe, John and Nadel, Lynn},
    publisher = {Clarendon Press},
         year = {1978},
}
@article{OReillyNC-96,
       author = {O'Reilly, Randall C.},
        title = {Biologically Plausible Error-driven Learning Using Local Activation Differences: The Generalized Recirculation Algorithm},
      journal = {Neural Computation},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
       volume = {8},
       number = {5},
         year = {1996},
        pages = {895-938},
     abstract = {The error backpropagation learning algorithm (BP) is generally considered biologically implausible because it does not use locally available, activation-based variables. A version of BP that can be computed locally using bidirectional activation recirculation (Hinton and McClelland 1988) instead of backpropagated error derivatives is more biologically plausible. This paper presents a generalized version of the recirculation algorithm (GeneRec), which overcomes several limitations of the earlier algorithm by using a generic recurrent network with sigmoidal units that can learn arbitrary input/output mappings. However, the contrastive Hebbian learning algorithm (CHL, also known as DBM or mean field learning) also uses local variables to perform error-driven learning in a sigmoidal recurrent network. CHL was derived in a stochastic framework (the Boltzmann machine), but has been extended to the deterministic case in various ways, all of which rely on problematic approximations and assumptions, leading some to conclude that it is fundamentally flawed. This paper shows that CHL can be derived instead from within the BP framework via the GeneRec algorithm. CHL is a symmetry-preserving version of GeneRec that uses a simple approximation to the midpoint or second-order accurate Runge-Kutta method of numerical integration, which explains the generally faster learning speed of CHL compared to BI. Thus, all known fully general error-driven learning algorithms that use local activation-based variables in deterministic networks can be considered variations of the GeneRec algorithm (and indirectly, of the backpropagation algorithm). GeneRec therefore provides a promising framework for thinking about how the brain might perform error-driven learning. To further this goal, an explicit biological mechanism is proposed that would be capable of implementing GeneRec-style learning. This mechanism is consistent with available evidence regarding synaptic modification in neurons in the neocortex and hippocampus, and makes further predictions.}
}
@article{OReillySCIENCE-06,
     title = {Biologically Based Computational Models of High-Level Cognition},
    author = {O'Reilly, Randall C.},
   journal = {Science},
    volume = 314,
     issue = 5796,
      year = 2006,
     pages = {91-94},
  abstract = {Computer models based on the detailed biology of the brain can help us understand the myriad complexities of human cognition and intelligence. Here, we review models of the higher level aspects of human intelligence, which depend critically on the prefrontal cortex and associated subcortical areas. The picture emerging from a convergence of detailed mechanistic models and more abstract functional models represents a synthesis between analog and digital forms of computation. Specifically, the need for robust active maintenance and rapid updating of information in the prefrontal cortex appears to be satisfied by bistable activation states and dynamic gating mechanisms. These mechanisms are fundamental to digital computers and may be critical for the distinctive aspects of human intelligence.},
}
@article{OReillyTiCS-98,
       author = {O'Reilly, R. C.},
        title = {{S}ix principles for biologically based computational models of cortical cognition},
      journal = {Trends Cognitive Science)},
         year = {1998},
       volume = {2},
       number = {11},
        pages = {455-462},
     abstract = {This review describes and motivates six principles for computational cognitive neuroscience models: biological realism, distributed representations, inhibitory competition, bidirectional activation propagation, error-driven task learning, and Hebbian model learning. Although these principles are supported by a number of cognitive, computational and biological motivations, the prototypical neural-network model (a feedforward back-propagation network) incorporates only two of them, and no widely used model incorporates all of them. It is argued here that these principles should be integrated into a coherent overall framework, and some potential synergies and conflicts in doing so are discussed.}
} 
@article{OReillyandFrankNC-06,
       author = {O'Reilly, Randall C. and Frank, Michael J.},
        title = {Making Working Memory Work: A Computational Model of Learning in the Prefrontal Cortex and Basal Ganglia},
      journal = {Neural Computation},
       volume = 18,
        issue = 2,
         year = 2006,
        pages = {283-328},
          url = {http://psych.colorado.edu/~oreilly/papers/OReillyFrank06_pbwm.pdf},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
     abstract = {The prefrontal cortex has long been thought to subserve both working memory (the holding of information online for processing) and executive functions (deciding how to manipulate working memory and perform processing). Although many computational models of working memory have been developed, the mechanistic basis of executive function remains elusive, often amounting to a homunculus. This article presents an attempt to deconstruct this homunculus through powerful learning mechanisms that allow a computational model of the prefrontal cortex to control both itself and other brain areas in a strategic, task-appropriate manner. These learning mechanisms are based on subcortical structures in the midbrain, basal ganglia, and amygdala, which together form an actor-critic architecture. The critic system learns which prefrontal representations are task relevant and trains the actor, which in turn provides a dynamic gating mechanism for controlling working memory updating. Computationally, the learning mechanism is designed to simultaneously solve the temporal and structural credit assignment problems. The model's performance compares favorably with standard backpropagation-based temporal learning mechanisms on the challenging 1-2-AX working memory task and other benchmark working memory tasks.}
} 
@book{OReillyandMunakata00,
     title = {Computational Explorations in Cognitive Neuroscience: Understanding the Mind by Simulating the Brain},
    author = {Randall O'Reilly and Yuko Munakata},
 publisher = {MIT Press},
   address = {Cambridge, Massachusetts},
      year = 2000,
}
@article{OReillyandRudyHIPPOCAMPUS-00,
       author = {O'Reilly, Randall C. and Rudy, Jerry W.},
        title = {Computational principles of learning in the neocortex and hippocampus},
      journal = {Hippocampus},
       volume = {10},
       number = {4},
         year = {2000},
        pages = {389-397},
     abstract = {We present an overview of our computational approach towards understanding the different contributions of the neocortex and hippocampus in learning and memory. The approach is based on a set of principles derived from converging biological, psychological, and computational constraints. The most central principles are that the neocortex employs a slow learning rate and overlapping distributed representations to extract the general statistical structure of the environment, while the hippocampus learns rapidly, using separated representations to encode the details of specific events while suffering minimal interference. Additional principles concern the nature of learning (error-driven and Hebbian), and recall of information via pattern completion. We summarize the results of applying these principles to a wide range of phenomena in conditioning, habituation, contextual learning, recognition memory, recall, and retrograde amnesia, and we point to directions of current development.},
}
@incollection{OReillyandSotoNIPS-02,
        title = {A Model of the Phonological Loop: Generalization and Binding},
       author = {Randall C. O'Reilly and R. Soto},
    booktitle = {Advances in Neural Information Processing Systems 14},
       editor = {T. G. Dietterich and S. Becker and Z. Ghahramani},
        pages = {83-90},
         year = {2002},
    publisher = {MIT Press},
     abstract = {We present a neural network model that shows how the prefrontal cortex, interacting with the basal ganglia, can maintain a sequence of phonological information in activation-based working memory (i.e., the phonological loop). The primary function of this phonological loop may be to transiently encode arbitrary bindings of information necessary for tasks the combinatorial expressive power of language enables very flexible binding of essentially arbitrary pieces of information. Our model takes advantage of the closed-class nature of phonemes, which allows different neural representations of all possible phonemes at each sequential position to be encoded. To make this work, we suggest that the basal ganglia provide a region-specific update signal that allocates phonemes to the appropriate sequential coding slot. To demonstrate that flexible, arbitrary binding of novel sequences can be supported by this mechanism, we show that the model can generalize to novel sequences after moderate amounts of training.},
}
@article{OReillyetalBEHAVIORAL-NEUROSCIENCE-07,
        title = {{PVLV}: the primary value and learned value Pavlovian learning algorithm},
       author = {Randall O'Reilly and Michael Joshua Frank and Thomas E. Hazy and Brandon Watz},
      journal = {Behavioral neuroscience},
         year = {2007},
       volume = {121},
       number = {1},
        pages = {31-49},
     abstract = {The authors present their primary value learned value (PVLV) model for understanding the reward- predictive firing properties of dopamine (DA) neurons as an alternative to the temporal-differences (TD) algorithm. PVLV is more directly related to underlying biology and is also more robust to variability in the environment. The primary value (PV) system controls performance and learning during primary rewards, whereas the learned value (LV) system learns about conditioned stimuli. The PV system is essentially the Rescorla-Wagner/delta-rule and comprises the neurons in the ventral striatum/nucleus accumbens that inhibit DA cells. The LV system comprises the neurons in the central nucleus of the amygdala that excite DA cells. The authors show that the PVLV model can account for critical aspects of the DA firing data, making a number of clear predictions about lesion effects, several of which are consistent with existing data. For example, first- and second-order conditioning can be anatomically dissociated, which is consistent with PVLV and not TD. Overall, the model provides a biologically plausible framework for understanding the neural basis of reward learning.}
}
@book{OReillyetalCCN-12,
       author = {Randall C. O'Reilly and Yuko Munakata and Michael J. Frank and Thomas E. Hazy Contributors},
        title = {Computational Cognitive Neuroscience},
         year = {2012},
    publisher = {Wiki Book, 1st Edition},
}
@book{OReillyetalCCN-16,
       author = {Randall C. O'Reilly and Yuko Munakata and Michael J. Frank and Thomas E. Hazy Contributors},
        title = {Computational Cognitive Neuroscience},
         year = {2016},
    publisher = {Wiki Book, Third Edition},
}
@book{OReillyetalCCN-20,
       author = {Randall C. O'Reilly and Yuko Munakata and Michael J. Frank and Thomas E. Hazy Contributors},
        title = {Computational Cognitive Neuroscience},
         year = {2020},
    publisher = {Wiki Book, 4th Edition},
}
@article{OReillyetalCON-10,
     title = {Computational models of cognitive control},
    author = {Randall C. O'Reilly and Seth A. Herd and Wolfgang M. Pauli},
   journal = {Current Opinion in Neurobiology},
    volume = 20,
     issue = 2,
      year = 2010,
     pages = {257-261},
     abstract = {Cognitive control refers to the ability to perform task-relevant processing in the face of other distractions or other forms of interference, in the absence of strong environmental support. It depends on the integrity of the prefrontal cortex and associated biological structures (e.g., the basal ganglia). Computational models have played an influential role in developing our understanding of this system, and we review current developments in three major areas: dynamic gating of prefrontal representations, hierarchies in the prefrontal cortex, and reward, motivation, and goal-related processing in prefrontal cortex. Models in these and other areas are advancing the field further forward.},
}
@article {OReillyetalCS-15,
       author = {O'Reilly, Randall C. and Bhattacharyya, Rajan and Howard, Michael D. and Ketz, Nicholas},
        title = {Complementary Learning Systems},
      journal = {Cognitive Science},
    publisher = {Blackwell Publishing Ltd},
       volume = {38},
       number = {6},
         year = {2014},
        pages = {1229-1248},
     abstract = {This paper reviews the fate of the central ideas behind the complementary learning systems (CLS) framework as originally articulated in McClelland, McNaughton, and O'Reilly (1995). This framework explains why the brain requires two differentially specialized learning and memory systems, and it nicely specifies their central properties (i.e., the hippocampus as a sparse, pattern-separated system for rapidly learning episodic memories, and the neocortex as a distributed, overlapping system for gradually integrating across episodes to extract latent semantic structure). We review the application of the CLS framework to a range of important topics, including the following: the basic neural processes of hippocampal memory encoding and recall, conjunctive encoding, human recognition memory, consolidation of initial hippocampal learning in cortex, dynamic modulation of encoding versus recall, and the synergistic interactions between hippocampus and neocortex. Overall, the CLS framework remains a vital theoretical force in the field, with the empirical data over the past 15 years generally confirming its key principles.},
}
@article{OReillyetalCoRR-17,
        title = {Deep Predictive Learning: A Comprehensive Model of Three Visual Streams},
       author = {Randall C. O'Reilly and Dean R. Wyatte and John Rohrlich},
         year = {2017},
       volume = {arXiv:1709.04654},
      journal = {CoRR},
     abstract = {How does the neocortex learn and develop the foundations of all our high-level cognitive abilities? We present a comprehensive framework spanning biological, computational, and cognitive levels, with a clear theoretical continuity between levels, providing a coherent answer directly supported by extensive data at each level. Learning is based on making predictions about what the senses will report at 100 msec (alpha frequency) intervals, and adapting synaptic weights to improve prediction accuracy. The pulvinar nucleus of the thalamus serves as a projection screen upon which predictions are generated, through deep-layer 6 corticothalamic inputs from multiple brain areas and levels of abstraction. The sparse driving inputs from layer 5 intrinsic bursting neurons provide the target signal, and the temporal difference between it and the prediction reverberates throughout the cortex, driving synaptic changes that approximate error backpropagation, using only local activation signals in equations derived directly from a detailed biophysical model. In vision, predictive learning requires a carefully-organized developmental progression and anatomical organization of three pathways (What, Where, and What * Where), according to two central principles: top-down input from compact, high-level, abstract representations is essential for accurate prediction of low-level sensory inputs; and the collective, low-level prediction error must be progressively and opportunistically partitioned to enable extraction of separable factors that drive the learning of further high-level abstractions. Our model self-organized systematic invariant object representations of 100 different objects from simple movies, accounts for a wide range of data, and makes many testable predictions.}
}
@incollection{OReillyetalLEABRA-16,
       author = {Randall C. O'Reilly and Thomas E. Hazy and Seth A. Herd},
        title = {The {Leabra} Cognitive Architecture: {H}ow to Play 20 Principles with Nature and Win!},
    booktitle = {The Oxford Handbook of Cognitive Science},
       editor = {Susan E. F. Chipman},
       series = {Oxford Handbooks},
         year = {2016},
        pages = {91-115},
    publisher = {Oxford University Press}
}
@incollection{OReillyetalTACO-14,
       author = {Randall C. O'Reilly and Alex A. Petrov and Jonathan D. Cohen and Christian J. Lebiere and  Seth A. Herd and Trent Kriete},
        title = {How Limited Systematicity Emerges: A  Computational Cognitive Neuroscience Approach},
       editor = {Paco Calvo and John Symons},
    booktitle = {The Architecture of Cognition},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
         year = {2014},
        pages = {191-224},
     abstract = {In this chapter, we address the claims made by Fodor and Pylyshyn (1988). We strike a middle ground between classic symbolic and connectionist perspectives, arguing that cognition is less systematic than classicists claim, but that connectionist, neural-processing-based theories have yet to explain the extent to which cognition is systematic. We offer a sketch of an emerging understanding of the basis of human systematicity in terms of interactions between specialized brain systems, leveraging the computational principles identified and empirical work done in the quarter-century since the target work was published. We identify a full spectrum of processing mechanisms, arrayed along the continuum between context-sensitivity and combinatorial, systematic processing, each associated with different parts of the human brain. We find that attempting to understand the role of these different brain areas through the lens of systematicity results in a rich picture of human cognitive abilities.}
}
@article{ORourkeetalNATURE-12,
     title = {Deep Molecular Diversity of Mammalian Synapses: Why it matters and how to measure it},
    author = {Nancy A. O'Rourke and Nicholas C. Weiler and Kristina D. Micheva and Stephen J. Smith},
   journal = {Nature Review Neuroscience},
    volume = 10,
      year = 2012,
     pages = {365-379},
}
@article{OberauerJoC-19,
       author = {Oberauer, Klaus},
        title = {{{W}orking {M}emory and {A}ttention - {A} {C}onceptual {A}nalysis and {R}eview}},
      journal = {Journal of Cognition},
         year = {2019},
       volume = {2},
       number = {1},
        pages = {36},
     abstract = {There is broad agreement that working memory is closely related to attention. This article delineates several theoretical options for conceptualizing this link, and evaluates their viability in light of their theoretical implications and the empirical support they received. A first divide exists between the concept of attention as a limited resource, and the concept of attention as selective information processing. Theories conceptualizing attention as a resource assume that this resource is responsible for the limited capacity of working memory. Three versions of this idea have been proposed: Attention as a resource for storage and processing, a shared resource for perceptual attention and memory maintenance, and a resource for the control of attention. The first of these three is empirically well supported, but the other two are not. By contrast, when attention is understood as a selection mechanism, it is usually not invoked to explain the capacity limit of working memory - rather, researchers ask how different forms of attention interact with working memory, in two areas. The first pertains to attentional selection of the contents of working memory, controlled by mechanisms of filtering out irrelevant stimuli, and removing no-longer relevant representations from working memory. Within working memory contents, a single item is often selected into the focus of attention for processing. The second area pertains to the role of working memory in cognitive control. Working memory contributes to controlling perceptual attention - by holding templates for targets of perceptual selection - and controlling action - by holding task sets to implement our current goals.}
}
@incollection{OberauerPLM-09,
        title = {Chapter 2 Design for a Working Memory},
       series = {Psychology of Learning and Motivation},
    booktitle = {The Psychology of Learning and Motivation},
    publisher = {Academic Press},
       volume = {51},
        pages = {45-100},
         year = {2009},
       author = {Klaus Oberauer},
     abstract = {Starting from the premise that working memory is a system for providing access to representations for complex cognition, six requirements for a working memory system are delineated: (1) maintaining structural representations by dynamic bindings, (2) manipulating structural representations, (3) flexible reconfiguration, (4) partial decoupling from long-term memory, (5) controlled retrieval from long-term memory, and (6) encoding of new structures into long-term memory. The chapter proposes an architecture for a system that meets these requirements. The working memory system consists of a declarative and a procedural part, each of which has three embedded components: the activated part of long-term memory, a component for creating new structural representations by dynamic bindings (the 'region of direct access' for declarative working memory, and the 'bridge' for procedural working memory), and a mechanism for selecting a single element ('focus of attention' for declarative working memory, and 'response focus' for procedural working memory). The architecture affords two modes of information processing, an analytical and an associative mode. This distinction provides a theoretically founded formulation of a dual-process theory of reasoning.}
}
@article{ObesoetalMOVEMENT-DISORDERS-08,
       author = {Obeso, Jose A. and Rodriguez-Oroz, Maria Cruz and Benitez-Temino, Beatriz and Blesa, Franscisco J. and Guridi, Jorge and Marin, Concepci and Rodriguez, Manuel},
        title = {Functional organization of the basal ganglia: Therapeutic implications for Parkinson's disease},
      journal = {Movement Disorders},
       volume = {23},
       number = {S3},
         year = {200},
        pages = {S548-S559},
     abstract = {Abstract The basal ganglia (BG) are a highly organized network, where different parts are activated for specific functions and circumstances. The BG are involved in movement control, as well as associative learning, planning, working memory, and emotion. We concentrate on the "motor circuit" because it is the best understood anatomically and physiologically, and because Parkinson's disease is mainly thought to be a movement disorder. Normal function of the BG requires fine tuning of neuronal excitability within each nucleus to determine the exact degree of movement facilitation or inhibition at any given moment. This is mediated by the complex organization of the striatum, where the excitability of medium spiny neurons is controlled by several pre- and postsynaptic mechanisms as well as interneuron activity, and secured by several recurrent or internal BG circuits. The motor circuit of the BG has two entry points, the striatum and the subthalamic nucleus (STN), and an output, the globus pallidus pars interna (GPi), which connects to the cortex via the motor thalamus. Neuronal afferents coding for a given movement or task project to the BG by two different systems: (1) Direct disynaptic projections to the GPi via the striatum and STN. (2) Indirect trisynaptic projections to the GPi via the globus pallidus pars externa (GPe). Corticostriatal afferents primarily act to inhibit medium spiny neurons in the "indirect circuit" and facilitate neurons in the "direct circuit." The GPe is in a pivotal position to regulate the motor output of the BG. Dopamine finely tunes striatal input as well as neuronal striatal activity, and modulates GPe, GPi, and STN activity. Dopaminergic depletion in Parkinson's disease disrupts the corticostriatal balance leading to increased activity the indirect circuit and reduced activity in the direct circuit. The precise chain of events leading to increased STN activity is not completely understood, but impaired dopaminergic regulation of the GPe, GPi, and STN may be involved. The parkinsonian state is characterized by disruption of the internal balance of the BG leading to hyperactivity in the two main entry points of the network (striatum and STN) and excessive inhibitory output from the GPi. Replacement therapy with standard levodopa creates a further imbalance, producing an abnormal pattern of neuronal discharge and synchronization of neuronal firing that sustain the "off" and "on with dyskinesia" states. The effect of levodopa is robust but short-lasting and converts the parkinsonian BG into a highly unstable system, where pharmacological and compensatory effects act in opposing directions. This creates a scenario that substantially departs from the normal physiological state of the BG. © 2008 Movement Disorder Society},
}
@article{ObokataetalNATURE-14a,
       author = {Obokata, Haruko and Wakayama, Teruhiko and Sasai, Yoshiki and Kojima, Koji and Vacanti, Martin P. and Niwa, Hitoshi and Yamato, Masayuki and Vacanti, Charles A.},
        title = {Stimulus-triggered fate conversion of somatic cells into pluripotency},
      journal = {Nature},
         year = 2014,
    publisher = {Nature Publishing Group},
       volume = 505,
        issue = 7485,
        pages = {641-647},
     abstract = {Here we report a unique cellular reprogramming phenomenon, called stimulus-triggered acquisition of pluripotency (STAP), which requires neither nuclear transfer nor the introduction of transcription factors. In STAP, strong external stimuli such as a transient low-pH stressor reprogrammed mammalian somatic cells, resulting in the generation of pluripotent cells. Through real-time imaging of STAP cells derived from purified lymphocytes, as well as gene rearrangement analysis, we found that committed somatic cells give rise to STAP cells by reprogramming rather than selection. STAP cells showed a substantial decrease in DNA methylation in the regulatory regions of pluripotency marker genes. Blastocyst injection showed that STAP cells efficiently contribute to chimaeric embryos and to offspring via germline transmission. We also demonstrate the derivation of robustly expandable pluripotent cell lines from STAP cells. Thus, our findings indicate that epigenetic fate determination of mammalian cells can be markedly converted in a context-dependent manner by strong environmental cues.},
}
@article{ObokataetalNATURE-14b,
       author = {Obokata, Haruko and Sasai, Yoshiki and Niwa, Hitoshi and Kadota, Mitsutaka and Andrabi, Munazah and Takata, Nozomu and Tokoro, Mikiko and Terashita, Yukari and Yonemura, Shigenobu and Vacanti, Charles A. and Wakayama, Teruhiko},
        title = {Bidirectional developmental potential in reprogrammed cells with acquired pluripotency},
      journal = {Nature},
         year = 2014,
    publisher = {Nature Publishing Group},
       volume = 505,
        issue = 7485,
        pages = {676-680},
}
@article {OckoaetalPNAS-18,
       author = {Ocko, Samuel A. and Hardcastle, Kiah and Giocomo, Lisa M. and Ganguli, Surya},
        title = {Emergent elasticity in the neural code for space},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {115},
       number = {50},
         year = {2018},
        pages = {E11798--E11806},
     abstract = {We develop a theoretical model, grounded in known properties of neural dynamics and synaptic plasticity, that can fuse information gathered from the past history of velocity and sequence of encountered landmarks during exploratory behavior, to construct a self-consistent internal representation of space. Moreover, through model reduction techniques, we obtain conceptual insights into how consistent internal spatial representations naturally emerge through an elastic relaxation process in an effective spring{\textendash}particle system. We verify several experimentally testable predictions of our model involving the spatial behavior of grid cells in the medial entorhinal cortex, as well as suggest additional experiments.Upon encountering a novel environment, an animal must construct a consistent environmental map, as well as an internal estimate of its position within that map, by combining information from two distinct sources: self-motion cues and sensory landmark cues. How do known aspects of neural circuit dynamics and synaptic plasticity conspire to accomplish this feat? Here we show analytically how a neural attractor model that combines path integration of self-motion cues with Hebbian plasticity in synaptic weights from landmark cells can self-organize a consistent map of space as the animal explores an environment. Intriguingly, the emergence of this map can be understood as an elastic relaxation process between landmark cells mediated by the attractor network. Moreover, our model makes several experimentally testable predictions, including (i) systematic path-dependent shifts in the firing fields of grid cells toward the most recently encountered landmark, even in a fully learned environment; (ii) systematic deformations in the firing fields of grid cells in irregular environments, akin to elastic deformations of solids forced into irregular containers; and (iii) the creation of topological defects in grid cell firing patterns through specific environmental manipulations. Taken together, our results conceptually link known aspects of neurons and synapses to an emergent solution of a fundamental computational problem in navigation, while providing a unified account of disparate experimental observations.},
}
@inproceedings{OgasawaraandRussellIJCAI-93,
       author = {Ogasawara, Gary H. and Russell, Stuart},
        title = {Planning Using Multiple Execution Architectures},
    booktitle = {Proceedings {IJCAI} 13},
 organization = {IJCAII},
         year = 1993,
}
@article{OgawaetalPNAS-90,
    author = {Ogawa, S and Lee, T M and Kay, A R and Tank, D W}, 
     title = {Brain magnetic resonance imaging with contrast dependent on blood oxygenation},
   journal = {Proceedings of the National Academy of Sciences}, 
    volume = 87, 
     issue = 24, 
     pages = {9868-9872}, 
      year = 1990, 
  abstract = {Paramagnetic deoxyhemoglobin in venous blood is a naturally occurring contrast agent for magnetic resonance imaging (MRI). By accentuating the effects of this agent through the use of gradient-echo techniques in high fields, we demonstrate in vivo images of brain microvasculature with image contrast reflecting the blood oxygen level. This blood oxygenation level-dependent (BOLD) contrast follows blood oxygen changes induced by anesthetics, by insulin-induced hypoglycemia, and by inhaled gas mixtures that alter metabolic demand or blood flow. The results suggest that BOLD contrast can be used to provide in vivo real-time maps of blood oxygenation in the brain under normal physiological conditions. BOLD contrast adds an additional feature to magnetic resonance imaging and complements other techniques that are attempting to provide positron emission tomography-like measurements related to regional neural activity.}, 
}
@inproceedings{OhmeretalDICTA-08,
    author = {Ohmer, Julius Fabian and Redding, Nicholas J.},
     title = {{GPU}-Accelerated {KLT} Tracking with {M}onte-{C}arlo-Based Feature Reselection},
 booktitle = {DICTA '08: Proceedings of the 2008 Digital Image Computing: Techniques and Applications},
      year = 2008,
     pages = {234-241},
 publisher = {IEEE Computer Society},
   address = {Washington, DC, USA},
}
@article{OishietalNEUROIMAGE-08,
       author = {Kenichi Oishi and Karl Zilles and Katrin Amunts and Andreia Faria and Hangyi Jiang and Xin Li and Kazi Akhter and Kegang Hua and Roger Woods and Arthur W. Toga and G. Bruce Pike and Pedro Rosa-Neto and Alan Evans and Jiangyang Zhang and Hao Huang and Michael I. Miller and Peter C.M. van Zijl and John Mazziotta and Susumu Mori},
        title = {Human brain white matter atlas: Identification and assignment of common anatomical structures in superficial white matter},
      journal = {NeuroImage},
       volume = {43},
       number = {3},
        pages = {447-457},
         year = {2008},
     abstract = {Structural delineation and assignment are the fundamental steps in understanding the anatomy of the human brain. The white matter has been structurally defined in the past only at its core regions (deep white matter). However, the most peripheral white matter areas, which are interleaved between the cortex and the deep white matter, have lacked clear anatomical definitions and parcellations. We used axonal fiber alignment information from diffusion tensor imaging (DTI) to delineate the peripheral white matter, and investigated its relationship with the cortex and the deep white matter. Using DTI data from 81 healthy subjects, we identified nine common, blade-like anatomical regions, which were further parcellated into 21 subregions based on the cortical anatomy. Four short association fiber tracts connecting adjacent gyri (U-fibers) were also identified reproducibly among the healthy population. We anticipate that this atlas will be useful resource for atlas-based white matter anatomical studies.}
}
@article{OjalaetalPR-96,
    author = {Ojala, T. and Pietik\"{a}inen, M. and Harwood, D.},
     title = {A comparative study of texture measures with classification based on feature distributions},
   journal = {Pattern Recognition},
    volume = 29,
     pages = {51-59},
      year = 1996,
}
@article{OkunetalNATURE-15,
       author = {Okun, M. and Steinmetz, N. A. and Cossell, L. and Iacaruso, M. F. and Ko, H. and Bartho, P. and Moore, T. and Hofer, S. B. and Mrsic-Flogel, T. D. and Carandini, M. and Harris, K. D.},
        title = {Diverse coupling of neurons to populations in sensory cortex},
      journal = {Nature},
       volume = {521},
       number = {7553},
         year = {2015},
        pages = {511-515},
     abstract = {A large population of neurons can, in principle, produce an astronomical number of distinct firing patterns. In cortex, however, these patterns lie in a space of lower dimension, as if individual neurons were "obedient members of a huge orchestra". Here we use recordings from the visual cortex of mouse (Mus musculus) and monkey (Macaca mulatta) to investigate the relationship between individual neurons and the population, and to establish the underlying circuit mechanisms. We show that neighbouring neurons can differ in their coupling to the overall firing of the population, ranging from strongly coupled 'choristers' to weakly coupled 'soloists'. Population coupling is largely independent of sensory preferences, and it is a fixed cellular attribute, invariant to stimulus conditions. Neurons with high population coupling are more strongly affected by non-sensory behavioural variables such as motor intention. Population coupling reflects a causal relationship, predicting the response of a neuron to optogenetically driven increases in local activity. Moreover, population coupling indicates synaptic connectivity; the population coupling of a neuron, measured in vivo, predicted subsequent in vitro estimates of the number of synapses received from its neighbours. Finally, population coupling provides a compact summary of population activity; knowledge of the population couplings of n neurons predicts a substantial portion of their n(2) pairwise correlations. Population coupling therefore represents a novel, simple measure that characterizes the relationship of each neuron to a larger population, explaining seemingly complex network firing patterns in terms of basic circuit variables.}
}
@article{OlafsdottirCURRENT-BIOLOGY-18,
        title = {The Role of Hippocampal Replay in Memory and Planning},
       author = {H. Freyja \'{O}lafsd\'{o}ttir and Daniel Bush and Caswell Barry},
      journal = {Current Biology},
       volume = {28},
       number = {1},
        pages = {R37-R50},
         year = {2018},
     abstract = {The mammalian hippocampus is important for normal memory function, particularly memory for places and events. Place cells, neurons within the hippocampus that have spatial receptive fields, represent information about an animal's position. During periods of rest, but also during active task engagement, place cells spontaneously recapitulate past trajectories. Such 'replay' has been proposed as a mechanism necessary for a range of neurobiological functions, including systems memory consolidation, recall and spatial working memory, navigational planning, and reinforcement learning. Focusing mainly, but not exclusively, on work conducted in rodents, we describe the methodologies used to analyse replay and review evidence for its putative roles. We identify outstanding questions as well as apparent inconsistencies in existing data, making suggestions as to how these might be resolved. In particular, we find support for the involvement of replay in disparate processes, including the maintenance of hippocampal memories and decision making. We propose that the function of replay changes dynamically according to task demands placed on an organism and its current level of arousal.}
}
@article{OlafsdottiretalCURRENT-BIOLOGY-18,
       author = {H. Freyja \'{O}lafsd\'{o}ttir and Daniel Bush and Caswell Barry},
        title = {The Role of Hippocampal Replay in Memory and Planning},
      journal = {Current Biology},
       volume = {28},
       number = {1},
        pages = {R37-R50},
         year = {2018},
     abstract = {The mammalian hippocampus is important for normal memory function, particularly memory for places and events. Place cells, neurons within the hippocampus that have spatial receptive fields, represent information about an animal's position. During periods of rest, but also during active task engagement, place cells spontaneously recapitulate past trajectories. Such 'replay' has been proposed as a mechanism necessary for a range of neurobiological functions, including systems memory consolidation, recall and spatial working memory, navigational planning, and reinforcement learning. Focusing mainly, but not exclusively, on work conducted in rodents, we describe the methodologies used to analyse replay and review evidence for its putative roles. We identify outstanding questions as well as apparent inconsistencies in existing data, making suggestions as to how these might be resolved. In particular, we find support for the involvement of replay in disparate processes, including the maintenance of hippocampal memories and decision making. We propose that the function of replay changes dynamically according to task demands placed on an organism and its current level of arousal.}
}
@misc{OlahLSTM-NEURAL-NETWORK-TUTORIAL-15,
        title = {Understanding LSTM Networks},
       author = {Christopher Olah},
 howpublished = {http://colah.github.io/posts/2015-08-Understanding-LSTMs},
         year = {2015},
}
@article{OlahandCarterATTENTION-TUTORIAL-16,
       author = {Olah, Chris and Carter, Shan},
        title = {Attention and Augmented Recurrent Neural Networks},
      journal = {Distill},
          url = {http://distill.pub/2016/augmented-rnns},
         year = {2016},
}
@article{OliveiraCSH-16,
       author = {Oliveira, A. M.},
        title = {{DNA} methylation: A permissive mark in memory formation and maintenance},
      journal = {Learning and Memory},
         year = {2016},
       volume = {23},
       number = {10},
        pages = {587-593},
     abstract = {DNA methylation was traditionally viewed as a static mechanism required during cell fate determination. This view has been challenged and it is now accepted that DNA methylation is involved in the regulation of genomic responses in mature neurons, particularly in cognitive functions. The evidence for a role of DNA methylation in memory formation and maintenance comes from the increasing number of studies that have assessed the effects of manipulation of DNA methylation modifiers in the ability to form and maintain memories. Moreover, insights from genome-wide analyses of the hippocampal DNA methylation status after neuronal activity show that DNA methylation is dynamically regulated. Despite all the experimental evidence, we are still far from having a clear picture of how DNA methylation regulates long-term adaptations. This review aims on one hand to describe the findings that led to the confirmation of DNA methylation as an important player in memory formation. On the other hand, it tries to integrate these discoveries into the current views of how memories are formed and maintained.}
}
@book{OliverandSmith90,
       editor = {Oliver, Robert M. and Smith, James Q.},
        title = {Influence Diagrams, Belief Nets, and Decision Analysis},
    publisher = {John Wiley and Sons},
         year = 1990,
}
@article{OlsenandWilsonTiN-08,
       author = {Olsen, Shawn R. and Wilson, Rachel I.},
        title = {Cracking neural circuits in a tiny brain: new approaches for understanding the neural circuitry of Drosophila},
      journal = {Trends in Neurosciences},
         year = {2008},
       volume = {31},
        issue = {10},
        pages = {512-520},
     abstract = {Genetic screens in Drosophila have identified many genes involved in neural development and function. However, until recently, it has been impossible to monitor neural signals in Drosophila central neurons, and it has been difficult to make specific perturbations to central neural circuits. This has changed in the past few years with the development of new tools for measuring and manipulating neural activity in the fly. Here we review how these new tools enable novel conceptual approaches to 'cracking circuits' in this important model organism. We discuss recent studies aimed at defining the cognitive demands on the fly brain, identifying the cellular components of specific neural circuits, mapping functional connectivity in those circuits, and defining causal relationships between neural activity and behavior.},
}
@incollection{Olshausen2012,
       author = {Bruno A. Olshausen},
        title = {20 years of learning about vision: Questions answered, Questions unanswered, and Questions not yet asked},
    booktitle = {20 Years of Computational Neuroscience},
       editor = {James M. Bower},
    publisher = {Springer-Verlag},
      address = {Berlin, Heidelberg},
 series       = {Springer Series in Computational Neuroscience, Vol.\ 9},
         year = 2012
}
@inproceedings{OlshausenICIP-03, 
     author = {Bruno Olshausen},
      title = {Learning Sparse, Overcomplete Representations of Time-varying Natural Images},
  booktitle = {Proceedings of the {IEEE} International Conference on Image Processing}, 
      month = {September},
   location = {Barcelona, Spain},
  publisher = {IEEE Computer Society},
     volume = 1,
       year = 2003,
      pages = {41-44},
}
@article{OlshausenandCadieuJV-07,
    author = {Olshausen, Bruno and Cadieu, Charles},
     title = {Learning invariant and variant components of time-varying natural images},
   journal = {Journal of Vision},
    volume = 7,
     issue = 9,
     pages = {964-964},
      year = 2007,
}
@inproceedings{OlshausenandField-95,
    author = {B. A. Olshausen and D. J. Field},
     title = {Natural image statistics and efficient coding},
 booktitle = {Workshop on Information Theory and the Brain},
      year = 1995
}
@article{OlshausenandFieldNC-05,
  author = {B. A. Olshausen and D. J. Field},
   title = {How close are we to understanding {V1}?},
 journal = {Neural Computation},
    year = 2005,
  volume = 17,
   pages = {1665-1699},
}
@article{OlshausenandFieldNC-96,
    author = {B. A. Olshausen and D. J. Field},
     title = {Natural image statistics and efficient coding},
   journal = {Computation in Neural Systems},
    volume = 7,
     issue = 2,
     pages = {333-339},
      year = 1996,
}
@article{OlshausenandFieldVR-97,
   author = {B. A. Olshausen and D. J. Field}, 
    title = {Sparse coding with an overcomplete basis set: {A} strategy employed by {V1}?},
  journal = {Vision Research},
     year = 1997,
   volume = 37,
    issue = 23,
    pages = {3311-3325},
}
@article{OlshausenetalJN-93,
    author = {Olshausen, B. A. and Anderson, A. and Van Essen, D. C.},
     title = {A neurobiological model of visual attention and pattern recognition based on dynamic routing of information},
   journal = {Journal of Neuroscience},
    volume = 13,
     issue = 11,
     pages = {4700-4719},
      year = 1993,
}
@article{OosterwijketalSCAN-16,
        title = {The neural representation of typical and atypical experiences of negative images: comparing fear, disgust and morbid fascination},
       author = {Oosterwijk, Suzanne and Lindquist, Kristen A and Adebayo, Morenikeji and Barrett, Lisa Feldman},
      journal = {Social cognitive and affective neuroscience},
    publisher = {Oxford University Press},
       volume = {11},
       number = {1},
         year = {2015},
        pages = {11-22},
     abstract = {Negative stimuli do not only evoke fear or disgust, but can also evoke a state of 'morbid fascination' which is an urge to approach and explore a negative stimulus. In the present neuroimaging study, we applied an innovative method to investigate the neural systems involved in typical and atypical conceptualizations of negative images. Participants received false feedback labeling their mental experience as fear, disgust or morbid fascination. This manipulation was successful; participants judged the false feedback correct for 70\% of the trials on average. The neuroimaging results demonstrated differential activity within regions in the 'neural reference space for discrete emotion' depending on the type of feedback. We found robust differences in the ventrolateral prefrontal cortex, the dorsomedial prefrontal cortex and the lateral orbitofrontal cortex comparing morbid fascination to control feedback. More subtle differences in the dorsomedial prefrontal cortex and the lateral orbitofrontal cortex were also found between morbid fascination feedback and the other emotion feedback conditions. This study is the first to forward evidence about the neural representation of the experimen- tally unexplored state of morbid fascination. In line with a constructionist framework, our findings suggest that neural resources associated with the process of conceptualization contribute to the neural representation of this state.}
}
@article{OosterwijketalSCAN-17,
        title = {Shared states: using MVPA to test neural overlap between self-focused emotion imagery and other-focused emotion understanding},
       author = {Oosterwijk, Suzanne and Snoek, Lukas and Rotteveel, Mark and Barrett, Lisa Feldman and Scholte, H Steven},
      journal = {Social Cognitive and Affective Neuroscience},
    publisher = {Oxford University Press},
        pages = {nsx037},
         year = {2017},
     abstract = {he present study tested whether the neural patterns that support imagining 'performing an action', 'feeling a bodily sensation' or 'being in a situation' are directly involved in understanding other people's actions, bodily sensations and situations. Subjects imagined the content of short sentences describing emotional actions, interoceptive sensations and situations (self-focused task), and processed scenes and focused on how the target person was expressing an emotion, what this person was feeling, and why this person was feeling an emotion (other-focused task). Using a linear support vector machine classifier on brain-wide multi-voxel patterns, we accurately decoded each individual class in the self-focused task. When generalizing the classifier from the self-focused task to the other-focused task, we also accurately decoded whether sub- jects focused on the emotional actions, interoceptive sensations and situations of others. These results show that the neural patterns that underlie self-imagined experience are involved in understanding the experience of other people. This supports the theoretical assumption that the basic components of emotion experience and understanding share resources in the brain.}
}
@inproceedings{OpeltetalCVPR-06, 
      title = {Incremental learning object detectors using a visual shape alphabet}, 
     author = {A. Opelt and A. Pinz and A. Zisserman},
  booktitle = {Proceedings of the 2006 {IEEE} Conference on Computer Vision and Pattern Recognition}, 
  publisher = {IEEE Computer Society},
     volume = 1, 
      pages = {3-10}, 
       year = 2006, 
}
@inproceedings{OpeltetalECCV-06,
    author = {Andreas Opelt and Axel Pinz and Andrew Zisserman},
     title = {A Boundary-Fragment-Model for Object Detection},
 booktitle = {Proceedings of the 9th European Conference on Computer Vision},
    volume = 2,
     pages = {575-588},
      year = 2006,
}
@book{Oppenheim83,
       author = {Oppenheim, Alan V. and Willsky, Allan S. and Young, Ian T.},
        title = {Signals and Systems},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1983,
}
@book{OramandTalbott91,
       author = {Andrew Oram and Steve Talbott},
        title = {Managing Projects with make},
    publisher = {Second edition, O'Reilly},
      address = {Sebastopol, CA},
         year = 1991,
}
@article{OranetalSCIENCE-18,
       author = {Oran, Daniel and Rodriques, Samuel G. and Gao, Ruixuan and Asano, Shoh and Skylar-Scott, Mark A. and Chen, Fei and Tillberg, Paul W. and Marblestone, Adam H. and Boyden, Edward S.},
        title = {3D nanofabrication by volumetric deposition and controlled shrinkage of patterned scaffolds},
      journal = {Science},
       volume = {362},
       number = {6420},
        pages = {1281-1285},
         year = {2018},
    publisher = {American Association for the Advancement of Science},
     abstract = {Although a range of materials can now be fabricated using additive manufacturing techniques, these usually involve assembly of a series of stacked layers, which restricts three-dimensional (3D) geometry. Oran et al. developed a method to print a range of materials, including metals and semiconductors, inside a gel scaffold (see the Perspective by Long and Williams). When the hydrogels were dehydrated, they shrunk 10-fold, which pushed the feature sizes down to the nanoscale.Science, this issue p. 1281; see also p. 1244Lithographic nanofabrication is often limited to successive fabrication of two-dimensional (2D) layers. We present a strategy for the direct assembly of 3D nanomaterials consisting of metals, semiconductors, and biomolecules arranged in virtually any 3D geometry. We used hydrogels as scaffolds for volumetric deposition of materials at defined points in space. We then optically patterned these scaffolds in three dimensions, attached one or more functional materials, and then shrank and dehydrated them in a controlled way to achieve nanoscale feature sizes in a solid substrate. We demonstrate that our process, Implosion Fabrication (ImpFab), can directly write highly conductive, 3D silver nanostructures within an acrylic scaffold via volumetric silver deposition. Using ImpFab, we achieve resolutions in the tens of nanometers and complex, non{\textendash}self-supporting 3D geometries of interest for optical metamaterials.},
}
@article{OrlitskyetalTIT-06,
     author = {Alon Orlitsky and Narayana P. Santhanam and Krishnamurthy Viswanathan and Junan Zhang},
      title = {Limit Results on Pattern Entropy},
    journal = {IEEE Transactions on Information Theory},
     volume = 52,
      issue = 7,
       year = 2006,
      pages = {2954-2964},
}
@incollection{OrmoneitetalNIPS-01,
       author = {Ormoneit, D. and Sidenbladh, S. and Black, M. J. and Hastie, T.},
        title = {Learning and tracking cyclic human motion},
    booktitle = {Advances in Neural Information Processing Systems 13},
       editor = {Leen, Todd K. and Dietterich, Thomas G. and Tresp, Volker},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2001,
        pages = {894-900},
}
@book{Ornish2010,
        title = {Dr. Dean Ornish's Program for Reversing Heart Disease: The Only System Scientifically Proven to Reverse Heart Disease Without Drugs Or Surgery},
       author = {Ornish, D.},
         year = {2010},
    publisher = {Random House Publishing Group}
}
@book{Ortonyetal1988,
    author = {Ortony, A. and Clore, G. and Collins, A.},
     title = {The cognitive structure of emotions},
      year = 1988,
  publisher = {Cambridge University Press},
    address = {New York, NY},
}
@article{OsaetalCoRR-19,
       author = {Takayuki Osa and Voot Tangkaratt and Masashi Sugiyama},
        title = {Hierarchical Reinforcement Learning via Advantage-Weighted Information Maximization},
      journal = {CoRR},
       volume = {arXiv:1901.01365},
         year = {2019},
     abstract = {Real-world tasks are often highly structured. Hierarchical reinforcement learning (HRL) has attracted research interest as an approach for leveraging the hierarchical structure of a given task in reinforcement learning (RL). However, identifying the hierarchical policy structure that enhances the performance of RL is not a trivial task. In this paper, we propose an HRL method that learns a latent variable of a hierarchical policy using mutual information maximization. Our approach can be interpreted as a way to learn a discrete and latent representation of the state-action space. To learn option policies that correspond to modes of the advantage function, we introduce advantage-weighted importance sampling. In our HRL method, the gating policy learns to select option policies based on an option-value function, and these option policies are optimized based on the deterministic policy gradient method. This framework is derived by leveraging the analogy between a monolithic policy in standard RL and a hierarchical policy in HRL by using a deterministic option policy. Experimental results indicate that our HRL approach can learn a diversity of options and that it can enhance the performance of RL in continuous control tasks.},
}
@incollection{OsinderoandHintonNIPS-08,
      title = {Modeling image patches with a directed hierarchy of Markov random fields},
     author = {Simon Osindero and Geoffrey Hinton},
  booktitle = {Advances in Neural Information Processing Systems 20},
     editor = {J.C. Platt and D. Koller and Y. Singer and S. Roweis},
  publisher = {MIT Press},
    address = {Cambridge, MA},
      pages = {1121-1128},
       year = 2008
}
@article{OstendorfetalASSP-96,
       author = {M. Ostendorf and V. Digalakis and O. Kimball},
        title = {From {HMM}s to segment models: a unified view of stochastic modeling for speech recognition},
      journal = {{IEEE} Transactions on Acoustics, Speech and Signal Processing},
       volume = 4,
        pages = {360-378},
         year = 1996,
}
@book{Ott2002,
        title = {Chaos in Dynamical Systems},
       author = {Ott, Edward},
         year = {2002},
    publisher = {Cambridge University Press}
}
@article{OttandNiederTiCS-19,
       author = {Torben Ott and Andreas Nieder},
        title = {Dopamine and Cognitive Control in Prefrontal Cortex},
      journal = {Trends in Cognitive Sciences},
       volume = {23},
       number = {3},
        pages = {213-234},
         year = {2019},
     abstract = {Cognitive control, the ability to orchestrate behavior in accord with our goals, depends on the prefrontal cortex. These cognitive functions are heavily influenced by the neuromodulator dopamine. We review here recent insights exploring the influence of dopamine on neuronal response properties in prefrontal cortex (PFC) during ongoing behaviors in primates. This review suggests three major computational roles of dopamine in cognitive control: (i) gating sensory input, (ii) maintaining and manipulating working memory contents, and (iii) relaying motor commands. For each of these roles, we propose a neuronal microcircuit based on known mechanisms of action of dopamine in PFC, which are corroborated by computational network models. This conceptual approach accounts for the various roles of dopamine in prefrontal executive functioning.}
}
@article{OuchietalCRN-17,
       author = {Ayako Ouchi and Nobuyoshi Matsumoto and Yuji Ikegaya},
        title = {Activation of Hilar Mossy Cells and Dentate Granule Cells During Sharp Wave/Ripples},
      journal = {Current Research in Neuroscience},
       volume = {7},
        issue = {1},
         year = {2017},
        pages = {1-8},
     abstract = {Background and Objective: Sharp waves (SWs)/ripples are high-frequency oscillations emitted by the hippocampus mainly during slow-wave sleep or quiet rest states and contribute to offline memory consolidation. They are thought to be initiated in the CA3 subregion and propagate to the downstream regions, including the CA1 area, the subiculum and the entorhinal cortex. At the same time, they also propagate from the CA3 subregion back to the dentate gyrus. However, neither the role of the CA3-to-DG backpropagation nor its circuit mechanism has been fully understood. Hilar mossy cells receive direct synaptic inputs from CA3 pyramidal cells and make synaptic connections with granule cells in the dentate gyrus. Methodology: The synaptic inputs to mossy cells and granule cells in response to SWs were investigated by using whole-cell recordings and extracellular recordings from acute brain slices that spontaneously emit SWs, therefore, mossy cells or granule cells were patch-clamped, while recording the local field potentials from the CA3 stratum pyramidale. Results: Results of this study showed that the mossy cells exhibited excitatory synaptic responses to SWs earlier than granule cells. The time lags of EPSPs relative to the SW onsets were 7.3±3.9 msec in mossy cells and 15.7±7.8 msec in granule cells (SDs of 10 and 3 cells, respectively, p = 0.023, studentʼs t-test). Furthermore, the ratio of the SW responsive cells was significantly higher in the mossy cells than the granule cells (mossy cell: 10/17 cells, granule cell: 3/15 cells, p = 0.036, Fisherʼs exact test). Conclusion: It can be concluded that the synaptic responses of mossy cells to SWs were likely stochastic but more reliable than those of granule cells. Our findings shed light on a novel mechanism underlying the SW backpropagation and provide a new perspective about memory consolidation.}
}
@book{Ousterhout2018,
        title = {A Philosophy of Software Design},
       author = {Ousterhout, John K.},
    publisher = {Yaknyam Press},
         year = {2018},
}
@article{OuzounovetalNATURE-METHODS-17,
       author = {Ouzounov, Dimitre G. and Wang, Tianyu and Wang, Mengran and Feng, Danielle D. and Horton, Nicholas G. and Cruz-Hernandez, Jean C. and Cheng, Yu-Ting and Reimer, Jacob and Tolias, Andreas S. and Nishimura, Nozomi and Xu, Chris},
        title = {In vivo three-photon imaging of activity of GCaMP6-labeled neurons deep in intact mouse brain},
      journal = {Nature Methods},
         year = {2017},
    publisher = {Nature Publishing Group},
       volume = {14},
        issue = {4},
        pages = {388--390},
     abstract = {high-resolution optical imaging is critical to understanding brain function. We demonstrate that three-photon microscopy at 1,300-nm excitation enables functional imaging of GcamP6s-labeled neurons beyond the depth limit of two- photon microscopy. We record spontaneous activity from up to 150 neurons in the hippocampal stratum pyramidale at ~1-mm depth within an intact mouse brain. our method creates opportunities for noninvasive recording of neuronal activity with high spatial and temporal resolution deep within scattering brain tissues.}
}
@article{OztasNEUROANATOMY-03,
      title = {Neuronal Tracing},
     author = {Emin Oztas},
    journal = {Neuroanatomy},
     volume = 2,
       year = 2003,
      pages = {2-5},
}
@article{PachitariuetalBIORXIV-16,
       author = {Pachitariu, Marius and Stringer, Carsen and Schr{\"{o}}der, Sylvia and Dipoppa, Mario and Rossi, L. Federico and Carandini, Matteo and Harris, Kenneth D.},
        title = {Suite2p: beyond 10,000 neurons with standard two-photon microscopy},
      journal = {bioRxiv},
         year = {2016},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {The combination of two-photon microscopy recordings and powerful calcium-dependent fluorescent sensors enables simultaneous recording of unprecedentedly large populations of neurons. While these sensors have matured over several generations of development, computational methods to process their fluorescence remain inefficient and the results hard to interpret. Here, we introduce a set of practical methods based on novel clustering algorithms, and provide a complete pipeline from raw image data to neuronal calcium traces to inferred spike times. We formulate a generative model of the fluorescence image, incorporating spike times and a spatially smooth neuropil signal, and solve the inference and learning problems using a fast algorithm. This implementation scales linearly with the number of recorded cells, and the complete pipeline runs in approximately one hour for typical two-hour long recordings, on commodity GPUs. Furthermore, this method recovers twice as many cells as a previous standard method. This allowed us to routinely record and detect \~10,000 cells simultaneously from the visual cortex of awake mice using standard two-photon resonant-scanning microscopes. The software is publicly available at github.com/cortex-lab/Suite2P, together with a graphical user interface that allows rapid manual curation of the results.}
}
@incollection{PachitariuetalNIPS-13,
        title = {Recurrent linear models of simultaneously-recorded neural populations},
       author = {Pachitariu, Marius and Petreska, Biljana and Sahani, Maneesh},
    booktitle = {Advances in Neural Information Processing Systems 26},
       editor = {C. J. C. Burges and L. Bottou and M. Welling and Z. Ghahramani and K. Q. Weinberger},
    publisher = {Curran Associates, Inc.},
         year = {2013},
        pages = {3138-3146},
     abstract = {Population neural recordings with long-range temporal structure are often best understood in terms of a common underlying low-dimensional dynamical process. Advances in recording technology provide access to an ever-larger fraction of the population, but the standard computational approaches available to identify the collective dynamics scale poorly with the size of the dataset. We describe a new, scalable approach to discovering low-dimensional dynamics that underlie simultaneously recorded spike trains from a neural population. We formulate the Recurrent Linear Model (RLM) by generalising the Kalman-filter-based likelihood calculation for latent linear dynamical systems to incorporate a generalised-linear observation process. We show that RLMs describe motor-cortical population data better than either directly-coupled generalised-linear models or latent linear dynamical system models with generalised-linear observations. We also introduce the cascaded generalised-linear model (CGLM) to capture low-dimensional instantaneous correlations in neural populations. The CGLM describes the cortical recordings better than either Ising or Gaussian models and, like the RLM, can be fit exactly and quickly. The CGLM can also be seen as a generalisation of a low-rank Gaussian model, in this case factor analysis. The computational tractability of the RLM and CGLM allow both to scale to very high-dimensional neural data.}
}
@article{PackardetalPHYSICAL-REVIEW-LETTERS-80,
       author = {Packard, N. H. and Crutchfield, J. P. and Farmer, J. D. and Shaw, R. S.},
        title = {Geometry from a Time Series},
      journal = {Physical Review Letters},
       volume = 45,
        issue = 9,
        pages = {712-716},
         year = 1980,
}
@article{PackerandYusteJoN-11,
        title = {Dense, unspecific connectivity of neocortical parvalbumin-positive interneurons: a canonical microcircuit for inhibition?},
       author = {Packer, Adam M and Yuste, Rafael},
      journal = {The Journal of Neuroscience},
       volume = 31,
        issue = 37,
        pages = {13260-13271},
         year = 2011,
    publisher = {Neuroscience Society},
     abstract = {GABAergic interneurons play a major role in the function of the mammalian neocortex, but their circuit connectivity is still poorly understood. We used two-photon RuBi-Glutamate uncaging to optically map how the largest population of cortical interneurons, the parvalbumin-positive cells (PV+), are connected to pyramidal cells (PCs) in mouse neocortex. We found locally dense connectivity from PV+ interneurons onto PCs across cortical areas and layers. In many experiments, all nearby PV+ cells were connected to every local PC sampled. In agreement with this, we found no evidence for connection specificity, as PV+ interneurons contacted PC pairs similarly regardless of whether they were synaptically connected or not. We conclude that the microcircuit architecture for PV+ interneurons, and probably neocortical inhibition in general, is an unspecific, densely homogenous matrix covering all nearby pyramidal cells.}
}
@article{PackeretalNATURE-13,
     title = {Targeting neurons and photons for optogenetics},
    author = {Packer, Adam M. and Roska, Botond and Hausser, Michael},
   journal = {Nature Neuroscience},
 publisher = {Nature Publishing Group},
    volume = 16,
     issue = 7,
      year = 2013,
     pages = {805-815},
}
@article{PackeretalNATURE-METHODS-15,
       author = {Packer, Adam M. and Russell, Lloyd E. and Dalgleish, Henry W. P. and Hausser, Michael},
        title = {Simultaneous all-optical manipulation and recording of neural circuit activity with cellular resolution in vivo},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = {12},
        issue = {2},
         year = {2015},
        pages = {140-146},
     abstract = {We describe an all-optical strategy for simultaneously manipulating and recording the activity of multiple neurons with cellular resolution in vivo. We performed simultaneous two-photon optogenetic activation and calcium imaging by coexpression of a red-shifted opsin and a genetically encoded calcium indicator. A spatial light modulator allows tens of user-selected neurons to be targeted for spatiotemporally precise concurrent optogenetic activation, while simultaneous fast calcium imaging provides high-resolution network-wide readout of the manipulation with negligible optical cross-talk. Proof-of-principle experiments in mouse barrel cortex demonstrate interrogation of the same neuronal population during different behavioral states and targeting of neuronal ensembles based on their functional signature. This approach extends the optogenetic toolkit beyond the specificity obtained with genetic or viral approaches, enabling high-throughput, flexible and long-term optical interrogation of functionally defined neural circuits with single-cell and single-spike resolution in the mouse brain in vivo.},
}
@book{Pagden2013,
       author = {Anthony Pagden},
        title = {The Enlightenment and Why It Still Matters},
         year = 2013,
    publisher = {Random House},
      address = {New York, NY},
     abstract = {One of our most renowned and brilliant historians takes a fresh look at the revolutionary intellectual movement that laid the foundation for the modern world. Liberty and equality. Human rights. Freedom of thought and expression. Belief in reason and progress. The value of scientific inquiry. These are just some of the ideas that were conceived and developed during the Enlightenment, and which changed forever the intellectual landscape of the Western world. Spanning hundreds of years of history, Anthony Pagden traces the origins of this seminal movement, showing how Enlightenment concepts directly influenced modern culture, making possible a secular, tolerant, and, above all, cosmopolitan world. Everyone can agree on its impact. But in the end, just what was Enlightenment? A cohesive philosophical project? A discrete time period in the life of the mind when the superstitions of the past were overthrown and reason and equality came to the fore? Or an open-ended intellectual process, a way of looking at the world and the human condition, that continued long after the eighteenth century ended? To address these questions, Pagden introduces us to some of the unforgettable characters who defined the Enlightenment, including David Hume, the Scottish skeptic who advanced the idea of a universal "science of man"; Francois-Marie Arouet, better known to the world as Voltaire, the acerbic novelist and social critic who challenged the authority of the Catholic Church; and Immanuel Kant, the reclusive German philosopher for whom the triumph of a cosmopolitan world represented the final stage in mankind's evolution. Comprehensive in his analysis of this heterogeneous group of scholars and their lasting impact on the world, Pagden argues that Enlightenment ideas go beyond the "empire of reason" to involve the full recognition of the emotional ties that bind all human beings together. The "human science" developed by these eminent thinkers led to a universalizing vision of humanity, a bid to dissolve the barriers past generations had attempted to erect between the different cultures of the world.}
}
@book{PagelandKirshtein2017,
        title = {Machine Dreaming and Consciousness},
       author = {Pagel, J.F. and Kirshtein, P.},
         year = {2017},
    publisher = {Elsevier Science}
}
@article{PagesetalFiN-15,
       author = {Pag\`{e}s, St\`{e}phane and Cane, Michele and Randall, J\`{e}r\^{o}me and Capello, Luca and Holtmaat, Anthony},
        title = {Single cell electroporation for longitudinal imaging of synaptic structure and function in the adult mouse neocortex in vivo},
      journal = {Frontiers in Neuroanatomy},
       volume = {9},
        pages = {36},
         year = {2015},
     abstract = {Longitudinal imaging studies of neuronal structures in vivo have revealed rich dynamics in dendritic spines and axonal boutons. Spines and boutons are considered to be proxies for synapses. This implies that synapses display similar dynamics. However, spines and boutons do not always bear synapses, some may contain more than one, and dendritic shaft synapses have no clear structural proxies. In addition, synaptic strength is not always accurately revealed by just the size of these structures. Structural and functional dynamics of synapses could be studied more reliably using fluorescent synaptic proteins as markers for size and function. These proteins are often large and possibly interfere with circuit development, which renders them less suitable for conventional transfection or transgenesis methods such as viral vectors, in utero electroporation and germline transgenesis. Single cell electroporation has been shown to be a potential alternative for transfection of recombinant fluorescent proteins in adult cortical neurons. Here we provide proof of principle for the use of single cell electroporation to express and subsequently image fluorescently tagged synaptic proteins over days to weeks in vivo.}
}
@article{PagetandLonstaffTIP-98,
       author = {R. Paget and I.D. Longstaff},
        title = {Texture synthesis via a noncausal nonparametric multiscale {M}arkov random field},
      journal = {IEEE Transactions Image Processing},
         year = 1998,
       volume = 7,
        issue = 6,
        pages = {925-931},
}
@article{PaietalDEVELOPMENT-11,
     author = {Pai, Vaibhav P. and Aw, Sherry and Shomrat, Tal and Lemire, Joan M. and Levin, Michael}, 
      title = {Transmembrane voltage potential controls embryonic eye patterning in Xenopus laevis}, 
    journal = {Development},
       year = 2011, 
   abstract = {Uncovering the molecular mechanisms of eye development is crucial for understanding the embryonic morphogenesis of complex structures, as well as for the establishment of novel biomedical approaches to address birth defects and injuries of the visual system. Here, we characterize change in transmembrane voltage potential (Vmem) as a novel biophysical signal for eye induction in Xenopus laevis. During normal embryogenesis, a striking hyperpolarization demarcates a specific cluster of cells in the anterior neural field. Depolarizing the dorsal lineages in which these cells reside results in malformed eyes. Manipulating Vmem of non-eye cells induces well-formed ectopic eyes that are morphologically and histologically similar to endogenous eyes. Remarkably, such ectopic eyes can be induced far outside the anterior neural field. A Ca2+ channel-dependent pathway transduces the Vmem signal and regulates patterning of eye field transcription factors. These data reveal a new, instructive role for membrane voltage during embryogenesis and demonstrate that Vmem is a crucial upstream signal in eye development. Learning to control bioelectric initiators of organogenesis offers significant insight into birth defects that affect the eye and might have significant implications for regenerative approaches to ocular diseases.}, 
}
@inproceedings{PaietalSENSORS-14,
       author = {P. Pai and L. Chen and M. Tabib-Azar},
    booktitle = {IEEE SENSORS 2014 Proceedings},
        title = {Fiber optic magnetometer with sub-pico Tesla sensitivity for magneto-encephalography},
         year = {2014},
        pages = {722-725},
     abstract = {This work reports a highly sensitive resonant Faraday force magnetometer with a minimum detectable rms magnetic field of 277 fT at 17.2 Hz suitable for magneto-encephalography and brain circuit imaging at room temperature. At this level of sensitivity, our sensor can favorably compete with superconducting quantum interference devices (SQUIDs) without requiring liquid helium cooling and with the possibility of monitoring brain activity in humans behaving in our natural environment. In addition to room temperature operation, our sensor is more compact (~4 mm2 device area) compared to SQUIDs and has power dissipation comparable to MEMS devices. The light intensity coupled to the fiber optic waveguide varies exponentially with the displacement of the sensing cantilever and in contrast to electron tunneling detectors, the optical magnetometer has a large dynamic range and high sensitivity.},
}
@article{PainetalIJNME-99,
       author = {Pain, C. C. and De Oliveira, C. R. E and Goddard, A. J. H.},
        title = {A neural network graph partitioning procedure for grid-based domain decomposition},
      journal = {International Journal for Numerical Methods in Engineering},
       volume = {44},
       number = {5},
    publisher = {John Wiley \& Sons, Ltd.},
         year = {1999},
        pages = {593-613},
     abstract = {This paper describes a neural network graph partitioning algorithm which partitions unstructured finite element/volume meshes as a precursor to a parallel domain decomposition solution method. The algorithm works by first constructing a coarse graph approximation using an automatic graph coarsening method. The coarse graph is partitioned and the results are interpolated onto the original graph to initialize an optimization of the graph partition problem. In practice, a hierarchy of (usually more than two) graphs are used to help obtain the final graph partition. A mean field theorem neural network is used to perform all partition optimization. The partitioning method is applied to graphs derived from unstructured finite element meshes and in this context it can be viewed as a multi-grid partitioning method.},
}
@article{PalangietalCoRR-15,
       author = {Hamid Palangi and Li Deng and Yelong Shen and Jianfeng Gao and Xiaodong He and Jianshu Chen and Xinying Song and Rabab K. Ward},
        title = {Deep Sentence Embedding Using the Long Short Term Memory Network: Analysis and Application to Information Retrieval},
      journal = {CoRR},
       volume = {arXiv:1502.06922},
         year = {2015},
     abstract = {This paper develops a model that addresses sentence embedding, a hot topic in current natural language processing research, using recurrent neural networks with Long Short-Term Memory (LSTM) cells. Due to its ability to capture long term memory, the LSTM-RNN accumulates increasingly richer information as it goes through the sentence, and when it reaches the last word, the hidden layer of the network provides a semantic representation of the whole sentence. In this paper, the LSTM-RNN is trained in a weakly supervised manner on user click-through data logged by a commercial web search engine. Visualization and analysis are performed to understand how the embedding process works. The model is found to automatically attenuate the unimportant words and detects the salient keywords in the sentence. Furthermore, these detected keywords are found to automatically activate different cells of the LSTM-RNN, where words belonging to a similar topic activate the same cell. As a semantic representation of the sentence, the embedding vector can be used in many different applications. These automatic keyword detection and topic allocation abilities enabled by the LSTM-RNN allow the network to perform document retrieval, a difficult language processing task, where the similarity between the query and documents can be measured by the distance between their corresponding sentence embedding vectors computed by the LSTM-RNN. On a web search task, the LSTM-RNN embedding is shown to significantly outperform several existing state of the art methods. We emphasize that the proposed model generates sentence embedding vectors that are specially useful for web document retrieval tasks. A comparison with a well known general sentence embedding method, the Paragraph Vector, is performed. The results show that the proposed method in this paper significantly outperforms it for web document retrieval task.},
} 
@article{PallyJAP3-10,
       author = {Pally, Regina},
        title = {The brain's shared circuits of interpersonal understanding: implications for psychoanalysis and psychodynamic psychotherapy},
      journal = {Journal of the Academy of Psychoanalysis and Psychodynamic Psychotherapy},
         year = {2010},
       volume = {38},
       number = {3},
        pages = {381-411},
     abstract = {Social Neuroscience maintains that human survival depends on interpersonal relations, and that shared circuits evolved to enhance our ability to interact with and understand other people. Shared circuits operate by re-creating the Other's experience in the same brain regions used for Self experience. The interpersonal understanding made possible by shared circuits is, for the most part, outside conscious awareness and plays a role in the transference-counter transference interaction. The brain mechanisms of shared circuits are presented and clinical vignettes illustrate the use of the concept of shared circuits in the clinical setting.}
}
@article{PalmetalBC-14,
       author = {Palm, G{\"{u}}nther and Knoblauch, Andreas and Hauser, Florian and Sch{\"{u}}z, Almut},
        title = {Cell assemblies in the cerebral cortex},
      journal = {Biological Cybernetics},
         year = {2014},
       volume = {108},
        issue = {5},
        pages = {559-572},
     abstract = {Donald Hebb's concept of cell assemblies is a physiology-based idea for a distributed neural representation of behaviorally relevant objects, concepts, or constellations. In the late 70s Valentino Braitenberg started the endeavor to spell out the hypothesis that the cerebral cortex is the structure where cell assemblies are formed, maintained and used, in terms of neuroanatomy (which was his main concern) and also neurophysiology. This endeavor has been carried on over the last 30 years corroborating most of his findings and interpretations. This paper summarizes the present state of cell assembly theory, realized in a network of associative memories, and of the anatomical evidence for its location in the cerebral cortex.},
}
@inproceedings{PalpanasetalICDM-04,
       author = {E. Palpanas and D. Gunopulos and W. Truppel and E. Keogh and M. Vlachos},
        title = {Online amnesic approximation of streaming time series},
    booktitle = {Proceedings of the IEEE International Conference on Data Mining},
        pages = {338-349},
         year = 2004,
}
@article{PandarinathetalBIORXIV-17,
       author = {Pandarinath, Chethan and O{\textquoteright}Shea, Daniel J. and Collins, Jasmine and Jozefowicz, Rafal and Stavisky, Sergey D. and Kao, Jonathan C. and Trautmann, Eric M. and Kaufman, Matthew T. and Ryu, Stephen I. and Hochberg, Leigh R. and Henderson, Jaimie M. and Shenoy, Krishna V. and Abbott, Larry F. and Sussillo, David},
        title = {Inferring single-trial neural population dynamics using sequential auto-encoders},
         year = {2017},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {Neuroscience is experiencing a data revolution in which simultaneous recording of many hundreds or thousands of neurons is revealing structure in population activity that is not apparent from single-neuron responses. This structure is typically extracted from trial-averaged data. Single-trial analyses are challenging due to incomplete sampling of the neural population, trial-to-trial variability, and fluctuations in action potential timing. Here we introduce Latent Factor Analysis via Dynamical Systems (LFADS), a deep learning method to infer latent dynamics from single-trial neural spiking data. LFADS uses a nonlinear dynamical system (a recurrent neural network) to infer the dynamics underlying observed population activity and to extract {\textquoteleft}de-noised{\textquoteright} single-trial firing rates from neural spiking data. We apply LFADS to a variety of monkey and human motor cortical datasets, demonstrating its ability to predict observed behavioral variables with unprecedented accuracy, extract precise estimates of neural dynamics on single trials, infer perturbations to those dynamics that correlate with behavioral choices, and combine data from non-overlapping recording sessions (spanning months) to improve inference of underlying dynamics. In summary, LFADS leverages all observations of a neural population{\textquoteright}s activity to accurately model its dynamics on single trials, opening the door to a detailed understanding of the role of dynamics in performing computation and ultimately driving behavior.},
}
@article{PandeyandNicholsPR-11,
       author = {Pandey, Udai Bhan and Nichols, Charles D.},
       editor = {Barker, Eric L.},
        title = {Human Disease Models in Drosophila melanogaster and the Role of the Fly in Therapeutic Drug Discovery},
      journal = {Pharmacological Review},
         year = {2011},
    publisher = {The American Society for Pharmacology and Experimental Therapeutics},
      address = {Bethesda, MD},
       volume = {63},
        issue = {2},
        pages = {411-436},
     abstract = {The common fruit fly, Drosophila melanogaster, is a well studied and highly tractable genetic model organism for understanding molecular mechanisms of human diseases. Many basic biological, physiological, and neurological properties are conserved between mammals and D. melanogaster, and nearly 75\% of human disease-causing genes are believed to have a functional homolog in the fly. In the discovery process for therapeutics, traditional approaches employ high-throughput screening for small molecules that is based primarily on in vitro cell culture, enzymatic assays, or receptor binding assays. The majority of positive hits identified through these types of in vitro screens, unfortunately, are found to be ineffective and/or toxic in subsequent validation experiments in whole-animal models. New tools and platforms are needed in the discovery arena to overcome these limitations. The incorporation of D. melanogaster into the therapeutic discovery process holds tremendous promise for an enhanced rate of discovery of higher quality leads. D. melanogaster models of human diseases provide several unique features such as powerful genetics, highly conserved disease pathways, and very low comparative costs. The fly can effectively be used for low- to high-throughput drug screens as well as in target discovery. Here, we review the basic biology of the fly and discuss models of human diseases and opportunities for therapeutic discovery for central nervous system disorders, inflammatory disorders, cardiovascular disease, cancer, and diabetes. We also provide information and resources for those interested in pursuing fly models of human disease, as well as those interested in using D. melanogaster in the drug discovery process.},
}
@article{PanieretalFiNC-13,
       author = {Panier, Thomas and Romano, Sebasti\'{a}n and Olive, Rapha\"{e}l and Pietri, Thomas and Sumbre, Germ\'{a}n and Candelier, Rapha\"{e}l and Debr\'{e}geas, Georges},
        title = {Fast functional imaging of multiple brain regions in intact zebrafish larvae using Selective Plane Illumination Microscopy},
      journal = {Frontiers in Neural Circuits},
       volume = {7},
       number = {65},
        pages = {1-11},
         year = {2013},
     abstract = {The optical transparency and the small dimensions of zebrafish at the larval stage make it a vertebrate model of choice for brain-wide in-vivo functional imaging. However, current point-scanning imaging techniques, such as two-photon or confocal microscopy, impose a strong limit on acquisition speed which in turn sets the number of neurons that can be simultaneously recorded. At 5Hz, this number is of the order of one thousand, i.e. approximately 1-2\% of the brain. Here we demonstrate that this limitation can be greatly overcome by using Selective-plane Illumination Microscopy (SPIM). Zebrafish larvae expressing the genetically encoded calcium indicator GCaMP3 were illuminated with a scanned laser sheet and imaged with a camera whose optical axis was oriented orthogonally to the illumination plane. This optical sectioning approach was shown to permit functional imaging of a very large fraction of the brain volume of 5-9 day old larvae with single- or near single-cell resolution. The spontaneous activity of up to 5000 neurons was recorded at 20Hz for 20-60min. By rapidly scanning the specimen in the axial direction, the activity of  25000 individual neurons from 5 different z-planes (approximately 30\% of the entire brain) could be simultaneously monitored at 4Hz. Compared to point-scanning techniques, this imaging strategy thus yields a ~20-fold increase in data throughput (number of recorded neurons times a`cquisition rate) without compromising the signal-to-noise ratio. The extended field of view offered by the SPIM method allowed us to directly identify large scale ensembles of neurons, spanning several brain regions, that displayed correlated activity and were thus likely to participate in common neural processes. The benefits and limitations of SPIM for functional imaging in zebrafish as well as future developments are briefly discussed.}
}
@article{PanseretalCURRENT-BIOLOGY-16,
       author = {Panser, Karin and Tirian, Laszlo and Schulze, Florian and Villalba, Santiago and Jefferis, Gregory S.X.E. and B\"{u}hler, Katja and Straw, Andrew D.},
        title = {Automatic Segmentation of Drosophila Neural Compartments Using {GAL4} Expression Data Reveals Novel Visual Pathways},
      journal = {Current Biology},
    publisher = {Cell Press},
       volume = {26},
        issue = {15},
         year = {2016},
        pages = {1943-1954},
     abstract = {Identifying distinct anatomical structures within the brain and developing genetic tools to target them are fundamental steps for understanding brain function. We hypothesize that enhancer expression patterns can be used to automatically identify functional units such as neuropils and fiber tracts. We used two recent, genome-scale Drosophila GAL4 libraries and associated confocal image datasets to segment large brain regions into smaller subvolumes. Our results (available at https://strawlab.org/braincode) support this hypothesis because regions with well-known anatomy, namely the antennal lobes and central complex, were automatically segmented into familiar compartments. The basis for the structural assignment is clustering of voxels based on patterns of enhancer expression. These initial clusters are agglomerated to make hierarchical predictions of structure. We applied the algorithm to central brain regions receiving input from the optic lobes. Based on the automated segmentation and manual validation, we can identify and provide promising driver lines for 11 previously identified and 14 novel types of visual projection neurons and their associated optic glomeruli. The same strategy can be used in other brain regions and likely other species, including vertebrates.},
}
@inproceedings{Pantetal02,
       author = {Gautam Pant and Padmini Srinivasan and Filippo Menczer},
        title = {Exploration versus Exploitation in Topic Driven Crawlers},
    booktitle = {Proceedings of the Second International Workshop on Web Dynamics},
         year = 2002,
}
@article{PanzeriPRS_B-99,
       author = {Panzeri, S. and Schultz, S. R. and Treves, A. and Rolls, E. T.},
        title = {Correlations and the encoding of information in the nervous system},
      journal = {Proceedings of the Royal Society B: Biological Sciences},
       volume = {266},
       number = {1423},
         year = {1999},
        pages = {1001-1012},
     abstract = {Is the information transmitted by an ensemble of neurons determined solely by the number of spikes fired by each cell, or do correlations in the emission of action potentials also play a significant role? We derive a simple formula which enables this question to be answered rigorously for short time-scales. The formula quantifies the corrections to the instantaneous information rate which result from correlations in spike emission between pairs of neurons. The mutual information that the ensemble of neurons conveys about external stimuli can thus be broken down into firing rate and correlation components. This analysis provides fundamental constraints upon the nature of information coding, showing that over short time-scales correlations cannot dominate information representation, that stimulus-independent correlations may lead to synergy (where the neurons together convey more information than they would if they were considered independently), but that only certain combinations of the different sources of correlation result in significant synergy rather than in redundancy or in negligible effects. This analysis leads to a new quantification procedure which is directly applicable to simultaneous multiple neuron recordings.}
}
@book{Papadimitriou94,
       author = {Christos Papadimitriou},
        title = {Computational Complexity},
    publisher = {Addison-Wesley},
      address = {Reading, MA},
         year = 1994,
}
@book{PapadimitriouandSteiglitz82,
       author = {Papadimitriou, Christos H. and Steiglitz, Kenneth},
        title = {Combinatorial Optimization: Algorithms and Complexity},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1982,
}
@article{PapadimitriouandTsitsiklisMOR-87,
       author = {Papadimitriou, Christos H. and Tsitsiklis, John N.},
        title = {The Complexity of {M}arkov Chain Decision Processes},
      journal = {Mathematics of Operations Research},
       volume = 12,
        issue = 3,
        pages = {441-450},
         year = 1987,
}
@article{PapaetalASC-16,
        title = {Fine-tuning deep belief networks using harmony search},
       author = {Joao Paulo Papa and Walter Scheirer and David Cox},
      journal = {Applied Soft Computing},
    publisher = {Elsevier},
       volume = {46},
         year = {2016},
        pages = {875-885},
     abstract = {In this paper, we deal with the problem of Deep Belief Networks (DBNs) pa- rameters fine-tuning by means of a fast meta-heuristic approach named Har- mony Search (HS). Although such deep learning-based technique has been widely used in the last years, more detailed studies about how to set its parameters may not be observed in the literature. We have shown we can obtain more accurate results comparing HS against with several of its vari- ants, a random search and two variants of the well-known Hyperopt library. The experimental results were carried out in two public datasets considering the task of binary image reconstruction, three DBN learning algorithms and three layers.}
}
@book{Papert80,
      author = {Seymour Papert},
       title = {Mindstorms: Children, Computers and Powerful Ideas},
   publisher = {Basic Books},
     address = {New York, NY},
        year = 1980,
}
@article{PapyanetalIEEE-SIGNAL-PROCESSING-18,
       author = {Vardan Papyan and Yaniv Romano and Jeremias Sulam and Michael Elad},
        title = {Theoretical Foundations of Deep Learning via Sparse Representations: {A} Multilayer Sparse Model and Its Connection to Convolutional Neural Networks},
      journal = {{IEEE} Signal Processing Magazine},
       volume = {35},
       number = {4},
         year = {2018},
        pages = {72-89},
     abstract = {Our prime objective in this article is to review a recently introduced [1] model-based explanation of deep learning, which relies on sparse modeling of data. We start by presenting the general story of Sparseland, describing its key achievements. We then turn to describe the convolutional-sparse-coding (CSC) model and present a multilayered (ML) extension of it, both being special cases of the general sparsity-inspired model. We show how ML-CSC leads to a solid and systematic theoretical justification of the architectures used in deep learning, along with a novel ability of analyzing the performance of the result- ing machines. As such, this work offers a unique and first of its kind theoretical view for a field that has been, until recently, considered as purely heuristic.}
}
@article{ParatalaetalPLoS-12,
     title = {Physicochemical characterization, and relaxometry studies of micro-graphite oxide, graphene nanoplatelets, and nanoribbons},
    author = {Paratala, B.S. and Jacobson, B.D. and Kanakia, S. and Francis, L.D. and Sitharaman, B.},
   journal = {{PLoS} One},
    volume = 7,
     issue = 6,
     pages = {e38185},
      year = 2012,
  abstract = {The chemistry of high-performance magnetic resonance imaging contrast agents remains an active area of research. In this work, we demonstrate that the potassium permanganate-based oxidative chemical procedures used to synthesize graphite oxide or graphene nanoparticles leads to the confinement (intercalation) of trace amounts of Mn(2+) ions between the graphene sheets, and that these manganese intercalated graphitic and graphene structures show disparate structural, chemical and magnetic properties, and high relaxivity (up to 2 order) and distinctly different nuclear magnetic resonance dispersion profiles compared to paramagn`etic chelate compounds. The results taken together with other published reports on confinement of paramagnetic metal ions within single-walled carbon nanotubes (a rolled up graphene sheet) show that confinement (encapsulation or intercalation) of paramagnetic metal ions within graphene sheets, and not the size, shape or architecture of the graphitic carbon particles is the key determinant for increasing relaxivity, and thus, identifies nano confinement of paramagnetic ions as novel general strategy to develop paramagnetic metal-ion graphitic-carbon complexes as high relaxivity MRI contrast agents.}
}
@article{ParikhetalNEUROIMAGE-18,
       author = {Parikh, N. and Ruzic, L. and Stewart, G. W. and Spreng, R. N. and De Brigard, F.},
        title = {{W}hat if? {N}eural activity underlying semantic and episodic counterfactual thinking},
      journal = {Neuroimage},
       volume = {178},
         year = {2018},
        pages = {332-345},
     abstract = {Counterfactual thinking (CFT) is the process of mentally simulating alternative versions of known facts. In the past decade, cognitive neuroscientists have begun to uncover the neural underpinnings of CFT, particularly episodic CFT (eCFT), which activates regions in the default network (DN) also activated by episodic memory (eM) recall. However, the engagement of DN regions is different for distinct kinds of eCFT. More plausible counterfactuals and counterfactuals about oneself show stronger activity in DN regions compared to implausible and other- or object-focused counterfactuals. The current study sought to identify a source for this difference in DN activity. Specifically, self-focused counterfactuals may also be more plausible, suggesting that DN core regions are sensitive to the plausibility of a simulation. On the other hand, plausible and self-focused counterfactuals may involve more episodic information than implausible and other-focused counterfactuals, which would imply DN sensitivity to episodic information. In the current study, we compared episodic and semantic counterfactuals generated to be plausible or implausible against episodic and semantic memory reactivation using fMRI. Taking multivariate and univariate approaches, we found that the DN is engaged more during episodic simulations, including eM and all eCFT, than during semantic simulations. Semantic simulations engaged more inferior temporal and lateral occipital regions. The only region that showed strong plausibility effects was the hippocampus, which was significantly engaged for implausible CFT but not for plausible CFT, suggestive of binding more disparate information. Consequences of these findings for the cognitive neuroscience of mental simulation are discussed.}
}
@article{ParisottoetalCoRR-16,
       author = {Emilio Parisotto and Abdel{-}rahman Mohamed and Rishabh Singh and Lihong Li and Dengyong Zhou and Pushmeet Kohli},
        title = {Neuro-Symbolic Program Synthesis},
      journal = {CoRR},
       volume = {arXiv:1611.01855},
         year = {2016},
     abstract = {Recent years have seen the proposal of a number of neural architectures for the problem of Program Induction. Given a set of input-output examples, these architectures are able to learn mappings that generalize to new test inputs. While achieving impressive results, these approaches have a number of important limitations: (a) they are computationally expensive and hard to train, (b) a model has to be trained for each task (program) separately, and (c) it is hard to interpret or verify the correctness of the learnt mapping (as it is defined by a neural network). In this paper, we propose a novel technique, Neuro-Symbolic Program Synthesis, to overcome the above-mentioned problems. Once trained, our approach can automatically construct computer programs in a domain-specific language that are consistent with a set of input-output examples provided at test time. Our method is based on two novel neural modules. The first module, called the cross correlation I/O network, given a set of input-output examples, produces a continuous representation of the set of I/O examples. The second module, the Recursive-Reverse-Recursive Neural Network (R3NN), given the continuous representation of the examples, synthesizes a program by incrementally expanding partial programs. We demonstrate the effectiveness of our approach by applying it to the rich and complex domain of regular expression based string transformations. Experiments show that the R3NN model is not only able to construct programs from new input-output examples, but it is also able to construct new programs for tasks that it had never observed before during training.}
}
@incollection{ParkerandPoveyLEEDS-09,
     title = {Scanning acoustic microscopy for mapping the microstructure of soft materials},
    author = {Parker and M. J. W. Povey},
 booktitle = {Proceedings of "What Where When: Multi-dimensional Advances for Industrial Process Monitoring"},
 publisher = {University of Leeds Press},
  location = {Leeds, UK},
      year = 2009,
     pages = {363-374},
}
@article{ParketalJoN-15,
       author = {Park, Eun Hye and Burghardt, Nesha S. and Dvorak, Dino and Hen, Ren\'{e} and Fenton, Andr\'{e} A.},
      journal = {Journal of Neuroscience},
        title = {Experience-Dependent Regulation of Dentate Gyrus Excitability by Adult-Born Granule Cells},
    publisher = {Society for Neuroscience},
       volume = {35},
       number = {33},
         year = {2015},
        pages = {11656-11666},
     abstract = {Behavioral studies have established a role for adult-born dentate granule cells in discriminating between similar memories. However, it is unclear how these cells mediate memory discrimination. Excitability is enhanced in maturing adult-born neurons, spurring the hypothesis that the activity of these cells 'directly' encodes and stores memories. An alternative hypothesis posits that maturing neurons 'indirectly' contribute to memory encoding by regulating excitation-inhibition balance. We evaluated these alternatives by using dentate-sensitive active place avoidance tasks to assess experience-dependent changes in dentate field potentials in the presence and absence of neurogenesis. Before training, X-ray ablation of adult neurogenesis-reduced dentate responses to perforant-path stimulation and shifted EPSP-spike coupling leftward. These differences were unchanged after place avoidance training with the shock zone in the initial location, which both groups learned to avoid equally well. In contrast, sham-treated mice decreased dentate responses and shifted EPSP-spike coupling leftward after the shock zone was relocated, whereas X-irradiated mice failed to show these changes in dentate function and were impaired on this test of memory discrimination. During place avoidance, excitation{\textendash}inhibition coupled neural synchrony in dentate local field potentials was reduced in X-irradiated mice, especially in the θ band. The difference was most prominent during conflict learning, which is impaired in the X-irradiated mice. These findings indicate that maturing adult-born neurons regulate both functional network plasticity in response to memory discrimination and dentate excitation-inhibition coordination. The most parsimonious interpretation of these results is that adult neurogenesis indirectly regulates hippocampal information processing.},
}
@article{ParkhurstetalCELL-13,
       author = {Parkhurst, C. N. and Yang, G.  and Ninan, I. and Savas, J. N. and Yates, J. R. and Lafaille, J. J. and Hempstead, B. L. and Littman, D. R. and Gan, W. B.},
        title = {Microglia promote learning-dependent synapse formation through brain-derived neurotrophic factor},
      journal = {Cell},
       volume = {155},
       number = {7},
         year = {2013},
        pages = {1596-1609},
     abstract = {Microglia are the resident macrophages of the CNS, and their functions have been extensively studied in various brain pathologies. The physiological roles of microglia in brain plasticity and function, however, remain unclear. To address this question, we generated CX3CR1(CreER) mice expressing tamoxifen-inducible Cre recombinase that allow for specific manipulation of gene function in microglia. Using CX3CR1(CreER) to drive diphtheria toxin receptor expression in microglia, we found that microglia could be specifically depleted from the brain upon diphtheria toxin administration. Mice depleted of microglia showed deficits in multiple learning tasks and a significant reduction in motor-learning-dependent synapse formation. Furthermore, Cre-dependent removal of brain-derived neurotrophic factor (BDNF) from microglia largely recapitulated the effects of microglia depletion. Microglial BDNF increases neuronal tropomyosin-related kinase receptor B phosphorylation, a key mediator of synaptic plasticity. Together, our findings reveal that microglia serve important physiological functions in learning and memory by promoting learning-related synapse formation through BDNF signaling.}
}
@article{Parnas72,
       author = {David Parnas},
        title = {On the Criteria to Be Used in Decomposing Systems into Modules},
      journal = {Communications of the {ACM}},
       volume = 15,
        issue = 12,
        pages = {1059-1062},
         year = 1972,
}
@inproceedings{ParrandRussellNIPS-97,
       author = {Parr, Ronald and Russell, Stuart},
        title = {Reinforcement Learning with Hierarchies of Machines},
    booktitle = {Advances in Neural Information Processing 9},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1997,
}
@article{ParthasarathiandPineauCoRR-18,
        title = {Extending Neural Generative Conversational Model using External Knowledge Sources},
       author = {Prasanna Parthasarathi and Joelle Pineau},
      journal = {CoRR},
       volume = {2018},
         year = {arXiv:1705.00673},
     abstract = {The use of connectionist approaches in conversational agents has been progressing rapidly due to the availability of large corpora. However current generative dialogue models often lack coherence and are content poor. This work proposes an architecture to incorporate unstructured knowledge sources to enhance the next utterance prediction in chit-chat type of generative dialogue models. We focus on Sequence-to-Sequence (Seq2Seq) conversational agents trained with the Reddit News dataset, and consider incorporating external knowledge from Wikipedia summaries as well as from the NELL knowledge base. Our experiments show faster training time and improved perplexity when leveraging external knowledge.},
}
@article{ParvizietalPNAS-06,
       author = {Parvizi, Josef and Van Hoesen, Gary W. and Buckwalter, Joseph and Damasio, Antonio},
        title = {Neural connections of the posteromedial cortex in the macaque},
      journal = {Proceedings of the National Academy of Sciences of the United States of America},
       volume = 103,
        issue = 5,
        pages = {1563-1568},
         year = 2006,
     abstract = {The posterior cingulate and the medial parietal cortices constitute an ensemble known as the posteromedial cortex (PMC   which consists of Brodmann areas 23, 29, 30, 31, and 7m. To understand the neural relationship of the PMC with the rest of the brain, we injected its component areas with four different anterograde and retrograde tracers in the cynomolgus monkey and found that all PMC areas are interconnected with each other and with the anterior cingulate, the mid-dorsolateral prefrontal, the lateral parietal cortices, and area TPO, as well as the thalamus, where projections from some of the PMC areas traverse in an uninterrupted bar-like manner, the dorsum of this structure from the posteriormost nuclei to its rostralmost tip. All PMC regions also receive projections from the claustrum and the basal forebrain and project to the caudate, the basis pontis, and the zona incerta. Moreover  the posterior cingulate areas are interconnected with the parahippocampal regions, whereas the medial parietal cortex projects only sparsely to the presubiculum. Although local interconnections and shared remote connections of all PMC components suggest a functional relationship among them, the distinct connections of each area with different neural structures suggests that distinct functional modules may be operating within the PMC. Our study provides a large-scale map of the PMC connections with the rest of the brain, which may serve as a useful tool for future studies of this cortical region and may contribute to elucidating its intriguing pattern of activity seen in recent functional imaging studies.},
}
@article{PascaNATURE-18,
       author = {Pasca, Sergiu P.},
        title = {The rise of three-dimensional human brain cultures},
      journal = {Nature},
    publisher = {Macmillan Publishers Limited},
         year = {2018},
       volume = {553},
        pages = {437-445},
     abstract = {Pluripotent stem cells show a remarkable ability to self-organize and differentiate in vitro in three-dimensional aggregates, known as organoids or organ spheroids, and to recapitulate aspects of human brain development and function. Regionspecific 3D brain cultures can be derived from any individual and assembled to model complex cell-cell interactions and to generate circuits in human brain assembloids. Here I discuss how this approach can be used to understand unique features of the human brain and to gain insights into neuropsychiatric disorders. In addition, I consider the challenges faced by researchers in further improving and developing methods to probe and manipulate patient-derived 3D brain cultures.}
}
@article{PascaetalNATURE-METHODS-15,
       author = {Pasca, Anca M. and Sloan, Steven A. and Clarke, Laura E. and Tian, Yuan and Makinson, Christopher D. and Huber, Nina and Kim, Chul Hoon and Park, Jin-Young and O'Rourke, Nancy A. and Nguyen, Khoa D. and Smith, Stephen J. and Huguenard, John R. and Geschwind, Daniel H. and Barres, Ben A. and Pasca, Sergiu P.},
        title = {Functional cortical neurons and astrocytes from human pluripotent stem cells in 3D culture},
      journal = {Nature Methods},
       volume = {12},
        issue = {7},
         year = {2015},
        pages = {671-678},
     abstract = {The human cerebral cortex develops through an elaborate succession of cellular events that, when disrupted, can lead to neuropsychiatric disease. The ability to reprogram somatic cells into pluripotent cells that can be differentiated in vitro provides a unique opportunity to study normal and abnormal corticogenesis. Here, we present a simple and reproducible 3D culture approach for generating a laminated cerebral cortex-like structure, named human cortical spheroids (hCSs), from pluripotent stem cells. hCSs contain neurons from both deep and superficial cortical layers and map transcriptionally to in vivo fetal development. These neurons are electrophysiologically mature, display spontaneous activity, are surrounded by nonreactive astrocytes and form functional synapses. Experiments in acute hCS slices demonstrate that cortical neurons participate in network activity and produce complex synaptic events. These 3D cultures should allow a detailed interrogation of human cortical development, function and disease, and may prove a versatile platform for generating other neuronal and glial subtypes in vitro.},
}
@article{PascanuetalCoRR-14,
       author = {Razvan Pascanu and Yann N. Dauphin and Surya Ganguli and Yoshua Bengio},
        title = {On the saddle point problem for non-convex optimization},
      journal = {CoRR},
       volume = {arXiv:1405.4604},
         year = 2014,
     abstract = {A central challenge to many fields of science and engineering involves minimizing non-convex error functions over continuous, high dimensional spaces. Gradient descent or quasi-Newton methods are almost ubiquitously used to perform such minimizations, and it is often thought that a main source of difficulty for the ability of these local methods to find the global minimum is the proliferation of local minima with much higher error than the global minimum. Here we argue, based on results from statistical physics, random matrix theory, and neural network theory, that a deeper and more profound difficulty originates from the proliferation of saddle points, not local minima, especially in high dimensional problems of practical interest. Such saddle points are surrounded by high error plateaus that can dramatically slow down learning, and give the illusory impression of the existence of a local minimum. Motivated by these arguments, we propose a new algorithm, the saddle-free Newton method, that can rapidly escape high dimensional saddle points, unlike gradient descent and quasi-Newton methods. We apply this algorithm to deep neural network training, and provide preliminary numerical evidence for its superior performance.}
}
@article{PascanuetalCoRR-17,
       author = {Razvan Pascanu and Yujia Li and Oriol Vinyals and Nicolas Heess and Lars Buesing and S{\'{e}}bastien Racani{\`{e}}re and David P. Reichert and Theophane Weber and Daan Wierstra and Peter Battaglia},
        title = {Learning model-based planning from scratch},
      journal = {CoRR},
       volume = {arXiv:1707.06170},
         year = {2017},
     abstract = {Conventional wisdom holds that model-based planning is a powerful approach to sequential decision-making. It is often very challenging in practice, however, because while a model can be used to evaluate a plan, it does not prescribe how to construct a plan. Here we introduce the "Imagination-based Planner", the first model-based, sequential decision-making agent that can learn to construct, evaluate, and execute plans. Before any action, it can perform a variable number of imagination steps, which involve proposing an imagined action and evaluating it with its model-based imagination. All imagined actions and outcomes are aggregated, iteratively, into a "plan context" which conditions future real and imagined actions. The agent can even decide how to imagine: testing out alternative imagined actions, chaining sequences of actions together, or building a more complex "imagination tree" by navigating flexibly among the previously imagined states using a learned policy. And our agent can learn to plan economically, jointly optimizing for external rewards and computational costs associated with using its imagination. We show that our architecture can learn to solve a challenging continuous control problem, and also learn elaborate planning strategies in a discrete maze-solving task. Our work opens a new direction toward learning the components of a model-based planning system and how to use them.}
}
@article{PashevichetalCoRR-18,
       author = {Alexander Pashevich and Danijar Hafner and James Davidson and Rahul Sukthankar and Cordelia Schmid},
        title = {Modulated Policy Hierarchies},
      journal = {CoRR},
       volume = {arXiv:1812.00025},
         year = {2018},
     abstract = {Solving tasks with sparse rewards is a main challenge in reinforcement learning. While hierarchical controllers are an intuitive approach to this problem, current methods often require manual reward shaping, alternating training phases, or manually defined sub tasks. We introduce modulated policy hierarchies (MPH), that can learn end-to-end to solve tasks from sparse rewards. To achieve this, we study different modulation signals and exploration for hierarchical controllers. Specifically, we find that communicating via bit-vectors is more efficient than selecting one out of multiple skills, as it enables mixing between them. To facilitate exploration, MPH uses its different time scales for temporally extended intrinsic motivation at each level of the hierarchy. We evaluate MPH on the robotics tasks of pushing and sparse block stacking, where it outperforms recent baselines.}
}
@article{PasleyetalPLoS-12,
    author = {Pasley, B. N. and David, S. V. and Mesgarani, N. and Flinker, A. and Shamma, S. A. and Crone, N. E. and Knight, R. T. and Chang, E. F.},
     title = {Reconstructing speech from human auditory cortex},
   journal = {{PLoS} Biology},
      year = 2012,
    volume = 10,
     issue = 1,
     pages = {e1001251},
}
@article{PassagliaetalPNAS-97,
        title = {Deciphering a neural code for vision},
       author = {Passaglia, C. and Dodge, F. and Herzog, E. and Jackson, S. and Barlow, R.},
      journal = {Proceedings of the National Academy of Science},
       volume = 94,
        issue = 23,
         year = 1997,
        pages = {12649-54},
     abstract = {Deciphering the information that eyes, ears, and other sensory organs transmit to the brain is important for understanding the neural basis of behavior. Recordings from single sensory nerve cells have yielded useful insights, but single neurons generally do not mediate behavior; networks of neurons do. Monitoring the activity of all cells in a neural network of a behaving animal, however, is not yet possible. Taking an alternative approach, we used a realistic cell-based model to compute the ensemble of neural activity generated by one sensory organ, the lateral eye of the horseshoe crab, Limulus polyphemus. We studied how the neural network of this eye encodes natural scenes by presenting to the model movies recorded with a video camera mounted above the eye of an animal that was exploring its underwater habitat. Model predictions were confirmed by simultaneously recording responses from single optic nerve fibers of the same animal. We report here that the eye transmits to the brain robust "neural images" of objects having the size, contrast, and motion of potential mates. The neural code for such objects is not found in ambiguous messages of individual optic nerve fibers but rather in patterns of coherent activity that extend over small ensembles of nerve fibers and are bound together by stimulus motion. Integrative properties of neurons in the first synaptic layer of the brain appear well suited to detecting the patterns of coherent activity. Neural coding by this relatively simple eye helps explain how horseshoe crabs find mates and may lead to a better understanding of how more complex sensory organs process information.}
}
@article{PassinoandAnsaklis89,
       author = {Passino, K. M. and Ansaklis, P. J.},
        title = {A System and Control Theoretic Perspective on Artificial Intelligence Systems},
      journal = {Applied Artificial Intelligence},
       volume = 3,
         year = 1989,
        pages = {1-32},
}
@book{PatelandChellappa2013,
        title = {Sparse Representations and Compressive Sensing for Imaging and Vision},
       author = {Patel, Vishal M., Chellappa, Rama},
    publisher = {Springer},
         year = {2013},
     abstract = {Compressed sensing or compressive sensing is a new concept in signal processing where one measures a small number of non-adaptive linear combinations of the signal.  These measurements are usually much smaller than the number of samples that define the signal.  From these small numbers of measurements, the signal is then reconstructed by non-linear procedure.  Compressed sensing has recently emerged as a powerful tool for efficiently processing data in non-traditional ways.  In this book, we highlight some of the key mathematical insights underlying sparse representation and compressed sensing and illustrate the role of these theories in classical vision, imaging and biometrics problems.}
}
@inproceedings{PatelandChellappaACPR-11,
       author = {V. M. Patel and R. Chellappa},
    booktitle = {The First Asian Conference on Pattern Recognition},
        title = {Sparse Representations, Compressive Sensing and dictionaries for pattern recognition},
         year = {2011},
        pages = {325-329},
     abstract = {In recent years, the theories of Compressive Sensing (CS), Sparse Representation (SR) and Dictionary Learning (DL) have emerged as powerful tools for efficiently processing data in non-traditional ways. An area of promise for these theories is object recognition. In this paper, we review the role of SR, CS and DL for object recognition. Algorithms to perform object recognition using these theories are reviewed. An important aspect in object recognition is feature extraction. Recent works in SR and CS have shown that if sparsity in the recognition problem is properly harnessed then the choice of features is less critical. What becomes critical, however, is the number of features and the sparsity of representation. This issue is discussed in detail.},
}
@article{PateletalCoRR-15,
       author = {Ankit B. Patel, Tan Nguyen, Richard G. Baraniuk},
        title = {A Probabilistic Theory of Deep Learning},
      journal = {CoRR},
       volume = {arXiv:1504.00641},
         year = {2015},
          url = {http://arXiv:1504.00641},
     abstract = {A grand challenge in machine learning is the development of computational algorithms that match or outperform humans in perceptual inference tasks that are complicated by nuisance variation. For instance, visual object recognition involves the unknown object position, orientation, and scale in object recognition while speech recognition involves the unknown voice pronunciation, pitch, and speed. Recently, a new breed of deep learning algorithms have emerged for high-nuisance inference tasks that routinely yield pattern recognition systems with near- or super-human capabilities. But a fundamental question remains: Why do they work? Intuitions abound, but a coherent framework for understanding, analyzing, and synthesizing deep learning architectures has remained elusive. We answer this question by developing a new probabilistic framework for deep learning based on the Deep Rendering Model: a generative probabilistic model that explicitly captures latent nuisance variation. By relaxing the generative model to a discriminative one, we can recover two of the current leading deep learning systems, deep convolutional neural networks and random decision forests, providing insights into their successes and shortcomings, as well as a principled route to their improvement.}
}
@article{PathaketalCoRR-17,
       author = {Deepak Pathak and Pulkit Agrawal and Alexei A. Efros and Trevor Darrell},
        title = {Curiosity-driven Exploration by Self-supervised Prediction},
      journal = {CoRR},
       volume = {arXiv:1705.05363},
         year = {2017},
     abstract = {In many real-world scenarios, rewards extrinsic to the agent are extremely sparse, or absent altogether. In such cases, curiosity can serve as an intrinsic reward signal to enable the agent to explore its environment and learn skills that might be useful later in its life. We formulate curiosity as the error in an agent's ability to predict the consequence of its own actions in a visual feature space learned by a self-supervised inverse dynamics model. Our formulation scales to high-dimensional continuous state spaces like images, bypasses the difficulties of directly predicting pixels, and, critically, ignores the aspects of the environment that cannot affect the agent. The proposed approach is evaluated in two environments: VizDoom and Super Mario Bros. Three broad settings are investigated: 1) sparse extrinsic reward, where curiosity allows for far fewer interactions with the environment to reach the goal; 2) exploration with no extrinsic reward, where curiosity pushes the agent to explore more efficiently; and 3) generalization to unseen scenarios (e.g. new levels of the same game) where the knowledge gained from earlier experience helps the agent explore new places much faster than starting from scratch.}
}
@article{PathaketalICCV-15,
       author = {Deepak Pathak and Philipp Kr{\"{a}}henb{\"{u}}hl and Trevor Darrell},
        title = {Constrained Convolutional Neural Networks for Weakly Supervised Segmentation},
    booktitle = {Proceedings of the {IEEE} International Conference on Computer Vision},
         year = {2015},
     abstract = {We present an approach to learn a dense pixel-wise labeling from image-level tags. Each image-level tag imposes constraints on the output labeling of a Convolutional Neural Network (CNN) classifier. We propose Constrained CNN (CCNN), a method which uses a novel loss function to optimize for any set of linear constraints on the output space (i.e. predicted label distribution) of a CNN. Our loss formulation is easy to optimize and can be incorporated directly into standard stochastic gradient descent optimization. The key idea is to phrase the training objective as a biconvex optimization for linear models, which we then relax to nonlinear deep networks. Extensive experiments demonstrate the generality of our new learning framework. The constrained loss yields state-of-the-art results on weakly supervised semantic image segmentation. We further demonstrate that adding slightly more supervision can greatly improve the performance of the learning algorithm.}
}
@misc{PattaniBRAINLESS-16,
        title = {Preliminary Research -- Living Without Brain Structures},
       author = {Aneri Pattani},
         year = {2016},
 howpublished = {StoryLab Posting},
          url = {http://www.northeastern.edu/storylab/sp16/2016/01/27/preliminary-research-living-without-brain-structures/},
     abstract = {How much of our brain do we actually need? A number of stories have appeared in the news in recent months about people with chunks of their brains missing or damaged. These cases tell a story about the mind that goes deeper than their initial shock factor. It isn't just that we don't understand how the brain works, but that we may be thinking about it in the entirely wrong way.}
}
@book{Paul81,
       author = {Paul, Richard P.},
        title = {Robot Manipulators: Mathematics, Programming, and Control},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1981,
}
@book{Paulos95,
       author = {John Allen Paulos},
        title = {A Mathematician Reads the Newspaper},
    publisher = {Basic Books},
      address = {New York, NY},
         year = 1995,
}
@inproceedings{PaulusetalNIPS-14,
       author = {Romain Paulus and Richard Socher and Christopher D. Manning},
        title = {Global Belief Recursive Neural Networks},
    booktitle = {Advances in Neural Information Processing Systems 27},
         year = 2014,
     abstract = {Recursive Neural Networks have recently obtained state of the art performance on several natural language processing tasks. However, because of their feedforward architecture they cannot correctly predict phrase or word labels that are determined by context. This is a problem in tasks such as aspect-specific sentiment classification which tries to, for instance, predict that the word Android is positive in the sentence Android beats iOS. We introduce global belief recursive neural networks (GB-RNNs) which are based on the idea of extending purely feedforward neural networks to include one feedbackward step during inference. This allows phrase level predictions and representations to give feedback to words. We show the effectiveness of this model on the task of contextual sentiment analysis. We also show that dropout can improve RNN training and that a combination of unsupervised and supervised word vector representations performs better than either alone. The feedbackward step improves F1 performance by 3\% over the standard RNN on this task, obtains state-of-the-art performance on the SemEval 2013 challenge and can accurately predict the sentiment of specific entities.},
}
@book{Paz71,
       author = {Paz, Azaria},
        title = {Introduction to Probabilistic Automata},
    publisher = {Academic Press},
      address = {New York},
         year = 1971,
}
@article{PazetalNATURE-12,
        title = {Closed-loop optogenetic control of thalamus as a tool for interrupting seizures after cortical injury},
       author = {Paz, J.T. and Davidson, T.J. and Frechette, E.S. and Delord, B. and Parada, I. and Peng, K. and Deisseroth, K. and Huguenard, J.R.},
      journal = {Nature Neuroscience},
         year = 2012,
     abstract = {Cerebrocortical injuries such as stroke are a major source of disability. Maladaptive consequences can result from post-injury local reorganization of cortical circuits. For example, epilepsy is a common sequela of cortical stroke, but the mechanisms responsible for seizures following cortical injuries remain unknown. In addition to local reorganization, long-range, extra-cortical connections might be critical for seizure maintenance. In rats, we found that the thalamus, a structure that is remote from, but connected to, the injured cortex, was required to maintain cortical seizures. Thalamocortical neurons connected to the injured epileptic cortex underwent changes in HCN channel expression and became hyperexcitable. Targeting these neurons with a closed-loop optogenetic strategy revealed that reducing their activity in real-time was sufficient to immediately interrupt electrographic and behavioral seizures. This approach is of therapeutic interest for intractable epilepsy, as it spares cortical function between seizures, in contrast with existing treatments, such as surgical lesioning or drugs.}
}
@book{Pearl85,
       author = {Pearl, Judea},
        title = {Heuristics},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1985,
}
@book{Pearl88,
       author = {Pearl, Judea},
        title = {Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1988,
}
@article{PearlAIJ-86a,
       author = {Pearl, Judea},
        title = {On Evidential Reasoning in a Hierarchy of Hypotheses},
      journal = AIJ,
       volume = 28,
        pages = {9-16},
         year = 1986,
}
@article{PearlAIJ-86b,
       author = {Pearl, Judea},
        title = {Fusion, Propagation, and Structuring in Belief Networks},
      journal = AIJ,
       volume = 29,
        pages = {241-288},
         year = 1986,
}
@article{PearlAIJ-87a,
       author = {Pearl, Judea},
        title = {Distributed Revision of Composite Beliefs},
      journal = AIJ,
       volume = 33,
         year = 1987,
        pages = {173-216},
}
@article{PearlAIJ-87b,
       author = {Judea Pearl},
        title = {Evidential Reasoning Using Stochastic Simulation of Causal Models},
      journal = AIJ,
       volume = 32,
         year = 1987,
        pages = {245-257},
}
@article{PearlCACM-82,
       author = {Pearl, Judea},
        title = {A Solution for the Branching Factor of the Alpha-Beta Pruning Algorithm and its Optimality},
      journal = {Communications of the {ACM}},
       volume = 25,
        issue = 8,
        pages = {559-564},
         year = 1982,
}
@inproceedings{PearlIJCAI-71,
       author = {Pearl, Judea},
        title = {The Reverend {B}ayes on inference engines: a distributed
                   hierarchical approach},
    booktitle = {Proceedings {IJCAI} 2},
     location = {London},
 organization = {IJCAII},
         year = 1971,
}
@article{PearlKYBERNETIKA-89,
       author = {Pearl, Judea and Geiger, Dan and Verma, Thomas},
        title = {Conditional independence and its representations},
      journal = {Kybernetika},
       volume = 25,
         year = 1989,
        pages = {33-44},
}
@incollection{PearlUAI-88,
       author = {Pearl, Judea},
        title = {A Constraint Propagation Approach to Probabilistic Reasoning},
     crossref = UAI-88,
}
@article{PearlmutterNC-94,
       author = {Barak A. Pearlmutter},
        title = {Fast Exact Multiplication by the Hessian},
      journal = {Neural Computation},
         year = {1994},
       volume = {6},
        pages = {147-160},
     abstract = {Just storing the Hessian (the matrix of second derivatives) of a large neural network is difficult. Since a common use of a large matrix like is to compute its product with various vectors, we derive a technique that directly calculates Hv, where v is an arbitrary vector. To calculate Hv, we first define a differential operator, and then apply the operator to the equations used to compute Hv. The result is an exact and numerically stable procedure for computing Hv, which takes about as much computation, and is about as local, as a gradient evaluation. We then apply the technique to a one pass gradient calculation algorithm (backpropagation), a relaxation gradient calculation algorithm (recurrent backpropagation), and two stochastic gradient calculation algorithms(Boltzmann Machines and weight perturbation). Finally, we show that this technique can be used at the heart of many iterative techniques for computing various properties of, obviating any need to calculate the full Hessian.}
}
@article{PecevskietalPLoS-11,
       author = {Pecevski, Dejan and Buesing, Lars and Maass, Wolfgang},
      journal = {PLoS Computational Biology},
    publisher = {Public Library of Science},
        title = {Probabilistic Inference in General Graphical Models through Sampling in Stochastic Networks of Spiking Neurons},
       volume = {7},
         year = {2011},
        pages = {1-25},
     abstract = {Author Summary Experimental data from neuroscience have provided substantial knowledge about the intricate structure of cortical microcircuits, but their functional role, i.e. the computational calculus that they employ in order to interpret ambiguous stimuli, produce predictions, and derive movement plans has remained largely unknown. Earlier assumptions that these circuits implement a logic-like calculus have run into problems, because logical inference has turned out to be inadequate to solve inference problems in the real world which often exhibits substantial degrees of uncertainty. In this article we propose an alternative theoretical framework for examining the functional role of precisely structured motifs of cortical microcircuits and dendritic computations in complex neurons, based on probabilistic inference through sampling. We show that these structural details endow cortical columns and areas with the capability to represent complex knowledge about their environment in the form of higher order dependencies among salient variables. We show that it also enables them to use this knowledge for probabilistic inference that is capable to deal with uncertainty in stored knowledge and current observations. We demonstrate in computer simulations that the precisely structured neuronal microcircuits enable networks of spiking neurons to solve through their inherent stochastic dynamics a variety of complex probabilistic inference tasks.},
}
@inproceedings{PecketalICANN-05,
        title = {A Biologically Motivated Classifier that Preserves Implicit Relationship Information in Layered Networks},
       author = {Charles C. Peck and James Kozloski and A. Ravishankar Rao and Guillermo A. Cecchi},
    booktitle = {Proceedings of the 17th International Conference on Adaptive and Natural Computing Algorithms},
     location = {Coimbra, Portugal},
         year = 2005,
}
@inproceedings{PedersolietalCVPR-11,
      title = {A coarse-to-fine approach for fast deformable object detection},
     author = {Marco Pedersoli and Andrea Vedaldi and Jordi Gonz{\`a}lez},
  booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
  publisher = {IEEE},
       year = 2011,
      pages = {1353-1360},
}
@article{PednaultAIJ-81,
       author = {Pednault, E. D. P. and Zucker, S. W. and Muresan, L. V.},
        title = {On the independence assumption underlying subjective {B}ayesian updating},
      journal = AIJ,
       volume = 16,
         year = 1981,
        pages = {213-222},
}
@article{PednaultCIJ-88,
       author = {Pednault, Edwin P. D.},
        title = {Synthesizing Plans that Contain Actions with Context-DependentEffects},
      journal = CIJ,
       volume = 4,
        issue = 4,
         year = 1988,
        pages = {356-372},
}
@techreport{PednaultTR-85,
       author = {Pednault, Edwin P. D.},
        title = {Preliminary Report on a Theory of Plan Synthesis},
        issue = 358,
 institution  = {Artificial Intelligence Center SRI International},
         year = 1985,
}
@article{PeikonetalBIORXIV-17,
       author = {Peikon, Ian D and Kebschull, Justus M and Vagin, Vasily V and Ravens, Diana I and Brouzes, Eric and Corr{\^e}a, Ivan R and Bressan, Dario and Zador, Anthony},
        title = {Using high-throughput barcode sequencing to efficiently map connectomes},
      journal = {bioRxiv},
         year = {2017},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {The function of a neural circuit is determined by the details of its synaptic connections. At present, the only available method for determining a neural wiring diagram with single synapse precision--a "connectome"--is based on imaging methods that are slow, labor-intensive and expensive. Here we present SYNseq, a method for converting the connectome into a form that can exploit the speed and low cost of modern high-throughput DNA sequencing. In SYNseq, each neuron is labeled with a unique random nucleotide sequence--an RNA "barcode"--which is targeted to the synapse using engineered proteins. Barcodes in pre- and postsynaptic neurons are then associated through protein-protein crosslinking across the synapse, extracted from the tissue, and then joined into a form suitable for sequencing. Although at present the inefficiency in our hands of barcode joining precludes the widespread application of this approach, we expect that with further development SYNseq will enable tracing of complex circuits at high speed and low cost.},
}
@book{Peirce1958collected,
        title = {Collected Papers},
       author = {Peirce, Charles Sanders},
       series = {Collected Papers},
         year = {1958},
    publisher = {Belknap Press of Harvard University Press}
}
@article{PelkowskietalBBR-11,
       author = {Pelkowski, S. D. and Kapoor, M. and Richendrfer, H. A. and Wang, X. and Colwill, R. M. and Creton, R.},
        title = {{{A} novel high-throughput imaging system for automated analyses of avoidance behavior in zebrafish larvae}},
      journal = {Behavioural Brain Research},
       volume = {223},
       number = {1},
         year = {2011},
        pages = {135-144},
     abstract = {Early brain development can be influenced by numerous genetic and environmental factors, with longlasting effects on brain function and behavior. The identification of these factors is facilitated by recent innovations in high-throughput screening. However, large-scale screening in whole organisms remains challenging, in particular when studying changes in brain function or behavior in vertebrate model systems. In this study, we present a novel imaging system for high-throughput analyses of behavior in zebrafish larvae. The three-camera system can image 12 multiwell plates simultaneously and is unique in its ability to provide local visual stimuli in the wells of a multiwell plate. The acquired images are converted into a series of coordinates, which characterize the location and orientation of the larvae. The developed imaging techniques were tested by measuring avoidance behaviors in seven-day-old zebrafish larvae. The system effectively quantified larval avoidance and revealed an increased edge preference in response to a blue or red 'bouncing ball' stimulus. Larvae also avoid a bouncing ball stimulus when it is counter-balanced with a stationary ball, but do not avoid blinking balls counter-balanced with a stationary ball. These results indicate that the seven-day-old larvae respond specifically to movement, rather than color, size, or local changes in light intensity. The imaging system and assays for measuring avoidance behavior may be used to screen for genetic and environmental factors that cause developmental brain disorders and for novel drugs that could prevent or treat these disorders.}
}
@inproceedings{PembertonandKorfSSS-94,
   author =    {Pemberton, Joseph and Korf, Richard},
   title =     {An Incremental Search Approach to Real-Time Decision Making},
   booktitle = {Proceedings of the {AAAI} Spring Symposium on Decision Theoretic Planning},
   location =  {Stanford, California},
   year =      1994,
}
@book{PemmarajuandSkiena2003,
       author = {Pemmaraju, Sriram and Skiena, Steven},
        title = {Computational Discrete Mathematics: Combinatorics and Graph Theory with Mathematica},
         year = 2003,
    publisher = {Cambridge University Press},
      address = {New York, NY, USA},
}
@inproceedings{PenberthyandWeldKR-92,
       author = {Penberthy, J. S. and Weld, Daniel S.},
        title = {{UCPOP}: A Sound, Complete, Partial Order Planner for {ADL}},
    booktitle = {Proceedings of the 1992 International Conference on Principles of Knowledge Representation and Reasoning},
        pages = {103-114},
         year = 1992,
}
@article{PenfieldandBoldreyBRAIN-37,
       author = {Penfield, Wilder and Boldrey, Edwin},
        title = {Somatic motor and sensory representation in the cerebral cortex of man as studied by electrical stimulation},
      journal = {Brain},
       volume = {60},
       number = {4},
         year = {1937},
        pages = {389-443},
}
@article{PengandWilliams,
       author = {Peng, Jing and Williams, Ronald J.},
        title = {Efficient Learning and Planning Within the Dyna Framework},
      journal = {Adaptive Behavior},
    publisher = {{MIT} Press},
       volume = 1,
        issue = 4,
        pages = {437-454},
         year = 1993,
}
@article{PengetalCoRR-16,
       author = {Zhanglin Peng and Ruimao Zhang and Xiaodan Liang and Xiaobai Liu and Liang Lin},
        title = {Geometric Scene Parsing with Hierarchical {LSTM}},
      journal = {CoRR},
       volume = {arXiv:1604.01931},
         year = {2016},
     abstract = {This paper addresses the problem of geometric scene parsing, i.e. simultaneously labeling geometric surfaces (e.g. sky, ground and vertical plane) and determining the interaction relations (e.g. layering, supporting, siding and affinity) between main regions. This problem is more challenging than the traditional semantic scene labeling, as recovering geometric structures necessarily requires the rich and diverse contextual information. To achieve these goals, we propose a novel recurrent neural network model, named Hierarchical Long Short-Term Memory (H-LSTM). It contains two coupled sub-networks: the Pixel LSTM (P-LSTM) and the Multi-scale Super-pixel LSTM (MS-LSTM) for handling the surface labeling and relation prediction, respectively. The two sub-networks provide complementary information to each other to exploit hierarchical scene contexts, and they are jointly optimized for boosting the performance. Our extensive experiments show that our model is capable of parsing scene geometric structures and outperforming several state-of-the-art methods by large margins. In addition, we show promising 3D reconstruction results from the still images based on the geometric parsing.}
}
@article{PengetalCoRR-17,
       author = {Baolin Peng and Xiujun Li and Lihong Li and Jianfeng Gao and Asli {\c{C}}elikyilmaz and Sungjin Lee and Kam-Fai Wong},
        title = {Composite Task-Completion Dialogue System via Hierarchical Deep Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1704.03084},
         year = {2017},
     abstract = {Building a dialogue agent to fulfill complex tasks, such as travel planning, is challenging because the agent has to learn to collectively complete multiple subtasks. For example, the agent needs to reserve a hotel and book a flight so that there leaves enough time for commute between arrival and hotel check-in. This paper addresses this challenge by formulating the task in the mathematical framework of options over Markov Decision Processes (MDPs), and proposing a hierarchical deep reinforcement learning approach to learning a dialogue manager that operates at different temporal scales. The dialogue manager consists of: (1) a top-level dialogue policy that selects among subtasks or options, (2) a low-level dialogue policy that selects primitive actions to complete the subtask given by the top-level policy, and (3) a global state tracker that helps ensure all cross-subtask constraints be satisfied. Experiments on a travel planning task with simulated and real users show that our approach leads to significant improvements over three baselines, two based on handcrafted rules and the other based on flat deep reinforcement learning.}
}
@article{PengetalNATURE-METHODS-11,
       author = {Peng, Hanchuan and Chung, Phuong and Long, Fuhui and Qu, Lei and Jenett, Arnim and Seeds, Andrew M. and Myers, Eugene W. and Simpson, Julie H.},
        title = {BrainAligner: 3D Registration Atlases of Drosophila Brains},
      journal = {Nat Methods},
         year = {2011},
       volume = {8},
        issue = {6},
        pages = {493-500},
     abstract = {Analyzing Drosophila neural expression patterns in thousands of 3D image stacks of individual brains requires registering them into a canonical framework based on a fiducial reference of neuropil morphology. Given a target brain labeled with predefined landmarks, the BrainAligner program automatically finds the corresponding landmarks in a subject brain and maps it to the coordinate system of the target brain via a deformable warp. Using a neuropil marker (the antibody nc82) as a reference of the brain morphology and a target brain that is itself a statistical average of 295 brains, we achieved a registration accuracy of 2 microns on average, permitting assessment of stereotypy, potential connectivity, and functional mapping of the adult fruitfly brain. We used BrainAligner to generate an image pattern atlas of 2,954 registered brains containing 470 different expression patterns that cover all the major compartments of the fly brain.},
}
@book{Pennington71,
       editor = {Pennington, R. H.},
        title = {Introductory Computer Methods and Numerical Analysis},
    publisher = {Macmillan},
         year = 1971,
}
@inproceedings{PenningtonetalEMNLP-14,
       author = {Jeffrey Pennington and Richard Socher and Christopher D. Manning},
    booktitle = {Empirical Methods in Natural Language Processing (EMNLP)},
        title = {GloVe: Global Vectors for Word Representation},
         year = {2014},
        pages = {1532-1543},
     abstract = {Recent methods for learning vector space representations of words have succeeded in capturing fine-grained semantic and syntactic regularities using vector arithmetic, but the origin of these regularities has remained opaque. We analyze and make explicit the model properties needed for such regularities to emerge in word vectors. The result is a new global logbilinear regression model that combines the advantages of the two major model families in the literature: global matrix factorization and local context window methods. Our model efficiently leverages statistical information by training only on the nonzero elements in a word-word cooccurrence matrix, rather than on the entire sparse matrix or on individual context windows in a large corpus. The model produces a vector space with meaningful substructure, as evidenced by its performance of 75\% on a recent word analogy task. It also outperforms related models on similarity tasks and named entity recognition},
}
@book{Pentland2008,
       author = {Pentland, Alex (Sandy)},
        title = {Honest Signals: How They Shape Our World},
         year = {2008},
    publisher = {The MIT Press},
}
@book{Pentland2014,
        title = {Social Physics: How Good Ideas Spread the Lessons from a New Science},
       author = {Pentland, Alex (Sandy)},
         year = {2014},
    publisher = {Penguin Press}
}
@inproceedings{PentlandetalCVPR-94,
       author = {Pentland, A. and  Moghaddam, B. and Starner, T.},
        title = {View-based and Modular Eigenspaces for Face Recognition},
    booktitle = {Proceedings of the 1994 {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 1994,
}
@article{PeotandShachterAIJ-91,
       author = {Peot, Mark and Shachter, Ross},
        title = {Fusion and Propagation with Multiple Observations in Belief Networks},
      journal = AIJ,
       volume = 48,
        issue = 3,
        pages = {299-318},
         year = 1991,
}
@inproceedings{PeotandSmithAIPS-92,
       author = {Peot, Mark A. and Smith, David E.},
        title = {Conditional Nonlinear Planning},
    booktitle = {Proceedings of the First International Conference on Artificial Intelligence Planning Systems},
        pages = {189-197},
         year = 1992,
}
@article{PepinoPHYSIOLOGY_BEHAVIOR-15,
       author = {Pepino, M. Yanina},
        title = {Metabolic effects of non-nutritive sweeteners},
      journal = {Physiology \& behavior},
         year = {2015},
       volume = {152},
        pages = {450-455},
     abstract = {Until recently, the general belief was that non-nutritive sweeteners (NNSs) were healthy sugar substitutes because they provide sweet taste without calories or glycemic effects. However, data from several epidemiological studies have found that consumption of NNSs, mainly in diet sodas, is associated with increased risk to develop obesity, metabolic syndrome, and type 2 diabetes. The main purpose of this article is to review recent scientific evidence supporting potential mechanisms that explain how ``metabolically inactive'' NNSs, which have few, if any, calories, might promote metabolic dysregulation. Three potential mechanisms, which are not mutually exclusive, are presented: 1) NNSs interfere with learned responses that contribute to control glucose and energy homeostasis, 2) NNSs interfere with gut microbiota and induce glucose intolerance, and 3) NNSs interact with sweet-taste receptors expressed throughout the digestive system that play a role in glucose absorption and trigger insulin secretion. In addition, recent findings from our laboratory showing an association between individual taste sensitivity to detect sucralose and sucralose's acute effects on metabolic response to an oral glucose load are reported. Taken as a whole, data support the notion that NNSs have metabolic effects. More research is needed to elucidate the mechanisms by which NNSs may drive metabolic dysregulation and better understand potential effects of these commonly used food additives.},
}
@inproceedings{PereiraetalACL-95,
       author = {Fernando Pereira and Yoram Singer and Naftali Tishby},
        title = {Beyond Word {N}-Grams},
    booktitle = {Proceedings of the Third Workshop on Very Large Corpora},
    publisher = {Association for Computational Linguistics},
      address = {Somerset, New Jersey},
       editor = {David Yarovsky and Kenneth Church},
        pages = {95-106},
         year = 1995,
}
@misc{PerezANALOGY-18,
       author = {Carlos Perez},
        title = {Intuitive Relational Reasoning for Deep Learning},
         year = {2018},
 howpublished = {{\tt{https://medium.com/intuitionmachine/intuitive-relational-reasoning-for-deep-learning-3ae164f9f5cd}}},
     abstract = {Recently, a class of models has arisen at the intersection of deep learning and structured approaches, which focuses on approaches for reasoning about explicitly structured data, in particular graphs ... What these approaches all have in common is a capacity for performing computation over discrete entities and the relations between them.}
}
@article{PerezetalCoRR-17,
       author = {Ethan Perez and Harm de Vries and Florian Strub and Vincent Dumoulin and Aaron C. Courville},
        title = {Learning Visual Reasoning Without Strong Priors},
      journal = {CoRR},
       volume = {arXiv:1707.03017},
         year = {2017},
     abstract = {Achieving artificial visual reasoning - the ability to answer image-related questions which require a multi-step, high-level process - is an important step towards artificial general intelligence. This multi-modal task requires learning a question-dependent, structured reasoning process over images from language. Standard deep learning approaches tend to exploit biases in the data rather than learn this underlying structure, while leading methods learn to visually reason successfully but are hand-crafted for reasoning. We show that a general-purpose, Conditional Batch Normalization approach achieves state-of-the-art results on the CLEVR Visual Reasoning benchmark with a 2.4\% error rate. We outperform the next best end-to-end method (4.5\%) and even methods that use extra supervision (3.1\%). We probe our model to shed light on how it reasons, showing it has learned a question-dependent, multi-step process. Previous work has operated under the assumption that visual reasoning calls for a specialized architecture, but we show that a general architecture with proper conditioning can learn to visually reason effectively.}
}
@article{PerinetalPNAS-11,
       author = {Perin, Rodrigo and Berger, Thomas K. and Markram, Henry},
        title = {A synaptic organizing principle for cortical neuronal groups},
      journal = {Proceedings of the National Academy of Sciences},
       volume = 108,
        issue = 13,
         year = 2011,
        pages = {5419-5424},
     abstract = {Neuronal circuitry is often considered a clean slate that can be dynamically and arbitrarily molded by experience. However, when we investigated synaptic connectivity in groups of pyramidal neurons in the neocortex, we found that both connectivity and synaptic weights were surprisingly predictable. Synaptic weights follow very closely the number of connections in a group of neurons, saturating after only 20\% of possible connections are formed between neurons in a group. When we examined the network topology of connectivity between neurons, we found that the neurons cluster into small world networks that are not scale-free, with less than 2 degrees of separation. We found a simple clustering rule where connectivity is directly proportional to the number of common neighbors, which accounts for these small world networks and accurately predicts the connection probability between any two neurons. This pyramidal neuron network clusters into multiple groups of a few dozen neurons each. The neurons composing each group are surprisingly distributed, typically more than 100 \hmmu{}m apart, allowing for multiple groups to be interlaced in the same space. In summary, we discovered a synaptic organizing principle that groups neurons in a manner that is common across animals and hence, independent of individual experiences. We speculate that these elementary neuronal groups are prescribed Lego-like building blocks of perception and that acquired memory relies more on combining these elementary assemblies into higher-order constructs.},
}
@article{Perlis82,
      author = {Alan J. Perlis},
       title = {Epigrams on Programming},
     journal = {ACM SIGPLAN Notices},
      volume = 17,
       issue = 19,
       pages = {7-13},
        year = 1982,
}
@article{PeronaandMalikPAMI-90,
       author = {Pietro Perona and Jitendra Malik},
        title = {Scale-Space and Edge Detection Using Anisotropic Diffusion},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
       volume = 12,
        issue = 7,
        pages = {629-639},
         year = 1990,
}
@article{PeronetalCOIN-15,
        title = {Comprehensive imaging of cortical networks},
       author = {Simon Peron and Tsai-Wen Chen and Karel Svoboda},
      journal = {Current Opinion in Neurobiology},
       volume = 32,
         year = 2015,
        pages = {115-123},
     abstract = {Neural computations are implemented by activity in spatially distributed neural circuits. Cellular imaging fills a unique niche in linking activity of specific types of neurons to behavior, over spatial scales spanning single neurons to entire brain regions, and temporal scales from milliseconds to months. Imaging may soon make it possible to track activity of all neurons in a brain region, such as a cortical column. We review recent methodological advances that facilitate optical imaging of neuronal populations in vivo, with an emphasis on calcium imaging using protein indicators in mice. We point out areas that are particularly ripe for future developments.}
}
@article{PeronetalNEURON-15,
       author = {Simon P. Peron and Jeremy Freeman and Vijay Iyer and Caiying Guo and Karel Svoboda},
        title = {A Cellular Resolution Map of Barrel Cortex Activity during Tactile Behavior},
      journal = {Neuron},
       volume = {86},
       number = {3},
        pages = {783-799},
         year = {2015},
     abstract = {Summary Comprehensive measurement of neural activity remains challenging due to the large numbers of neurons in each brain area. We used volumetric two-photon imaging in mice expressing \{GCaMP6s\} and nuclear red fluorescent proteins to sample activity in 75\% of superficial barrel cortex neurons across the relevant cortical columns, approximately 12,000 neurons per animal, during performance of a single whisker object localization task. Task-related activity peaked during object palpation. An encoding modelÂ related activity to behavioral variables. In the column corresponding to the spared whisker, 300 layer (L) 2/3 pyramidal neurons (17\%) each encoded touch and whisker movements. Touch representation declined by half in surrounding columns; whisker movement representation was unchanged. Following the emergence of stereotyped task-related movement, sensory representations showed no measurable plasticity. Touch direction was topographically organized, with distinct organization for passive and active touch. Our work reveals sparse and spatially intermingled representations of multiple tactile features.},
}
@article{PerraultetalACS-14,
       author = {Perrault, Steven D. and Shih, William M.},
        title = {Virus-Inspired Membrane Encapsulation of DNA Nanostructures To Achieve In Vivo Stability},
      journal = {ACS Nano},
       volume = 000,
        issue = {000-000},
         year = 2014,
     abstract = {DNA nanotechnology enables engineering of molecular-scale devices with exquisite control over geometry and site-specific functionalization. This capability promises compelling advantages in advancing nanomedicine; nevertheless, instability in biological environments and innate immune activation remain as obstacles for in vivo application. Natural particle systems (i.e., viruses) have evolved mechanisms to maintain structural integrity and avoid immune recognition during infection, including encapsulation of their genome and protein capsid shell in a lipid envelope. Here we introduce virus-inspired enveloped DNA nanostructures as a design strategy for biomedical applications. Achieving a high yield of tightly wrapped unilamellar nanostructures, mimicking the morphology of enveloped virus particles, required precise control over the density of attached lipid conjugates and was achieved at 1 per \~180 nm2. Envelopment of DNA nanostructures in PEGylated lipid bilayers conferred protection against nuclease digestion. Immune activation was decreased 2 orders of magnitude below controls, and pharmacokinetic bioavailability improved by a factor of 17. By establishing a design strategy suitable for biomedical applications, we have provided a platform for the engineering of sophisticated, translation-ready DNA nanodevices.}
}
@article{PerronetalPBR-12,
       author = {Perron, A. and Akemann, W. and Mutoh, H. and Knopfel, T.},
        title = {Genetically encoded probes for optical imaging of brain electrical activity},
      journal = {Progress Brain Research},
       volume = 196,
         year = 2012,
        pages = {63-77}
}
@article{PerryandFallahFiCN,
       author = {Perry, C. J. and Fallah, M.},
        title = {{{F}eature integration and object representations along the dorsal stream visual hierarchy}},
      journal = {Frontiers in Computational Neuroscience},
         year = {2014},
       volume = {8},
        pages = {84},
     abstract = {The visual system is split into two processing streams: a ventral stream that receives color and form information and a dorsal stream that receives motion information. Each stream processes that information hierarchically, with each stage building upon the previous. In the ventral stream this leads to the formation of object representations that ultimately allow for object recognition regardless of changes in the surrounding environment. In the dorsal stream, this hierarchical processing has classically been thought to lead to the computation of complex motion in three dimensions. However, there is evidence to suggest that there is integration of both dorsal and ventral stream information into motion computation processes, giving rise to intermediate object representations, which facilitate object selection and decision making mechanisms in the dorsal stream. First we review the hierarchical processing of motion along the dorsal stream and the building up of object representations along the ventral stream. Then we discuss recent work on the integration of ventral and dorsal stream features that lead to intermediate object representations in the dorsal stream. Finally we propose a framework describing how and at what stage different features are integrated into dorsal visual stream object representations. Determining the integration of features along the dorsal stream is necessary to understand not only how the dorsal stream builds up an object representation but also which computations are performed on object representations instead of local features.}
}
@article{PetersenNEURON-07,
       author = {Petersen, Carl C. H.},
        title = {The Functional Organization of the Barrel Cortex},
      journal = {Neuron},
         year = 2007,
    publisher = {Elsevier},
       volume = 56,
        issue = 2,
        pages = {339-355},
     abstract = {The tactile somatosensory pathway from whisker to cortex in rodents provides a well-defined system for exploring the link between molecular mechanisms, synaptic circuits, and behavior. The primary somatosensory cortex has an exquisite somatotopic map where each individual whisker is represented in a discrete anatomical unit, the ?barrel,? allowing precise delineation of functional organization, development, and plasticity. Sensory information is actively acquired in awake behaving rodents and processed differently within the barrel map depending upon whisker-related behavior. The prominence of state-dependent cortical sensory processing is likely to be crucial in our understanding of active sensory perception, experience-dependent plasticity and learning.},
}
@article{PetersenNRN-19,
       author = {Petersen, C. C. H.},
        title = {Sensorimotor processing in the rodent barrel cortex},
      journal = {Nature Reviews Neuroscience},
         year = {2019},
       volume = {20},
       number = {9},
        pages = {533-546},
     abstract = {Tactile sensory information from facial whiskers provides nocturnal tunnel-dwelling rodents, including mice and rats, with important spatial and textural information about their immediate surroundings. Whiskers are moved back and forth to scan the environment (whisking), and touch signals from each whisker evoke sparse patterns of neuronal activity in whisker-related primary somatosensory cortex (wS1; barrel cortex). Whisking is accompanied by desynchronized brain states and cell-type-specific changes in spontaneous and evoked neuronal activity. Tactile information, including object texture and location, appears to be computed in wS1 through integration of motor and sensory signals. wS1 also directly controls whisker movements and contributes to learned, whisker-dependent, goal-directed behaviours. The cell-type-specific neuronal circuitry in wS1 that contributes to whisker sensory perception is beginning to be defined.}
}
@article{PetersetalCoRR-18,
       author = {Matthew E. Peters and Mark Neumann and Mohit Iyyer and Matt Gardner and Christopher Clark and Kenton Lee and Luke Zettlemoyer},
        title = {Deep contextualized word representations},
      journal = {CoRR},
       volume = {arXiv:1802.05365},
         year = {2018},
     abstract = {We introduce a new type of deep contextualized word representation that models both (1) complex characteristics of word use (e.g., syntax and semantics), and (2) how these uses vary across linguistic contexts (i.e., to model polysemy). Our word vectors are learned functions of the internal states of a deep bidirectional language model (biLM), which is pre-trained on a large text corpus. We show that these representations can be easily added to existing models and significantly improve the state of the art across six challenging NLP problems, including question answering, textual entailment and sentiment analysis. We also present an analysis showing that exposing the deep internals of the pre-trained network is crucial, allowing downstream models to mix different types of semi-supervision signals.},
}
@article{PetersetalPLoS_ONE-14,
       author = {Peters, Shanan E. and Zhang, Ce and Livny, Miron and R\`{e}, Christopher},
      journal = {{PLoS} ONE},
    publisher = {Public Library of Science},
        title = {A Machine Reading System for Assembling Synthetic Paleontological Databases},
         year = {2014},
       volume = {9},
        issue = {12},
          url = {http://dx.doi.org/10.1371\%2Fjournal.pone.0113523},
     abstract = {Many aspects of macroevolutionary theory and our understanding of biotic responses to global environmental change derive from literature-based compilations of paleontological data. Existing manually assembled databases are, however, incomplete and difficult to assess and enhance with new data types. Here, we develop and validate the quality of a machine reading system, PaleoDeepDive, that automatically locates and extracts data from heterogeneous text, tables, and figures in publications. PaleoDeepDive performs comparably to humans in several complex data extraction and inference tasks and generates congruent synthetic results that describe the geological history of taxonomic diversity and genus-level rates of origination and extinction. Unlike traditional databases, PaleoDeepDive produces a probabilistic database that systematically improves as information is added. We show that the system can readily accommodate sophisticated data types, such as morphological data in biological illustrations and associated textual descriptions. Our machine reading approach to scientific data integration and synthesis brings within reach many questions that are currently underdetermined and does so in ways that may stimulate entirely new modes of inquiry.},
}
@article{PetersonetalCoRR-16,
       author = {Joshua C. Peterson and Joshua T. Abbott and Thomas L. Griffiths},
        title = {Adapting Deep Network Features to Capture Psychological Representations},
      journal = {CoRR},
       volume = {arXiv:1608.02164},
         year = {2016},
     abstract = {Deep neural networks have become increasingly successful at solving classic perception problems such as object recognition, semantic segmentation, and scene understanding, often reaching or surpassing human-level accuracy. This success is due in part to the ability of DNNs to learn useful representations of high-dimensional inputs, a problem that humans must also solve. We examine the relationship between the representations learned by these networks and human psychological representations recovered from similarity judgments. We find that deep features learned in service of object classification account for a significant amount of the variance in human similarity judgments for a set of animal images. However, these features do not capture some qualitative distinctions that are a key part of human representations. To remedy this, we develop a method for adapting deep features to align with human similarity judgments, resulting in image representations that can potentially be used to extend the scope of psychological experiments.}
}
@article{PetersonetalCoRR-17,
       author = {Joshua C. Peterson and Joshua T. Abbott and Thomas L. Griffiths},
        title = {Leveraging deep neural networks to capture psychological representations},
      journal = {CoRR},
       volume = {arXiv:1706.02417},
         year = {2017},
     abstract = {Decades of psychological research have been aimed at modeling how people learn features and categories. The empirical validation of these theories is often based on artificial stimuli with simple representations. Recently, deep neural networks have reached or surpassed human accuracy on tasks such as identifying objects in natural images. These networks learn representations of real-world stimuli that can potentially be leveraged to capture psychological representations. We find that state-of-the-art object classification networks provide surprisingly accurate predictions of human similarity judgments for natural images, but fail to capture some of the structure represented by people. We show that a simple transformation that corrects these discrepancies can be obtained through convex optimization. We use the resulting representations to predict the difficulty of learning novel categories of natural images. Our results extend the scope of psychological experiments and computational modeling by enabling tractable use of large natural stimulus sets.}
}
@article{PetillaNATURE-08,
       author = {Petilla Interneuron Nomenclature Group},
        title = {Petilla terminology: nomenclature of features of {GABA}ergic interneurons of the cerebral cortex},
      journal = {Nature Reviews Neuroscience},
       volume = 9,
        issue = 7,
         year = 2008,
        pages = {557-568},
     abstract = {Neuroscience produces a vast amount of data from an enormous diversity of neurons. A neuronal classification system is essential to organize such data and the knowledge that is derived from them. Classification depends on the unequivocal identification of the features that distinguish one type of neuron from another. The problems inherent in this are particularly acute when studying cortical interneurons. To tackle this, we convened a representative group of researchers to agree on a set of terms to describe the anatomical, physiological and molecular features of GABAergic interneurons of the cerebral cortex. The resulting terminology might provide a stepping stone towards a future classification of these complex and heterogeneous cells. Consistent adoption will be important for the success of such an initiative, and we also encourage the active involvement of the broader scientific community in the dynamic evolution of this project.}
}
@article{PetittoandMarentetteSCIENCE-91,
       author = {Petitto, L.A. and Marentette, P.F.},
        title = {Babbling in the manual mode: evidence for the ontogeny of language},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {251},
       number = {5000},
        pages = {1493-1496},
         year = {1991},
     abstract = {Infant vocal babbling has been assumed to be a speech-based phenomenon that reflects the maturation of the articulatory apparatus responsible for spoken language production. Manual babbling has now been reported to occur in deaf children exposed to signed languages from birth. The similarities between manual and vocal babbling suggest that babbling is a product of an amodal, brain-based language capacity under maturational control, in which phonetic and syllabic units are produced by the infant as a first step toward building a mature linguistic system. Contrary to prevailing accounts of the neurological basis of babbling in language ontogeny, the speech modality is not critical in babbling. Rather, babbling is tied to the abstract linguistic structure of language and to an expressive capacity capable of processing different types of signals (signed or spoken).},
}
@article{PetittoetalNATURE-01,
       author = {Petitto, Laura Ann and Holowka, Siobhan and Sergio, Lauren E. and Ostry, David},
        title = {Language rhythms in baby hand movements},
      journal = {Nature},
         year = {2001},
       volume = {413},
        issue = {6851},
        pages = {35-36},
     abstract = {Hearing babies born to deaf parents babble silently with their hands.},
}
@article{PetreanuetalNATURE-07,
      title = {Channelrhodopsin-2-assisted circuit mapping of long-range callosal projections},
     author = {Petreanu, Leopoldo and Huber, Daniel and Sobczyk, Aleksander and Svoboda, Karel},
    journal = {Nature Neuroscience},
  publisher = {Nature Publishing Group},
     volume = 10,
      issue = 5,
       year = 2007,
      pages = {663-668},
}
@article{PetrofandShermanPHYSIOLOGY-13,
       author = {Petrof, Iraklis and Sherman, S. Murray},
        title = {Functional significance of synaptic terminal size in glutamatergic sensory pathways in thalamus and cortex},
      journal = {Journal of Physiology},
    publisher = {Blackwell Science},
       volume = 591,
        issue = 13,
         year = 2013,
        pages = {3125-3131},
     abstract = {Glutamatergic pathways are a major information-carrying and -processing network of inputs in the brain. There is considerable evidence suggesting that glutamatergic pathways do not represent a homogeneous group and that they can be segregated into at least two broad categories. Class 1 glutamatergic inputs, which are suggested to be the main information carriers, are characterized by a number of unique synaptic and anatomical features, such as the large synaptic boutons with which they often terminate. On the other hand, Class 2 inputs, which are thought to play a modulatory role, are associated, amongst other features, with exclusively small terminal boutons. Here we summarize and briefly discuss these two classes of glutamatergic input and how their unique features, including their terminal bouton size and anatomy, are related to their suggested function},
}
@phdthesis{PetrovPhD-09,
     author = {Petrov, Slav Orlinov},
      title = {Coarse-to-Fine Natural Language Processing},
     school = {EECS Department, University of California, Berkeley},
      month = {Aug},
      issue = {UCB/EECS-2009-116},
       year = 2009,
}
@inproceedings{PetrovandKleinACL-08,
     author = {Petrov, Slav and Klein, Dan},
      title = {Sparse multi-scale grammars for discriminative latent variable parsing},
  booktitle = {Proceedings of the Conference on Empirical Methods in Natural Language Processing},
   location = {Honolulu, Hawaii},
      pages = {867-876},
       year = 2008,
}
@article{PetrovicetalTIP-04,
    author = {Ana Petrovic and {\`O}scar Divorra Escoda and Pierre Vandergheynst},
     title = {Multiresolution segmentation of natural images: from linear
              to nonlinear scale-space representations},
   journal = {IEEE Transactions on Image Processing},
    volume = 13,
     issue = 8,
     pages = {1104-1114},
      year = 2004,
}
@techreport{PewCollege02,
      author = {Steve Jones},
       title = {The Internet Goes to College: How students are
                living in the future with today's technology},
 institution = {Pew Internet \& American Life Project},
       month = {September},
        year = 2002,
         url = {http://www.pewinternet.org/reports/},
}
@techreport{PewSearch02,
      author = {Susannah Fox},
       title = {Search Engines},
 institution = {Pew Internet \& American Life Project},
       month = {July},
        year = 2002,
         url = {http://www.pewinternet.org/reports/},
}
@article{PfeiletalFiN-13,
       author = {Pfeil, Thomas and Gr\"ubl, Andreas and Jeltsch, Sebastian and M\"uller, Eric and M\"uller, Paul and Petrovici, Mihai A. and Schmuker, Michael and Br\"uderle, Daniel and Schemmel, Johannes and Meier, Karlheinz},
        title = {Six networks on a universal neuromorphic computing substrate},
      journal = {Frontiers in Neuroscience},
       volume = {7},
         year = {2013},
        pages = {11},
     abstract = {In this study, we present a highly configurable neuromorphic computing substrate and use it for emulating several types of neural networks. At the heart of this system lies a mixed-signal chip, with analog implementations of neurons and synapses and digital transmission of action potentials. Major advantages of this emulation device, which has been explicitly designed as a universal neural network emulator, are its inherent parallelism and high acceleration factor compared to conventional computers. Its configurability allows the realization of almost arbitrary network topologies and the use of widely varied neuronal and synaptic parameters. Fixed-pattern noise inherent to analog circuitry is reduced by calibration routines. An integrated development environment allows neuroscientists to operate the device without any prior knowledge of neuromorphic circuit design. As a showcase for the capabilities of the system, we describe the successful emulation of six different neural networks which cover a broad spectrum of both structure and functionality.},
}
@article{PfenningetalSCIENCE-14,
       author = {Pfenning, Andreas R. and Hara, Erina and Whitney, Osceola and Rivas, Miriam V. and Wang, Rui and Roulhac, Petra L. and Howard, Jason T. and Wirthlin, Morgan and Lovell, Peter V. and Ganapathy, Ganeshkumar and Mouncastle, Jacquelyn and Moseley, M. Arthur and Thompson, J. Will and Soderblom, Erik J. and Iriki, Atsushi and Kato, Masaki and Gilbert, M. Thomas P. and Zhang, Guojie and Bakken, Trygve and Bongaarts, Angie and Bernard, Amy and Lein, Ed and Mello, Claudio V. and Hartemink, Alexander J. and Jarvis, Erich D.},
        title = {Convergent transcriptional specializations in the brains of humans and song-learning birds},
      journal = {Science},
       volume = 346,
        issue = 6215,
         year = 2014,
     abstract = {Song-learning birds and humans share independently evolved similarities in brain pathways for vocal learning that are essential for song and speech and are not found in most other species. Comparisons of brain transcriptomes of song-learning birds and humans relative to vocal nonlearners identified convergent gene expression specializations in specific song and speech brain regions of avian vocal learners and humans. The strongest shared profiles relate bird motor and striatal song-learning nuclei, respectively, with human laryngeal motor cortex and parts of the striatum that control speech production and learning. Most of the associated genes function in motor control and brain connectivity. Thus, convergent behavior and neural connectivity for a complex trait are associated with convergent specialized expression of multiple genes.},
}
@inproceedings{PflegerAAAI-FS-04,
    author = {Karl Pfleger},
     title = {On-Line Learning of Predictive Compositional Hierarchies by {H}ebbian Chunking},
 booktitle = {AAAI Fall Symposium: Compositional Connectionism in Cognitive Science},
      year = 2004,
}
@inproceedings{PflegerICDL-04,
    author = {Karl Pfleger},
     title = {On-line cumulative learning of hierarchical sparse $n$-grams},
 booktitle = {Proceedings of the International Conference on Development and Learning},
      year = 2004,
}
@article{PhametalCoRR-18,
       author = {Hieu Pham and Melody Y. Guan and Barret Zoph and Quoc V. Le and Jeff Dean},
        title = {Efficient Neural Architecture Search via Parameter Sharing},
      journal = {CoRR},
       volume = {arXiv:1802.03268},
         year = {2018},
     abstract = {We propose Efficient Neural Architecture Search (ENAS), a fast and inexpensive approach for automatic model design. In ENAS, a controller learns to discover neural network architectures by searching for an optimal subgraph within a large computational graph. The controller is trained with policy gradient to select a subgraph that maximizes the expected reward on the validation set. Meanwhile the model corresponding to the selected subgraph is trained to minimize a canonical cross entropy loss. Thanks to parameter sharing between child models, ENAS is fast: it delivers strong empirical performances using much fewer GPU-hours than all existing automatic model design approaches, and notably, less expensive than standard Neural Architecture Search. On the Penn Treebank dataset, ENAS discovers a novel architecture that achieves a test perplexity of 55.8, establishing a new state-of-the-art among all methods without post-training processing. On the CIFAR-10 dataset, ENAS designs novel architectures that achieve a test error of 2.89\%, which is on par with NASNet (Zoph et al., 2018), whose test error is 2.65\%.},
}
@book{Piaget1936origin,
        title = {Origin of Intelligence in the Child},
       author = {Piaget, Jean},
         year = {1936},
    publisher = {Routledge \& Kegan Paul}
}
@book{Picard1997,
    author = {Picard, Rosalind W.},
     title = {Affective computing},
      year = 1997,
      isbn = {0-262-16170-2},
 publisher = {MIT Press},
   address = {Cambridge, MA},
}
@article{PicardTiN-15,
       author = {Picard, M.},
        title = {Mitochondrial synapses: intracellular communication and signal integration},
      journal = {Trends in Neuroscience},
         year = 2015,
       volume = 38,
        issue = 8,
        pages = {468-474},
     abstract = {Communication is a central theme in biology. Consequently, specialized structures have evolved to permit rapid communication among cells, tissues, organs, and physiological systems, thus enhancing the overall function and adaptation of the organism. A prime example is the neuronal synapse. In the brain, synaptic communication establishes neuronal networks with the capacity to integrate, process, and store information, giving rise to complex output signals capable of orchestrating functions across the organism. At the intracellular level, discoveries now reveal the existence of 'mitochondrial synapses' establishing mitochondrial networks, with defined chromatin-modifying mitochondrial output signals capable of orchestrating gene expression across the genome. These discoveries raise the possibility that in addition to their role as powerhouses and neuromodulators, mitochondria behave as intracellular signal-processing network.},
}
@article{PickettandWilliamsNANOTECHNOLOGY-12,
       author = {Matthew D. Pickett and R. Stanley Williams},
        title = {Sub-100 fJ and sub-nanosecond thermally driven threshold switching in niobium oxide crosspoint nanodevices},
      journal = {Nanotechnology},
       volume = 23,
        issue = 21,
         year = 2012,
        pages = 215202,
     abstract = {We built and measured the dynamical current versus time behavior of nanoscale niobium oxide crosspoint devices which exhibited threshold switching (current-controlled negative differential resistance). The switching speeds of 110 \hmtimes{} 110 nm$^2$ devices were found to be \Delta{} t ONs = 700 ps and \Delta{} t OFF = 2.3 ns while the switching energies were of the order of 100 fJ. We derived a new dynamical model based on the Joule heating rate of a thermally driven insulator-to-metal phase transition that accurately reproduced the experimental results, and employed the model to estimate the switching time and energy scaling behavior of such devices down to the 10 nm scale. These results indicate that threshold switches could be of practical interest in hybrid CMOS nanoelectronic circuits.}
}
@article{PickettetalNATURE-MATERIALS-13,
       author = {Pickett, Matthew D. and Medeiros-Ribeiro, Gilberto and Williams, R. Stanley},
        title = {A scalable neuristor built with Mott memristors},
      journal = {Nature Materials},
    publisher = {Nature Publishing Group},
       volume = 12,
        issue = 2,
         year = 2013,
        pages = {114-117},
     abstract = {The Hodgkin-Huxley model for action potential generation in biological axons is central for understanding the computational capability of the nervous system and emulating its functionality. Owing to the historical success of silicon complementary metal-oxide-semiconductors, spike-based computing is primarily confined to software simulations and specialized analogue metal-oxide-semiconductor field-effect transistor circuits. However, there is interest in constructing physical systems that emulate biological functionality more directly, with the goal of improving efficiency and scale. The neuristor was proposed as an electronic device with properties similar to the Hodgkin-Huxley axon, but previous implementations were not scalable. Here we demonstrate a neuristor built using two nanoscale Mott memristors, dynamical devices that exhibit transient memory and negative differential resistance arising from an insulating-to-conducting phase transition driven by Joule heating. This neuristor exhibits the important neural functions of all-or-nothing spiking with signal gain and diverse periodic spiking, using materials and structures that are amenable to extremely high-density integration with or without silicon transistors.}
}
@book{Pickvance1984vangogh,
        title = {Van Gogh in Arles},
       author = {van Gogh, V. and Pickvance, R. and Metropolitan Museum of Art (New York, N.Y.)},
         year = {1984},
    publisher = {Metropolitan Museum of Art}
}
@article{PiechetalCoRR-15,
       author = {Chris Piech and Jonathan Huang and Andy Nguyen and Mike Phulsuksombati and Mehran Sahami and Leonidas J. Guibas},
        title = {Learning Program Embeddings to Propagate Feedback on Student Code},
      journal = {CoRR},
       volume = {arXiv:1505.05969},
         year = {2015},
     abstract = {Providing feedback, both assessing final work and giving hints to stuck students, is difficult for open-ended assignments in massive online classes which can range from thousands to millions of students. We introduce a neural network method to encode programs as a linear mapping from an embedded precondition space to an embedded postcondition space and propose an algorithm for feedback at scale using these linear maps as features. We apply our algorithm to assessments from the Code.org Hour of Code and Stanford University's CS1 course, where we propagate human comments on student assignments to orders of magnitude more submissions.}
}
@inproceedings{PiechetalICML-15,
       author = {Piech, Chris and Huang, Jonathan and Nguyen, Andy and Phulsuksombati, Mike and Sahami, Mehran and Guibas, Leonidas},
        title = {Learning Program Embeddings to Propagate Feedback on Student Code},
    booktitle = {Proceedings of the 32nd International Conference on International Conference on Machine Learning - Volume 37},
         year = {2015},
     location = {Lille, France},
        pages = {1093-1102},
     abstract = {Providing feedback, both assessing final work and giving hints to stuck students, is difficult for open-ended assignments in massive online classes which can range from thousands to millions of students. We introduce a neural network method to encode programs as a linear mapping from an embedded precondition space to an embedded postcondition space and propose an algorithm for feedback at scale using these linear maps as features. We apply our algorithm to assessments from the Code.org Hour of Code and Stanford University's CS1 course, where we propagate human comments on student assignments to orders of magnitude more submissions.}
}
@article{PierrotetalCoRR-19,
       author = {Thomas Pierrot and Guillaume Ligner and Scott E. Reed and Olivier Sigaud and Nicolas Perrin and Alexandre Laterre and David Kas and Karim Beguir and Nando de Freitas},
        title = {Learning Compositional Neural Programs with Recursive Tree Search and Planning},
      journal = {CoRR},
       volume = {arXiv:1905.12941},
         year = {2019},
     abstract = {We propose a novel reinforcement learning algorithm, AlphaNPI, that incorporates the strengths of Neural Programmer-Interpreters (NPI) and AlphaZero. NPI contributes structural biases in the form of modularity, hierarchy and recursion, which are helpful to reduce sample complexity, improve generalization and increase interpretability. AlphaZero contributes powerful neural network guided search algorithms, which we augment with recursion. AlphaNPI only assumes a hierarchical program specification with sparse rewards: 1 when the program execution satisfies the specification, and 0 otherwise. Using this specification, AlphaNPI is able to train NPI models effectively with RL for the first time, completely eliminating the need for strong supervision in the form of execution traces. The experiments show that AlphaNPI can sort as well as previous strongly supervised NPI variants. The AlphaNPI agent is also trained on a Tower of Hanoi puzzle with two disks and is shown to generalize to puzzles with an arbitrary number of disk}
}
@article{PierrotetalCoRR-20,
        title = {Learning Compositional Neural Programs for Continuous Control},
       author = {Thomas Pierrot and Nicolas Perrin and Feryal Behbahani and Alexandre Laterre and Olivier Sigaud and Karim Beguir and Nando de Freitas},
         year = {2020},
       volume = {arXiv:2007.13363},
     abstract = {We propose a novel solution to challenging sparse-reward, continuous control problems that require hierarchical planning at multiple levels of abstraction. Our solution, dubbed AlphaNPI-X, involves three separate stages of learning. First, we use off-policy reinforcement learning algorithms with experience replay to learn a set of atomic goal-conditioned policies, which can be easily repurposed for many tasks. Second, we learn self-models describing the effect of the atomic policies on the environment. Third, the self-models are harnessed to learn recursive compositional programs with multiple levels of abstraction. The key insight is that the self-models enable planning by imagination, obviating the need for interaction with the world when learning higher-level compositional programs. To accomplish the third stage of learning, we extend the AlphaNPI algorithm, which applies AlphaZero to learn recursive neural programmer-interpreters. We empirically show that AlphaNPI-X can effectively learn to tackle challenging sparse manipulation tasks, such as stacking multiple blocks, where powerful model-free baselines fail.}
}
@article{PietalNATURE-13,
       author = {Pi, Hyun-Jae and Hangya, Bal\`{a}zs and Kvitsiani, Duda and Sanders, Joshua I. and Huang, Z. Josh and Kepecs, Adam},
        title = {Cortical interneurons that specialize in disinhibitory control},
      journal = {Nature},
    publisher = {Nature Publishing Group, a division of Macmillan Publishers Limited. All Rights Reserved.},
       volume = {503},
         year = {2013},
        pages = {521-524},
     abstract = {In the mammalian cerebral cortex the diversity of interneuronal subtypes underlies a division of labour subserving distinct modes of inhibitory control. A unique mode of inhibitory control may be provided by inhibitory neurons that specifically suppress the firing of other inhibitory neurons. Such disinhibition could lead to the selective amplification of local processing and serve the important computational functions of gating and gain modulation. Although several interneuron populations are known to target other interneurons to varying degrees, little is known about interneurons specializing in disinhibition and their in vivo function. Here we show that a class of interneurons that express vasoactive intestinal polypeptide (VIP) mediates disinhibitory control in multiple areas of neocortex and is recruited by reinforcement signals. By combining optogenetic activation with single-cell recordings, we examined the functional role of VIP interneurons in awake mice, and investigated the underlying circuit mechanisms in vitro in auditory and medial prefrontal cortices. We identified a basic disinhibitory circuit module in which activation of VIP interneurons transiently suppresses primarily somatostatin- and a fraction of parvalbumin-expressing inhibitory interneurons that specialize in the control of the input and output of principal cells, respectively. During the performance of an auditory discrimination task, reinforcement signals (reward and punishment) strongly and uniformly activated VIP neurons in auditory cortex, and in turn VIP recruitment increased the gain of a functional subpopulation of principal neurons. These results reveal a specific cell type and microcircuit underlying disinhibitory control in cortex and demonstrate that it is activated under specific behavioural conditions.}
}
@incollection{PiginoetalCYTOSKELETON-12,
       author = {Gustavo Pigino and Yuyu Song and Laura L. Kirkpatrick},
        title = {Chapter 6: Cytoskeleton of Neurons and Glia},
    booktitle = {Basic Neurochemistry: Principles of Molecular, Cellular, and Medical Neurobiology, Edition 8},
    publisher = {Academic Press},
      address = {Waltham, MA},
       editor = {Scott Brady and George Siegel and R. Wayne Albers and Donald Price},
         year = 2012,
        pages = {126-143},
}
@article{PikeetalSCIPROG-05,
       author = {Rob Pike and Sean Dorward and Robert Griesemer and Sean Quinlan},
        title = {Interpreting the Data: Parallel Analysis with Sawzall},
      journal = {Scientific Programming},
       volume = 13,
        issue = 4,
        pages = {277-298},
         year = 2005,
          url = {http://labs.google.com/papers/sawzall-sciprog.pdf},
}
@article{PillowandRubinNEURON-02,
       author = {Jonathan Pillow and Rubin},
        title = {Perceptual Completion across the Vertical Meridian and the Role of Early Visual Cortex},
      journal = {Neuron},
       volume = 33,
         pages = {805-813},
         year = 2002,
          url = {PillowandRubinNEURON-02.pdf},
}
@incollection{PillowetalNIPS-04,
       author = {Jonathan W. Pillow and Liam Paninski and Eero P. Simoncelli},
        title = {Maximum Likelihood Estimation of a Stochastic Integrate-and-Fire Neural Model},
    publisher = {MIT Press},
    booktitle = {Advances in Neural Information Processing Systems},
       editor = {S. Thrun and L. Saul and B. Sch\"{o}lkopf},
       volume = 16,
      address = {Cambridge, MA},
         year = 2004,
        pages = {1311-1318},
          url = {PillowetalNIPS-04},
}
@book{Pinker02,
       author = {Steven Pinker},
        title = {The Blank Slate: The Modern Denial of Human Nature},
    publisher = {Viking},
      address = {New York, NY},
         year = 2002,
}
@book{Pinker94,
       author = {Steven Pinker},
        title = {The Language Instinct: How the Mind Creates Language},
         year = {1994},
      address = {New York},
    publisher = {HarperCollins Publishers},
}
@book{Pinker94instinct,
       author = {Steven Pinker},
        title = {The Language Instinct: {H}ow the Mind Creates Language},
    publisher = {HarperCollins Publishers},
      address = {New York, NY},
         year = {1994},
}
@book{Pinker97,
       author = {Steven Pinker},
        title = {How the Mind Works},
    publisher = {W. W. Norton},
      address = {New York, NY},
         year = 1997,
}
@book{PinkerANGELS-12,
        title = {The Better Angels of Our Nature: Why Violence Has Declined},
       author = {Pinker, Steven},
         year = {2012},
    publisher = {Penguin Books}
}
@incollection{PinkerBOOTSTRAP-87,
       author = {Pinker, Steven},
        title = {The bootstrapping problem in language acquisition},
       editor = {Brian MacWhinney},
    booktitle = {Mechanisms of Language Aquisition},
      address = {Hillsdale, NJ, US},
    publisher = {Lawrence Erlbaum Associates},
         year = {1987},
        pages = {399-441},
     abstract = {This chapter is about the fundamental problem of getting the child started in forming the correct types of rules for natural languages. This "bootstrapping problem" is the first problem one must solve in designing models of language acquisition. I first discuss three proposed solutions that clearly will not work / then I review a solution that worked well enough to lie at the foundation of an explicit theory of language acquisition I outlined in Pinker, 1984a ("Language Learnability and Language Development"), then I try to diagnose the fundamental problem that seems to torpedo all the solutions discussed. I sketch out an alternative class of language acquisition mechanisms, based on "constraint satisfaction" models taken from artificial intelligence research, and show how these models might solve the bootstrapping problem. Finally, I spell out the implication of this class of models for the debate over language bootstrapping}
}
@inproceedings{PintoandCoxAFGR-11,
       author = {Pinto, Nicolas and Cox, David D.},
        title = {{Beyond Simple Features: A Large-Scale Feature Search Approach to Unconstrained Face Recognition}},
    booktitle = {{IEEE} Automatic Face and Gesture Recognition},
         year = 2011,
        pages = {8-15},
      comment = {http://www.rowland.harvard.edu/cox/pdfs/fg2011_lfw.pdf},
}
@inproceedings{PintoetalCVPR-09,
       author = {Pinto, N. and DiCarlo, J.~J. and Cox, D.~D.},
        title = {How far can you get with a modern face recognition test set using only simple features?},
    booktitle = {{IEEE} Computer Society Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
        pages = {2591-2598},
         year = 2009,
}
@inproceedings{PintoetalCVPR-11,
    author = {Pinto, Nicolas and Stone, Zac and Zickler, Todd and Cox, David D.},
     title = {{Scaling-up Biologically-Inspired Computer Vision: A Case-Study on Facebook}},
 booktitle = {{IEEE} Computer Vision and Pattern Recognition, Workshop on Biologically Consistent Vision},
      year = 2011,
     pages = {35-42},
}
@article{PintoetalPLoS-08,
     author = {Pinto, Nicolas and Cox, David D and DiCarlo, James J},
    journal = {{PLoS} Computational Biology},
  publisher = {Public Library of Science},
      title = {Why is Real-World Visual Object Recognition Hard?},
      month = {January},
     volume = 4,
      issue = 1,
      pages = 27,
       year = 2008,
   abstract = {Progress in understanding the brain mechanisms underlying vision requires the construction of computational models that not only emulate the brain's anatomy and physiology, but ultimately match its performance on visual tasks. In recent years, "natural" images have become popular in the study of vision and have been used to show apparently impressive progress in building such models. Here, we challenge the use of uncontrolled "natural" images in guiding that progress. In particular, we show that a simple V1-like model - a neuroscientist's "null" model, which should perform poorly at real-world visual object recognition tasks-outperforms state-of-the-art object recognition systems (biologically inspired and otherwise) on a standard, ostensibly natural image recognition test. As a counterpoint, we designed a "simpler" recognition test to better span the real-world variation in object pose, position, and scale, and we show that this test correctly exposes the inadequacy of the V1-like model. Taken together, these results demonstrate that tests based on uncontrolled natural images can be seriously misleading, potentially guiding progress in the wrong direction. Instead, we reexamine what it means for images to be natural and argue for a renewed focus on the core problem of object recognition-real-world image variation.},
}
@article{PintoetalPLoS-09,
     author = {Pinto, Nicolas and Doukhan, David and DiCarlo, James and Cox, David},
    journal = {{PLoS} Computational Biology},
  publisher = {Public Library of Science},
      title = {A High-Throughput Screening Approach to Discovering Good Forms of Biologically Inspired Visual Representation},
      month = {November},
     volume = 5,
      issue = 11,
      pages = {e1000579},
       year = 2009,
}
@inproceedings{PippengerAMS-90,
       author = {Nicholas Pippenger},
        title = {Developments in 'The Synthesis of Reliable Organisms from Unreliable Components'},
    booktitle = {{Proceedings of Symposia in Pure Mathematics: The Legacy of John von Neumann}},
       volume = {50},
       editor = {James Glimm and John Impagiazo and Isadore Singer},
    publisher = {American Mathematical Society},
         year = {1990},
        pages = {311-324},
}
@article{PippengerSPM-90,
     title = {Developments in "The Synthesis of Reliable Organisms from Unreliable Components"}, 
    author = {Nicholas John Pippenger}, 
    volume = 50,
     issue = 5,
   journal = {Proceedings of Symposia in Pure Mathematics},
      year = 1990, 
}
@article{PirinoetalPHYSICAL_BIOLOGY-15,
        title = {A topological study of repetitive co-activation networks in in vitro cortical assemblies},
       author = {Pirino, Virginia and Riccomagno, Eva and Martinoia, Sergio and Massobrio, Paolo},
      journal = {Physical Biology},
       volume = {12},
       number = {1},
         year = {2014},
        pages = {016007-016007},
     abstract = {To address the issue of extracting useful information from large data-set of large scale networks of neurons, we propose an algorithm that involves both algebraic-statistical and topological tools. We investigate the electrical behavior of in vitro cortical assemblies both during spontaneous and stimulus-evoked activity coupled to Micro-Electrode Arrays (MEAs). Our goal is to identify core sub-networks of repetitive and synchronous patterns of activity and to characterize them. The analysis is performed at different resolution levels using a clustering algorithm that reduces the network dimensionality. To better visualize the results, we provide a graphical representation of the detected sub-networks and characterize them with a topological invariant, i.e. the sequence of Betti numbers computed on the associated simplicial complexes. The results show that the extracted sub-populations of neurons have a more heterogeneous firing rate with respect to the entire network. Furthermore, the comparison of spontaneous and stimulus-evoked behavior reveals similarities in the identified clusters of neurons, indicating that in both conditions similar activation patterns drive the global network activity.}
}
@inproceedings{PirsiavashandRamananCVPR-12,
     title = {Steerable part models},
    author = {Hamed Pirsiavash and Deva Ramanan},
 booktitle = {Proceedings of {IEEE} Conference on Computer Vision and Pattern Recognition},
 publisher = {{IEEE} Computer Society},
      year = 2012,
     pages = {3226-3233},
}
@inproceedings{PittandWarmuthSTOC-89,
       author = {Pitt, Leonard and Warmuth, Manfred K.},
        title = {The Minimum Consistent {DFA} Problem Cannot Be Approximated Within Any Polynomial},
    booktitle = {Proceedings of the Twenty First Annual {ACM} Symposium on Theoretical Computing},
        pages = {421-432},
         year = 1989,
}
@book{PlateCSLI-03a,
       author = {Plate, Tony A.},
        title = {Holographic Reduced Representation: Distributed Representation for Cognitive Structures},
         year = 2003,
    publisher = {CSLI Publications},
          url = {https://web.stanford.edu/group/cslipublications/cslipublications/site/1575864304.shtml},
      comment = {Tony A. Plate is a research scientist at Black Mesa Capital in Santa Fe, New Mexico.},
      address = {Stanford, CA, USA},
     abstract = {While neuroscientists garner success in identifying brain regions and in analyzing individual neurons, ground is still being broken at the intermediate scale of understanding how neurons combine to encode information. This book proposes a method of representing information in a computer that would be suited for modeling the brain's methods of processing information. Holographic Reduced Representations (HRRs) are introduced here to model how the brain distributes each piece of information among thousands of neurons. It had been previously thought that the grammatical structure of a language cannot be encoded practically in a distributed representation, but HRRs can overcome the problems of earlier proposals. Thus this work has implications for psychology, neuroscience, linguistics, and computer science, and engineering.}
}
@inproceedings{PlateIJCAI-91,
       author = {Tony Plate},
        title = {Holographic Reduced Representations: Convolution Algebra for Compositional Distributed Representations},
    booktitle = {International Joint Conference on Artificial Intelligence},
         year = 1991,
        pages = {30-35},
    publisher = {Morgan Kaufmann},
     abstract = {A solution to the problem of representing compositional structure using distributed representations is described. The method uses circular convolution to associate items, which are represented by vectors. Arbitrary variable bindings, short sequences of various lengths, frames, and reduced representations can be compressed into a fixed width vector. These representations are items in their own right, and can be used in constructing compositional structures. The noisy reconstructions given by convolution memories can be cleaned up by using a separate associative memory that has good recon-
structive properties.}
}
@article{PlathandLowryNATURE-11,
       author = {Plath, K. and Lowry, W. E.},
        title = {{P}rogress in understanding reprogramming to the induced pluripotent state},
      journal = {Nature Reviews Genetics},
         year = 2011,
       volume = 12,
        issue = 4,
        pages = {253-265},
}
@incollection{Platt1967,
        title = {Diversity in Science and Technology},
    booktitle = {The Knowledge Most Worth Having},
       author = {John Rader Platt},
       editor = {Booth, Wayne C.},
         year = {1967},
        pages = {109-140},
}
@article{PlattandSpelkeCOiN-09,
       author = {Platt, M.L. and Spelke, E. S},
        title = {What can developmental and comparative cognitive neuroscience tell us about the adult human brain?},
      journal = {Current Opinion in Neurobiology},
       volume = {19},
       number = {1},
         year = {2009},
        pages = {1-5},
     abstract = {An oft-repeated dictum in biology holds that ontogeny recapitulates phylogeny. Despite failures of Haeckl's principle to predict in strict form the morphological, behavioral, or cognitive stages an organism passes through during development, the underlying sentiment continues to inform con- temporary approaches to mind and brain. For example, studies of behavior and cognition in developing humans provide an anchor for understanding both intellectual primitives and the nature of representation in animal minds. Conversely, comparative studies of cognition in animals serve to address seemingly intractable questions concerning the roles of innate capacities, culture, and language in the development of mature human cognition. And both comparative and developmental studies shed light on the neural mechanisms that subserve the most complex and distinctive cognitive abilities of adult humans, such as the capacity for abstract thought.},
}
@article{PlazaetalCoRR-14,
        title = {Annotating Synapses in Large {EM} Datasets},
       author = {Stephen M. Plaza and Toufiq Parag and Gary B. Huang and Donald J. Olbris and Mathew A. Saunders and Patricia K. Rivlin},
         year = {2014},
      journal = {CoRR},
       volume = {arXiv:1409.1801},
     abstract = {Reconstructing neuronal circuits at the level of synapses is a central problem in neuroscience and becoming a focus of the emerging field of connectomics. To date, electron microscopy (EM) is the most proven technique for identifying and quantifying synaptic connections. As advances in EM make acquiring larger datasets possible, subsequent manual synapse identification ({\em i.e.}, proofreading) for deciphering a connectome becomes a major time bottleneck. Here we introduce a large-scale, high-throughput, and semi-automated methodology to efficiently identify synapses. We successfully applied our methodology to the Drosophila medulla optic lobe, annotating many more synapses than previous connectome efforts. Our approaches are extensible and will make the often complicated process of synapse identification accessible to a wider-community of potential proofreaders.}
}
@incollection{PlesseretalEP-07,
        title = {Efficient Parallel Simulation of Large-Scale Neuronal Networks on Clusters of Multiprocessor Computers},
    booktitle = {Euro-Par 2007 Parallel Processing},
       author = {Hans Plesser and Jochen Eppler and Abigail Morrison and Markus Diesmann and Marc-Oliver Gewaltig},
       series = {Lecture Notes in Computer Science},
    publisher = {Springer},
      address = {Berlin},
       volume = 4641,
         year = 2007,
}
@article{PletcheretalBMC-04,
       author = {Pletcher, Mark J. and Tice, Jeffrey A. and Pignone, Michael and McCulloch, Charles and Callister, Tracy Q. and Browner, Warren S.},
        title = {What does my patient's coronary artery calcium score mean?: Combining information from the coronary artery calcium score with information from conventional risk factors to estimate coronary heart disease risk},
      journal = {BMC Medicine},
         year = {2004},
    publisher = {BioMed Central},
      address = {London},
       volume = {2},
        pages = {31-31},
     abstract = {BACKGROUND: The coronary artery calcium (CAC) score is an independent predictor of coronary heart disease. We sought to combine information from the CAC score with information from conventional cardiac risk factors to produce post-test risk estimates, and to determine whether the score may add clinically useful information. METHODS: We measured the independent cross-sectional associations between conventional cardiac risk factors and the CAC score among asymptomatic persons referred for non-contrast electron beam computed tomography. Using the resulting multivariable models and published CAC score-specific relative risk estimates, we estimated post-test coronary heart disease risk in a number of different scenarios. RESULTS: Among 9341 asymptomatic study participants (age 35-88 years, 40\% female), we found that conventional coronary heart disease risk factors including age, male sex, self-reported hypertension, diabetes and high cholesterol were independent predictors of the CAC score, and we used the resulting multivariable models for predicting post-test risk in a variety of scenarios. Our models predicted, for example, that a 60-year-old non-smoking non-diabetic women with hypertension and high cholesterol would have a 47\% chance of having a CAC score of zero, reducing her 10-year risk estimate from 15\% (per Framingham) to 6-9\%; if her score were over 100, however (a 17-25\% chance), her risk estimate would be markedly higher (25\% in 10 years). In low risk scenarios, the CAC score is very likely to be zero or low, and unlikely to change management. CONCLUSION: Combining information from the CAC score with information from conventional risk factors can change assessment of coronary heart disease risk to an extent that may be clinically important, especially when the pre-test 10-year risk estimate is intermediate. The attached spreadsheet makes these calculations easy.},
}
@inproceedings{PlotkinFOCS-91,
        title = {Fast approximation algorithms for fractional packing and covering problems},
       author = {Serge A. Plotkin and David B. Shmoys and Eva Tardos},
    booktitle = {32nd Annual Symposium on Foundations of Computer Science},
        pages = {495-504},
         year = 1991,
}
@techreport{PlotkinTR-92,
        title = {Fast approximation algorithms for fractional packing and covering problems},
       author = {Serge A. Plotkin and David B. Shmoys and Eva Tardos},
        issue = 999,
  institution = {School of Operations Research and Industrial Engineering, College of Engineering, Cornell University},
      address = {Ithaca, NY},
         year = 1992,
}
@article{PlumbleyetalIEEE-10,
       author = {M. D. Plumbley and T. Blumensath and L. Daudet and R. Gribonval and M. E. Davies},
      journal = {Proceedings of the {IEEE}},
        title = {Sparse Representations in Audio and Music: From Coding to Source Separation},
       volume = {98},
       number = {6},
         year = {2010},
        pages = {995-1005},
     abstract = {Sparse representations have proved a powerful tool in the analysis and processing of audio signals and already lie at the heart of popular coding standards such as MP3 and Dolby AAC. In this paper we give an overview of a number of current and emerging applications of sparse representations in areas from audio coding, audio enhancement and music transcription to blind source separation solutions that can solve the "cocktail party problem". In each case we will show how the prior assumption that the audio signals are approximately sparse in some time-frequency representation allows us to address the associated signal processing task.},
}
@inproceedings{PnueliFOCS-77,
       author = {Pnueli, Amir},
        title = {The Temporal Logic of Programs},
    booktitle = {Proceedings of the 18th Annual Symposium on Foundations of Computer Science},
         year = 1977,
        pages = {46-57},
}
@article{PoeppeletalJoN-12,
       author = {Poeppel, David and Emmorey, Karen and Hickok, Gregory and Pylkk{\:{a}}nen, Liina},
        title = {Towards a new neurobiology of language},
      journal = {The Journal of Neuroscience: The official journal of the Society for Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {32},
        issue = {41},
         year = {2012},
        pages = {14125-14131},
     abstract = {Theoretical advances in language research and the availability of increasingly high-resolution experimental techniques in the cognitive neurosciences are profoundly changing how we investigate and conceive of the neural basis of speech and language processing. Recent work closely aligns language research with issues at the core of systems neuroscience, ranging from neurophysiological and neuroanatomic characterizations to questions about neural coding. Here we highlight, across different aspects of language processing (perception, production, sign language, meaning construction), new insights and approaches to the neurobiology of language, aiming to describe promising new areas of investigation in which the neurosciences intersect with linguistic research more closely than before. This paper summarizes in brief some of the issues that constitute the background for talks presented in a symposium at the Annual Meeting of the Society for Neuroscience. It is not a comprehensive review of any of the issues that are discussed in the symposium.},
}
@techreport{PoggioandGirosiTR-89,
       author = {Poggio, Tomaso and Girosi, Federico},
        title = {A Theory of Networks for Approximation and Learning},
        issue = {AI Memo No. 1140},
  institution = {{MIT} {AI} Laboratory},
         year = 1989,
}
@book{Poincare1905,
        title = {Science and Hypothesis},
       author = {Poincar\'{e}, H. and Halsted, G.B.},
       series = {Library of philosophy, psychology and scientific methods},
         year = {1905},
    publisher = {Science Press}
}
@article{PoirazietalNEURON-03,
        title = {Pyramidal Neuron as Two-Layer Neural Network},
       author = {Panayiota Poirazi and Terrence Brannon and Bartlett W. Mel},
      journal = {Neuron},
       volume = 37,
        issue = 6,
         year = 2003,
        pages = {989-999},
}
@book{PoldermanandWillems1998,
       author = {Polderman, Jan Willem and Willems, Jan C.},
        title = {Introduction to Mathematical Systems Theory: A Behavioral Approach},
    publisher = {Springer-Verlag},
      address = {Berlin, Heidelberg},
         year = {1998},
}
@book{Poldrack2018,
        title = {The New Mind Readers: What Neuroimaging Can and Cannot Reveal about Our Thoughts},
       author = {Poldrack, Russell A.},
         year = {2018},
    publisher = {Princeton University Press}
}
@article{PollackAIJ-90,
       author = {Jordan B. Pollack},
        title = {Recursive Distributed Representations},
      journal = {Artificial Intelligence},
         year = 1990,
       volume = 46,
        pages = {77-105},
     abstract = {A long-standing difficulty for connectionist modeling has been how to represent variable-sized recursive data structures, such as trees and lists, in fixed-width patterns. This paper presents a connectionist architecture which automatically develops compact distributed representations for such compositional structures, as well as efficient accessing mechanisms for them. Patterns which stand for the internal nodes of fixed-valence trees are devised through the recursive use of back-propagation on three-layer autoassociative encoder networks. The resulting representations are novel, in that they combine apparently immiscible aspects of features, pointers, and symbol structures. They form a bridge between the data structures necessary for high-level cognitive tasks and the associative, pattern recognition machinery provided by neural networks.},
}
@article{PollocketalTiN-13,
       author = {Pollock, J. D. and Wu, D. Y. and Satterlee, J. S.},
        title = {Molecular neuroanatomy: a generation of progress},
      journal = {Trends in Neurosciences},
         year = 2013,
        pages = {000-000},
     abstract = {The neuroscience research landscape has changed dra- matically over the past decade. Specifically, an impressive array of new tools and technologies have been generated, including but not limited to: brain gene expression atlases, genetically encoded proteins to monitor and manipulate neuronal activity, and new methods for imaging and mapping circuits. However, despite these technological advances, several significant challenges must be overcome to enable a better understanding of brain function and to develop cell type-targeted therapeutics to treat brain disorders. This review provides an overview of some of the tools and technologies currently being used to advance the field of molecular neuroanatomy, and also discusses emerging technologies that may enable neuroscientists to address these crucial scientific challenges over the coming decade.}
}
@misc{PolosukhinNPI-REVIEW-18,
        title = {Review of Neural Programmer-Interpreters},
       author = {Illia Polosukhin},
 howpublished = {\url{https://medium.com/near-ai/review-of-neural-programmer-interpreters-854a14a494fb}},
         year = {2018},
}
@article{PolosukhinandSkidanovCoRR-18,
       author = {Illia Polosukhin and Alexander Skidanov},
        title = {Neural Program Search: Solving Programming Tasks from Description and Examples},
      journal = {CoRR},
       volume = {arXiv:1802.04335},
         year = {2018},
     abstract = {We present a Neural Program Search, an algorithm to generate programs from natural language description and a small number of input/output examples. The algorithm combines methods from Deep Learning and Program Synthesis fields by designing rich domain-specific language (DSL) and defining efficient search algorithm guided by a Seq2Tree model on it. To evaluate the quality of the approach we also present a semi-synthetic dataset of descriptions with test examples and corresponding programs. We show that our algorithm significantly outperforms a sequence-to-sequence model with attention baseline.}
}
@article{PolosukhinandSkidanovICLR-18,
        title = {Neural Program Search: Solving Programming Tasks from Description and Examples},
       author = {Illia Polosukhin and Alexander Skidanov},
      journal = {International Conference on Learning Representations (Accepted)},
         year = {2018},
          url = {https://openreview.net/forum?id=B1KJJf-R-},
     abstract = {We present a Neural Program Search, an algorithm to generate programs from natural language description and a small number of input / output examples. The algorithm combines methods from Deep Learning and Program Synthesis fields by designing rich domain-specific language (DSL) and defining efficient search algorithm guided by a Seq2Tree model on it. To evaluate the quality of the approach we also present a semi-synthetic dataset of descriptions with test examples and corresponding programs. We show that our algorithm significantly outperforms sequence-to-sequence model with attention baseline.}
} 
@article{PolskyetalNN-04,
        title = {Computational subunits in thin dendrites of pyramidal cells},
       author = {Polsky A and Mel B.W. and Schiller J.},
      journal = {Nature Neuroscience},
       volume = 37,
        issue = 6,
         year = 2004,
        pages = {621-617},
}
@misc{Polymer_Data_Handbook-99,
        title = {Polymer Data Handbook},
       author = {James E. Mark},
         year = 1999,
          url = {file:///u/tld/Drive/write/articles/POLYMER_DATA_HANDBOOK/POLYMER_DATA_HANDBOOK-99.pdf}, 
 howpublished = {Oxford University Press},
}
@incollection{PomerleauNIPS-89,
       author = {Pomerleau, Dean A.},
        title = {{ALVINN}: An Autonomous Land Vehicle In a Neural Network},
       editor = {Touretsky, David},
    booktitle = {Advances in Neural Information Processing 1},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1989,
}
@incollection{PomerleauNIPS-91,
       author = {Pomerleau, Dean A.},
        title = {Rapidly Adapting Artificial Neural Network for Autonomonus Navgation},
       editor = {Lippmann, R. P. and Moody, J. E. and Touretsky, D. S.},
    booktitle = {Advances in Neural Information Processing 3},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1991,
}
@book{Pontryaginetal62,
       author = {Pontryagin, L. S. and Boltyanskii, V. G. and Gamkrelidze, R. V.
                and Msichenko, E. F.},
        title = {The Mathematical Theory of Optimal Processes},
    publisher = {John Wiley and Sons},
      address = {New York},
         year = 1962,
}
@article{PontzeretalNATURE-16,
       author = {Pontzer, Herman and Brown, Mary H. and Raichlen, David A. and Dunsworth, Holly and Hare, Brian and Walker, Kara and Luke, Amy and Dugas, Lara R. and Durazo-Arvizu, Ramon and Schoeller, Dale and Plange-Rhule, Jacob and Bovet, Pascal and Forrester, Terrence E. and Lambert, Estelle V. and Thompson, Melissa Emery and Shumaker, Robert W. and Ross, Stephen R.},
        title = {Metabolic acceleration and the evolution of human brain size and life history},
      journal = {Nature},
       volume = {533},
        issue = {7603},
         year = {2016},
        pages = {390-392},
     abstract = {Humans are distinguished from the other living apes in having larger brains and an unusual life history that combines high reproductive output with slow childhood growth and exceptional longevity. This suite of derived traits suggests major changes in energy expenditure and allocation in the human lineage, but direct measures of human and ape metabolism are needed to compare evolved energy strategies among hominoids. Here we used doubly labelled water measurements of total energy expenditure (TEE; kcal / day) in humans, chimpanzees, bonobos, gorillas and orangutans to test the hypothesis that the human lineage has experienced an acceleration in metabolic rate, providing energy for larger brains and faster reproduction without sacrificing maintenance and longevity. In multivariate regressions including body size and physical activity, human TEE exceeded that of chimpanzees and bonobos, gorillas and orangutans by approximately 400, 635 and 820 kcal / day, respectively, readily accommodating the cost of humans' greater brain size and reproductive output. Much of the increase in TEE is attributable to humans' greater basal metabolic rate (kcal / day), indicating increased organ metabolic activity. Humans also had the greatest body fat percentage. An increased metabolic rate, along with changes in energy allocation, was crucial in the evolution of human brain size and life history.},
}
@inproceedings{PopleIJCAI-73,
       author = {Pople, H.},
        title = {On the Mechanization of Abductive Logic},
    booktitle = {Proceedings {IJCAI} 3},
     location = {Menlo Park, California},
 organization = {IJCAII},
         year = 1973,
        pages = {147-152},
}
@article{PopovicetalJoVE-12,
       author = {Popovic M. and Gao X. and Zecevic D.},
        title = {Voltage-sensitive dye recording from axons, dendrites and dendritic spines of individual neurons in brain slices},
      journal = {Journal of Visualized Experiments},
       volume = 29,
        issue = 69,
         year = 2012,
        pages = {e4261},
}
@misc{Portilla-Web,
        title = {http://decsai.ugr.es/\ {}javier/denoise/test\_images/index.htm},
}
@article{PortillaetalTIP-04,
       author = {J. Portilla and V. Strela and Martin J. Wainwright and E. P. Simoncelli},
        title = {Image denoising using {Gaussian} scale mixtures in the wavelet domain},
      journal = {IEEE Transactions Image Processing},
       volume = 12,
        issue = 11,
        pages = {1338-1351},
         year = 2004,
}
@article{PortuguesetalNEURON-14,
       author = {Portugues, Ruben and Feierstein, Claudia E. and Engert, Florian and Orger, Michael B.},
        title = {Whole-brain activity maps reveal stereotyped, distributed networks for visuomotor behavior},
      journal = {Neuron},
         year = {2014},
       volume = {81},
        issue = {6},
        pages = {1328-1343},
     abstract = {Most behaviors, even simple innate reflexes, are mediated by circuits of neurons spanning areas throughout the brain. However, in most cases, the distribution and dynamics of firing patterns of these neurons during behavior are not known. We imaged activity, with cellular resolution, throughout the whole brains of zebrafish performing the optokinetic response. We found a sparse, broadly distributed network that has an elaborate but ordered pattern, with a bilaterally symmetrical organization. Activity patterns fell into distinct clusters reflecting sensory and motor processing. By correlating neuronal responses with an array of sensory and motor variables, we find that the network can be clearly divided into distinct functional modules. Comparing aligned data from multiple fish, we find that the spatiotemporal activity dynamics and functional organization are highly stereotyped across individuals. These experiments systematically reveal the functional architecture of neural circuits underlying a sensorimotor behavior in a vertebrate brain.},
}
@article{PosaDM-76,
       author = {P\'{o}sa, L.},
        title = {Hamiltonian Circuits in Random Graphs},
      journal = {Discrete Mathematics},
       volume = 14,
         year = 1976,
        pages = 359,
}
@article{PosnerATTENTIONAL-NETWORKS-12,
       author = {Posner, Michael I.},
        title = {Attentional Networks and Consciousness},
      journal = {Frontiers in Psychology},
         year = {2012},
    publisher = {Frontiers Research Foundation},
       volume = {3},
        pages = {64},
     abstract = {The term consciousness is an important one in the vernacular of the western literature in many fields. It is no wonder that scientists have assumed that consciousness will be found as a component of the human brain and that we will come to understand its neural basis. However, there is rather little in common between consciousness as the neurologist would use it to diagnose the vegetative state, how the feminist would use it to support raising male consciousness of the economic plight of women and as the philosopher would use it when defining the really hard question of the subjective state of awareness induced by sensory qualities. When faced with this kind of problem it is usual to subdivide the term into more manageable perhaps partly operational definitions. Three meanings that capture aspects of consciousness are: (1) the neurology of the state of mind allowing coherent orientation to time and place (2) the selection of sensory or memorial information for awareness and (3) the voluntary control over overt responses. In each of these cases the mechanisms of consciousness overlap with one or more of the attentional networks that have been studied with the methods of cognitive neuroscience. In this paper we explore the overlap and discuss how to exploit the growing knowledge of attentional networks to constrain ideas of consciousness.},
}
@article{PotterandFaulconerNATURE-75,
     title = {Time to understand pictures and words},
    author = {Potter, M. C. and Faulconer, B. A.},
   journal = {Nature},
    volume = 253,
     issue = 5491,
      year = 1975,
     pages = {437-438},
}
@article{PouilleandScanzianiNATURE-04,
       author = {Pouille, Frederic and Scanziani, Massimo},
        title = {Routing of spike series by dynamic circuits in the hippocampus},
      journal = {Nature},
       volume = {429},
        issue = {6993},
         year = {2004},
        pages = {717-723},
     abstract = {Recurrent inhibitory loops are simple neuronal circuits found in the central nervous system, yet little is known about the physiological rules governing their activity. Here we use simultaneous somatic and dendritic recordings in rat hippocampal slices to show that during a series of action potentials in pyramidal cells recurrent inhibition rapidly shifts from their soma to the apical dendrites. Two distinct inhibitory circuits are sequentially recruited to produce this shift: one, time-locked with submillisecond precision to the onset of the action potential series, transiently inhibits the somatic and perisomatic regions of pyramidal cells; the other, activated in proportion to the rate of action potentials in the series, durably inhibits the distal apical dendrites. These two operating modes result from the synergy between pre- and postsynaptic properties of excitatory synapses onto recurrent inhibitory neurons with distinct projection patterns. Thus, the onset of a series of action potentials and the rate of action potentials in the series are selectively captured and transformed into different spatial patterns of recurrent inhibition.}
}
@article{PoulteretalCURRENT-BIOLOGY-18,
        title = {The Neurobiology of Mammalian Navigation},
       author = {Steven Poulter and Tom Hartley and Colin Lever},
      journal = {Current Biology},
       volume = {28},
       number = {17},
        pages = {R1023-R1042},
         year = {2018},
     abstract = {Mammals have evolved specialized brain systems to support efficient navigation within diverse habitats and over varied distances, but while navigational strategies and sensory mechanisms vary across species, core spatial components appear to be widely shared. This review presents common elements found in mammalian spatial mapping systems, focusing on the cells in the hippocampal formation representing orientational and locational spatial information, and 'core' mammalian hippocampal circuitry. Mammalian spatial mapping systems make use of both allothetic cues (space-defining cues in the external environment) and idiothetic cues (cues derived from self-motion). As examples of each cue type, we discuss: environmental boundaries, which control both orientational and locational neuronal activity and behaviour; and 'path integration', a process that allows the estimation of linear translation from velocity signals, thought to depend upon grid cells in the entorhinal cortex. Building cognitive maps entails sampling environments: we consider how the mapping system controls exploration to acquire spatial information, and how exploratory strategies may integrate idiothetic with allothetic information. We discuss how 'replay' may act to consolidate spatial maps, and simulate trajectories to aid navigational planning. Finally, we discuss grid cell models of vector navigation.}
}
@article{PourdamghanietalPRL-12,
        title = {Graph based semi-supervised human pose estimation: When the output space comes to help},
      journal = {Pattern Recognition Letters},
       volume = {33},
       number = {12},
        pages = {1529-1535},
         year = {2012},
     abstract = {In this letter, we introduce a semi-supervised manifold regularization framework for human pose estimation. We utilize the unlabeled data to compensate for the complexities in the input space and model the underlying manifold by a nearest neighbor graph. We argue that the optimal graph is a subgraph of the k nearest neighbors (k-NN) graph. Then, we estimate distances in the output space to approximate this subgraph. In addition, we use the underlying manifold of the points in the output space to introduce a novel regularization term which captures the correlation among the output dimensions. The modified graph and the proposed regularization term are utilized for a smooth regression over both the learned input and output manifolds. Experimental results on various human activities demonstrate the superiority of the proposed algorithm compared to the current state of the art methods.}
} 
@incollection{Powers03,
      author = {Richard Powers},
       title = {Literary Devices},
   booktitle = {Living With the Genie: Essays on Technology and the Quest for Human Mastery},
      editor = {Alan Lightman and Daniel Sarewitz and Christina Desser},
   publisher = {Island Press},
     address = {Washington, DC},
        year = 2003,
}
@book{Powers95,
      author = {Richard Powers},
       title = {Galatea 2.2},
   publisher = {Farrar, Straus \& Giroux},
     address = {New York, NY},
        year = 1995,
}
@inproceedings{PradhamandDagumUAI-96,
       author = {Pradhan, Malcolm and Dagum, Paul},
        title = {Optimal Monte Carlo Estimation of Belief Network Inference},
    booktitle = {Twelfth Conference on Uncertainty in Artificial Intelligence},
     location = {Portland, Oregon},
        pages = {446-453},
         year = 1996,
}
@article{PratchettetalNARATIVIUM-13,
       author = {A S. Byatt},
        title = {Terry {P}ratchett, science and story telling},
      journal = {The New Statesman},
         year = {2013},
       volume = {April},
      comment = {https://www.newstatesman.com/culture/culture/2013/04/terry-pratchett-science-and-story-telling}
}
@article{Pratt,
       author = {Pratt, John W.},
        title = {Risk aversion in the small and in the large},
      journal = {Econometrica},
       volume = 32,
         year = 1964,
        pages = {122-136},
}
@inproceedings{PrattPOPL-69,
       author = {Pratt, V. R.},
        title = {Process Logic},
    booktitle = {Proceedings of the 6th {POPL}},
 organization = {ACM},
        pages = {93-100},
         year = 1969,
}
@article{PremackPNAS-07,
       author = {Premack, David},
        title = {{{H}uman and animal cognition: continuity and discontinuity}},
      journal = {Proceedings of the National Academy of Science},
         year = {2007},
       volume = {104},
       number = {35},
        pages = {13861-13867},
     abstract = {Microscopic study of the human brain has revealed neural structures, enhanced wiring, and forms of connectivity among nerve cells not found in any animal, challenging the view that the human brain is simply an enlarged chimpanzee brain. On the other hand, cognitive studies have found animals to have abilities once thought unique to the human. This suggests a disparity between brain and mind. The suggestion is misleading. Cognitive research has not kept pace with neural research. Neural findings are based on microscopic study of the brain and are primarily cellular. Because cognition cannot be studied microscopically, we need to refine the study of cognition by using a different approach. In examining claims of similarity between animals and humans, one must ask: What are the dissimilarities? This approach prevents confusing similarity with equivalence. We follow this approach in examining eight cognitive cases--teaching, short-term memory, causal reasoning, planning, deception, transitive inference, theory of mind, and language--and find, in all cases, that similarities between animal and human abilities are small, dissimilarities large. There is no disparity between brain and mind.}
}
@article{PremackandWoodruffBBS-78,
        title = {Does the chimpanzee have a theory of mind?},
       author = {Premack, David and Woodruff, Guy},
       volume = {1},
       number = {4},
      journal = {Behavioral and Brain Sciences},
    publisher = {Cambridge University Press},
         year = {1978},
        pages = {515-526},
     abstract = {An individual has a theory of mind if he imputes mental states to himself and others. A system of inferences of this kind is properly viewed as a theory because such states are not directly observable, and the system can be used to make predictions about the behavior of others. As to the mental states the chimpanzee may infer, consider those inferred by our own species, for example, purpose or intention, as well as knowledge, belief, thinking, doubt, guessing, pretending, liking, and so forth. To determine whether or not the chimpanzee infers states of this kind, we showed an adult chimpanzee a series of videotaped scenes of a human actor struggling with a variety of problems. Some problems were simple, involving inaccessible food \emdash{} bananas vertically or horizontally out of reach, behind a box, and so forth \emdash{} as in the original Kohler problems; others were more complex, involving an actor unable to extricate himself from a locked cage, shivering because of a malfunctioning heater, or unable to play a phonograph because it was unplugged. With each videotape the chimpanzee was given several photographs, one a solution to the problem, such as a stick for the inaccessible bananas, a key for the locked up actor, a lit wick for the malfunctioning heater. The chimpanzee's consistent choice of the correct photographs can be understood by assuming that the animal recognized the videotape as representing a problem, understood the actor's purpose, and chose alternatives compatible with that purpose.}
}
@article{PrescottetalNEUROPHYSIOLOGY-08,
       author = {Prescott, Steven A. and Ratt\`{e}, St\`{e}phanie and De Koninck, Yves and Sejnowski, Terrence J.},
        title = {Pyramidal Neurons Switch From Integrators In Vitro to Resonators Under In Vivo-Like Conditions},
      journal = {Journal of Neurophysiology},
    publisher = {American Physiological Society},
       volume = 100,
        issue = 6,
         year = 2008,
        pages = {3030-3042},
     abstract = {During wakefulness, pyramidal neurons in the intact brain are bombarded by synaptic input that causes tonic depolarization, increased membrane conductance (i.e., shunting), and noisy fluctuations in membrane potential; by comparison, pyramidal neurons in acute slices typically experience little background input. Such differences in operating conditions can compromise extrapolation of in vitro data to explain neuronal operation in vivo. For instance, pyramidal neurons have been identified as integrators (i.e., class 1 neurons according to Hodgkin's classification of intrinsic excitability) based on in vitro experiments but that classification is inconsistent with the ability of hippocampal pyramidal neurons to oscillate/resonate at theta frequency since intrinsic oscillatory behavior is limited to class 2 neurons. Using long depolarizing stimuli and dynamic clamp to reproduce in vivo-like conditions in slice experiments, we show that CA1 hippocampal pyramidal cells switch from integrators to resonators, i.e., from class 1 to class 2 excitability. The switch is explained by increased outward current contributed by the M-type potassium current I(M), which shifts the balance of inward and outward currents active at perithreshold potentials and thereby converts the spike-initiating mechanism as predicted by dynamical analysis of our computational model. Perithreshold activation of I(M) is enhanced by the depolarizing shift in spike threshold caused by shunting and/or sodium channel inactivation secondary to tonic depolarization. Our conclusions were validated by multiple comparisons between simulation and experimental data. Thus even so-called intrinsic properties may differ qualitatively between in vitro and in vivo conditions.},
}
@book{Pressetal92,
       author = {Press, William H. and Teukolsky, Saul A. and Vettering, William T. and Flannery, Brian P.},
        title = {Numerical Recipes in {C}: The Art of Scientific Computing (Second Edition)},
    publisher = {Cambridge University Press},
      address = {New York},
         year = 1992,
}
@article{PrestonandEichenbaumCURRENT-BIOLOGY-13,
       author = {Preston, Alison R. and Eichenbaum, Howard},
        title = {Interplay of hippocampus and prefrontal cortex in memory},
      journal = {Current Biology},
       volume = {23},
        issue = {17},
         year = {2013},
        pages = {R764-R773},
     abstract = {Recent studies on the hippocampus and the prefrontal cortex have considerably advanced our understanding of the distinct roles of these brain areas in the encoding and retrieval of memories, and of how they interact in the prolonged process by which new memories are consolidated into our permanent storehouse of knowledge. These studies have led to a new model of how the hippocampus forms and replays memories and how the prefrontal cortex engages representations of the meaningful contexts in which related memories occur, as well as how these areas interact during memory retrieval. Furthermore, they have provided new insights into how interactions between the hippocampus and prefrontal cortex support the assimilation of new memories into pre-existing networks of knowledge, called schemas, and how schemas are modified in this process as the foundation of memory consolidation.},
}
@incollection{Preuss2001,
     title = {The discovery of cerebral diversity: {A}n unwelcome scientific revolution},
    author = {Todd M. Preuss},
    editor = {Dean Falk and Kathleen R. Gibson},
 booktitle = {Evolutionary Anatomy of the Primate Cerebral Cortex},
 publisher = {Cambridge University Press},
      year = 2001,
     pages = {138-164},
}
@incollection{PreussCN-09, 
     title = {The Cognitive Neuroscience of Human Uniquness},
    author = {Todd M. Preuss}, 
 booktitle = {The Cognitive Neurosciences, 4th Edition}, 
 publisher = {MIT Press}, 
   address = {Cambridge, MA},
    editor = {Gazzaniga, Michael}, 
      year = 2009, 
     pages = {49-66},
}
@article{PreussetalNATURE-04,
    author = {Preuss, T. M. and C\'{a}ceres, M. and Oldham, M. C. and Geschwind, D. H.},
   journal = {Nature Review Genetics},
    volume = 5,
     issue = 11,
     pages = {850-60},
     title = {Human brain evolution: {I}nsights from microarrays},
      year = 2004,
  abstract = {Several recent microarray studies have compared gene-expression patterns n humans, chimpanzees and other non-human primates to identify evolutionary changes that contribute to the distinctive cognitive and behavioural characteristics of humans. These studies support the surprising conclusion that the evolution of the human brain involved an upregulation of gene expression relative to non-human primates  a finding that could be relevant to understanding human cerebral physiology and function. These results show how genetic and genomic methods can shed light on the basis of human neural and cognitive specializations, and have important implications for neuroscience, anthropology and medicine.}, 
}
@article{PrevedeletalCoRR-13,
       author = {R. Prevedel and Y.G. Yoon and M. Hoffmann and N. Pak and G. Wetzstein and S. Kato and T. Schr\"{o}del and R. Raskar and M. Zimmer and E.S. Boyden and A. Vaziri},
        title = {Simultaneous whole-animal {3D}-imaging of neuronal activity using light field microscopy},
      journal = {CoRR},
       volume = {arXiv:1401.5333},
         year = 2013,
     abstract = {3D functional imaging of neuronal activity in entire organisms at single cell level and physiologically relevant time scales faces major obstacles due to trade-offs between the size of the imaged volumes, and spatial and temporal resolution. Here, using light-field microscopy in combination with {3D} deconvolution, we demonstrate intrinsically simultaneous volumetric functional imaging of neuronal population activity at single neuron resolution for an entire organism, the nematode Caenorhabditis elegans. The simplicity of our technique and possibility of the integration into epi-fluoresence microscopes makes it an attractive tool for high-speed volumetric calcium imaging.}
}  
@article{PrevedeletalNATURE-METHODS-14,
       author = {Prevedel, Robert and Yoon, Young-Gyu and Hoffmann, Maximilian and Pak, Nikita and Wetzstein, Gordon and Kato, Saul and Schrodel, Tina and Raskar, Ramesh and Zimmer, Manuel and Boyden, Edward S. and Vaziri, Alipasha},
        title = {Simultaneous whole-animal 3D imaging of neuronal activity using light-field microscopy},
      journal = {Nature Methods},
         year = {2014},
    publisher = {Nature Publishing Group},
       volume = {11},
        issue = {7},
        pages = {727-730},
     abstract = {High-speed, large-scale three-dimensional (3d) imaging of neuronal activity poses a major challenge in neuroscience. here we demonstrate simultaneous functional imaging of neuronal activity at single-neuron resolution in an entire Caenorhabditis elegans and in larval zebrafish brain. our technique captures the dynamics of spiking neurons in volumes of \approx{}700 mm \times{}
700 mm \times{} 200 mm at 20 hz. its simplicity makes it an attractive tool for high-speed volumetric calcium imaging.},
}
@article{PrevedeletalNATURE-METHODS-16,
       author = {Prevedel, Robert and Verhoef, Aart J. and Pernia-Andrade, Alejandro J. and Weisenburger, Siegfried and Huang, Ben S. and Nobauer, Tobias and Fernandez, Alma and Delcour, Jeroen E. and Golshani, Peyman and Baltuska, Andrius and Vaziri, Alipasha},
        title = {Fast volumetric calcium imaging across multiple cortical layers using sculpted light},
      journal = {Nature Methods},
         year = {2016},
    publisher = {Nature Publishing Group},
     abstract = {Although whole-organism calcium imaging in small and semi-transparent animals has been demonstrated, capturing the functional dynamics of large-scale neuronal circuits in awake behaving mammals at high speed and resolution has remained one of the main frontiers in systems neuroscience. Here we present a method based on light sculpting that enables unbiased single- and dual-plane high-speed (up to 160 Hz) calcium imaging as well as in vivo volumetric calcium imaging of a mouse cortical column (0.5 mm [times] 0.5 mm [times] 0.5 mm) at single-cell resolution and fast volume rates (3-6 Hz). We achieved this by tailoring the point-spread function of our microscope to the structures of interest while maximizing the signal-to-noise ratio using a home-built fiber laser amplifier with pulses that are synchronized to the imaging voxel speed. This enabled in vivo recording of calcium dynamics of several thousand neurons across cortical layers and in the hippocampus of awake behaving mice.},
}
@article{PreziosoetalNATURE-15,
       author = {Prezioso, M. and Merrikh-Bayat, F. and Hoskins, B. D. and Adam, G. C. and Likharev, K. K. and Strukov, D. B.},
        title = {Training and operation of an integrated neuromorphic network based on metal-oxide memristors},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = 521,
        issue = 7550,
         year = 2015,
        pages = {61-64},
     anstract = {Despite all the progress of semiconductor integrated circuit technology, the extreme complexity of the human cerebral cortex, featuring in particular approximately $10^{14}$ synapses, makes the hardware implementation of neuromorphic networks with a comparable number of devices exceptionally challenging. One of the most prospective candidates to provide comparable complexity, while operating much faster and with manageable power dissipation, are so-called CrossNets based on hybrid CMOS/memristor circuits. In these circuits, the usual complementary metal-oxide semiconductor (CMOS) stack is augmented with one or several crossbar layers, with adjustable two-terminal resistive devices ("memristors") at each crosspoint. Recently, there was a significant progress in improvement of technology of fabrication of such memristive crossbars and their integration with CMOS circuits, including first demonstrations of their vertical integration. Separately, there have been several demonstrations of discrete memristors as artificial synapses for neuromorphic networks. Very recently such experiments were extended to crossbar arrays of phase-change memristive devices. The adjustment of such devices, however, requires an additional transistor at each crosspoint, and hence the prospects of their scaling are much less impressive than those of metal-oxide memristors, whose nonlinear I-V curves enable transistor-free operation. Here we report the first experimental implementation of a transistor-free metal-oxide memristor crossbar with device variability lowered sufficiently to demonstrate a successful operation of a simple integrated neural network, a single-layer perceptron. The network could be taught in situ using a coarse-grain variety of the delta-rule algorithm to perform the perfect classification of $3 \times 3$-pixel black/white images into 3 classes. We believe that this demonstration is an important step towards the implementation of much larger and more complex memristive neuromorphic networks.},
}
@article{PrietoetalVR-02,
        title = {Development of new expression vector based on Pseudorabies virus amplicons: application to human insulin expression},
       author = {Jes\'{u}s Prieto and Jes\'{u}s Solera and Enrique Tabar\'{e}s},
      journal = {Virus Research},
       volume = 89,
        issue = 1,
         year = 2002,
        pages = {123-129},
}
@incollection{Prinz1984,
       author = {Prinz, Wolfgang},
       editor = {Prinz, Wolfgang and Sanders, Andries F.},
        title = {Modes of Linkage Between Perception and Action},
    booktitle = {Cognition and Motor Processes},
         year = {1984},
    publisher = {Springer Berlin Heidelberg},
      address = {Berlin, Heidelberg},
        pages = {185-193},
     abstract = {The problem of understanding the nature of the linkage between perception and action can be raised at two different levels: execution and initiation. At the level of execution the linkage problem is mainly related to the conditions and mechanisms of interaction between efferent and afferent information during the execution of a given movement or response. Relevant issues such as open vs. closed loop control or feed backward vs. feed forward have recently attracted considerable interest in the field of motor learning and motor control.},
}
@article{PrinzetalNATURE-IMMUNOLOGY-17,
       author = {Prinz, Marco and Erny, Daniel and Hagemeyer, Nora},
        title = {Ontogeny and homeostasis of {CNS} myeloid cells},
      journal = {Nature Immunology},
    publisher = {Nature Publishing Group},
       volume = {18},
         year = {2017},
        pages = {385-392},
     abstract = {Myeloid cells in the central nervous system (CNS) represent a heterogeneous class of innate immune cells that contribute to the maintenance of tissue homeostasis differentially during development and adulthood. The subsets of CNS myeloid cells identified so far, including parenchymal microglia and non-parenchymal meningeal, perivascular and choroid-plexus macrophages, as well as disease-associated monocytes, have classically been distinguished on the basis of their surface epitope expression, localization and morphology. However, studies using cell-specific targeting, in vivo imaging, single-cell expression analysis and other sophisticated tools have now increased the depth of knowledge of this immune-cell compartment and call for reevaluation of the traditional views on the origin, fate and function of distinct CNS myeloid subsets. The concepts of CNS macrophage biology that are emerging from these new insights have broad implications for the understanding and treatment of CNS diseases.},
}
@book{Prior67,
       author = {Prior, A. N.},
        title = {Past, Present, and Future},
    publisher = {Clarendon Press},
     location = {Oxford},
         year = 1967,
}
@article{PritzeletalCoRR-17,
       author = {Alexander Pritzel and Benigno Uria and Sriram Srinivasan and Adri\`{a} Puigdom\`{e}nech Badia and Oriol Vinyals and Demis Hassabis and Daan Wierstra and Charles Blundell},
        title = {Neural Episodic Control},
      journal = {CoRR},
       volume = {arXiv:1703.01988},
         year = {2017},
     abstract = {Deep reinforcement learning methods attain super-human performance in a wide range of environments. Such methods are grossly inefficient, often taking orders of magnitudes more data than humans to achieve reasonable performance. We propose Neural Episodic Control: a deep reinforcement learning agent that is able to rapidly assimilate new experiences and act upon them. Our agent uses a semi-tabular representation of the value function: a buffer of past experience containing slowly changing state representations and rapidly updated estimates of the value function. We show across a wide range of environments that our agent learns significantly faster than other state-of-the-art, general purpose deep reinforcement learning agents.}
}
@inproceedings{PritzeletalICML-17,
        title = {Neural Episodic Control},
       author = {Alexander Pritzel and Benigno Uria and Sriram Srinivasan and Adri{\`a} Puigdom{\`e}nech Badia and Oriol Vinyals and Demis Hassabis and Daan Wierstra and Charles Blundell},
    booktitle = {Proceedings of the 34th International Conference on Machine Learning},
       editor = {Doina Precup and Yee Whye Teh},
       volume = {70},
       series = {Proceedings of Machine Learning Research},
      address = {International Convention Centre, Sydney, Australia},
    publisher = {PMLR},
         year = {2017},
        pages = {2827-2836},
     abstract = {Deep reinforcement learning methods attain super-human performance in a wide range of environments. Such methods are grossly inefficient, often taking orders of magnitudes more data than humans to achieve reasonable performance. We propose Neural Episodic Control: a deep reinforcement learning agent that is able to rapidly assimilate new experiences and act upon them. Our agent uses a semi-tabular representation of the value function: a buffer of past experience containing slowly changing state representations and rapidly updated estimates of the value function. We show across a wide range of environments that our agent learns significantly faster than other state-of-the-art, general purpose deep reinforcement learning agents.}
}
@article{ProskovecetalSCIENTIFIC-REPORTS-18,
       author = {Proskovec, Amy L. and Wiesman, Alex I. and Heinrichs-Graham, Elizabeth and Wilson, Tony W.},
        title = {Beta Oscillatory Dynamics in the Prefrontal and Superior Temporal Cortices Predict Spatial Working Memory Performance},
      journal = {Scientific Reports},
         year = {2018},
       volume = {8},
        issue = {1},
        pages = {8488},
     abstract = {The oscillatory dynamics serving spatial working memory (SWM), and how such dynamics relate to performance, are poorly understood. To address these topics, the present study recruited 22 healthy adults to perform a SWM task during magnetoencephalography (MEG). The resulting MEG data were transformed into the time-frequency domain, and significant oscillatory responses were imaged using a beamformer. Voxel time series data were extracted from the cluster peaks to quantify the dynamics, while whole-brain partial correlation maps were computed to identify regions where oscillatory strength varied with accuracy on the SWM task. The results indicated transient theta oscillations in spatially distinct subregions of the prefrontal cortices at the onset of encoding and maintenance, which may underlie selection of goal-relevant information. Additionally, strong and persistent decreases in alpha and beta oscillations were observed throughout encoding and maintenance in parietal, temporal, and occipital regions, which could serve sustained attention and maintenance processes during SWM performance. The neuro-behavioral correlations revealed that beta activity within left dorsolateral prefrontal control regions and bilateral superior temporal integration regions was negatively correlated with SWM accuracy. Notably, this is the first study to employ a whole-brain approach to significantly link neural oscillations to behavioral performance in the context of SWM.},
}
@article{PruskyetalVR-00,
       author = {Glen T. Prusky and Paul W.R. West and Robert M. Douglas},
        title = {Behavioral assessment of visual acuity in mice and rats},
      journal = {Vision Research},
       volume = {40},
       number = {16},
         year = {2000},
        pages = {2201-2209},
     abstract = {We have developed a simple computer-based discrimination task that enables the quick determination of visual acuities in rodents. A grating is displayed randomly on one of two monitors at the wide end of a trapezoidal-shaped tank containing shallow water. Animals are trained to swim toward the screens, and at a fixed distance, choose the screen displaying the grating and escape to a submerged platform hidden below it. Both mice and rats learn the task quickly. Performance falls below 70\% when the spatial frequency is increased beyond 0.5 cycles in most \{C57BU6\} mice, and around 1.0 cycles per degree (cpd) in Long-Evans rats.}
}
@article{PucaketalJoCN-96,
       author = {Pucak, Michele L. and Levitt, Jonathan B. and Lund, Jennifer S. and Lewis, David A.},
        title = {Patterns of intrinsic and associational circuitry in monkey prefrontal cortex},
      journal = {Journal of Comparative Neurology},
       volume = {376},
       number = {4},
         year = {1996},
        pages = {614-630},
     abstract = {Abstract Both local and long-range connections are critical mediators of information processing in the cerebral cortex, but little is known about the relationships among these types of connections, especially in higher-order cortical regions. We used quantitative reconstructions of the label arising from discrete (approximately 350 $\mu$m diameter) injections of biotinylated dextran amine and cholera toxin B to determine the spatial organization of the axon collaterals and principal axon projections furnished by pyramidal neurons in the supragranular layers of monkey prefrontal cortex (areas 9 and 46). Both terminals and cell bodies labeled by transport along axon collaterals in the gray matter formed intrinsic clusters which were arrayed as a series of discontinuous stripes of similar size and shape. The co-registration of anterograde and retrograde transport confirmed that these convergent and divergent intrinsic connections also were reciprocal. Transport from the same injection sites along principal axons through the white matter formed associational clusters which were also arrayed as a series of discontinuous stripes. The dimensions of the anterogradely- and retrogradely-labeled associational stripes were very similar to each other and to the intrinsic stripes. These findings demonstrate that divergence, convergence, and reciprocity characterize both the intrinsic and associational excitatory connections in the prefrontal cortex. These patterns of connections provide an anatomical substrate by which activation of a discrete group of neurons would lead to the recruitment of a specific neuronal network comprised of both local and distant groups of cells. Furthermore, the consistent size of the intrinsic and associational stripes (approximately 275 by 1,800 m) suggests that they may represent basic functional units in the primate prefrontal cortex.},
}
@article{PudasainietalFiMB-15,
       author = {Pudasaini, Ashutosh and El-Arab, Kaley K. and Zoltowski, Brian D.},
        title = {{LOV}-based optogenetic devices: light-driven modules to impart photoregulated control of cellular signaling},
      journal = {Frontiers in Molecular Bioscience},
    publisher = {Frontiers Media},
       volume = {2},
         year = {2015},
        pages = {18},
     abstract = {The Light-Oxygen-Voltage domain family of proteins is widespread in biology where they impart sensory responses to signal transduction domains. The small, light responsive LOV modules offer a novel platform for the construction of optogenetic tools. Currently, the design and implementation of these devices is partially hindered by a lack of understanding of how light drives allosteric changes in protein conformation to activate diverse signal transduction domains. Further, divergent photocycle properties amongst LOV family members complicate construction of highly sensitive devices with fast on/off kinetics. In the present review we discuss the history of LOV domain research with primary emphasis on tuning LOV domain chemistry and signal transduction to allow for improved optogenetic tools.},
}
@article{PulvermullerPiN-18,
       author = {Friedemann Pulverm{\"{u}}ller},
        title = {Neural reuse of action perception circuits for language, concepts and communication},
      journal = {Progress in Neurobiology},
       volume = {160},
         year = {2018},
        pages = {1-44},
     abstract = {Neurocognitive and neurolinguistics theories make explicit statements relating specialized cognitive and linguistic processes to specific brain loci. These linking hypotheses are in need of neurobiological justification and explanation. Recent mathematical models of human language mechanisms constrained by fundamental neuroscience principles and established knowledge about comparative neuroanatomy offer explanations for where, when and how language is processed in the human brain. In these models, network structure and connectivity along with action- and perception-induced correlation of neuronal activity co-determine neurocognitive mechanisms. Language learning leads to the formation of action perception circuits (APCs) with specific distributions across cortical areas. Cognitive and linguistic processes such as speech production, comprehension, verbal working memory and prediction are modeled by activity dynamics in these APCs, and combinatorial and communicative-interactive knowledge is organized in the dynamics within, and connections between APCs. The network models and, in particular, the concept of distributionally-specific circuits, can account for some previously not well understood facts about the cortical 'hubs' for semantic processing and the motor system's role in language understanding and speech sound recognition. A review of experimental data evaluates predictions of the APC model and alternative theories, also providing detailed discussion of some seemingly contradictory findings. Throughout, recent disputes about the role of mirror neurons and grounded cognition in language and communication are assessed critically.}
}
@article{PulvermullerTiCS-13,
        title = {How neurons make meaning: brain mechanisms for embodied and abstract-symbolic semantics},
       author = {Friedemann Pulverm{\"{u}}ller},
      journal = {Trends in Cognitive Sciences},
       volume = {17},
       number = {9},
         year = {2013},
        pages = {458-470},
     abstract = {How brain structures and neuronal circuits mechanistically underpin symbolic meaning has recently been elucidated by neuroimaging, neuropsychological, and neurocomputational research. Modality-specific 'embodied' mechanisms anchored in sensorimotor systems appear to be relevant, as are 'disembodied' mechanisms in multimodal areas. In this paper, four semantic mechanisms are proposed and spelt out at the level of neuronal circuits: referential semantics, which establishes links between symbols and the objects and actions they are used to speak about; combinatorial semantics, which enables the learning of symbolic meaning from context; emotional-affective semantics, which establishes links between signs and internal states of the body; and abstraction mechanisms for generalizing over a range of instances of semantic meaning. Referential, combinatorial, emotional-affective, and abstract semantics are complementary mechanisms, each necessary for processing meaning in mind and brain.}
}
@article{PulvermulleretalBC-14,
       author = {Pulverm{\"{u}}ller, F. and Garagnani, M. and Wennekers, T.},
        title = {Thinking in circuits: toward neurobiological explanation in cognitive neuroscience},
      journal = {Biological Cybernetics},
         year = {2014},
       volume = {108},
       number = {5},
        pages = {573-593},
     abstract = {Cognitive theory has decomposed human mental abilities into cognitive (sub) systems, and cognitive neuroscience succeeded in disclosing a host of relationships between cognitive systems and specific structures of the human brain. However, an explanation of why specific functions are located in specific brain loci had still been missing, along with a neurobiological model that makes concrete the neuronal circuits that carry thoughts and meaning. Brain theory, in particular the Hebb-inspired neurocybernetic proposals by Braitenberg, now offers an avenue toward explaining brain-mind relationships and to spell out cognition in terms of neuron circuits in a neuromechanistic sense. Central to this endeavor is the theoretical construct of an elementary functional neuronal unit above the level of individual neurons and below that of whole brain areas and systems: the distributed neuronal assembly (DNA) or thought circuit (TC). It is shown that DNA/TC theory of cognition offers an integrated explanatory perspective on brain mechanisms of perception, action, language, attention, memory, decision and conceptual thought. We argue that DNAs carry all of these functions and that their inner structure (e.g., core and halo subcomponents), and their functional activation dynamics (e.g., ignition and reverberation processes) answer crucial localist questions, such as why memory and decisions draw on prefrontal areas although memory formation is normally driven by information in the senses and in the motor system. We suggest that the ability of building DNAs/TCs spread out over different cortical areas is the key mechanism for a range of specifically human sensorimotor, linguistic and conceptual capacities and that the cell assembly mechanism of overlap reduction is crucial for differentiating a vocabulary of actions, symbols and concepts.}
}
@article{PunjanietalNATURE-METHODS-17,
       author = {Punjani, Ali and Rubinstein, John L. and and Fleet, David J.},
        title = {{cryoSPARC}: algorithms for rapid unsupervised cryo-{EM} structure determination},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = {Advanced Publication},
         year = {2017},
     abstract = {Single-particle electron cryomicroscopy (cryo-EM) is a powerful method for determining the structures of biological macromolecules. With automated microscopes, cryo-EM data can often be obtained in a few days. However, processing cryo-EM image data to reveal heterogeneity in the protein structure and to refine 3D maps to high resolution frequently becomes a severe bottleneck, requiring expert intervention, prior structural knowledge, and weeks of calculations on expensive computer clusters. Here we show that stochastic gradient descent (SGD) and branch-and-bound maximum likelihood optimization algorithms permit the major steps in cryo-EM structure determination to be performed in hours or minutes on an inexpensive desktop computer. Furthermore, SGD with Bayesian marginalization allows ab initio 3D classification, enabling automated analysis and discovery of unexpected structures without bias from a reference map. These algorithms are combined in a user-friendly computer program named cryoSPARC (http://www.cryosparc.com).}
}
@incollection{Purvesetal2001,
       author = {Purves. D.},
        title = {Excitatory and Inhibitory Postsynaptic Potentials},
    booktitle = {Neuroscience. 2nd edition},
    publisher = {Sinauer Associates},
       editor = {Purves, D. and Augustine, G. J. and Fitzpatrick D. and et al},
         year = {2001},
}
@book{PurvesetalNEUROSCIENCE-17,
        title = {Neuroscience, Sixth Edition},
       author = {Dale Purves, George J. Augustine and David Fitzpatrick and William C. Hall and Anthony-Samuel LaMantia and Richard D. Mooney and Michael L. Platt and Leonard E. White},
    publisher = {Oxford University Press},
         year = {2017},
}
@article{PurwinsetalDEEP-SPEECH-19,
       author = {Hendrik Purwins and Bo Li and Tuomas Virtanen and Jan Schl{\"{u}}ter and Shuo{-}Yiin Chang and Tara N. Sainath},
        title = {Deep Learning for Audio Signal Processing},
      journal = {CoRR},
       volume = {arXiv:1905.00078},
         year = {2019},
     abstract = {Given the recent surge in developments of deep learning, this article provides a review of the state-of-the-art deep learning techniques for audio signal processing. Speech, music, and environmental sound processing are considered side-by-side, in order to point out similarities and differences between the domains, highlighting general methods, problems, key references, and potential for cross-fertilization between areas. The dominant feature representations (in particular, log-mel spectra and raw waveform) and deep learning models are reviewed, including convolutional neural networks, variants of the long short-term memory architecture, as well as more audio-specific neural network models. Subsequently, prominent deep learning application areas are covered, i.e. audio recognition (automatic speech recognition, music information retrieval, environmental sound detection, localization and tracking) and synthesis and transformation (source separation, audio enhancement, generative models for speech, sound, and music synthesis). Finally, key issues and future questions regarding deep learning applied to audio signal processing are identified.},
}
@incollection{PuscheletalSPIRAL-11,
        title = {Spiral},
       author = {Markus P\"{u}schel and Franz Franchetti and Yevgen Voronenko},
    booktitle = {Encyclopedia of Parallel Computing},
       editor = {David Padua},
    publisher = {Springer},
      address = {Berlin},
         year = 2011,
}
@book{Puterman94,
       author = {Martin L. Puterman},
        title = {{M}arkov Decision Processes},
    publisher = {John Wiley \& Sons},
      address = {New York},
         year = 1994,
}
@article{PutermanMS-78,
       author = {Martin L. Puterman and Shin, Moon Chirl},
        title = {Modified Policy Iteration algorithms for Discounted {M}arkov Decision Processes},
      journal = {Management Science},
       volume = 24,
        pages = {1127-1137},
         year = 1978,
}
@article{PyersandSenghasPSYCHOLOGICAL-SCIENCE-09,
       author = {Pyers, Jennie E. and Senghas, Ann},
        title = {Language Promotes False-Belief Understanding: Evidence From Learners of a New Sign Language},
      journal = {Psychological Science},
       volume = {20},
        issue = {7},
         year = {2009},
        pages = {805-812},
     abstract = {Developmental studies have identified a strong correlation in the timing of language development and false-belief understanding. However, the nature of this relationship remains unresolved. Does language promote false-belief understanding, or does it merely facilitate development that could occur independently, albeit on a delayed timescale? We examined language development and false-belief understanding in deaf learners of an emerging sign language in Nicaragua. The use of mental-state vocabulary and performance on a low-verbal false-belief task were assessed, over 2 years, in adult and adolescent users of Nicaraguan Sign Language. Results show that those adults who acquired a nascent form of the language during childhood produce few mental-state signs and fail to exhibit false-belief understanding. Furthermore, those whose language developed over the period of the study correspondingly developed in false-belief understanding. Thus, language learning, over and above social experience, drives the development of a mature theory of mind.},
}
@incollection{PynadathetalCAGI-14,
       author = {Pynadath, David V. and Rosenbloom, Paul S. and Marsella, Stacy C.},
       editor = {Goertzel, Ben and Orseau, Laurent and Snaider, Javier},
        title = {Reinforcement Learning for Adaptive Theory of Mind in the Sigma Cognitive Architecture},
    booktitle = {Artificial General Intelligence: 7th International Conference, AGI 2014, Quebec City, QC, Canada, August 1-4, 2014. Proceedings},
    publisher = {Springer International Publishing},
         year = {2014},
        pages = {143-154},
     abstract = {One of the most common applications of human intelligence is social interaction, where people must make effective decisions despite uncertainty about the potential behavior of others around them. Reinforcement learning (RL) provides one method for agents to acquire knowledge about such interactions. We investigate different methods of multiagent reinforcement learning within the Sigma cognitive architecture. We leverage Sigma's architectural mechanism for gradient descent to realize four different approaches to multiagent learning: (1) with no explicit model of the other agent, (2) with a model of the other agent as following an unknown stationary policy, (3) with prior knowledge of the other agent's possible reward functions, and (4) through inverse reinforcement learning (IRL) of the other agent's reward function. While the first three variations re-create existing approaches from the literature, the fourth represents a novel combination of RL and IRL for social decision-making. We show how all four styles of adaptive Theory of Mind are realized through Sigma's same gradient descent algorithm, and we illustrate their behavior within an abstract negotiation task.},
}
@book{QR-85,
       editor = {Bobrow, Daniel G.},
        title = {Qualitative Reasoning and Physical Systems},
    booktitle = {Qualitative Reasoning and Physical Systems},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1985,
}
@article{QuadratoetalNATURE-17,
       author = {Quadrato, Giorgia and Nguyen, Tuan and Macosko, Evan Z. and Sherwood, John L. and Min Yang, Sung and Berger, Daniel R. and Maria, Natalie and Scholvin, Jorg and Goldman, Melissa and Kinney, Justin P. and Boyden, Edward S. and Lichtman, Jeff W. and Williams, Ziv M. and McCarroll, Steven A. and Arlotta, Paola},
        title = {Cell diversity and network dynamics in photosensitive human brain organoids},
      journal = {Nature},
         year = {2017},
    publisher = {Macmillan Publishers Limited},
       volume = {545},
        issue = {7652},
        pages = {48-53},
     abstract = {In vitro models of the developing brain such as three-dimensional brain organoids offer an unprecedented opportunity to study aspects of human brain development and disease. However, the cells generated within organoids and the extent to which they recapitulate the regional complexity, cellular diversity and circuit functionality of the brain remain undefined. Here we analyse gene expression in over 80,000 individual cells isolated from 31 human brain organoids. We find that organoids can generate a broad diversity of cells, which are related to endogenous classes, including cells from the cerebral cortex and the retina. Organoids could be developed over extended periods (more than 9 months), allowing for the establishment of relatively mature features, including the formation of dendritic spines and spontaneously active neuronal networks. Finally, neuronal activity within organoids could be controlled using light stimulation of photosensitive cells, which may offer a way to probe the functionality of human neuronal circuits using physiological sensory stimuli.},
}
@book{Quammen2018,
        title = {The Tangled Tree: A Radical New History of Life},
       author = {Quammen, David},
         year = {2018},
    publisher = {Simon \& Schuster}
}
@article{QuammenHARPERS-98,
       author = {David Quammen},
        title = {Planet of Weeds},
      journal = {Harper's},
       volume = {October},
         year = {1998},
}
@book{Quigley01,
      author = {Ellie Quigley},
       title = {Unix Shells by Example},
   publisher = {Third edition, Prentice Hall},
     address = {Upper Saddle River, NJ},
        year = 2001,
}
@inproceedings{QuinlanIJCAI-69,
       author = {Quinlan, J. R.},
        title = {A task-independent experience-gathering scheme for a problem solver},
    booktitle = {Proceedings {IJCAI} 1},
     location = {Washington, DC, USA},
 organization = {IJCAII},
         year = 1969,
        pages = {193-197},
}
@article{QuinlanML-86,
       author = {Quinlan, J. R.},
        title = {Induction of Decision Trees},
      journal = {Machine Learning},
       volume = 1,
        issue = 1,
         year = 1986,
        pages = {81-106},
}
@inproceedings{QuirketalACL-04,
       author = {Chris Quirk and Chris Brockett and William Dolan},
        title = {Monolingual machine translation for paraphrase generation},
    booktitle = {Proceedings of the 2004 Conference on Empirical Methods in Natural Language Processing},
    publisher = {Association for Computational Linguistics},
         year = 2004,
}
@article{QuirogaNATURE-12,
       author = {Quiroga, Rodrigo Quian},
        title = {Concept cells: the building blocks of declarative memory functions},
      journal = {Nature Reviews Neuroscience},
         year = {2012},
       volume = {13},
        issue = {8},
        pages = {587-597},
     abstract = {Neurons in the human medial temporal lobe respond in a selective and abstract manner to particular persons or objects. Rodrigo Quian Quiroga argues that these 'concept cells' are crucial for memory functions and the transition between related concepts that leads to the flow of consciousness.},
}
@article{QuirogaetalNATURE-05,
       author = {Quiroga, R. Quian and Reddy, L. and Kreiman, G. and Koch, C. and Fried, I.},
        title = {Invariant visual representation by single neurons in the human brain},
      journal = {Nature},
       volume = 435,
         year = 2005,
        pages = {1102-1107},
}
@inproceedings{QureshietalICLR-20,
        title = {Composing Task-Agnostic Policies with Deep Reinforcement Learning},
       author = {Ahmed H. Qureshi and Jacob J. Johnson and Yuzhe Qin and Taylor Henderson and Byron Boots and Michael C. Yip},
    booktitle = {International Conference on Learning Representations},
         year = {2020},
     abstract = {The composition of elementary behaviors to solve challenging transfer learning problems is one of the key elements in building intelligent machines. To date, there has been plenty of work on learning task-specific policies or skills but almost no focus on composing necessary, task-agnostic skills to find a solution to new problems. In this paper, we propose a novel deep reinforcement learning-based skill transfer and composition method that takes the agent's primitive policies to solve unseen tasks. We evaluate our method in difficult cases where training policy through standard reinforcement learning (RL) or even hierarchical RL is either not feasible or exhibits high sample complexity. We show that our method not only transfers skills to new problem settings but also solves the challenging environ- ments requiring both task planning and motion control with high data efficiency.}
}
@book{Rabaeyetal1999,
       author = {Jan M. Rabaey and Anantha Chandrakasan and Borivoje Nikolic},
        title = {Digital Integrated Circuits - First Edition},
    publisher = {Prentice Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1999,
}
@book{Rabaeyetal2002,
       author = {Jan M. Rabaey and Anantha Chandrakasan and Borivoje Nikolic},
        title = {Digital Integrated Circuits - Second Edition},
    publisher = {Prentice Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 2003,
}
@book{Rabenstein75,
       author = {Rabenstein, Albert L.},
        title = {Elementary Differential Equations with Linear Algebra},
    publisher = {Academic Press},
      address = {New York},
         year = 1975,
}
@article{RabinerIEEE-89,
       author = {L. R. Rabiner},
        title = {A tutorial on hidden {M}arkov models and selected applications in speech recognition},
      journal = {Proceedings of the IEEE},
       volume = 77,
        issue = 2,
        pages = {257-286},
         year = 1989,
}
@article{RabinerandJuangASSP-86,
       author = {Rabiner, L. R. and Juang, B. H.},
        title = {An introduction to hidden {M}arkov models},
      journal = {{IEEE} {ASSP} Magazine},
        pages = {4-15},
         year = 1986,
}
@techreport{Rabinovichetal07,
       author = {Andrew Rabinovich and Andrea Vedaldi and Serge Belongie},
        title = {Does Image Segmentation Improve Object Categorization?},
        issue = {Technical Report CS2007-0908},
  institution = {University of California, San Diego},
         year = 2007,
}
@article{RabinowitzetalCoRR-18,
       author = {Neil C. Rabinowitz and Frank Perbet and H. Francis Song and Chiyuan Zhang and S.M. Ali Eslami and Matthew Botvinick},
        title = {Machine Theory of Mind},
      journal = {CoRR},
       volume = {arXiv:1802.07740},
         year = {2018},
     abstract = {Theory of mind (ToM; Premack and Woodruff, 1978) broadly refers to humans' ability to represent the mental states of others, including their desires, beliefs, and intentions. We propose to train a machine to build such models too. We design a Theory of Mind neural network -- a ToMnet -- which uses meta-learning to build models of the agents it encounters, from observations of their behaviour alone. Through this process, it acquires a strong prior model for agents' behaviour, as well as the ability to bootstrap to richer predictions about agents' characteristics and mental states using only a small number of behavioural observations. We apply the ToMnet to agents behaving in simple gridworld environments, showing that it learns to model random, algorithmic, and deep reinforcement learning agents from varied populations, and that it passes classic ToM tasks such as the "Sally-Anne" test (Wimmer and Perner, 1983; Baron-Cohen et al., 1985) of recognising that others can hold false beliefs about the world. We argue that this system -- which autonomously learns how to model other agents in its world -- is an important step forward for developing multi-agent AI systems, for building intermediating technology for machine-human interaction, and for advancing the progress on interpretable AI.}
}
@article{RachmuthetalPNAS-11,
       author = {Rachmuth, Guy and Shouval, Harel Z. and Bear, Mark F. and Poon, Chi-Sang},
        title = {A biophysically-based neuromorphic model of spike rate- and timing-dependent plasticity},
      journal = {Proceedings of the National Academy of Sciences},
         year = 2011,
     abstract = {Current advances in neuromorphic engineering have made it possible to emulate complex neuronal ion channel and intracellular ionic dynamics in real time using highly compact and power-efficient complementary metal-oxide-semiconductor (CMOS) analog very-large-scale-integrated circuit technology. Recently, there has been growing interest in the neuromorphic emulation of the spike-timing-dependent plasticity (STDP) Hebbian learning rule by phenomenological modeling using CMOS, memristor or other analog devices. Here, we propose a CMOS circuit implementation of a biophysically grounded neuromorphic (iono-neuromorphic) model of synaptic plasticity that is capable of capturing both the spike rate-dependent plasticity (SRDP, of the Bienenstock-Cooper-Munro or BCM type) and STDP rules. The iono-neuromorphic model reproduces bidirectional synaptic changes with NMDA receptor-dependent and intracellular calcium-mediated long-term potentiation or long-term depression assuming retrograde endocannabinoid signaling as a second coincidence detector. Changes in excitatory or inhibitory synaptic weights are registered and stored in a nonvolatile and compact digital format analogous to the discrete insertion and removal of AMPA or GABA receptor channels. The versatile Hebbian synapse device is applicable to a variety of neuroprosthesis, brain-machine interface, neurorobotics, neuromimetic computation, machine learning, and neural-inspired adaptive control problems.},
}
@article{RadfordetalOPENAI-19,
        title = {Language Models are Unsupervised Multitask Learners},
       author = {Radford, Alec and Wu, Jeff and Child, Rewon and Luan, David and Amodei, Dario and Sutskever, Ilya},
         year = {2019},
      comment = {https://openai.com/blog/better-language-models/},
     abstract = {Natural language processing tasks, such as question answering, machine translation, reading comprehension, and summarization, are typically approached with supervised learning on taskspecific datasets. We demonstrate that language models begin to learn these tasks without any explicit supervision when trained on a new dataset of millions of webpages called WebText. When conditioned on a document plus questions, the answers generated by the language model reach 55 F1 on the CoQA dataset - matching or exceeding the performance of 3 out of 4 baseline systems without using the 127,000+ training examples. The capacity of the language model is essential to the success of zero-shot task transfer and increasing it improves performance in a log-linear fashion across tasks. Our largest model, GPT-2, is a 1.5B parameter Transformer that achieves state of the art results on 7 out of 8 tested language modeling datasets in a zero-shot setting but still underfits WebText. Samples from the model reflect these improvements and contain coherent paragraphs of text. These findings suggest a promising path towards building language processing systems which learn to perform tasks from their naturally occurring demonstrations.}
}
@inproceedings{RadulescuetalCMSB-15,
       author = {Ovidiu Radulescu and Satya Swarup Samal and Dima Grigoriev and Andreas Weber},
        title = {Symbolic dynamics of biochemical pathways as finite states machines},
    booktitle = {Computational Methods in Systems Biology: 13th International Conference, CMSB 2015,},
       editor = {Roux, Olivier and Bourdon, J\'{e}r\'{e}mie},
    publisher = {Springer International Publishing},
         year = {2015},
        pages = {104-120},
     abstract = {We discuss the symbolic dynamics of biochemical networks with separate timescales. We show that symbolic dynamics of monomolecular reaction net- works with separated rate constants can be described by deterministic, acyclic automata with a number of states that is inferior to the number of biochemical species. For nonlinear pathways, we propose a general approach to approximate their dynamics by finite state machines working on the metastable states of the network (long life states where the system has slow dynamics). For net- works with polynomial rate functions we propose to compute metastable states as solutions of the tropical equilibration problem. Tropical equilibrations are defined by the equality of at least two dominant monomials of opposite signs in the differential equations of each dynamic variable. In algebraic geometry, tropical equilibrations are tantamount to tropical prevarieties, that are finite intersections of tropical hypersurfaces.}
}
@article{RaghavanCSS-88,
       author = {Raghavan, Prabhakar},
        title = {Probabilistic Construction of Deterministic Algprithms: Approximating Packing Integer Programs},
      journal = {Computer and Systems Sciences},
       volume = 37,
         year = 1988,
        pages = {130-143},
}
@article{RaghavanandThompsonCOMBINATORICA-87,
       author = {Raghavan, Prabhakar and Thompson, Clark},
        title = {Randomized Rounding: A Technique for Provably Good Algorithms and Algorithmic Proofs},
      journal = {Combinatorica},
       volume = 7,
        issue = 4,
         year = 1987,
        pages = {365-374},
}
@inproceedings{RahtuetalICCV-11,
       author = {Rahtu, E. and Kannala, J. and Blaschko, M. B.},
        title = {Learning a Category Independent Object Detection Cascade},
    booktitle = {{IEEE} International Conference on Computer Vision},
    publisher = {IEEE},
         year = 2011,
        pages = {1052-1059},
 abststract   = {Cascades are a popular framework to speed up object detection systems. Here we focus on the first layers of a category independent object detection cascade in which we sample a large number of windows from an objectness prior, and then discriminatively learn to filter these candidate windows by an order of magnitude. We make a number of contributions to cascade design that substantially improve over the state of the art: (i) our novel objectness prior gives much higher recall than competing methods, (ii) we propose objectness features that give high performance with very low computational cost, and (iii) we make use of a structured output ranking approach to learn highly effective, but inexpensive linear feature combinations by directly optimizing cascade performance. Thorough evaluation on the PASCAL VOC data set shows consistent improvement over the current state of the art, and over alternative discriminative learning strategies.}
}
@book{Raibert86,
       author = {Raibert, Marc H.},
        title = {Legged Robots That Balance},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1986,
}
@techreport{RaibertTR-81,
       author = {Raibert, M. H. and B. Brown, Jr. and M. Chepponis and E. Hastings and S. E. Shreve and F. C. Wimberly},
        title = {Dynamically Stable Legged Locomotion},
        issue = {CMU-RI-TR-81-9},
  institution = {Carnegie-Mellon University Robotics Institute},
         year = 1981,
}
@article{RaichleandSnyderNEUROIMAGE-07,
        title = {A default mode of brain function: A brief history of an evolving idea},
       author = {Marcus E. Raichle and Abraham Z. Snyder},
      journal = {NeuroImage},
       volume = 37,
        issue = 4,
        pages = {1083-1090},
         year = 2007,
     abstract = {The concept of a default mode of brain function arose out of a focused need to explain the appearance of activity decreases in functional neuroimaging data when the control state was passive visual fixation or eyes closed resting. The problem was particularly compelling because these activity decreases were remarkably consistent across a wide variety of task conditions. Using PET, we determined that these activity decreases did not arise from activations in the resting state. Hence, their presence implied the existence of a default mode. While the unique constellation of brain areas provoking this analysis has come to be known as the default system, all areas of the brain have a high level of organized default functional activity. Most critically, this work has called attention to the importance of intrinsic functional activity in assessing brain behavior relationships.}
}
@book{Raiffa68,
       author = {Raiffa, Howard},
        title = {Decision Analysis: Introductory Lectures on Choices Under Uncertainty},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1968,
}
@book{RaiffaandSchlaifer61,
       author = {Raiffa, Howard and Schlaifer, R.},
        title = {Applied Statistical Decision Theory},
    publisher = {Harvard University Press},
         year = 1961,
}
@article{RaileanuetalCoRR-18,
       author = {Roberta Raileanu, Emily Denton, Arthur Szlam, Rob Fergus},
        title = {Modeling Others using Oneself in Multi-Agent Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1802.09640},
         year = {2018},
     abstract = {We consider the multi-agent reinforcement learning setting with imperfect information in which each agent is trying to maximize its own utility. The reward function depends on the hidden state (or goal) of both agents, so the agents must infer the other players' hidden goals from their observed behavior in order to solve the tasks. We propose a new approach for learning in these domains: Self Other-Modeling (SOM), in which an agent uses its own policy to predict the other agent's actions and update its belief of their hidden state in an online manner. We evaluate this approach on three different tasks and show that the agents are able to learn better policies using their estimate of the other players' hidden states, in both cooperative and adversarial settings.}
}
@inproceedings{RainaetalICML-07,
       author = {Rajat Raina and Alexis Battle and Honglak Lee and Benjamin Packer and Andrew Ng},
        title = {Self-taught learning: transfer learning from unlabeled data},
    booktitle = {Proceedings of the 24th Annual International Conference on Machine Learning},
     location = {Corvallis, OR},
       editor = {Zoubin Ghahramani},
    publisher = {Omnipress},
         year = 2007,
        pages = {759-766}
}
@inproceedings{RainaetalICML-09,
       author = {Rajat Raina and Anand Madhavan and Andrew Ng},
        title = {Large-scale Deep Unsupervised Learning using Graphics Processors},
    booktitle = {Proceedings of the 25th Annual International Conference on Machine Learning},
         year = 2009,
}
@techreport{Rainer_FriedrichREPORT-14,
        title = {Neuronal circuits and computations},
       author = {Rainer Friedrich},
         year = 2014,
  institution = {Friedrich Miescher Institute for Biomedical Research},
}
@article{RainsetalBGER-10,
    author = {Rains G and Kulasiri D and Zhou Z and Samarasinghe S and Tomberlin J and Olson D.},
     title = {Synthesizing neurophysiology, genetics, behaviour and learning to produce whole-insect programmable sensors to detect volatile chemicals},
   journal = {Biotechnology \& Genetic Engineering Reviews},
    volume = {26},
      year = {2010},
      pages = {179-204},
      abstract = {Insects have extremely sensitive systems of olfaction. These systems have been explored as potential sensors for odourants associated with forensics, medicine, security, and agriculture application. Most sensors based on insect olfaction utilize associative learning to "program" the insects to exhibit some form of behavioural response to a target odourant. To move to the next stage of development with whole-insect programmable sensors, an examination of how odourants are captured, processed and used to create behaviour is necessary. This review article examines how the neurophysiological, molecular, genetic and behavioural system of olfaction works and how an understanding of these systems should lead the way to future developments in whole-insect programmable sensors.}
}
@article{RaisetalNATURE-13,
       author = {Rais, Yoach and Zviran, Asaf and Geula, Shay and Gafni, Ohad and Chomsky, Elad and Viukov, Sergey and Mansour, Abed AlFatah and Caspi, Inbal and Krupalnik, Vladislav and Zerbib, Mirie and Maza, Itay and Mor, Nofar and Baran, Dror and Weinberger, Leehee and Jaitin, Diego A. and Lara-Astiaso, David and Blecher-Gonen, Ronnie and Shipony, Zohar and Mukamel, Zohar and Hagai, Tzachi and Gilad, Shlomit and Amann-Zalcenstein, Daniela and Tanay, Amos and Amit, Ido and Novershtern, Noa and Hanna, Jacob H.},
        title = {Deterministic direct reprogramming of somatic cells to pluripotency},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = 502,
        issue = 7469,
         year = 2013,
        pages = {65-70},
     abstract = {Somatic cells can be inefficiently and stochastically reprogrammed into induced pluripotent stem (iPS) cells by exogenous expression of Oct4 (also called Pou5f1), Sox2, Klf4 and Myc (hereafter referred to as OSKM). The nature of the predominant rate-limiting barrier(s) preventing the majority of cells to successfully and synchronously reprogram remains to be defined. Here we show that depleting Mbd3, a core member of the Mbd3/NuRD (nucleosome remodelling and deacetylation) repressor complex, together with OSKM transduction and reprogramming in naive pluripotency promoting conditions, result in deterministic and synchronized iPS cell reprogramming (near 100\% efficiency within seven days from mouse and human cells). Our findings uncover a dichotomous molecular function for the reprogramming factors, serving to reactivate endogenous pluripotency networks while simultaneously directly recruiting the Mbd3/NuRD repressor complex that potently restrains the reactivation of OSKM downstream target genes. Subsequently, the latter interactions, which are largely depleted during early pre-implantation development in vivo, lead to a stochastic and protracted reprogramming trajectory towards pluripotency in vitro. The deterministic reprogramming approach devised here offers a novel platform for the dissection of molecular dynamics leading to establishing pluripotency at unprecedented flexibility and resolution.},
}
@article {RajalinghamJoN-15,
       author = {Rajalingham, Rishi and Schmidt, Kailyn and James J. {DiCarlo}},
        title = {Comparison of Object Recognition Behavior in Human and Monkey.},
      journal = {The Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {35},
         year = {2015},
        pages = {12127-12136},
     abstract = {Although the rhesus monkey is used widely as an animal model of human visual processing, it is not known whether invariant visual object recognition behavior is quantitatively comparable across monkeys and humans. To address this question, we systematically compared the core object recognition behavior of two monkeys with that of human subjects. To test true object recognition behavior (rather than image matching), we generated several thousand naturalistic synthetic images of 24 basic-level objects with high variation in viewing parameters and image background. Monkeys were trained to perform binary object recognition tasks on a match-to-sample paradigm. Data from 605 human subjects performing the same tasks on Mechanical Turk were aggregated to characterize "pooled human" object recognition behavior, as well as 33 separate Mechanical Turk subjects to characterize individual human subject behavior. Our results show that monkeys learn each new object in a few days, after which they not only match mean human performance but show a pattern of object confusion that is highly correlated with pooled human confusion patterns and is statistically indistinguishable from individual human subjects. Importantly, this shared human and monkey pattern of 3D object confusion is not shared with low-level visual representations (pixels, V1+; models of the retina and primary visual cortex) but is shared with a state-of-the-art computer vision feature representation. Together, these results are consistent with the hypothesis that rhesus monkeys and humans share a common neural shape representation that directly supports object perception.},
}
@article {RajalinghametalJoN-18,
      journal = {Journal of Neuroscience},
       author = {Rajalingham, Rishi and Issa, Elias B. and Bashivan, Pouya and Kar, Kohitij and Schmidt, Kailyn and DiCarlo, James J.},
        title = {Large-Scale, High-Resolution Comparison of the Core Visual Object Recognition Behavior of Humans, Monkeys, and State-of-the-Art Deep Artificial Neural Networks},
    publisher = {Society for Neuroscience},
       volume = {38},
       number = {33},
         year = {2018},
        pages = {7255-7269},
     abstract = {Primates, including humans, can typically recognize objects in visual images at a glance despite naturally occurring identity-preserving image transformations (e.g., changes in viewpoint). A primary neuroscience goal is to uncover neuron-level mechanistic models that quantitatively explain this behavior by predicting primate performance for each and every image. Here, we applied this stringent behavioral prediction test to the leading mechanistic models of primate vision (specifically, deep, convolutional, artificial neural networks; ANNs) by directly comparing their behavioral signatures against those of humans and rhesus macaque monkeys. Using high-throughput data collection systems for human and monkey psychophysics, we collected more than one million behavioral trials from 1472 anonymous humans and five male macaque monkeys for 2400 images over 276 binary object discrimination tasks. Consistent with previous work, we observed that state-of-the-art deep, feedforward convolutional ANNs trained for visual categorization (termed DCNNIC models) accurately predicted primate patterns of object-level confusion. However, when we examined behavioral performance for individual images within each object discrimination task, we found that all tested DCNNIC models were significantly nonpredictive of primate performance and that this prediction failure was not accounted for by simple image attributes nor rescued by simple model modifications. These results show that current DCNNIC models cannot account for the image-level behavioral patterns of primates and that new ANN models are needed to more precisely capture the neural mechanisms underlying primate object vision. To this end, large-scale, high-resolution primate behavioral benchmarks such as those obtained here could serve as direct guides for discovering such models.SIGNIFICANCE STATEMENT Recently, specific feedforward deep convolutional artificial neural networks (ANNs) models have dramatically advanced our quantitative understanding of the neural mechanisms underlying primate core object recognition. In this work, we tested the limits of those ANNs by systematically comparing the behavioral responses of these models with the behavioral responses of humans and monkeys at the resolution of individual images. Using these high-resolution metrics, we found that all tested ANN models significantly diverged from primate behavior. Going forward, these high-resolution, large-scale primate behavioral benchmarks could serve as direct guides for discovering better ANN models of the primate visual system.},
}
@book{Rajaniemi2016,
        title = {Invisible Planets},
       author = {Rajaniemi, H.},
         year = {2016},
    publisher = {Orion Publishing Group}
}
@book{RajaramanandUllman2011,
       author = {Rajaraman, Anand and Ullman, Jeffrey D.},
        title = {{Mining of Massive Datasets}},
         year = 2011,
    publisher = {Cambridge University Press},
}
@article {RakicSCIENCE-04,
       author = {Rakic, Pasko},
        title = {Genetic Control of Cortical Convolutions},
       volume = {303},
       number = {5666},
        pages = {1983-1984},
         year = {2004},
    publisher = {American Association for the Advancement of Science},
      journal = {Science},
} 
@article {RakicSCIENCE-88,
       author = {Rakic, P},
        title = {Specification of cerebral cortical areas},
      journal = {Science},
       volume = {241},
       number = {4862},
        pages = {170-176},
         year = {1988},
    publisher = {American Association for the Advancement of Science},
     abstract = {How the immense population of neurons that constitute the human cerebral neocortex is generated from progenitors lining the cerebral ventricle and then distributed to appropriate layers of distinctive cytoarchitectonic areas can be explained by the radial unit hypothesis. According to this hypothesis, the ependymal layer of the embryonic cerebral ventricle consists of proliferative units that provide a proto-map of prospective cytoarchitectonic areas. The output of the proliferative units is translated via glial guides to the expanding cortex in the form of ontogenetic columns, whose final number for each area can be modified through interaction with afferent input. Data obtained through various advanced neurobiological techniques, including electron microscopy, immunocytochemistry, [3H]thymidine and receptor autoradiography, retrovirus gene transfer, neural transplants, and surgical or genetic manipulation of cortical development, furnish new details about the kinetics of cell proliferation, their lineage relationships, and phenotypic expression that favor this hypothesis. The radial unit model provides a framework for understanding cerebral evolution, epigenetic regulation of the parcellation of cytoarchitectonic areas, and insight into the pathogenesis of certain cortical disorders in humans.},
}
@incollection{RakiketalTCN-09,
       author = {Pasko Rakik and Jon Arellano and Joshua Breunig},
        title = {Development of the primate cerebral cortex},
    booktitle = {The Cognitive Neurosciences, 4th Edition},
    publisher = {The MIT Press},
      address = {Cambridge, MA},
       editor = {Gazzaniga, Michael S.},
         year = {2009},
        pages = {7-28},
}
@article{RallandShepherdJM-68,
       author = {Rall, W. and Shepherd, G. M.},
        title = {Theoretical reconstruction of field potentials and dendrodendritic synaptic interactions in olfactory bulb},
      journal = {Journal of Neurophysiology},
         year = {1968},
       volume = {31},
       number = {6},
        pages = {884-915},
}
@book{Ralston78,
       author = {Ralston, A. and Rabinowitz, P.},
        title = {A First Course in Numerical Analysis},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1978,
}
@book{Ramachandran2011,
     author = {V.S. Ramachandran},
      title = {The Tell-Tale Brain: A Neuroscientist's Quest for What Makes Us Human},
  publisher = {W. W. Norton},
    address = {New York, NY},
       year = 2011,
}
@inproceedings{RamachandranandAmirIJCAI-07,
       author = {Ramachandran, Deepak and Amir, Eyal},
        title = {Bayesian Inverse Reinforcement Learning},
    booktitle = {Proceedings of the 20th International Joint Conference on Artifical Intelligence},
     location = {Hyderabad, India},
    publisher = {Morgan Kaufmann Publishers Inc.},
      address = {San Francisco, CA, USA},
         year = {2007},
        pages = {2586-2591},
     abstract = {Inverse Reinforcement Learning (IRL) is the problem of learning the reward function underlying a Markov Decision Process given the dynamics of the system and the behaviour of an expert. IRL is motivated by situations where knowledge of the rewards is a goal by itself (as in preference elicitation) and by the task of apprenticeship learning (learning policies from an expert). In this paper we show how to combine prior knowledge and evidence from the expert's actions to derive a probability distribution over the space of reward functions. We present efficient algorithms that find solutions for the reward learning and apprenticeship learning tasks that generalize well over these distributions. Experimental results show strong improvement for our methods over previous heuristic-based approaches.}
}
@inproceedings{RamachandranandLeICLR-19,
        title = {Diversity and Depth in Per-Example Routing Models},
       author = {Prajit Ramachandran and Quoc V. Le},
    booktitle = {International Conference on Learning Representations},
         year = {2019},
     abstract = {Routing models, a form of conditional computation where examples are routed through a subset of components in a larger network, have shown promising results in recent works. Surprisingly, routing models to date have lacked important properties, such as architectural diversity and large numbers of routing decisions. Both architectural diversity and routing depth can increase the representational power of a routing network. In this work, we address both of these deficiencies. We discuss the significance of architectural diversity in routing models, and explain the tradeoffs between capacity and optimization when increasing routing depth. In our experiments, we find that adding architectural diversity to routing models significantly improves performance, cutting the error rates of a strong baseline by 35\% on an Omniglot setup. However, when scaling up routing depth, we find that modern routing techniques struggle with optimization. We conclude by discussing both the positive and negative results, and suggest directions for future research.}
}
@article{RamadgeIEEE-89,
       author = {Ramadge, Peter and Wonham, Murray},
        title = {The Control of Discrete Event Systems},
      journal = {Proceedings of the IEEE},
       volume = 77,
        issue = 1,
         year = 1989,
        pages = {81-98},
}
@article{RamaekersandBosmanJoP-04,
       author = {Ramaekers, F. C. and Bosman, F. T.},
        title = {The cytoskeleton and disease},
      journal = {Journal of Pathology},
         year = 2004,
       volume = 204,
        issue = 4,
        pages = {351-354},
     abstract = {Cytoskeletal research in recent years has revolutionized cell biology and biomedicine. The cytoskeleton spans the cytoplasm and interconnects the cell nucleus with the extracellular matrix, thereby forming a structural link between molecules involved in cell communication on the one hand, and gene expression on the other. Since the cytoskeleton is involved in virtually all cellular processes, abnormalities in this essential cellular component frequently result in disease. In this introduction, the basic structure of the cytoskeleton is briefly outlined. Furthermore, the disease processes in which the cytoskeleton plays a decisive role, and which are reviewed in detail in the papers in this issue, are briefly introduced. The advances in our understanding of the cytoskeleton and its function in disease will lead to new diagnostic and therapeutic applications in the foreseeable future.}
}
@article{RamaswamyetalFiNC-15,
       author = {Ramaswamy, Srikanth and Courcol, Jean-Denis and Abdellah, Marwan and Adaszewski, Stanislaw R. and Antille, Nicolas and Arsever, Selim and Atenekeng, Guy and Bilgili, Ahmet and Brukau, Yury and Chalimourda, Athanassia and Chindemi, Giuseppe and Delalondre, Fabien and Dumusc, Raphael and Eilemann, Stefan and Gevaert, Michael Emiel and Gleeson, Padraig and Graham, Joe W. and Hernando, Juan B. and Kanari, Lida and Katkov, Yury and Keller, Daniel and King, James G. and Ranjan, Rajnish and Reimann, Michael W. and Rassert, Christian and Shi, Ying and Shillcock, Julian C. and Telefont, Martin and Van Geit, Werner and Villafranca Diaz, Jafet and Walker, Richard and Wang, Yun and Zaninetta, Stefano M. and DeFelipe, Javier and Hill, Sean L. and Muller, Jeffrey and Segev, Idan and Sch\"{u}rmann, Felix and Muller, Eilif B. and Markram, Henry},
        title = {The neocortical microcircuit collaboration portal: a resource for rat somatosensory cortex},
      journal = {Frontiers in Neural Circuits},
         year = {2015},
    publisher = {Frontiers Media S.A.},
       volume = {9},
        pages = {44},
     abstract = {We have established a multi-constraint, data-driven process to digitally reconstruct, and simulate prototypical neocortical microcircuitry, using sparse experimental data. We applied this process to reconstruct the microcircuitry of the somatosensory cortex in juvenile rat at the cellular and synaptic levels. The resulting reconstruction is broadly consistent with current knowledge about the neocortical microcircuit and provides an array of predictions on its structure and function. To engage the community in exploring, challenging, and refining the reconstruction, we have developed a collaborative, internet-accessible facility--the Neocortical Microcircuit Collaboration portal (NMC portal; https://bbp.epfl.ch/nmc-portal). The NMC portal allows users to access the experimental data used in the reconstruction process, download cellular and synaptic models, and analyze the predicted properties of the microcircuit: six layers, \\~31,000 neurons, 55 morphological types, 11 electrical types, 207 morpho-electrical types, 1941 unique synaptic connection types between neurons of specific morphological types, predicted properties for the anatomy and physiology of \\~40 million intrinsic synapses. It also provides data supporting comparison of the anatomy and physiology of the reconstructed microcircuit against results in the literature. The portal aims to catalyze consensus on the cellular and synaptic organization of neocortical microcircuitry (ion channel, neuron and synapse types and distributions, connectivity, etc.). Community feedback will contribute to refined versions of the reconstruction to be released periodically. We consider that the reconstructions and the simulations they enable represent a major step in the development of in silico neuroscience.},
}
@article{RameshetalCoRR-21,
        title = {Zero-Shot Text-to-Image Generation},
       author = {Aditya Ramesh and Mikhail Pavlov and Gabriel Goh and Scott Gray and Chelsea Voss and Alec Radford and Mark Chen and Ilya Sutskever},
         year = {2021},
       volume = {arXiv:2102.12092},
     abstract = {Text-to-image generation has traditionally focused on finding better modeling assumptions for training on a fixed dataset. These assumptions might involve complex architectures, auxiliary losses, or side information such as object part labels or segmentation masks supplied during training. We describe a simple approach for this task based on a transformer that autoregressively models the text and image tokens as a single stream of data. With sufficient data and scale, our approach is competitive with previous domain-specific models when evaluated in a zero-shot fashion.}
}
@article{RamirezetalNEUROSCIENCE-11,
       author = {Ramirez, Alexandro D. and Ahmadian, Yashar and Schumacher, Joseph and Schneider, David and Woolley, Sarah M. N. and Paninski, Liam},
        title = {Incorporating Naturalistic Correlation Structure Improves Spectrogram Reconstruction from Neuronal Activity in the Songbird Auditory Midbrain},
      journal = {Journal Neuroscience},
         year = 2011,
       volume = 31,
        issue = 10,
        pages = {3828-3842},
     abstract = {Birdsong is comprised of rich spectral and temporal organization, which might be used for vocal perception. To quantify how this structure could be used, we have reconstructed birdsong spectrograms by combining the spike trains of zebra finch auditory midbrain neurons with information about the correlations present in song. We calculated maximum a posteriori estimates of song spectrograms using a generalized linear model of neuronal responses and a series of prior distributions, each carrying different amounts of statistical information about zebra finch song. We found that spike trains from a population of mesencephalicus lateral dorsalis (MLd) neurons combined with an uncorrelated Gaussian prior can estimate the amplitude envelope of song spectrograms. The same set of responses can be combined with Gaussian priors that have correlations matched to those found across multiple zebra finch songs to yield song spectrograms similar to those presented to the animal. The fidelity of spectrogram reconstructions from MLd responses relies more heavily on prior knowledge of spectral correlations than temporal correlations. However, the best reconstructions combine MLd responses with both spectral and temporal correlations.},
}
@article{RamnaniNRN-06,
        title = {The primate cortico-cerebellar system: anatomy and function},
       author = {Ramnani, Narender},
      journal = {Nature Reviews Neuroscience},
       volume = {7},
       number = {7},
        pages = {511},
         year = {2006},
    publisher = {Nature Publishing Group}
}
@article{RamsaueretalCoRR-20,
       author = {Ramsauer, Hubert and Sch{\"{a}}fl, Bernhard and Lehner, Johannes and Seidl, Philipp and Widrich, Michael and Gruber, Lukas and Holzleitner, Markus and Pavlovi{\'{c}}, Milena and Sandve, Geir and Greiff, Victor and Kreil, David and Kopp, Michael and Klambauer, G{\"{u}}nter and Brandstetter, Johannes and Hochreiter, Sepp},
        title = {Hopfield Networks is All You Need},
         year = {2020},
      journal = {CoRR},
       volume = {arXiv:2008.02217},
     abstract = {The transformer and BERT models pushed the performance on NLP tasks to new levels via their attention mechanism. We show that this attention mechanism is the update rule of a modern Hopfield network with continuous states. This new Hopfield network can store exponentially (with the dimension) many patterns, converges with one update, and has exponentially small retrieval errors. The number of stored patterns must be traded off against convergence speed and retrieval error. The new Hopfield network has three types of energy minima (fixed points of the update): (1) global fixed point averaging over all patterns, (2) metastable states averaging over a subset of patterns, and (3) fixed points which store a single pattern. Transformers learn an attention mechanism by constructing an embedding of patterns and queries into an associative space. Transformer and BERT models operate in their first layers preferably in the global averaging regime, while they operate in higher layers in metastable states. The gradient in transformers is maximal in the regime of metastable states, is uniformly distributed when averaging globally, and vanishes when a fixed point is near a stored pattern. Based on the Hopfield network interpretation, we analyzed learning of transformer and BERT architectures. Learning starts with attention heads that average and then most of them switch to metastable states. However, the majority of heads in the first layers still averages and can be replaced by averaging operations like the Gaussian weighting that we propose. In contrast, heads in the last layers steadily learn and seem to use metastable states to collect information created in lower layers. These heads seem a promising target for improving transformers. Neural networks that integrate Hopfield networks, that are equivalent to attention heads, outperform other methods on immune repertoire classification, where the Hopfield net stores several hundreds of thousands of patterns. We provide a PyTorch implementation of a new layer called 'Hopfield' which allows to equip deep learning architectures with Hopfield networks as new memory concepts.}
}
@techreport{RandazzaMSc,
       author = {Randazza, Margaret J.},
        title = {The Feature Recognition Module of the LDP System for the Robot Huey},
         type = {{M.Sc.} Thesis},
  institution = {Brown University},
         year = 1989,
}
@article{RandlettetalNATURE-METHODS-15,
       author = {Randlett, Owen and Wee, Caroline L. and Naumann, Eva A. and Nnaemeka, Onyeka and Schoppik, David and Fitzgerald, James E. and Portugues, Ruben and Lacoste, Alix M.B. and Riegler, Clemens and Engert, Florian and Schier, Alexander F.},
        title = {Whole-brain activity mapping onto a zebrafish brain atlas},
      journal = {Nature Methods},
       volume = {12},
        issue = {11},
         year = {2015},
        pages = {1039-1046},
     abstract = {In order to localize the neural circuits involved in generating behaviors, it is necessary to assign activity onto anatomical maps of the nervous system. Using brain registration across hundreds of larval zebrafish, we have built an expandable open source atlas containing molecular labels and anatomical region definitions, the Z-Brain. Using this platform and immunohistochemical detection of phosphorylated-Extracellular signal-regulated kinase (ERK/MAPK) as a readout of neural activity, we have developed a system to create and contextualize whole brain maps of stimulus- and behavior-dependent neural activity. This MAP-Mapping (Mitogen Activated Protein kinase - Mapping) assay is technically simple, fast, inexpensive, and data analysis is completely automated. Since MAP-Mapping is performed on fish that are freely swimming, it is applicable to nearly any stimulus or behavior. We demonstrate the utility of our high-throughput approach using hunting/feeding, pharmacological, visual and noxious stimuli. The resultant maps outline hundreds of areas associated with behaviors.}
}
@inproceedings{RanzatoandHintonCVPR-10,
       author = {M. Ranzato and G. Hinton},
        title = {Deep learning with multiplicative interactions},
    booktitle = {{IEEE} Computer Society Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 2010,
}
@inproceedings{RanzatoetalCVPR-07,
       author = {M. Ranzato and F. J. Huang and Y. Boureau and Y. LeCun},
        title = {Unsupervised Learning of Invariant Feature Hierarchies with Applications to Object Recognition},
    booktitle = {Proceedings 2007 of the {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
         year = 2007,
}
@incollection{RanzatoetalNIPS-07,
       author = {M. Ranzato and Y. Boureau and Y. LeCun},
        title = {Sparse Feature Learning for Deep Belief Networks},
    booktitle = {Advances in Neural Information Processing Systems 20},
      address = {Cambridge, MA},
    publisher = {MIT Press},
         year = 2007,
}
@article{Rao82,
       author = {C. Radhakrishna Rao},
        title = {Diversity: Its measurement, decomposition, apportionment and analysis},
      journal = {Sankhya: The Indian Journal of Statistics},
       volume = 44,
        issue = {A},
        pages = {1-22},
         year = 1982,
}
@incollection{RaoNIPS-05,
       author = {Rao, Rajesh P. N.},
        title = {Hierarchical {B}ayesian Inference in Networks of Spiking Neurons},
    booktitle = {Advances in Neural Information Processing Systems 17},
       editor = {Lawrence K. Saul and Yair Weiss and {L\'{e}on} Bottou},
    publisher = {MIT Press},
      address = {Cambridge, MA},
      comment = {NIPS2005_0617.pdf},
         year = 2005,
}
@inproceedings{RaoRA-87,
       author = {Nageswara S. V. Rao, S. S. Iyengar, C. C. Jorgensen and C. R. Weisbin},
        title = {On Terrain Acquisition by a Finite-sized Mobile Robot in Plane},
    booktitle = {Proceedings of the {IEEE} International Conference on Robotics and Automation},
         year = 1987,
        pages = {1314-1319},
}
@article{RaoandBallardNATURE-NEUROSCIENCE-99,
       author = {Rao, Rajesh P. N. and Ballard, Dana H.},
        title = {Predictive coding in the visual cortex: a functional interpretation of some extra-classical receptive-field effects},
      journal = {Nature Neuroscience},
         year = {1999},
       volume = {2},
        issue = {1},
        pages = {79-87},
     abstract = {We describe a model of visual processing in which feedback connections from a higher- to a lower-order visual cortical area carry predictions of lower-level neural activities, whereas the feedforward connections carry the residual errors between the predictions and the actual lower-level activities. When exposed to natural images, a hierarchical network of model neurons implementing such a model developed simple-cell-like receptive fields. A subset of neurons responsible for carrying the residual errors showed endstopping and other extra-classical receptive-field effects. These results suggest that rather than being exclusively feedforward phenomena, nonclassical surround effects in the visual cortex may also result from cortico-cortical feedback as a consequence of the visual system using an efficient hierarchical strategy for encoding natural images.}
}
@article{RaoandBallardNC-97,
       author = {Rao, Rajesh P. N. and Ballard, Dana H.},
        title = {Dynamic Model of Visual Recognition Predicts Neural Response Properties in the Visual Cortex},
      journal = {Neural Computation},
       volume = 9,
       number = 5,
        pages = {721-763},
          url = {http://citeseer.ist.psu.edu/rpn95dynamic.html},
         year = 1996,
}
@techreport{RaoandBallardNRL-97,
    author = {Rajesh P. N. Rao and Dana H. Ballard},
     title = {Localized Receptive Fields May Mediate
              Transformation-Invariant Recognition in the Visual Cortex},
institution = {National Resource Laboratory for the Study of Brain and Behavior, University of Rochester},
     issue = {NRL TR-97-2},
      year = 1997,
}
@incollection{RaoandRudermanNIPS-99,
    author = {Rajesh P. N. Rao and Daniel L. Ruderman},
     title = {Learning Lie Groups for Invariant Visual Perception},
 booktitle = {Advances in Neural Information Processing Systems},
    volume = 11,
    editor = {M. S. Kearns and S. A. Solla and D. A. Cohn},
 publisher = {MIT Press},
   address = {Cambridge, MA},
      year = 1999,
}
@book{Raoetal2002,
    author = {Rao, R. P. N. and Olshausen, B. A. and Lewicki, M. S.},
     title = {Probabilistic Models of the Brain: Perception and Neural Function}, 
 publisher = {MIT Press}, 
   address = {Cambridge, MA},
      year = 2002,
}
@article{RaoetalCURRENT-BIOLOGY-18,
       author = {Vikram R. Rao and Kristin K. Sellers and Deanna L. Wallace and Morgan B. Lee and Maryam Bijanzadeh and Omid G. Sani and Yuxiao Yang and Maryam M. Shanechi and Heather E. Dawes and Edward F. Chang},
        title = {Direct Electrical Stimulation of Lateral Orbitofrontal Cortex Acutely Improves Mood in Individuals with Symptoms of Depression},
      journal = {Current Biology},
         year = {2018},
     abstract = {Mood disorders cause significant morbidity and mortality, and existing therapies fail 20\%-30\% of patients. Deep brain stimulation (DBS) is an emerging treatment for refractory mood disorders, but its success depends critically on target selection. DBS focused on known targets within mood-related frontostriatal and limbic circuits has been variably efficacious. Here, we examine the effects of stimulation in orbitofrontal cortex (OFC), a key hub for mood-related circuitry that has not been well characterized as a stimulation target. We studied 25 subjects with epilepsy who were implanted with intracranial electrodes for seizure localization. Baseline depression traits ranged from mild to severe. We serially assayed mood state over several days using a validated questionnaire. Continuous electrocorticography enabled investigation of neurophysiological correlates of mood-state changes. We used implanted electrodes to stimulate OFC and other brain regions while collecting verbal mood reports and questionnaire scores. We found that unilateral stimulation of the lateral OFC produced acute, dose-dependent mood-state improvement in subjects with moderate-to-severe baseline depression. Stimulation suppressed low-frequency power in OFC, mirroring neurophysiological features that were associated with positive mood states during natural mood fluctuation. Stimulation potentiated single-pulse-evoked responses in OFC and modulated activity within distributed structures implicated in mood regulation. Behavioral responses to stimulation did not include hypomania and indicated an acute restoration to non-depressed mood state. Together, these findings indicate that lateral OFC stimulation broadly modulates mood-related circuitry to improve mood state in depressed patients, revealing lateral OFC as a promising new target for therapeutic brain stimulation in mood disorders.}
}
@incollection{RaposoetalCSC-17,
        title = {A model of structure learning, inference, and generation for scene understanding},
       author = {David Raposo and Peter Dayan and Demis Hassabis and Peter Battaglia},
    booktitle = {Proceedings of the 39th Annual Meeting of the Cognitive Science Society},
    publisher = {Cognitive Science Society},
         year = {2017},
     abstract = {Humans possess rich knowledge of the structure of the world, including co-occurrences among entities, and covariation among their discrete and continuous features. But how people learn, infer and predict this structure is not well understood. Here we explore everyday scene understanding as a case study of people's structural knowledge and reasoning. We introduce a probabilistic model over scene graphs that can learn the relational structure of objects and their arrangements and support inference and generation. Our model was able to learn the underlying structure of real-world scenes, and use it for inference and compression. In two human psychophysical experiments we found that a corresponding computational cognitive model was able to explain how people learn novel scene distributions and use it for classification and construction. Our work represents the first computational theory of human scene understanding that can account for people's rich capacity for learning and reasoning about structure.}
}
@article{RaposoetalNATURE-NEUROSCIENCE-14,
       author = {Raposo, D. and Kaufman, M. T. and Churchland, A. K.},
        title = {{{A} category-free neural population supports evolving demands during decision-making}},
      journal = {Nature Neuroscience},
       volume = {17},
       number = {12},
         year = {2014},
        pages = {1784-1792},
     abstract = {The posterior parietal cortex (PPC) receives diverse inputs and is involved in a dizzying array of behaviors. These many behaviors could rely on distinct categories of neurons specialized to represent particular variables or could rely on a single population of PPC neurons that is leveraged in different ways. To distinguish these possibilities, we evaluated rat PPC neurons recorded during multisensory decisions. Newly designed tests revealed that task parameters and temporal response features were distributed randomly across neurons, without evidence of categories. This suggests that PPC neurons constitute a dynamic network that is decoded according to the animal's present needs. To test for an additional signature of a dynamic network, we compared moments when behavioral demands differed: decision and movement. Our new state-space analysis revealed that the network explored different dimensions during decision and movement. These observations suggest that a single network of neurons can support the evolving behavioral demands of decision-making.}
}
@article{RaposoetalNATURE-NEUROSCIENCE-2014,
       author = {Raposo, David and Kaufman, Matthew T. and Churchland, Anne K.},
        title = {A category-free neural population supports evolving demands during decision-making},
      journal = {Nature neuroscience},
    publisher = {Nature Publishing Group},
       volume = {17},
        issue = {12},
         year = {2014},
        pages = {1784-1792},
     abstract = {The posterior parietal cortex (PPC) receives diverse inputs and is involved in a dizzying array of behaviors. These many behaviors could rely on distinct categories of neurons specialized to represent particular variables or could rely on a single population of PPC neurons that is leveraged in different ways. To distinguish these possibilities, we evaluated rat PPC neurons recorded during multisensory decisions. Newly designed tests revealed that task parameters and temporal response features were distributed randomly across neurons, without evidence of categories. This suggests that PPC neurons constitute a dynamic network that is decoded according to the animal's present needs. To test for an additional signature of a dynamic network, we compared moments when behavioral demands differed: decision and movement. Our new state-space analysis revealed that the network explored different dimensions during decision and movement. These observations suggest that a single network of neurons can support the evolving behavioral demands of decision-making.},
}
@article{RaschandBornPR-13,
       author = {Rasch, B. and Born, J.},
        title = {{{A}bout sleep's role in memory}},
      journal = {Physiological Review},
       volume = {93},
       number = {2},
         year = {2013},
        pages = {681-766},
     abstract = {Over more than a century of research has established the fact that sleep benefits the retention of memory. In this review we aim to comprehensively cover the field of "sleep and memory" research by providing a historical perspective on concepts and a discussion of more recent key findings. Whereas initial theories posed a passive role for sleep enhancing memories by protecting them from interfering stimuli, current theories highlight an active role for sleep in which memories undergo a process of system consolidation during sleep. Whereas older research concentrated on the role of rapid-eye-movement (REM) sleep, recent work has revealed the importance of slow-wave sleep (SWS) for memory consolidation and also enlightened some of the underlying electrophysiological, neurochemical, and genetic mechanisms, as well as developmental aspects in these processes. Specifically, newer findings characterize sleep as a brain state optimizing memory consolidation, in opposition to the waking brain being optimized for encoding of memories. Consolidation originates from reactivation of recently encoded neuronal memory representations, which occur during SWS and transform respective representations for integration into long-term memory. Ensuing REM sleep may stabilize transformed memories. While elaborated with respect to hippocampus-dependent memories, the concept of an active redistribution of memory representations from networks serving as temporary store into long-term stores might hold also for non-hippocampus-dependent memory, and even for nonneuronal, i.e., immunological memories, giving rise to the idea that the offline consolidation of memory during sleep represents a principle of long-term memory formation established in quite different physiological systems.}
}
@article {RaslauetalAJN-14,
       author = {Raslau, F.D. and Klein, A.P. and Ulmer, J.L. and Mathews, V. and Mark, L.P.},
        title = {Memory Part 1: Overview},
      journal = {American Journal of Neuroradiology},
       volume = {35},
       number = {11},
         year = {2014},
        pages = {2058-2060},
}
@article{RaslauetalAJN-15,
       author = {Raslau, F.D. and Mark, I.T. and Klein, A.P. and Ulmer, J.L. and Mathews, V. and Mark, L.P.},
        title = {Memory Part 2: The Role of the Medial Temporal Lobe},
      journal = {American Journal of Neuroradiology},
       volume = {36},
       number = {5},
         year = {2015},
        pages = {846-849},
}
@article{RasmusetalCoRR-14,
       author = {Antti Rasmus and Tapani Raiko and Harri Valpola},
        title = {Denoising autoencoder with modulated lateral connections learns invariant representations of natural images},
      journal = {CoRR},
       volume = {arXiv:1412.7210},
         year = 2014,
     abstract = {Suitable lateral connections between encoder and decoder are shown to allow higher layers of a denoising autoencoder (dAE) to focus on invariant representations. In regular autoencoders, detailed information needs to be carried through the highest layers but lateral connections from encoder to decoder relieve this pressure. It is shown that abstract invariant features can be translated to detailed reconstructions when invariant features are allowed to modulate the strength of the lateral connection. Three dAE structures with modulated and additive lateral connections, and without lateral connections were compared in experiments using real-world images. The experiments verify that adding modulated lateral connections to the model 1) improves the accuracy of the probability model for inputs, as measured by denoising performance; 2) results in representations whose degree of invariance grows faster towards the higher layers; and 3) supports the formation of diverse invariant poolings.}
}
@article{RasmussenetalPLoS-ONE-17,
       author = {Rasmussen, Daniel and Voelker, Aaron and Eliasmith, Chris},
      journal = {PLOS ONE},
    publisher = {Public Library of Science},
        title = {A neural model of hierarchical reinforcement learning},
         year = {2017},
       volume = {12},
       number = {7},
        pages = {1-39},
     abstract = {We develop a novel, biologically detailed neural model of reinforcement learning (RL) processes in the brain. This model incorporates a broad range of biological features that pose challenges to neural RL, such as temporally extended action sequences, continuous environments involving unknown time delays, and noisy/imprecise computations. Most significantly, we expand the model into the realm of hierarchical reinforcement learning (HRL), which divides the RL process into a hierarchy of actions at different levels of abstraction. Here we implement all the major components of HRL in a neural model that captures a variety of known anatomical and physiological properties of the brain. We demonstrate the performance of the model in a range of different environments, in order to emphasize the aim of understanding the brain's general reinforcement learning ability. These results show that the model compares well to previous modelling work and demonstrates improved performance as a result of its hierarchical ability. We also show that the model's behaviour is consistent with available data on human hierarchical RL, and generate several novel predictions.},
}
@article{RatcliffPSYCHOLOGICAL-REVIEW-90,
       author = {Ratcliff, R.},
        title = {Connectionist models of recognition memory: constraints imposed by learning and forgetting functions},
      journal = {Psychological Review},
         year = {1990},
       volume = {97},
       number = {2},
        pages = {285-308},
     abstract = {Multilayer connectionist models of memory based on the encoder model using the backpropagation learning rule are evaluated. The models are applied to standard recognition memory procedures in which items are studied sequentially and then tested for retention. Sequential learning in these models leads to 2 major problems. First, well-learned information is forgotten rapidly as new information is learned. Second, discrimination between studied items and new items either decreases or is nonmonotonic as a function of learning. To address these problems, manipulations of the network within the multilayer model and several variants of the multilayer model were examined, including a model with prelearned memory and a context model, but none solved the problems. The problems discussed provide limitations on connectionist models applied to human memory and in tasks where information to be learned is not all available during learning.}
}
@article{RatushnyetalJCN-12,
        title = {From keratinocyte to cancer: the pathogenesis and modeling of cutaneous squamous cell carcinoma},
       author = {Ratushny, Vladimir and Gober, Michael D and Hick, Ryan and Ridky, Todd W and Seykora, John T},
      journal = {The Journal of Clinical Investigation},
       volume = {122},
       number = {2},
         year = {2012},
        pages = {464-72},
     abstract = {Cutaneous squamous cell carcinoma (cSCC) is the second most common human cancer with over 250,000 new cases annually in the US and is second in incidence only to basal cell carcinoma. cSCC typically manifests as a spectrum of progressively advanced malignancies, ranging from a precursor actinic keratosis (AK) to squamous cell carcinoma (SCC) in situ (SCCIS), invasive cSCC, and finally metastatic SCC. In this Review we discuss clinical and molecular parameters used to define this range of cutaneous neoplasia and integrate these with the multiple experimental approaches used to study this disease. Insights gained from modeling cSCCs have suggested innovative therapeutic targets for treating these lesions.}
}
@incollection{Ray91,
      author = {Tom Ray},
       title = {An Approach to the Synthesis of Life},
   booktitle = {Artificial Life {II}, Santa Fe Institute Studies in the Sciences of Complexity, Volume {XI}},
      editor = {C. G. Langton and C. Taylor and J. D. Farmer and S. Rasmussen},
   publisher = {Addison-Wesley},
     address = {Reading, MA},
       pages = {371-408},
        year = 1991,
}
@article{RaymondetalSCIENCE-96,
       author = {Jennifer L. Raymond and Stephen G. Lisberger and Michael D. Mauk},
        title = {The Cerebellum: A Neuronal Learning Machine?},
      journal = {Science},
       volume = 722,
        issue = 6158,
         year = 1996,
     abstract = {The comparison of two seemingly quite different behaviors yields a surprisingly consistent picture of the role of the cerebellum in motorlearning. Behavioral and physiological data about classical conditioning of the eye lid response and motor learning in the vestibulo-ocular reflex suggest that (i) plasticity is distributed between the cerebellar cortex and the deep cerebellar nuclei; (i) the cerebellar cortex plays a special role in learning the timing of movement; and (i) the cerebellar cortex guides learning in the deep nuclei, which may for movement coordination (1,2,5,7) allow learning to be transferred from the cortex to the deep nuclei. Because many of the similarities in the data from the two systems typify general features of cerebellar organization, the cerebellar mechanisms of learning in these two systems may represent principles that apply to many motor systems.}
}
@incollection{RaynerKR-89,
       author = {Rayner, Manny},
        title = {Did {N}ewton Solve the "Extended Prediction Problem?"},
     crossref = {KR-89},
        pages = {381-385},
}
@article{RazavietalCoRR-19,
       author = {Ali Razavi and A{\"{a}}ron van den Oord and Oriol Vinyals},
        title = {Generating Diverse High-Fidelity Images with {VQ-VAE-2}},
      journal = {CoRR},
       volume = {arXiv:1906.00446},
         year = {2019},
     abstract = {We explore the use of Vector Quantized Variational AutoEncoder (VQ-VAE) models for large scale image generation. To this end, we scale and enhance the autoregressive priors used in VQ-VAE to generate synthetic samples of much higher coherence and fidelity than possible before. We use simple feed-forward encoder and decoder networks, making our model an attractive candidate for applications where the encoding and/or decoding speed is critical. Additionally, VQ-VAE requires sampling an autoregressive model only in the compressed latent space, which is an order of magnitude faster than sampling in the pixel space, especially for large images. We demonstrate that a multi-scale hierarchical organization of VQ-VAE, augmented with powerful priors over the latent codes, is able to generate samples with quality that rivals that of state of the art Generative Adversarial Networks on multifaceted datasets such as ImageNet, while not suffering from GAN's known shortcomings such as mode collapse and lack of diversity.}
}
@article{RaznahanetalJN-11,
       author = {Raznahan, A. and Shaw, P. and Lalonde, F. and Stockman, M. and Wallace, G. L. and Greenstein, D. and Clasen, L. and Gogtay, N. and Giedd, J. N.},
        title = {How does your cortex grow?},
      journal = {The Journal of Neuroscience},
       volume = 31,
        issue = 19,
        pages = {7174-7177},
         year = 2011,
     abstract = {Understanding human cortical maturation is a central goal for developmental neuroscience. Significant advances toward this goal have come from two recent strands of in vivo structural magnetic resonance imaging research: (1) longitudinal study designs have revealed that factors such as sex, cognitive ability, and disease are often better related to variations in the tempo of anatomical change than to variations in anatomy at any one time point; (2) largely cross-sectional applications of new surface-based morphometry (SBM) methods have shown how the traditional focus on cortical volume (CV) can obscure information about the two evolutionarily and genetically distinct determinants of CV: cortical thickness (CT) and surface area (SA). Here, by combining these two strategies for the first time and applying SBM in [more than] 1250 longitudinally acquired brain scans from 647 healthy individuals aged 3-30 years, we deconstruct cortical development to reveal that distinct trajectories of anatomical change are hidden within, and give rise to, a curvilinear pattern of CV maturation. Developmental changes in CV emerge through the sexually dimorphic and age-dependent interaction of changes in CT and SA. Moreover, SA change itself actually reflects complex interactions between brain size-related changes in exposed cortical convex hull area, and changes in the degree of cortical gyrification, which again vary by age and sex. Knowing of these developmental dissociations, and further specifying their timing and sex-biases, provides potent new research targets for basic and clinical neuroscience.},
}
@article{RealetalCURRENT-BIOLOGY-17,
       author = {Real, E. and Asari, H. and Gollisch, T. and Meister, M.},
        title = {Neural Circuit Inference from Function to Structure},
      journal = {Current Biology},
       volume = {27},
       number = {2},
         year = {2017},
        pages = {189-198},
     abstract = {Advances in technology are opening new windows on the structural connectivity and functional dynamics of brain circuits. Quantitative frameworks are needed that integrate these data from anatomy and physiology. Here, we present a modeling approach that creates such a link. The goal is to infer the structure of a neural circuit from sparse neural recordings, using partial knowledge of its anatomy as a regularizing constraint. We recorded visual responses from the output neurons of the retina, the ganglion cells. We then generated a systematic sequence of circuit models that represents retinal neurons and connections and fitted them to the experimental data. The optimal models faithfully recapitulated the ganglion cell outputs. More importantly, they made predictions about dynamics and connectivity among unobserved neurons internal to the circuit, and these were subsequently confirmed by experiment. This circuit inference framework promises to facilitate the integration and understanding of big data in neuroscience.}
}
@article {RedondoetalJoN-10,
       author = {Redondo, Roger L. and Okuno, Hiroyuki and Spooner, Patrick A. and Frenguelli, Bruno G. and Bito, Haruhiko and Morris, Richard G. M.},
        title = {Synaptic Tagging and Capture: Differential Role of Distinct Calcium/Calmodulin Kinases in Protein Synthesis-Dependent Long-Term Potentiation},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {30},
       number = {14},
         year = {2010},
        pages = {4981-4989},
     abstract = {Weakly tetanized synapses in area CA1 of the hippocampus that ordinarily display long-term potentiation lasting \~{}3 h (called early-LTP) will maintain a longer-lasting change in efficacy (late-LTP) if the weak tetanization occurs shortly before or after strong tetanization of an independent, but convergent, set of synapses in CA1. The synaptic tagging and capture hypothesis explains this heterosynaptic influence on persistence in terms of a distinction between local mechanisms of synaptic tagging and cell-wide mechanisms responsible for the synthesis, distribution, and capture of plasticity-related proteins (PRPs). We now present evidence that distinct CaM kinase (CaMK) pathways serve a dissociable role in these mechanisms. Using a hippocampal brain-slice preparation that permits stable long-term recordings in vitro for \&gt;10 h and using hippocampal cultures to validate the differential drug effects on distinct CaMK pathways, we show that tag setting is blocked by the CaMK inhibitor KN-93 (2-[N-(2-hydroxyethyl)]-N-(4-methoxybenzenesulfonyl)amino-N-(4-chlorocinnamyl)-N-methylbenzylamine) that, at low concentration, is more selective for CaMKII. In contrast, the CaMK kinase inhibitor STO-609 [7H-benzimidazo(2,1-a)benz(de)isoquinoline-7-one-3-carboxylic acid] specifically limits the synthesis and/or availability of PRPs. Analytically powerful three-pathway protocols using sequential strong and weak tetanization in varying orders and test stimulation over long periods of time after LTP induction enable a pharmacological dissociation of these distinct roles of the CaMK pathways in late-LTP and so provide a novel framework for the molecular mechanisms by which synaptic potentiation, and possibly memories, become stabilized.},
}
@article{ReedandDeFreitasCoRR-15,
       author = {Scott E. Reed and Nando de Freitas},
        title = {Neural Programmer-Interpreters},
      journal = {CoRR},
       volume = {arXiv:1511.06279},
         year = {2015},
     abstract = {We propose the neural programmer-interpreter (NPI): a recurrent and compositional neural network that learns to represent and execute programs. NPI has three learnable components: a task-agnostic recurrent core, a persistent key-value program memory, and domain-specific encoders that enable a single NPI to operate in multiple perceptually diverse environments with distinct affordances. By learning to compose lower-level programs to express higher-level programs, NPI reduces sample complexity and increases generalization ability compared to sequence-to-sequence LSTMs. The program memory allows efficient learning of additional tasks by building on existing programs. NPI can also harness the environment (e.g. a scratch pad with read-write pointers) to cache intermediate results of computation, lessening the long-term memory burden on recurrent hidden units. In this work we train the NPI with fully-supervised execution traces; each program has example sequences of calls to the immediate subprograms conditioned on the input. Rather than training on a huge number of relatively weak labels, NPI learns from a small number of rich examples. We demonstrate the capability of our model to learn several types of compositional programs: addition, sorting, and canonicalizing 3D models. Furthermore, a single NPI learns to execute these programs and all 21 associated subprograms.}
}
@article{ReeseetalPLoS-12,
       author = {Reese, Tyler M. and Brzoska, Antoni and Yott, Dylan T. and Kelleher, Daniel J.},
      journal = {PLoS ONE},
    publisher = {Public Library of Science},
        title = {Analyzing Self-Similar and Fractal Properties of the C. elegans Neural Network},
       volume = {7},
       number = {10},
         year = {2012},
        pages = {1-10},
     abstract = {The brain is one of the most studied and highly complex systems in the biological world. While much research has concentrated on studying the brain directly, our focus is the structure of the brain itself: at its core an interconnected network of nodes (neurons). A better understanding of the structural connectivity of the brain should elucidate some of its functional properties. In this paper we analyze the connectome of the nematode Caenorhabditis elegans. Consisting of only 302 neurons, it is one of the better-understood neural networks. Using a Laplacian Matrix of the 279-neuron 'giant component, of the network, we use an eigenvalue counting function to look for fractal-like self similarity. This matrix representation is also used to plot visualizations of the neural network in eigenfunction coordinates. Small-world properties of the system are examined, including average path length and clustering coefficient. We test for localization of eigenfunctions, using graph energy and spacial variance on these functions. To better understand results, all calculations are also performed on random networks, branching trees, and known fractals, as well as fractals which have been 'rewired' to have small-world properties. We propose algorithms for generating Laplacian matrices of each of these graphs.},
}
@article{ReesetalTiN-17,
       author = {Rees, Christopher L. and Moradi, Keivan and Ascoli, Giorgio A.},
        title = {Weighing the Evidence in {Peters's Rule}: Does Neuronal Morphology Predict Connectivity?},
      journal = {Trends in Neurosciences},
         year = {2017},
    publisher = {Elsevier},
       volume = {40},
        issue = {2},
        pages = {63-71},
     abstract = {Although the importance of network connectivity is increasingly recognized, identifying synapses remains challenging relative to the routine characterization of neuronal morphology. Thus, researchers frequently employ axon-dendrite colocations as proxies of potential connections. This putative equivalence, commonly referred to as Peters? rule, has been recently studied at multiple levels and scales, fueling passionate debates regarding its validity. Our critical literature review identifies three conceptually distinct but often confused applications: inferring neuron type circuitry, predicting synaptic contacts among individual cells, and estimating synapse numbers within neuron pairs. Paradoxically, at the originally proposed cell-type level, Peters? rule remains largely untested. Leveraging Hippocampome.org, we validate and refine the relationship between axonal-dendritic colocations and synaptic circuits, clarifying the interpretation of existing and forthcoming data.},
}
@article{RegolinBL-11,
       author = {Regolin, Lucia  and Rugani, Rosa  and Stancher, Gionata  and Vallortigara, Giorgio},
        title = {Spontaneous discrimination of possible and impossible objects by newly hatched chicks},
      journal = {Biology Letters},
       volume = {7},
       number = {5},
        pages = {654-657},
         year = {2011},
     abstract = {Four-month-old infants can integrate local cues provided by two-dimensional pictures and interpret global inconsistencies in structural information to discriminate between possible and impossible objects. This leaves unanswered the issue of the relative contribution of maturation of biologically predisposed mechanisms and of experience with real objects, to the development of this capability. Here we show that, after exposure to objects in which junctions providing cues to global structure were occluded, day-old chicks selectively approach the two-dimensional image that depicted the possible rather than the impossible version of a three-dimensional object, after restoration of the junctions. Even more impressively, completely naive newly hatched chicks showed spontaneous preferences towards approaching two-dimensional depictions of structurally possible rather than impossible objects. These findings suggest that the vertebrate brain can be biologically predisposed towards approaching a two-dimensional image representing a view of a structurally possible three-dimensional object.}
}
@book{ReichSUCCESS,
       author = {Robert Reich},
        title = {The Future of Success},
    publisher = {Alfred A. Knopf, Inc.},
      address = {New York},
         year = 2001,
}
@article{ReidNEURON-12,
       author = {Reid, R. Clay},
        title = {From Functional Architecture to Functional Connectomics},
      journal = {Neuron},
         year = 2012,
    publisher = {Elsevier},
       volume = 75,
        issue = 2,
        pages = {209-217},
     abstract = {Receptive Fields, Binocular Interaction and Functional Architecture in the Cat's Visual Cortex? by Hubel and Wiesel (1962) reported several important discoveries: orientation columns, the distinct structures of simple and complex receptive fields, and binocular integration. But perhaps the paper's greatest influence came from the concept of functional architecture (the complex relationship between in vivo physiology and the spatial arrangement of neurons) and several models of functionally specific connectivity. They thus identified two distinct concepts, topographic specificity and functional specificity, which together with cell-type specificity constitute the major determinants of nonrandom cortical connectivity. Orientation columns are iconic examples of topographic specificity, whereby axons within a column connect with cells of a single orientation preference. Hubel and Wiesel also saw the need for functional specificity at a finer scale in their model of thalamic inputs to simple cells, verified in the 1990s. The difficult but potentially more important question of functional specificity between cortical neurons is only now becoming tractable with new experimental techniques,},
}
@article{ReiffetalNATURE-NEUROSCIENCE-10,
       author = {Reiff, Dierk F. and Plett, Johannes and Mank, Marco and Griesbeck, Oliver and Borst, Alexander},
        title = {Visualizing retinotopic half-wave rectified input to the motion detection circuitry of Drosophila},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {13},
        issue = {8},
         year = {2010},
        pages = {973-978},
     abstract = {In the visual system of Drosophila, photoreceptors R1-R6 relay achromatic brightness information to five parallel pathways. Two of them, the lamina monopolar cells L1 and L2, represent the major input lines to the motion detection circuitry. We devised a new method for optical recording of visually evoked changes in intracellular Ca2+ in neurons using targeted expression of a genetically encoded Ca2+ indicator. Ca2+ in single terminals of L2 neurons in the medulla carried no information about the direction of motion. However, we found that brightness decrements (light-OFF) induced a strong increase in intracellular Ca2+ but brightness increments (light-ON) induced only small changes, suggesting that half-wave rectification of the input signal occurs. Thus, L2 predominantly transmits brightness decrements to downstream circuits that then compute the direction of image motion.},
}
@article{ReimanetalFiCN-15,
       author = {Reimann, Michael Wolfgang  and  Muller, Eilif Benjamin  and  Ramaswamy, Srikanth  and  Markram, Henry},
        title = {An Algorithm to Predict the Connectome of Neural Microcircuits},
      journal = {Frontiers in Computational Neuroscience},
       volume = {9},
       number = {28},
         year = {2015},
     abstract = {Experimentally mapping synaptic connections, in terms of the numbers and locations of their synapses and estimating connection probabilities, is still not a tractable task, even for small volumes of tissue. In fact, the six layers of the neocortex contain thousands of unique types of synaptic connections between the many different types of neurons, of which only a handful have been characterized experimentally. Here we present a theoretical framework and a data-driven algorithmic strategy to digitally reconstruct the complete synaptic connectivity between the different types of neurons in a small well-defined volume of tissue--the micro-scale connectome of a neural microcircuit. By enforcing a set of established principles of synaptic connectivity, and leveraging interdependencies between fundamental properties of neural microcircuits to constrain the reconstructed connectivity, the algorithm yields three parameters per connection type that predict the anatomy of all types of biologically viable synaptic connections. The predictions reproduce a spectrum of experimental data on synaptic connectivity not used by the algorithm. We conclude that an algorithmic approach to the connectome can serve as a tool to accelerate experimental mapping, indicating the minimal dataset required to make useful predictions, identifying the datasets required to improve their accuracy, testing the feasibility of experimental measurements, and making it possible to test hypotheses of synaptic connectivity.}
}
@article{ReimannetalFiCN-15,
       author = {Reimann, Michael W. and King, James G. and Muller, Eilif B. and Ramaswamy, Srikanth and Markram, Henry},
        title = {An algorithm to predict the connectome of neural microcircuits},
      journal = {Frontiers in Compututational Neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = {9},
        pages = {120},
         year = {2015},
     abstract = {Experimentally mapping synaptic connections, in terms of the numbers and locations of their synapses and estimating connection probabilities, is still not a tractable task, even for small volumes of tissue. In fact, the six layers of the neocortex contain thousands of unique types of synaptic connections between the many different types of neurons, of which only a handful have been characterized experimentally. Here we present a theoretical framework and a data-driven algorithmic strategy to digitally reconstruct the complete synaptic connectivity between the different types of neurons in a small well-defined volume of tissue \emdash{} the micro-scale connectome of a neural microcircuit. By enforcing a set of established principles of synaptic connectivity, and leveraging interdependencies between fundamental properties of neural microcircuits to constrain the reconstructed connectivity, the algorithm yields three parameters per connection type that predict the anatomy of all types of biologically viable synaptic connections. The predictions reproduce a spectrum of experimental data on synaptic connectivity not used by the algorithm. We conclude that an algorithmic approach to the connectome can serve as a tool to accelerate experimental mapping, indicating the minimal dataset required to make useful predictions, identifying the datasets required to improve their accuracy, testing the feasibility of experimental measurements, and making it possible to test hypotheses of synaptic connectivity.},
}
@article{ReimannetalNEURON-13,
        title = {A Biophysically Detailed Model of Neocortical Local Field Potentials Predicts the Critical Role of Active Membrane Currents},
       author = {Michael W. Reimann and Costas A. Anastassiou and Rodrigo Perin and Sean L. Hill and Henry Markram and Christof Koch},
      journal = {Neuron},
       volume = 79,
        issue = 2,
        pages = {375-390},
         year = 2013,
     abstract = {Brain activity generates extracellular voltage fluctuations recorded as local field potentials (LFPs). It is known that the relevant microvariables, the ionic currents across membranes, jointly generate the macrovariables, the extracellular voltage, but neither the detailed biophysical knowledge nor the required computational power have been available to model these processes. We simulated the \{LFP\} in a model of the rodent neocortical column composed of > 12,000 reconstructed, multicompartmental, and spiking cortical layer 4 and 5 pyramidal neurons and basket cells, including five million dendritic and somatic compartments with voltage- and ion-dependent currents, realistic connectivity, and probabilistic AMPA, NMDA, and \{GABA\} synapses. We found that, depending on a number of factors, the \{LFP\} reflects local and cross-layer processing. Active currents dominate the generation of LFPs, not synaptic ones. Spike-related currents impact the \{LFP\} not only at higher frequencies but below 50 Hz. This work calls for re-evaluating the genesis of LFPs.}
}
@article{ReinetalCB-02,
       author = {Rein, Karlheinz and Z\"{o}ckler, Malte and Mader, Michael T. and Gr\"{o}bel, Cornelia and Heisenberg, Martin},
        title = {The Drosophila Standard Brain},
      journal = {Current Biology},
       volume = {12},
        issue = {3},
         year = {2002},
        pages = {227-231},
     abstract = {Organisms and organs come in sizes and shapes. With size, science has no problems, but how to quantify shape? How similar are two birds or two brains? This problem is particularly pressing in cases like brains where structure reflects function. The problem is not new, but satisfying solutions have yet to be worked out. For brain anatomy, no general methodology for a statistically secured quantitative description is available. Using the small brain of the fly Drosophila melanogaster, we have explored a new approach combining immunohistochemistry, high-resolution {3D} confocal microscopy, and advanced graphics computing. For a genetic model organism such as Drosophila, a quantitative assessment of brain structure is particularly rewarding, since it allows for the identification of genetic variants with subtle brain structure phenotypes and, even more importantly, the organization of the wealth of gene expression patterns in the brain into a genetic atlas linking molecular and organismic gene function. We now provide a representative standard for the brain of D. melanogaster wild-type with means and variances for several aspects of its shape. Its application to volumetry, mutants, and gene expression patterns is demonstrated.},
}
@article{ReiserandDickinsonJNM-08,
       author = {Michael B. Reiser and Michael H. Dickinson},
        title = {A modular display system for insect behavioral neuroscience},
      journal = {Journal of Neuroscience Methods},
       volume = {167},
       number = {2},
         year = {2008},
        pages = {127-139},
     abstract = {Flying insects exhibit stunning behavioral repertoires that are largely mediated by the visual control of flight. For this reason, presenting a controlled visual environment to tethered insects has been and continues to be a powerful tool for studying the sensory control of complex behaviors. To create an easily controlled, scalable, and customizable visual stimulus, we have designed a modular system, based on panels composed of an 8 \% 8 array of individual LEDs, that may be connected together to 'tile' an experimental environment with controllable displays. The panels have been designed to be extremely bright, with the added flexibility of individual-pixel brightness control, allowing experimentation over a broad range of behaviorally relevant conditions. Patterns to be displayed may be designed using custom software, downloaded to a controller board, and displayed on the individually addressed panels via a rapid communication interface. The panels are controlled by a microprocessor-based display controller which, for most experiments, will not require a computer in the loop, greatly reducing the experimental infrastructure. This technology allows an experimenter to build and program a visual arena with a customized geometry in a matter of hours. To demonstrate the utility of this system, we present results from experiments with tethered Drosophila melanogaster: (1) in a cylindrical arena composed of 44 panels, used to test the contrast dependence of object orientation behavior, and (2) above a 30-panel floor display, used to examine the effects of ground motion on orientation during flight.},
}
@techreport{ReissTR-86,
       author = {Reiss, Steve},
        title = {Visual Languages and the {GARDEN} System},
        issue = {CS-86-16},
  institution = {Brown University Department of Computer Science},
         year = 1986,
}
@article{ReiterAIJ-80,
       author = {Reiter, Raymond},
        title = {A Logic for Default Reasoning},
      journal = AIJ,
       volume = 13,
         year = 1980,
        pages = {81-132},
}
@techreport{ReiterTR-83,
       author = {Reiter, Raymond},
        title = {A Sound and Sometimes Complete Query Evaluation Algorithm for Relational Databases with Null Values},
        issue = {83-11},
  institution = {University of British Columbia},
         year = 1983,
}
@inproceedings{RenandMalikCVPR-07,
        title = {Tracking as Repeated Figure/Ground Segmentation},
       author = {Xiaofeng Ren and Jitendra Malik},
    booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
         year = 2007
}
@inproceedings{RenandMalikICCV-03,
       author = {Xiaofeng Ren and Jitendra Malik},
        title = {Learning a Classification Model for Segmentation},
    booktitle = {Proceedings of the Ninth {IEEE} International Conference on Computer Vision},
       volume = 1,
        pages = {10-17},
         year = 2003,
}
@inproceedings{RennieandMcCallum99,
       author = {Jason Rennie and Andrew Kachites McCallum},
        title = {Using reinforcement learning to spider the {W}eb efficiently},
    booktitle = {Proceedings of {ICML}-99, 16th International Conference on Machine Learning},
    publisher = {Morgan Kaufmann Publishers, San Francisco, US},
       editor = {Ivan Bratko and Saso Dzeroski},
        pages = {335-343},
         year = 1999,
}
@article{RensinkVR-00,
       author = {Rensink, R. A.},
        title = {Seeing, sensing, and scrutinizing},
      journal = {Vision Research},
       volume = 40,
        issue = {10-12},
        pages = {1469-1487},
         year = 2000,
}
@book{Renyi84,
       author = {Alfr\`{e}d R\`{e}nyi},
        title = {A Diary on Information Theory},
    publisher = {John Wiley \& Sons},
         year = 1984,
}
@misc{RepettoANALOGY-18,
       author = {Anthony Repetto},
        title = {Neural Networks: Forming Analogies {I}},
         year = {2018},
 howpublished = {{\tt{https://towardsdatascience.com/neural-networks-forming-analogies-587557c3b26e}}},
     abstract = {Our brains have a capacity which artificial neural networks still lack: we can form analogies, relating disparate inputs and processing them using the same heuristic. The official lingo is 'transfer learning'. Fundamentally, analogies are a form of compression, allowing out brains to simulate the dynamics of many different systems with minimal space devoted to each subsystem. Where two systems behave similarly, a single analogy describes them both.},
}
@misc{RepettoANALOGY-18b,
       author = {Anthony Repetto},
        title = {Neural Networks: Forming Analogies {II}},
         year = {2018},
 howpublished = {{\tt{https://towardsdatascience.com/neural-networks-forming-analogies-587557c3b26e}}},
}
@article{RereetalISIC-15,
        title = {Simulated Annealing Algorithm for Deep Learning},
       author = {L.M. Rasdi Rere and Mohamad Ivan Fanany and Aniati Murni Arymurthy},
      journal = {Proceedings of the 3rd Information Systems International Conference} ,
       volume = {72},
         year = {2015},
        pages = {137-144},
     abstract = {Deep learning (DL) is a new area of research in machine learning, in which the objective is moving us closer to the goal of artificial intelligent. This method can learn many levels of abstraction and representation to create a common sense of data such as text, sound and image. Although DL is useful for a variety of tasks, it's hard to train. Some methods in training deep learning to make it optimal have been proposed, including Stochastic Gradient Descent, Conjugate Gradient, Hessian-free optimization, and Krylov Subspace Descent. In this paper, we proposed Simulated Annealing (SA) to improve the performance of Convolution Neural Network (CNN), as an alternative approach for optimal DL using modern optimization technique, i.e. metaheuristic algorithm. MNIST dataset is used to ensure the accuracy and efficiency of the proposed method. Moreover, we also compare our proposed method with the original of CNN. Although there is an increase in computation time, the experiment results show that the proposed method can improve the performance of original CNN.}
}
@book{RescherandUrquhart71,
       author = {Rescher, Nicholas and Urquhart, Alasdair},
        title = {Temporal Logic},
    publisher = {Springer-Verlag},
      address = {New York},
         year = 1971,
}
@article{ResslerandMaybergNATURE-07,
     title = {Targeting abnormal neural circuits in mood and anxiety disorders: from the laboratory to the clinic},
    author = {Ressler, K.J. and Mayberg, H.S.},
   journal = {Nature Neuroscience},
    volume = 10,
     issue = 9,
     pages = {1116-1124},
      year = 2007,
  abstract = {Recent decades have witnessed tremendous advances in the neuroscience of emotion, learning and memory, and in animal models for understanding depression and anxiety. This review focuses on new rationally designed psychiatric treatments derived from preclinical human and animal studies. Nonpharmacological treatments that affect disrupted emotion circuits include vagal nerve stimulation, rapid transcranial magnetic stimulation and deep brain stimulation, all borrowed from neurological interventions that attempt to target known pathological foci. Other approaches include drugs that are given in relation to specific learning events to enhance or disrupt endogenous emotional learning processes. Imaging data suggest that common regions of brain activation are targeted with pharmacological and somatic treatments as well as with the emotional learning in psychotherapy. Although many of these approaches are experimental, the rapidly developing understanding of emotional circuit regulation is likely to provide exciting and powerful future treatments for debilitating mood and anxiety disorders.}
}
@article{ReynoldsandChelazzi04,
   author = {John H. Reynolds and Leonardo Chelazzi},
    title = {Attentional modulation of visual processing},
  journal = {Annual Review of Neuroscience},
   volume = 27,
    pages = {611-647}, 
     year = 2004,
}
@inproceedings{ReynoldsandMurphyCRV-07,
    author = {Jordan Reynolds and Kevin Murphy},
     title = {Figure-ground segmentation using a hierarchical conditional random field},
 booktitle = {Fourth Canadian Conference on Computer and Robot Vision},
      year = 2007,
}
@article{ReynoldsandOReillyCOGNITION-09,
       author = {Reynolds, J. R. and O'Reilly, R. C.},
        title = {Developing {PFC} representations using reinforcement learning},
      journal = {Cognition},
       volume = {113},
       number = {3},
         year = {2009},
        pages = {281-292},
     abstract = {From both functional and biological considerations, it is widely believed that action production, planning, and goal-oriented behaviors supported by the frontal cortex are organized hierarchically [Fuster (1991); Koechlin, E., Ody, C., \& Kouneiher, F. (2003). Neuroscience: The architecture of cognitive control in the human prefrontal cortex. Science, 424, 1181-1184; Miller, G. A., Galanter, E., \& Pribram, K. H. (1960). Plans and the structure of behavior. New York: Holt]. However, the nature of the different levels of the hierarchy remains unclear, and little attention has been paid to the origins of such a hierarchy. We address these issues through biologically-inspired computational models that develop representations through reinforcement learning. We explore several different factors in these models that might plausibly give rise to a hierarchical organization of representations within the PFC, including an initial connectivity hierarchy within PFC, a hierarchical set of connections between PFC and subcortical structures controlling it, and differential synaptic plasticity schedules. Simulation results indicate that architectural constraints contribute to the segregation of different types of representations, and that this segregation facilitates learning. These findings are consistent with the idea that there is a functional hierarchy in PFC, as captured in our earlier computational models of PFC function and a growing body of empirical data.}
}
@misc{RezchikovetalELEGANS-17,
       author = {Semon Rezchikov and Saul Kato and Manuel Zimmer and Edward S. Boyden and Adam H. Marblestone and Thomas L. Dean},
        title = {Hierarchical Dynamical Structures in {\it{Caenorhabditis elegans}} Neural Activity},
         year = {2017},
 howpublished = {http://www.arxiv.org/},
}
@article{RezendeetalCoRR-15,
       author = {Danilo Jimenez Rezende and Shakir Mohamed},
        title = {Variational Inference with Normalizing Flows},
      journal = {CoRR},
       volume = {arXiv:1505.05770},
         year = {2015},
     abstract = {The choice of approximate posterior distribution is one of the core problems in variational inference. Most applications of variational inference employ simple families of posterior approximations in order to allow for efficient inference, focusing on mean-field or other simple structured approximations. This restriction has a significant impact on the quality of inferences made using variational methods. We introduce a new approach for specifying flexible, arbitrarily complex and scalable approximate posterior distributions. Our approximations are distributions constructed through a normalizing flow, whereby a simple initial density is transformed into a more complex one by applying a sequence of invertible transformations until a desired level of complexity is attained. We use this view of normalizing flows to develop categories of finite and infinitesimal flows and provide a unified view of approaches for constructing rich posterior approximations. We demonstrate that the theoretical advantages of having posteriors that better match the true posterior, combined with the scalability of amortized variational approaches, provides a clear improvement in performance and applicability of variational inference.}
}
@inproceedings{RezendeetalCoRR-16,
       author = {Danilo Jimenez Rezende and Shakir Mohamed and Ivo Danihelka and Karol Gregor and Daan Wierstra},
        title = {One-Shot Generalization in Deep Generative Models},
    booktitle = {Proceedings of the 33nd International Conference on Machine Learning},
        pages = {1521-1529},
         year = {2016},
     abstract = {Humans have an impressive ability to reason about new concepts and experiences from just a single example. In particular, humans have an ability for one-shot generalization: an ability to encounter a new concept, understand its structure, and then be able to generate compelling alternative variations of the concept. We develop machine learning systems with this important capacity by developing new deep generative models, models that combine the representational power of deep learning with the inferential power of Bayesian reasoning. We develop a class of sequential generative models that are built on the principles of feedback and attention. These two characteristics lead to generative models that are among the state-of-the art in density estimation and image generation. We demonstrate the one-shot generalization ability of our models using three tasks: unconditional sampling, generating new exemplars of a given concept, and generating new exemplars of a family of concepts. In all cases our models are able to generate compelling and diverse samples---having seen new examples just once---providing an important class of general-purpose models for one-shot machine learning.}
}
@book{Rheingold2002,
       author = {Howard Rheingold},
        title = {Tools for Thought: The History and Future of Mind-Expanding Technology},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2002,
}
@book{Rhodes1986,
       author = {Rhodes, Richard},
        title = {The Making of the Atomic Bomb},
    publisher = {Simon and Schuster},
      address = {New York},
         year = 1986,
}
@inproceedings{RhodesIUI-00,
     author = {Rhodes, Bradley J.},
      title = {Margin notes: building a contextually aware associative memory},
  booktitle = {Proceedings of the 5th international conference on Intelligent user interfaces},
   location = {New Orleans, Louisiana, United States},
  publisher = {ACM},
       year = 2000,
      pages = {219-224},
}
@article{RhodesandAndersonFiN-12,
       author = {Rhodes, Paul A. and Anderson, Todd O.},
        title = {Evolving a Neural Olfactorimotor System in Virtual and Real Olfactory Environments},
      journal = {Frontiers in Neuroengineering},
       volume = {5},
        issue = {22},
         year = {2012},
        pages = {1-14},
     abstract = {To provide a platform to enable the study of simulated olfactory circuitry in context, we have integrated a simulated neural olfactorimotor system with a virtual world which simulates both computational fluid dynamics as well as a robotic agent capable of exploring the simulated plumes. A number of the elements which we developed for this purpose have not, to our knowledge, been previously assembled into an integrated system, including: control of a simulated agent by a neural olfactorimotor system; continuous interaction between the simulated robot and the virtual plume; the inclusion of multiple distinct odorant plumes and background odor; the systematic use of artificial evolution driven by olfactorimotor performance (e.g., time to locate a plume source) to specify parameter values; the incorporation of the realities of an imperfect physical robot using a hybrid model where a physical robot encounters a simulated plume. We close by describing ongoing work toward engineering a high dimensional, reversible, low power electronic olfactory sensor which will allow olfactorimotor neural circuitry evolved in the virtual world to control an autonomous olfactory robot in the physical world. The platform described here is intended to better test theories of olfactory circuit function, as well as provide robust odor source localization in realistic environments.}
}
@article{RhodesandLlinasJPS-05,
     author = {Paul Rhodes and Rodolfo Llin\'{a}s},
    journal = {Journal of the Physiological Society},
      title = {A model of thalamo-cortical relay cells},
       year = 2005
}
@article{RhodesandMaesIBM-00,
      title = {Just-in-time information retrieval agents}, 
     author = {Bradley Rhodes and Pattie Maes},
    journal = {IBM Systems Journal}, 
     volume = 39,
      issue = {3-4},
  publisher = {IBM}, 
       year = 2000, 
      pages = {685-704},
}
@inproceedings{RhodesandStarnerPAAM-96,
      title = {The Remembrance Agent: A continuously running automated information retrieval system}, 
     author = {Bradley Rhodes and Thad Starner}, 
  booktitle = {The Proceedings of The First International Conference on The Practical Application of Intelligent Agents and Multi Agent Technology}, 
   location = {London, UK},
       year = 1996,
      pages = {487-495},
}
@article{RiazetalSCIENTIFIC-REPORTS-17,
       author = {Riaz, Bushra and Pfeiffer, Christoph and Schneiderman, Justin F.},
        title = {Evaluation of realistic layouts for next generation on-scalp MEG: spatial information density maps},
      journal = {Scientific Reports},
         year = {2017},
       volume = {7},
        issue = {1},
        pages = {6974},
     abstract = {While commercial magnetoencephalography (MEG) systems are the functional neuroimaging state-of-the-art in terms of spatio-temporal resolution, MEG sensors have not changed significantly since the 1990s. Interest in newer sensors that operate at less extreme temperatures, e.g., high critical temperature (high-$T_c$) SQUIDs, optically-pumped magnetometers, etc., is growing because they enable significant reductions in head-to-sensor standoff (on-scalp MEG). Various metrics quantify the advantages of on-scalp MEG, but a single straightforward one is lacking. Previous works have furthermore been limited to arbitrary and/or unrealistic sensor layouts. We introduce spatial information density (SID) maps for quantitative and qualitative evaluations of sensor arrays. SID-maps present the spatial distribution of information a sensor array extracts from a source space while accounting for relevant source and sensor parameters. We use it in a systematic comparison of three practical on-scalp MEG sensor array layouts (based on high-$T_c$ SQUIDs) and the standard Elekta Neuromag TRIUX magnetometer array. Results strengthen the case for on-scalp and specifically high-$T_c$ SQUID-based MEG while providing a path for the practical design of future MEG systems. SID-maps are furthermore general to arbitrary magnetic sensor technologies and source spaces and can thus be used for design and evaluation of sensor arrays for magnetocardiography, magnetic particle imaging, etc.},
}
@article{Riba-FernandeNEURON-11,
        title = {A Neural Signature of Hierarchical Reinforcement Learning},
       author = {Jos\'{e} J.F. Ribas-Fernandes and Alec Solway and Carlos Diuk and Joseph T. McGuire and Andrew G. Barto and Yael Niv and Matthew M. Botvinick},
      journal = {Neuron},
       volume = {71},
       number = {2},
        pages = {370-379},
         year = {2011},
     abstract = {Human behavior displays hierarchical structure: simple actions cohere into subtask sequences, which work together to accomplish overall task goals. Although the neural substrates of such hierarchy have been the target of increasing research, they remain poorly understood. We propose that the computations supporting hierarchical behavior may relate to those in hierarchical reinforcement learning (HRL), a machine-learning framework that extends reinforcement-learning mechanisms into hierarchical domains. To test this, we leveraged a distinctive prediction arising from HRL. In ordinary reinforcement learning, reward prediction errors are computed when there is an unanticipated change in the prospects for accomplishing overall task goals. HRL entails that prediction errors should also occur in relation to task subgoals. In three neuroimaging studies we observed neural responses consistent with such subgoal-related reward prediction errors, within structures previously implicated in reinforcement learning. The results reported support the relevance of HRL to the neural processes underlying hierarchical behavior.}
}
@article{RibasFernandesNEURON-11,
       author = {Ribas-Fernandes, J. J. and Solway, A. and Diuk, C. and McGuire, J. T. and Barto, A. G. and Niv, Y. and Botvinick, M. M.},
        title = {A neural signature of hierarchical reinforcement learning},
      journal = {Neuron},
         year = {2011},
       volume = {71},
       number = {2},
        pages = {370-379},
     abstract = {Human behavior displays hierarchical structure: simple actions cohere into subtask sequences, which work together to accomplish overall task goals. Although the neural substrates of such hierarchy have been the target of increasing research, they remain poorly understood. We propose that the computations supporting hierarchical behavior may relate to those in hierarchical reinforcement learning (HRL), a machine-learning framework that extends reinforcement-learning mechanisms into hierarchical domains. To test this, we leveraged a distinctive prediction arising from HRL. In ordinary reinforcement learning, reward prediction errors are computed when there is an unanticipated change in the prospects for accomplishing overall task goals. HRL entails that prediction errors should also occur in relation to task subgoals. In three neuroimaging studies we observed neural responses consistent with such subgoal-related reward prediction errors, within structures previously implicated in reinforcement learning. The results reported support the relevance of HRL to the neural processes underlying hierarchical behavior.}
}
@article{RibasNEUROSURGERY-FOCUS-10,
       author = {Guilherme Carvalhal Ribas},
        title = {The cerebral sulci and gyri},
      journal = {Neurosurg Focus},
       volume = {28},
       number = {2},
         year = {2010},
     abstract = {The aim of this study was to describe in detail the microanatomy of the cerebral sulci and gyri, clarifying the nomenclature for microneurosurgical purposes. An extensive review of the literature regarding the historical, evolutionary, embryological, and anatomical aspects pertinent to human cerebral sulci and gyri was conducted, with a special focus on microneuroanatomy issues in the field of neurosurgery. An intimate knowledge of the cerebral sulci and gyri is needed to understand neuroimaging studies, as well as to plan and execute current microneurosurgical procedures.}
}
@article{RibraultetalNATURE_REVIEWS_NEUROSCIENCE-11,
       author = {Ribrault, Claire and Sekimoto, Ken and Triller, Antoine},
        title = {From the stochasticity of molecular processes to the variability of synaptic transmission},
      journal = {Nature Reviews Neuroscience},
         year = {2011},
    publisher = {Nature Publishing Group},
       volume = {12},
        issue = {7},
        pages = {375-387},
     abstract = {The variability of the postsynaptic response following a single action potential arises from two sources: the neurotransmitter release is probabilistic, and the postsynaptic response to neurotransmitter release has variable timing and amplitude. At individual synapses, the number of molecules of a given type that are involved in these processes is small enough that the stochastic (random) properties of molecular events cannot be neglected. How the stochasticity of molecular processes contributes to the variability of synaptic transmission, its sensitivity and its robustness to molecular fluctuations has important implications for our understanding of the mechanistic basis of synaptic transmission and of synaptic plasticity.},
}
@book{RichandKnight1991,
       author = {Rich, Elaine and Knight, Kevin},
        title = {Artificial Intelligence},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1991,
}
@article{RichandWatersAIM-87,
       author = {Rich, Charles and Waters, Richard C.},
        title = {The Programmer's Apprentice: A Research Overview},
      journal = {Computer},
       volume = {21},
       number = {11},
    publisher = {IEEE Computer Society Press},
      address = {Los Alamitos, CA, USA},
         year = {1988},
        pages = {10-25},
     abstract = {The long-term goal of the Programmer's Apprentice project is to develop a theory of how expert programmers analyze, synthesize, modify, explain, specify, verify, and document programs. The authors present their vision of the Programmer's Apprentice, the principles and techniques underlying it, and their progress toward it. The primary vehicle for this exposition is three scenarios illustrating the use of the Apprentice in three phases of the programming task: implementation, design, and requirements. The first scenario is taken from a completed working prototype. The second and third scenarios are the targets for prototype systems currently under construction}
}
@article{RichandWatersCOMPUTER-88,
        title = {Automatic programming: myths and prospects},
       author = {Charlies Rich and Richard C. Waters},
      journal = {{IEEE} Computer},
       volume = {21},
        issue = {8},
         year = {1988},
     abstract = {The authors consider five common myths about automatic programming and expose the fallacies on which they rest. They attempt to provide an accurate picture of these systems in terms of what the user sees, how the system works, and what the system knows. They describe commercially available systems and discuss what is on the horizon.}
}
@book{Richards1988,
       editor = {Richards, Whitman},
        title = {Natural Computation},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1988,
}
@article{RichardsetalNATURE-NEUROSCIENCE-19,
       author = {Richards, Blake A.
and Lillicrap, Timothy P.
and Beaudoin, Philippe
and Bengio, Yoshua
and Bogacz, Rafal
and Christensen, Amelia
and Clopath, Claudia
and Costa, Rui Ponte
and de Berker, Archy
and Ganguli, Surya
and Gillon, Colleen J.
and Hafner, Danijar
and Kepecs, Adam
and Kriegeskorte, Nikolaus
and Latham, Peter
and Lindsay, Grace W.
and Miller, Kenneth D.
and Naud, Richard
and Pack, Christopher C.
and Poirazi, Panayiota
and Roelfsema, Pieter
and Sacramento, Jo{\~A}{\textsterling}o
and Saxe, Andrew
and Scellier, Benjamin
and Schapiro, Anna C.
and Senn, Walter
and Wayne, Greg
and Yamins, Daniel
and Zenke, Friedemann
and Zylberberg, Joel
and Therien, Denis
and Kording, Konrad P.},
        title = {A deep learning framework for neuroscience},
      journal = {Nature Neuroscience},
         year = {2019},
       volume = {22},
        issue = {11},
        pages = {1761-1770},
     abstract = {Systems neuroscience seeks explanations for how the brain implements a wide variety of perceptual, cognitive and motor tasks. Conversely, artificial intelligence attempts to design computational systems based on the tasks they will have to solve. In artificial neural networks, the three components specified by design are the objective functions, the learning rules and the architectures. With the growing success of deep learning, which utilizes brain-inspired architectures, these three designed components have increasingly become central to how we model, engineer and optimize complex artificial learning systems. Here we argue that a greater focus on these components would also benefit systems neuroscience. We give examples of how this optimization-based framework can drive theoretical and experimental progress in neuroscience. We contend that this principled perspective on systems neuroscience will help to generate more rapid progress.},
}
@article{Richardson-BurnsetalJNE-07,
       author = {Sarah M Richardson-Burns and Jeffrey L Hendricks and David C Martin},
        title = {Electrochemical polymerization of conducting polymers in living neural tissue},
      journal = {Journal of Neural Engineering},
       volume = 4,
        issue = 2,
        pages = {L6-L13},
         year = 2007,
     abstract = {A number of biomedical devices require extended electrical communication with surrounding tissue. Significant improvements in device performance would be achieved if it were possible to maintain communication with target cells despite the reactive, insulating scar tissue that forms at the device-tissue interface. Here, we report that the conducting polymer poly(3,4-ethylenedioxythiophene) (PEDOT) can be polymerized directly within living neural tissue resulting in an electrically conductive network that is integrated within the tissue. Nano and microscale PEDOT filaments extend out from electrode sites, presumably forming within extracellular spaces. The cloud of PEDOT filaments penetrates out into the tissue far enough that it should be possible to bypass fibrous scar tissue and contact surrounding healthy neurons. These electrically functional, diffuse conducting polymer networks grown directly within tissue signify a new paradigm for creating soft, low impedance implantable electrodes.}
}
@inproceedings{RichardsonandDomingos02,
       author = {Mathew Richardson and Pedro Domingos},
        title = {The Intelligent Surfer: Probabilistic Combination of Link and Content Information in {P}age{R}ank},
    booktitle = {Advances in Neural Information Processing Systems 14},
    publisher = {MIT Press},
         year = 2002,
}
@inproceedings{RichiardietalICPR-10,
       author = {Richiardi, J. and Van De Ville, D. and Riesen, K. and Bunke, H.},
    booktitle = {International Conference on Pattern Recognition},
        title = {Vector Space Embedding of Undirected Graphs with Fixed-cardinality Vertex Sequences for Classification},
         year = 2010,
        pages = {902-905},
     abstract = {Simple weighted undirected graphs with a fixed number of vertices and fixed vertex orderings can be used to represent data and patterns in a wide variety of scientific and engineering domains. Classification of such graphs by existing graph matching methods perform rather poorly because they do not exploit their specificity. As an alternative, methods relying on vector-space embedding hold promising potential. We propose two such techniques that can be deployed as a front-end for any pattern recognition classifiers: one has low computational cost but generates high-dimensional spaces, while the other is more computationally demanding but can yield relatively low-dimensional vector space representations. We show experimental results on an fMRI brain state decoding task and discuss the shortfalls of graph edit distance for the type of graph under consideration.},
}
@article{RichterandGjorgjievaBIORXIV-17,
       author = {Richter, Leonidas M. A. and Gjorgjieva, Julijana},
        title = {Understanding neural circuit development through theory and models},
      journal = {bioRxiv},
         year = {2017},
    publisher = {Cold Spring Harbor Laboratory},
     abstract = {How are neural circuits organized and tuned to achieve stable function and produce robust behavior? The organization process begins early in development and involves a diversity of mechanisms unique to this period. We summarize recent progress in theoretical neuroscience that has substantially contributed to our understanding of development at the single neuron, synaptic and network level. We go beyond classical models of topographic map formation, and focus on the generation of complex spatiotemporal activity patterns, their role in refinements of particular circuit features, and the emergence of functional computations. Aided by the development of novel quantitative methods for data analysis, theoretical and computational models have enabled us to test the adequacy of specific assumptions, explain experimental data and propose testable hypotheses. With the accumulation of larger data sets, theory and models will likely play an even more important role in understanding the development of neural circuits.},
}
@article{RickgaueretalNATURE-NEUROSCIENCE-14,
       author = {Rickgauer, J. P. and Deisseroth, K. and Tank, D. W.},
        title = {Simultaneous cellular-resolution optical perturbation and imaging of place cell firing fields},
      journal = {Nature Neuroscience},
         year = {2014},
       volume = {17},
       number = {12},
        pages = {1816-1824},
     abstract = {Linking neural microcircuit function to emergent properties of the mammalian brain requires fine-scale manipulation and measurement of neural activity during behavior, where each neuron's coding and dynamics can be characterized. We developed an optical method for simultaneous cellular-resolution stimulation and large-scale recording of neuronal activity in behaving mice. Dual-wavelength two-photon excitation allowed largely independent functional imaging with a green fluorescent calcium sensor (GCaMP3, \hmlambda{} = 920 \hmplusminus{} 6 nm) and single-neuron photostimulation with a red-shifted optogenetic probe (C1V1, \hmlambda{} = 1,064 \hmplusminus{} 6 nm) in neurons coexpressing the two proteins. We manipulated task-modulated activity in individual hippocampal CA1 place cells during spatial navigation in a virtual reality environment, mimicking natural place-field activity, or 'biasing', to reveal subthreshold dynamics. Notably, manipulating single place-cell activity also affected activity in small groups of other place cells that were active around the same time in the task, suggesting a functional role for local place cell interactions in shaping firing fields.}
}
@article{RidderinkhofetalSCI-04,
        title = {The Role of Medial Frontal Cortex in Cognitive Control},
       author = {K. Ridderinkhof and Markus Ullsperger and Eveline A. Crone and Sander Nieuwenhuis},
      journal = {Science},
       volume = 306,
        issue = 5695,
        pages = {443-446},
         year = 2004,
}
@inproceedings{RidgewayetalAISTAT-99,
       author = {Greg Ridgeway and David Madigan and Thomas Richardson},
        title = {Boosting Methodology for Regression Problems},
    booktitle = {The Seventh International Workshop on Artificial Intelligence and Statistics},
    publisher = {Morgan Kaufmann},
         year = 1999,
        pages = {152-161},
}
@article{RidkeretalLANCET-12,
        title = {Cardiovascular benefits and diabetes risks of statin therapy in 
              primary prevention: an analysis from the {JUPITER} trial},
       author = {Ridker, Paul M and Pradhan, Aruna and MacFadyen, Jean G and Libby, Peter and Glynn, Robert J},
      journal = {The Lancet},
       volume = 380,
        issue = 9841,
         year = 2012,
        pages = {565-571},
}
@book{Ridley1996,
      author = {Matt Ridley},
       title = {The Origins of Virtue: Human Instincts and the Evolution of Cooperation},
   publisher = {Viking},
     address = {London},
        year = 1996,
}
@book{Ridley1999,
      author = {Matt Ridley},
       title = {Genome: The Autobiography of a Species in 23 Chapters},
   publisher = {Harper-Collins},
     address = {New York, NY},
        year = 1999,
}
@book{Ridley2015,
       author = {Matt Ridley},
        title = {The Evolution of Everything: How Ideas Emerge},
    publisher = {HarperCollins Publishers},
         year = 2015,
}
@book{Ridley2093,
      author = {Matt Ridley},
       title = {The Red Queen: Sex and the Evolution of Human Nature},
   publisher = {Viking},
     address = {London},
        year = 1993,
}
@article{RiedmilleretalCoRR-18,
       author = {Martin A. Riedmiller and Roland Hafner and Thomas Lampe and Michael Neunert and Jonas Degrave and Tom Van de Wiele and Volodymyr Mnih and Nicolas Heess and Jost Tobias Springenberg},
        title = {Learning by Playing - Solving Sparse Reward Tasks from Scratch},
      journal = {CoRR},
       volume = {arXiv:1802.10567},
         year = {2018},
     abstract = {We propose Scheduled Auxiliary Control (SAC-X), a new learning paradigm in the context of Reinforcement Learning (RL). SAC-X enables learning of complex behaviors - from scratch - in the presence of multiple sparse reward signals. To this end, the agent is equipped with a set of general auxiliary tasks, that it attempts to learn simultaneously via off-policy RL. The key idea behind our method is that active (learned) scheduling and execution of auxiliary policies allows the agent to efficiently explore its environment - enabling it to excel at sparse reward RL. Our experiments in several challenging robotic manipulation settings demonstrate the power of our approach.}
}
@article{RiesenhuberandPoggioNN-99,
       author = {Riesenhuber, M. and Poggio, T.},
        title = {Hierarchical models of object recognition in cortex},
      journal = {Nature Neuroscience},
       volume = 2,
        issue = 11,
         year = 1999,
        pages = {1019-1025},
}
@article{RijntjesetalFiEN-12,
       author = {Rijntjes, Michel and Weiller, Cornelius and Bormann, Tobias and Musso, Mariachristina},
        title = {The dual loop model: its relation to language and other modalities},
      journal = {Frontiers in Evolutionary Neuroscience},
       volume = {4},
        pages = {9},
         year = {2012},
     abstract = {The current neurobiological consensus of a general dual-loop system scaffolding human and primate brains gives evidence that the dorsal and ventral connections subserve similar functions, independent of the modality and species. However, most current commentators agree that, although bees dance and chimpanzees grunt, these systems of communication differ qualitatively from human language. So why is language in humans unique? We discuss anatomical differences between humans and other animals, the meaning of lesion studies in patients, the role of inner speech, and compare functional imaging studies in language with other modalities in respect to the dual loop model. These aspects might be helpful for understanding what kind of biological system the language faculty is, and how it relates to other systems in our own species and others.}
}
@inproceedings{RimeyIUW-92,
      author = {Rimey, Raymond D.},
       title = {Where to Look Next Using a {B}ayes Net: An Overview},
   booktitle = {Proceedings of the DARPA Image Understanding Workshop},
    location = {San Diego, California},
organization = {DARPA},
       pages = {927-932},
        year = 1992,
}
@article{RimeyandBrownIJCV-91,
     author = {Rimey, Raymond D. and Brown, Christopher M.},
      title = {Controlling Eye Movements with Hidden {M}arkov Models},
    journal = {International Journal of Computer Vision},
  publisher = {Kluwer Academic Publishers},
     volume = 7,
      issue = 1,
      pages = {47-65},
       year = 1991,
}
@article{RingachetalJV-02,
     author = {D. L. Ringach and M. J. Hawken and R. Shapley},
      title = {Receptive field structure of neurons in monkey primary visual
               cortex revealed by stimulation with natural iamge sequences},
    journal = {Journal of Vision},
       year = 2002,
     volume = 2,
      pages = {12-24},
}
@book{RiordanandHoddeson1977,
        title = {Crystal Fire: The Invention of the Transistor and the Birth of the Information Age},
       author = {Michael Riordan and Lillian Hoddeson},
         year = 1977,
    publisher = {W.W. Norton \& Co.},
}
@book{Ripley96,
     author = {Ripley, Brian D.},
      title = {Pattern Recognition and Neural Networks},
  publisher = {Cambridge University Press},
    address = {New York},
       year = 1996,
}
@article{Rissanen78,
       author = {Rissanen, J.},
        title = {Modeling by Shortest Data Description},
      journal = {Automatica},
       volume = 14,
        pages = {445-471},
         year = 1978,
}
@book{Rissanen89,
     author = {Jorma Rissanen},
      title = {Stochastic Complexity in Statistical Inquiry},
  publisher = {World Scientific},
    address = {River Edge, New Jersey},
       year = 1989,
}
@article{RissanenIEEE-83,
     author = {Jorma Rissanen},
      title =  {A Universal Data Compression System},
    journal = {{IEEE} Transactions on Information Theory},
     volume = 29,
      issue = 5,
       year = 1983,
      pages = {656-664},
}
@article{RitteretalBIORXIV-18,
       author = {Ritter, Samuel and Wang, Jane X. and Kurth-Nelson, Zeb and Botvinick, Matthew M.},
        title = {Episodic Control as Meta-Reinforcement Learning},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Laboratory},
         year = {2018},
     abstract = {Recent research has placed episodic reinforcement learning (RL) alongside model-free and model-based RL on the list of processes centrally involved in human reward-based learning. In the present work, we extend the unified account of model-free and model-based RL developed by Wang et al. (2018) to further integrate episodic learning. In this account, a generic model-free "meta-learner" learns to deploy and coordinate among all of these learning algorithms. The meta-learner learns through brief encounters with many novel tasks, so that it learns to learn about new tasks. We show that when equipped with an episodic memory system inspired by theories of reinstatement and gating, the meta-learner learns to use the episodic and model-based learning algorithms observed in humans in a task designed to dissociate among the influences of various learning strategies. We discuss implications and predictions of the model.},
}
@inproceedings{RivestAAAI-88,
       author = {Rivest, Ronald L. and Robert Sloan},
        title = {Learning Complicated Concepts Reliably and Usefully},
    booktitle = {Proceedings AAAI-88},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
        pages = {635-640},
         year = 1988,
}
@inproceedings{RivestandSchapireFOCS-87,
       author = {Rivest, Ronald L. and Robert E. Schapire},
        title = {Diversity-Based Inference of Finite Automata},
    booktitle = {Proceedings of the Twenty Eighth Annual Symposium on Foundations
                of Computer Science},
        pages = {78-87},
         year = 1987,
}
@inproceedings{RivestandSchapireSTOC-89,
       author = {Rivest, Ronald L. and Robert E. Schapire},
        title = {Inference of Finite Automata Using Homing Sequences},
    booktitle = {Proceedings of the Twenty First Annual {ACM} Symposium on
                Theoretical Computing},
        pages = {411-420},
         year = 1989,
}
@article{RivestandSloanIC-94,
       author = {R. L. Rivest and R. Sloan},
        title = {A Formal Model of Hierarchical Concept Learning},
      journal = {Information and Computation},
       volume = 114,
        issue = 1,
        pages = {88-114},
         year = 1994,
}
@incollection{RizzolattiandRozziNL-16,
       author = {Giacomo Rizzolatti and Stefano Rozzi},
        title = {Motor Cortex and Mirror System in Monkeys and Humans},
       editor = {Hickok, Gregory  and Small, Steven L.},
    booktitle = {Neurobiology of Language},
    publisher = {Academic Press},
      address = {San Diego},
         year = {2016},
        pages = {59-72},
     abstract = {The view of the functions of the motor system has radically changed in past years. It is now clear that the motor system not only is a "producer" of movements but also is involved in cognitive functions. In this chapter we review the anatomical and functional organization of the cortical motor system in monkeys and humans with particular emphasis on those areas that are endowed of the mirror mechanism and involved in communication. We discuss the possible evolutionary link between basic unintentional communication and voluntary communication. We conclude examining the hypothesis that purports that motor activation is involved in phonemes and understanding of action words.}
}
@book{RizzolattietalMIRRORS-07,
       author = {Giacomo Rizzolatti and Corrado Sinigaglia and Frances Anderson},
        title = {Mirrors in the Brain How Our Minds Share Actions, Emotions, and Experience},
    publisher = {Oxford University Press},
      address = {Oxford, UK},
         year = 2007,
}
@book{Robert86,
       author = {Robert, F.},
        title = {Discrete Iterations: A Metric Study},
    publisher = {Springer-Verlag},
      address = {New York},
         year = 1986,
}
@article{RobertsetalJCC-13,
       author = {Roberts, Elijah and Stone, John E. and Luthey-Schulten, Zaida},
        title = {{Lattice Microbes}: High-performance stochastic simulation method for the reaction-diffusion master equation},
      journal = {Journal of Computational Chemistry},
    publisher = {Wiley Subscription Services, Inc., A Wiley Company},
       volume = 34,
        issue = 3,
         year = 2013,
        pages = {245-255},
     abstract = {Spatial stochastic simulation is a valuable technique for studying reactions in biological systems. With the availability of high-performance computing (HPC), the method is poised to allow integration of data from structural, single-molecule and biochemical studies into coherent computational models of cells. Here, we introduce the Lattice Microbes software package for simulating such cell models on HPC systems. The software performs either well-stirred or spatially resolved stochastic simulations with approximated cytoplasmic crowding in a fast and efficient manner. Our new algorithm efficiently samples the reaction-diffusion master equation using NVIDIA graphics processing units and is shown to be two orders of magnitude faster than exact sampling for large systems while maintaining an accuracy of \~0.1\%. Display of cell models and animation of reaction trajectories involving millions of molecules is facilitated using a plug-in to the popular VMD visualization platform. The Lattice Microbes software is open source and available for download at: http://www.scs.illinois.edu/schulten/lm},
}
@article{RobieetalCELL-17,
       author = {Alice A. Robie and Jonathan Hirokawa and Austin W. Edwards and Lowell A. Umayam and Allen Lee and Mary L. Phillips and Gwyneth M. Card and Wyatt Korff and Gerald M. Rubin and Julie H. Simpson and Michael B. Reiser and Kristin Branson},
        title = {Mapping the Neural Substrates of Behavior},
      journal = {Cell},
       volume = {170},
       number = {2},
        pages = {393406},
         year = {2017},
     abstract = {Assigning behavioral functions to neural structures has long been a central goal in neuroscience and is a necessary first step toward a circuit-level understanding of how the brain generates behavior. Here, we map the neural substrates of locomotion and social behaviors for Drosophila melanogaster using automated machine-vision and machine-learning techniques. From videos of 400,000 flies, we quantified the behavioral effects of activating 2,204 genetically targeted populations of neurons. We combined a novel quantification of anatomy with our behavioral analysis to create brain-behavior correlation maps, which are shared as browsable web pages and interactive software. Based on these maps, we generated hypotheses of regions of the brain causally related to sensory processing, locomotor control, courtship, aggression, and sleep. Our maps directly specify genetic tools to target these regions, which we used to identify a small population of neurons with a role in the control of walking.}
}
@article{RobinsCONNECTION-SCIENCE-95,
       author = {Anthony Robins},
        title = {Catastrophic Forgetting, Rehearsal and Pseudorehearsal},
      journal = {Connection Science},
       volume = {7},
       number = {2},
        pages = {123-146},
         year = {1995},
    publisher = {Taylor \& Francis},
     abstract = {This paper reviews the problem of catastrophic forgetting (the loss or disruption of previously learned information when new information is learned) in neural networks, and explores rehearsal mechanisms (the retraining of some of the previously learned information as the new information is added) as a potential solution. We replicate some of the experiments described by Ratcliff (1990), including those relating to a simple 'recency' based rehearsal regime. We then develop further rehearsal regimes which are more effective than recency rehearsal. In particular, 'sweep rehearsal' is very successful at minimizing catastrophic forgetting. One possible limitation of rehearsal in general, however, is that previously learned information may not be available for retraining. We describe a solution to this problem, 'pseudorehearsal', a method which provides the advantages of rehearsal without actually requiring any access to the previously learned information (the original training population) itself. We then suggest an interpretation of these rehearsal mechanisms in the context of a function approximation based account of neural network learning. Both rehearsal and pseudorehearsal may have practical applications, allowing new information to be integrated into an existing network with minimum disruption of old information. }
}
@article{RobinsonJACM-65,
       author = {Robinson, J. A.},
        title = {A Machine-Oriented Logic Based on the Resolution Principle},
      journal = {Journal of the {ACM}},
       volume = 12,
         year = 1965,
        pages = {23-41},
}
@book{Rockland2016axonsandbrains,
       author = {Kathleen S. Rockland},
        title = {Axons and Brain Architecture},
    publisher = {Academic Press},
      address = {San Diego},
         year = {2016},
}
@misc{RodriguezALNP-17,
        title = {Advanced Natural Language Processing},
       author = {Horacio Rodr\`{i}guez},
         year = {2017},
 howpublished = {Facultat d'Informatica Barcelona, Universitat Politecnica Catalunya},
          url = {http://www.lsi.upc.edu/~ageno/anlp/},
}
@article{RodriguezetalJCN-04,
    author = {Rodriguez, A. and Whitson, J. and Granger, R.},
     title = {Derivation and Analysis of Basic Computational Operations of Thalamocortical Circuits},
   journal = {Journal of Cognitive Neuroscience},
    volume = 16,
     issue = 5,
 publisher = {MIT Press},
   address = {Cambridge, MA},
     pages = {856-877},
      year = 2004,
}
@article{RoeNETWORK-NEUROSCIENCE-19,
       author = {Roe, Anna Wang},
        title = {Columnar connectome: toward a mathematics of brain function},
      journal = {Network neuroscience},
         year = {2019},
    publisher = {MIT Press},
       volume = {3},
        issue = {3},
        pages = {779-791},
     abstract = {Understanding brain networks is important for many fields, including neuroscience, psychology, medicine, and artificial intelligence. To address this fundamental need, there are multiple ongoing connectome projects in the United States, Europe, and Asia producing brain connection maps with resolutions at macro- and microscales. However, still lacking is a mesoscale connectome. This viewpoint (1) explains the need for a mesoscale connectome in the primate brain (the columnar connectome), (2) presents a new method for acquiring such data rapidly on a large scale, and (3) proposes how one might use such data to achieve a mathematics of brain function.},
}
@article{RogersandMcClellandBBS-08,
       author = {Rogers, Timothy T. and McClelland, James L.},
        title = {A simple model from a powerful framework that spans levels of analysis},
       volume = {31},
       number = {6},
      journal = {Behavioral and Brain Sciences},
    publisher = {Cambridge University Press},
         year = {2008},
        pages = {729-749},
     abstract = {The commentaries reflect three core themes that pertain not just to our theory, but to the enterprise of connectionist modeling more generally. The first concerns the relationship between a cognitive theory and an implemented computer model. Specifically, how does one determine, when a model departs from the theory it exemplifies, whether the departure is a useful simplification or a critical flaw? We argue that the answer to this question depends partially upon the model's intended function, and we suggest that connectionist models have important functions beyond the commonly accepted goals of fitting data and making predictions. The second theme concerns perceived in-principle limitations of the connectionist approach to cognition, and the specific concerns these perceived limitations raise for our theory. We argue that the approach is not in fact limited in the ways our critics suggest. One common misconception, that connectionist models cannot address abstract or relational structure, is corrected through new simulations showing directly that such structure can be captured. The third theme concerns the relationship between parallel distributed processing (PDP) models and structured probabilistic approaches. In this case we argue that there the difference between the approaches is not merely one of levels. Our PDP approach differs from structured statistical approaches at all of Marr's levels, including the characterization of the goals of cognitive computations, and of the representations and algorithms used.}
}
@article{RogersetalCoRR-20,
        title = {{A Primer in BERTology: What we know about how BERT works}},
       author = {Anna Rogers and Olga Kovaleva and Anna Rumshisky},
         year = {2020},
      journal = {CoRR},
       volume = {arXiv:2002.12327},
     abstract = {Transformer-based models have pushed state of the art in many areas of NLP, but our understanding of what is behind their success is still limited. This paper is the first survey of over 150 studies of the popular BERT model. We review the current state of knowledge about how BERT works, what kind of information it learns and how it is represented, common modifications to its training objectives and architecture, the overparameterization issue and approaches to compression. We then outline directions for future research.}
}
@misc{RohlfingCMTK-11,
       author = {Torsten Rohlfing},
        title = {User Guide to The Computational Morphometry Toolkit},
         year = {2011},
 howpublished = {http://www.nitrc.org/projects/cmtk/},
     abstract = {This guide is intended as a very brief introduction of the main tools in the Computational Morphometry Toolkit (CMTK), which is available in source code and as pre-compiled binaries from http://www.nitrc.org/projects/cmtk/. The target audience of this document are CMTK users, who might use this document as a reference to the most common processing tasks, and prospective users, who may find this information useful to determine whether CMTK provides functionality that they can use. We focus in particular on a simplified workflow for deformation morphometry studies based on magnetic resonance images: DICOM conversion, artifact correction, affine and nonlinear image registration, reformatting, Jacobian determinant map generation, and statistical hypothesis testing.}
}
@article{RoknietalNEURON-07,
       author = {Rokni, Uri and Richardson, Andrew G. and Bizzi, Emilio and Seung, H. Sebastian},
        title = {Motor Learning with Unstable Neural Representations},
      journal = {Neuron},
    publisher = {Elsevier},
       volume = {54},
        issue = {4},
         year = {2007},
        pages = {653-666},
     abstract = {It is often assumed that learning takes place by changing an otherwise stable neural representation. To test this assumption, we studied changes in the directional tuning of primate motor cortical neurons during reaching movements performed in familiar and novel environments. During the familiar task, tuning curves exhibited slow random drift. During learning of the novel task, random drift was accompanied by systematic shifts of tuning curves. Our analysis suggests that motor learning is based on a surprisingly unstable neural representation. To explain these results, we propose that motor cortex is a redundant neural network, i.e., any single behavior can be realized by multiple configurations of synaptic strengths. We further hypothesize that synaptic modifications underlying learning contain a random component, which causes wandering among synaptic configurations with equivalent behaviors but different neural representations. We use a simple model to explore the implications of these assumptions.},
}
@article{RollsARP-01,
       author = {Edmund Rolls},
        title = {Memory systems in the brain},
      journal = {Annual Review Psychology},
       volume = {51},
         year = {2001},
        pages = {599-630},
     abstract = {The operation of different brain systems involved in different types of memory is described. One is a system in the primate orbitofrontal cortex and amygdala involved in representing rewards and punishers, and in learning stimulus-reinforcer associations. This system is involved in emotion and motivation. A second system in the temporal cortical visual areas is involved in learning invariant representations of objects. A third system in the hippocampus is implicated in episodic memory and in spatial function. Fourth, brain systems in the frontal and temporal cortices involved in short term memory are described. The approach taken provides insight into the neuronal operations that take place in each of these brain systems, and has the aim of leading to quantitative biologically plausible neuronal network models of how each of these memory systems actually operates.}
}
@article{RollsFiSN-13,
       author = {Rolls, Edmund},
        title = {The mechanisms for pattern completion and pattern separation in the hippocampus},
      journal = {Frontiers in Systems Neuroscience},
       volume = {7},
        pages = {74},
         year = {2013},
      abstract = {The mechanisms for pattern completion and pattern separation are described in the context of a theory of hippocampal function in which the hippocampal CA3 system operates as a single attractor or autoassociation network to enable rapid, one-trial, associations between any spatial location (place in rodents, or spatial view in primates) and an object or reward, and to provide for completion of the whole memory during recall from any part. The factors important in the pattern completion in CA3 together with a large number of independent memories stored in CA3 include a sparse distributed representation which is enhanced by the graded firing rates of CA3 neurons, representations that are independent due to the randomizing effect of the mossy fibers, heterosynaptic long-term depression as well as long-term potentiation in the recurrent collateral synapses, and diluted connectivity to minimize the number of multiple synapses between any pair of CA3 neurons which otherwise distort the basins of attraction. Recall of information from CA3 is implemented by the entorhinal cortex perforant path synapses to CA3 cells, which in acting as a pattern associator allow some pattern generalization. Pattern separation is performed in the dentate granule cells using competitive learning to convert grid-like entorhinal cortex firing to place-like fields. Pattern separation in CA3, which is important for completion of any one of the stored patterns from a fragment, is provided for by the randomizing effect of the mossy fiber synapses to which neurogenesis may contribute, by the large number of dentate granule cells each with a sparse representation, and by the sparse independent representations in CA3. Recall to the neocortex is achieved by a reverse hierarchical series of pattern association networks implemented by the hippocampo-cortical backprojections, each one of which performs some pattern generalization, to retrieve a complete pattern of cortical firing in higher-order cortical areas.}
}
@article{RollsHIPPOCAMPUS-00,
       author = {Rolls, Edmund T.},
        title = {Hippocampo-cortical and cortico-cortical backprojections},
      journal = {Hippocampus},
       volume = {10},
       number = {4},
         year = {2000},
        pages = {380-388},
     abstract = {First, the information represented in the primate hippocampus, and what is computed by the primate hippocampus, are considered. Then a theory is described of how the information represented in the hippocampus is able to influence the cerebral cortex by a hierarchy of hippocampo-cortical and cortico-cortical backprojection stages. The recalled backprojected information in the cerebral neocortex could then be used by the neocortex as part of memory recall, including that required in spatial working memory; to influence the processing that each cortical stage performs based on its forward inputs; to influence the formation of long-term memories; and/or in the selection of appropriate actions.},
}
@article{RollsNN-13,
       author = {Rolls, Edmund},
        title = {The mechanisms for pattern completion and pattern separation in the hippocampus},
      journal = {Frontiers in Systems Neuroscience},
       volume = 7,
         year = 2013,
        issue = 74,
     abstract = {The mechanisms for pattern completion and pattern separation are described in the context of a theory of hippocampal function in which the hippocampal CA3 system operates as a single attractor or autoassociation network to enable rapid, one-trial, associations between any spatial location (place in rodents, or spatial view in primates) and an object or reward, and to provide for completion of the whole memory during recall from any part. The factors important in the pattern completion in CA3 together with a large number of independent memories stored in CA3 include a sparse distributed representation which is enhanced by the graded firing rates of CA3 neurons, representations that are independent due to the randomizing effect of the mossy fibers, heterosynaptic long-term depression as well as long-term potentiation in the recurrent collateral synapses, and diluted connectivity to minimize the number of multiple synapses between any pair of CA3 neurons which otherwise distort the basins of attraction. Recall of information from CA3 is implemented by the entorhinal cortex perforant path synapses to CA3 cells, which acting as a pattern associator allow some pattern generalization. Pattern separation is performed in the dentate granule cells using competitive learning to convert grid-like entorhinal cortex firing to place-like fields. Pattern separation in CA3, which is important for completion of any one of the stored patterns from a fragment, is provided for by the randomizing effect of the mossy fiber synapses to which neurogenesis may contribute, by the large number of dentate granule cells each with a sparse representation, and by the sparse independent representations in CA3. Recall to the neocortex is achieved by a reverse hierarchical series of pattern association networks implemented by the hippocampo-cortical backprojections, each one of which performs some pattern generalization, to retrieve a complete pattern of cortical firing in higher-order cortical areas}
}
@article{RollsandStringer07,
       author = {E. T. Rolls and S. M. Stringer},
         year = 2007,
        title = {Invariant global motion recognition in the dorsal visual system: a unifying theory},
      journal = {Neural Computation},
       volume = 19,
        pages = {139-169},
}
@article{RollsetalNN-97,
       author = {Edmund T. Rolls and Alesandro Treves and David Foster and Conrado Perez-Vicente},
        title = {Simulation studies of the \{CA3\} hippocampal subfield modelled as an attractor neural network},
      journal = {Neural Networks},
       volume = 10,
        issue = 9,
        pages = {1559-1569},
         year = 1997,
     abstract = {Real neuronal networks in the brain utilize networks of neurons with graded not binary firing rates. A theoretical analysis of the operation of autoassociative networks with neurons with graded firing rates has therefore been developed. The present simulation study was performed in order to investigate the operation of such a network with values for the asymmetric diluted neuronal connectivity typical of some brain regions such as the hippocampus, which are outside the range to which the theoretical analysis strictly applies. We report that, in line with theoretical predictions, the amount of information that can be retrieved is relatively independent of the resolution of the stored patterns (binary, ternary, decimal, or fifty-fold). The implication of this is that if the network stores many graded patterns, which it can, then the retrieval quality of each of the patterns becomes low. The implications of this trade-off between the number of patterns stored and the retrieval quality of each pattern when graded firing rates are stored for understanding the operation of networks in the hippocampus are considered.}
}
@article{RomeraetalNATURE-18,
       author = {Romera, Miguel and Talatchian, Philippe and Tsunegi, Sumitoand Abreu Araujo, Flavioand Cros, Vincentand Bortolotti, Paoloand Trastoy, Juanand Yakushiji, Kayand Fukushima, Akioand Kubota, Hitoshiand Yuasa, Shinjiand Ernoult, Maxenceand Vodenicarevic, Damirand Hirtzlin, Tifennand Locatelli, Nicolasand Querlioz, Damienand Grollier, Julie},
        title = {Vowel recognition with four coupled spin-torque nano-oscillators},
      journal = {Nature},
       volume = {563},
        issue = {7730},
         year = {2018},
        pages = {230-234},
     abstract = {In recent years, artificial neural networks have become the flagship algorithm of artificial intelligence1. In these systems, neuron activation functions are static, and computing is achieved through standard arithmetic operations. By contrast, a prominent branch of neuroinspired computing embraces the dynamical nature of the brain and proposes to endow each component of a neural network with dynamical functionality, such as oscillations, and to rely on emergent physical phenomena, such as synchronization[2-6], for solving complex problems with small networks[7-11]. This approach is especially interesting for hardware implementations, because emerging nanoelectronic devices can provide compact and energy-efficient nonlinear auto-oscillators that mimic the periodic spiking activity of biological neurons[12-16]. The dynamical couplings between oscillators can then be used to mediate the synaptic communication between the artificial neurons. One challenge for using nanodevices in this way is to achieve learning, which requires fine control and tuning of their coupled oscillations[17]; the dynamical features of nanodevices can be difficult to control and prone to noise and variability18. Here we show that the outstanding tunability of spintronic nano-oscillators-that is, the possibility of accurately controlling their frequency across a wide range, through electrical current and magnetic field-can be used to address this challenge. We successfully train a hardware network of four spin-torque nano-oscillators to recognize spoken vowels by tuning their frequencies according to an automatic real-time learning rule. We show that the high experimental recognition rates stem from the ability of these oscillators to synchronize. Our results demonstrate that non-trivial pattern classification tasks can be achieved with small hardware neural networks by endowing them with nonlinear dynamical features such as oscillations and synchronization.},
}
@article{RomeroetalCoRR-14,
       author = {Adriana Romero and Nicolas Ballas and Samira Ebrahimi Kahou and Antoine Chassang and Carlo Gatta and Yoshua Bengio},
        title = {{FitNets}: Hints for Thin Deep Nets},
      journal = {CoRR},
       volume = {arXiv:1412.6550},
         year = 2014,
     abstract = {While depth tends to improve network performances, it also makes gradient-based training more difficult since deeper networks tend to be more non-linear. The recently proposed knowledge distillation approach is aimed at obtaining small and fast-to-execute models, and it has shown that a student network could imitate the soft output of a larger teacher network or ensemble of networks. In this paper, we extend this idea to allow the training of a student that is deeper and thinner than the teacher, using not only the outputs but also the intermediate representations learned by the teacher as hints to improve the training process and final performance of the student. Because the student intermediate hidden layer will generally be smaller than the teacher's intermediate hidden layer, additional parameters are introduced to map the student hidden layer to the prediction of the teacher hidden layer. This allows one to train deeper students that can generalize better or run faster, a trade-off that is controlled by the chosen student capacity. For example, on CIFAR-10, a deep student network with almost 10.4 times less parameters outperforms a larger, state-of-the-art teacher network.},
}
@inproceedings{RomeroetalWWW-11,
       author = {Romero, Daniel M. and Meeder, Brendan and Kleinberg, Jon},
        title = {Differences in the mechanics of information diffusion across topics: idioms, political hashtags, and complex contagion on twitter},
    booktitle = {Proceedings of the 20th international conference on World wide web},
     location = {Hyderabad, India},
    publisher = {ACM},
      address = {New York, NY, USA},
        pages = {695-704},
         year = 2011,
}
@article{RomoffetalCoRR-19,
       author = {Joshua Romoff and Peter Henderson and Ahmed Touati and Yann Ollivier and  Emma Brunskill and Joelle Pineau},
        title = {Separating value functions across time-scales},
      journal = {CoRR},
       volume = {arXiv:1902.01883},
         year = {2019},
     abstract = {In many finite horizon episodic reinforcement learning (RL) settings, it is desirable to optimize for the undiscounted return - in settings like Atari, for instance, the goal is to collect the most points while staying alive in the long run. Yet, it may be difficult (or even intractable) mathematically to learn with this target. As such, temporal discounting is often applied to optimize over a shorter effective planning horizon. This comes at the risk of potentially biasing the optimization target away from the undiscounted goal. In settings where this bias is unacceptable - where the system must optimize for longer horizons at higher discounts - the target of the value function approximator may increase in variance leading to difficulties in learning. We present an extension of temporal difference (TD) learning, which we call TD(\Delta), that breaks down a value function into a series of components based on the differences between value functions with smaller discount factors. The separation of a longer horizon value function into these components has useful properties in scalability and performance. We discuss these properties and show theoretic and empirical improvements over standard TD learning in certain settings.}
}
@inproceedings{RonetalCOLT-94,
       author = {Ron, Dana and Singer, Yoram and Tishby, Naftali},
        title = {Learning Probabilistic Automata with Variable Memory Length},
    booktitle = {Proceedings of the Seventh Annual {ACM} Conference on Computational Learning Theory},
        pages = {35-46},
         year = 1994,
}
@inproceedings{RonetalCOLT-95,
       author = {Ron, Dana and Singer, Yoram and Tishby, Naftali},
        title = {On the Learning and Usage of Acyclic Probabilistic Finite Automata},
    booktitle = {Proceedings of the Eighth Annual {ACM} Conference on Computational Learning Theory},
        pages = {31-40},
         year = 1995,
}
@article{RonetalML-96,
       author = {Dana Ron and Yoram Singer and Naftali Tishby},
        title = {The Power of Amnesia: Learning Probabilistic Automata with Variable Memory Length},
      journal = {Machine Learning},
       volume = 25,
        issue = {2-3},
        pages = {117-149},
         year = 1996,
}
@article{RonnebergeretalCoRR-15,
        title = {{U-Net}: {C}onvolutional Networks for Biomedical Image Segmentation},
       author = {Olaf Ronneberger and Philipp Fischer and Thomas Brox},
      journal = {CoRR},
       volume = {arXiv:1505.04597},
         year = 2015,
     abstract = {There is large consent that successful training of deep networks requires many thousand annotated training samples. In this paper, we present a network and training strategy that relies on the strong use of data augmentation to use the available annotated samples more efficiently. The architecture consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. We show that such a network can be trained end-to-end from very few images and outperforms the prior best method (a sliding-window convolutional network) on the ISBI challenge for segmentation of neuronal structures in electron microscopic stacks. Using the same network trained on transmitted light microscopy images (phase contrast and DIC) we won the ISBI cell tracking challenge 2015 in these categories by a large margin. Moreover, the network is fast. Segmentation of a 512x512 image takes less than a second on a recent GPU. The full implementation (based on Caffe) and the trained networks are available at this link address: http://lmb.informatik.uni-freiburg.de/people/ronneber/u-net}
}
@article{RoodbeenandVanHestBIOESSAYS-09,
        title = {Synthetic cells and organelles: compartmentalization strategies},
       author = {Roodbeen, R. and van Hest, J.C.M.},
      journal = {Bioessays},
         year = 2009,
     abstract = {The recent development of RNA replicating protocells and capsules that enclose complex biosynthetic cascade reactions are encouraging signs that we are gradually getting better at mastering the complexity of biological systems. The road to truly cellular compartments is still very long, but concrete progress is being made. Compartmentalization is a crucial natural methodology to enable control over biological processes occurring within the living cell. In fact, compartmentalization has been considered by some theories to be instrumental in the creation of life. With the advancement of chemical biology, artificial compartments that can mimic the cell as a whole, or that can be regarded as cell organelles, have recently received much attention. The membrane between the inner and outer environment of the compartment has to meet specific requirements, such as semi-permeability, to allow communication and molecular transport over the border. The membrane can either be built from natural constituents or from synthetic polymers, introducing robustness to the capsule.}
}
@article{RosatietalEVOLUTIONARY-PSYCHOLOGY-14,
       author = {Alexandra G. Rosati and Victoria Wobber and Kelly Hughes and Laurie R. Santos},
        title = {Comparative Developmental Psychology: How is Human Cognitive Development Unique?},
      journal = {Evolutionary Psychology},
       volume = {12},
       number = {2},
         year = {2014},
        pages = {147470491401200211},
     abstract = { The fields of developmental and comparative psychology both seek to illuminate the roots of adult cognitive systems. Developmental studies target the emergence of adult cognitive systems over ontogenetic time, whereas comparative studies investigate the origins of human cognition in our evolutionary history. Despite the long tradition of research in both of these areas, little work has examined the intersection of the two: the study of cognitive development in a comparative perspective. In the current article, we review recent work using this comparative developmental approach to study non-human primate cognition. We argue that comparative data on the pace and pattern of cognitive development across species can address major theoretical questions in both psychology and biology. In particular, such integrative research will allow stronger biological inferences about the function of developmental change, and will be critical in addressing how humans come to acquire species-unique cognitive abilities.}
}
@article{RoschNATURAL-CATEGORIES-91,
       author = {Eleanor H. Rosch},
        title = {Natural categories},
      journal = {Cognitive Psychology},
       volume = {4},
       number = {3},
         year = {1973},
        pages = {328-350},
     abstract = {The hypothesis of the study was that the domains of color and form are structured into nonarbitrary, semantic categories which develop around perceptually salient "natural prototypes." Categories which reflected such an organization (where the presumed natural prototypes were central tendencies of the categories) and categories which violated the organization (natural prototypes peripheral) were taught to a total of 162 members of a Stone Age culture which did not initially have hue or geometric-form concepts. In both domains, the presumed "natural" categories were consistently easier to learn than the "distorted" categories. Even when not central, natural prototype stimuli tended to be more rapidly learned and more often chosen as the most typical example of the category than were other stimuli. Implications for general differences between natural categories and the artificial categories of concept formation research were discussed.}
}
@article{RoschNATURAL-CATEGORIES-95,
       author = {Eleanor Rosch},
        title = {Cognitive reference points},
      journal = {Cognitive Psychology},
       volume = {7},
       number = {4},
         year = {1975},
        pages = {532-547},
     abstract = {Two methods were used to test the hypothesis that natural categories (such as colors, line orientations, and numbers) have reference point stimuli (such as focal colors, vertical and horizontal lines, and numbers that are multiples of 10) in relation to which other stimuli of the category are judged. In Experiment I, subjects placed pairs of stimuli into sentence frames consisting of linguistic "hedges" such as "- is essentially-." Results were that the supposed reference stimuli were most often placed in the second (reference) slot. In Experiment II, the subject placed a stimulus in physical space to represent his feeling of the psychological distance of that stimulus from another spatially fixed stimulus. Results showed that, when supposed reference stimuli were fixed, other stimuli were placed closer to them than vice versa. The results have substantive implications for the understanding of internal structure of categories and methodological implications for the mapping of reference points, quantification of linguistic intuitions, and the assumption of symmetry in psychological distance judgments.}
}
@article{RoseIEEE-98,
       author = {Kenneth Rose},
        title = {Deterministic annealing for clustering, compression, classification, regression, and related optimization problems},
    booktitle = {Proceedings of {IEEE}},
       volume = 86,
        issue = 11,
        pages = {2210-2239},
         year = 1998,
}
@article{RoseetalCHILD-DEVELOPMENT-09,
       author = {Rose, S. A. and Feldman, J. F. and Jankowski, J. J.},
        title = {A cognitive approach to the development of early language},
      journal = {Child Development},
       volume = {80},
       number = {1},
         year = {2009},
        pages = {134-150},
     abstract = {A controversial issue in the field of language development is whether language emergence and growth is dependent solely on processes specifically tied to language or could also depend on basic cognitive processes that affect all aspects of cognitive competence (domain-general processes). The present article examines this issue using a large battery of infant information-processing measures of memory, representational competence, processing speed, and attention, many of which have been shown to predict general cognition in a cohort of full-terms and preterms. Results showed that various aspects of infant memory and representational competence (a) related to language at both 12 and 36 months, (b) predicted similarly for the two groups, and (c) predicted 36-month language, independently of birth status, 12-month language, and the 12-month Bayley Mental Development Index. Additionally, the results established predictive validity for the MacArthur 12-month language measure. These findings support a domain-general view of language.}
}
@inproceedings{Rosen-ZvietalUAI-04,
       author = {Michal Rosen-Zvi and Thomas Griffiths and Mark Steyvers and Padhraic Smyth},
        title = {The author-topic model for authors and documents},
    booktitle = {Proceedings of the 20th Conference on Uncertainty in Artificial Intelligence},
     location = {Banff, Canada},
    publisher = {AUAI Press},
      address = {Arlington, Virginia, United States},
        pages = {487-494},
         year = 2004,
}
@article{RosenbaumetalCoRR-19,
        title = {Routing Networks and the Challenges of Modular and Compositional Computation},
       author = {Rosenbaum, Clemens and Cases, Ignacio and Riemer, Matthew and Klinger, Tim},
      journal = {CoRR},
       volume = {arXiv:1904.12774},
         year = {2019},
     abstract = {Compositionality is a key strategy for addressing combinatorial complexity and the curse of dimensionality. Recent work has shown that compositional solutions can be learned and offer substantial gains across a variety of domains, including multi-task learning, language modeling, visual question answering, machine comprehension, and others. However, such models present unique challenges during training when both the module parameters and their composition must be learned jointly. In this paper, we identify several of these issues and analyze their underlying causes. Our discussion focuses on routing networks, a general approach to this problem, and examines empirically the interplay of these challenges and a variety of design decisions. In particular, we consider the effect of how the algorithm decides on module composition, how the algorithm updates the modules, and if the algorithm uses regularization.},
}
@article{RosenbergandArielJoN-90,
        title = {Visual-response properties of neurons in turtle basal optic nucleus in vitro},
       author = {Rosenberg, A. F. and Ariel, M.},
      journal = {Journal of Neurophysiology},
    publisher = {American Physiological Society},
       volume = 63,
        issue = 5,
         year = 1990,
        pages = {1033-1045},
     abstract = {The spike responses of 105 cells to visual-stimulus movement were analyzed in the turtle{\textquoteright}s basal optic nucleus (BON) in vitro in the absence of the telencephalon. All cells were direction sensitive (DS) and were driven solely by stimulation of the contralateral eye. These cells had large receptive fields and had vigorous responses to moving, textured patterns. Small moving spots generated only weak responses from these cells, as did the onset or offset of diffuse light flashes. The direction tuning of BON cells was quite broad with most back and forth responses being DS. In fact, for 86\% of the cells, there were seven to nine axes (out of 9 total, in 20 degrees increments) for which response to movement in one direction was at least twice that for the opposite direction. In instances where spontaneous activity was relatively high, a suppression of that spike firing was evident when the stimulus moved in directions opposite to preferred stimulus directions. Cells preferring many different directions are found in the BON. More cells preferred inferior-temporal directed motion (49\%), compared to superior-temporal (35\%) and nasal stimuli (13\%). BON cells remained DS over 3 log units of velocity, with their strongest responses between 1 and 50 degrees/s. Responses were often non-DS for stimuli moving slower than 0.1 degrees/s. The receptive fields of BON cells were large and occupied different parts of the retina. When different subregions of a receptive field were stimulated, the cell{\textquoteright}s directional tuning always remained the same as the full field direction tuning. Thus, BON cells seem well-suited for the analysis of global, visual-field motion in any direction, performed by the accessory optic system. Other brain stem pathways necessary for optokinetic reflexes can be elucidated with the use of this whole-brain, eyes-attached in vitro preparation.},
}
@book{RosenbergandBirdzell1986,
     title = {How the West Grew Rich: The Economic Transformation of the Industrial World},
    author = {Nathan Rosenberg and L.E. Birdzell Jr.},
 publisher = {Basic Books},
   address = {New York, NY},
      year = 1986
}
@article{Rosenblatt58,
       author = {Rosenblatt, F.},
      journal = {Psychological Review},
        title = {The {P}erceptron: {A} probabilistic model for information storage and organization in the brain},
       volume = 65,
         year = 1958,
        pages = {386-408},
}
@book{Rosenblatt61,
       author = {Rosenblatt, F.},
        title = {Principles of Neurodynamics: Perceptrons and the Theory of Brain
                Mechanisms},
    publisher = {Spartan Books},
      address = {Washington, DC, USA},
         year = 1961,
}
@book{Rosenblatt62,
       author = {Rosenblatt, F.},
        title = {Principles of Neurodynamics: Perceptrons and the Theory of Brain Mechanisms},
      address = {Washington, DC},
    publisher = {Spartan Books},
         year = 1962,
}
@article{RosenbloomUBIQUITY-10,
       author = {Rosenbloom, Paul S.},
        title = {{Ubiquity Symposium '{What is Computation?}': Computing and Computation}},
      journal = {Ubiquity},
    publisher = {ACM},
      address = {New York, NY, USA},
       number = {December},
         year = {2010},
}
@article{RoseneggeretalPLoS-14,
       author = {Rosenegger, David G. and Tran, Cam Ha T. and LeDue, Jeffery and Zhou, Ning and Gordon, Grant R.},
        title = {A High Performance, Cost-Effective, Open-Source Microscope for Scanning Two-Photon Microscopy that is Modular and Readily Adaptable},
      journal = {PLoS ONE},
    publisher = {Public Library of Science},
       volume = {9},
       number = {10},
         year = {2014},
        pages = {1-16},
     abstract = {Two-photon laser scanning microscopy has revolutionized the ability to delineate cellular and physiological function in acutely isolated tissue and {\it{in vivo}}. However, there exist barriers for many laboratories to acquire two-photon microscopes. Additionally, if owned, typical systems are difficult to modify to rapidly evolving methodologies. A potential solution to these problems is to enable scientists to build their own high-performance and adaptable system by overcoming a resource insufficiency. Here we present a detailed hardware resource and protocol for building an upright, highly modular and adaptable two-photon laser scanning fluorescence microscope that can be used for {\it{in vitro}} or {\it{in vivo}} applications. The microscope is comprised of high-end componentry on a skeleton of off-the-shelf compatible opto-mechanical parts. The dedicated design enabled imaging depths close to 1 mm into mouse brain tissue and a signal-to-noise ratio that exceeded all commercial two-photon systems tested. In addition to a detailed parts list, instructions for assembly, testing and troubleshooting, our plan includes complete three dimensional computer models that greatly reduce the knowledge base required for the non-expert user. This open-source resource lowers barriers in order to equip more laboratories with high-performance two-photon imaging and to help progress our understanding of the cellular and physiological function of living systems.},
}
@article{RosenfeldetalCoRR-17,
        title = {Priming Neural Networks},
       author = {Amir Rosenfeld and Mahdi Biparva and John K. Tsotsos},
      journal = {CoRR},
       volume = {arXiv:1711.05918},
         year = 2017,
     abstract = {Visual priming is known to affect the human visual system to allow detection of scene elements, even those that may have been near unnoticeable before, such as the presence of camouflaged animals. This process has been shown to be an effect of top-down signaling in the visual system triggered by the said cue. In this paper, we propose a mechanism to mimic the process of priming in the context of object detection and segmentation. We view priming as having a modulatory, cue dependent effect on layers of features within a network. Our results show how such a process can be complementary to, and at times more effective than simple post-processing applied to the output of the network, notably so in cases where the object is hard to detect such as in severe noise. Moreover, we find the effects of priming are sometimes stronger when early visual layers are affected. Overall, our experiments confirm that top-down signals can go a long way in improving object detection and segmentation.}
}
@inproceedings{RosenscheinIJCAI-81,
       author = {Rosenschein, Stan},
        title = {Plan Synthesis: A Logical Perspective},
    booktitle = {Proceedings {IJCAI} 7},
 organization = {IJCAII},
         year = {1981},
        pages = {331-337},
}
@incollection{RosenscheinKR-89,
       author = {Rosenschein, Stan},
        title = {Synthesizing Information-Tracking Automata from Environment Descriptions},
    booktitle = {Proceedings of the First International Conference on Principles of Knowledge Representation and Reasoning},
       editor = {Hector J. Levesque},
    publisher = {Kaufmann},
      address = {Los Altos, CA.},
         year = {1989},
        pages = {386-393},
}
@techreport{RosenscheinTR-87,
       author = {Rosenschein, Stan},
        title = {Formal Theories of Knowledge in {AI} and Robotics},
        issue = {CSLI-87-84},
  institution = {Center for the Study of Language and Information},
         year = 1987,
}
@inproceedings{RosenscheinandKaelblingTARK-86,
       author = {Rosenschein, Stan and Leslie Pack Kaelbling},
        title = {The Synthesis of Digital Machines with Provable Epistemic Properties},
     crossref = {HalpernTARK-86a},
        pages = {83-98},
         year = 1986,
}
@book{Ross83,
       author = {Ross, Sheldon M.},
        title = {Introduction to Stochastic Dynamic Programming},
    publisher = {Academic Press},
      address = {New York},
         year = 1983,
}
@book{Ross94,
       author = {Ross, Sheldon M.},
        title = {A First Course in Probability, Fourth Edition},
    publisher = {Prentice Hall},
      address = {Englewood Cliffs, NJ},
         year = 1994,
}
@inproceedings{RossandZemelNIPS-2003,
       author = {D. Ross and R. Zemel},
        title = {Multiple cause vector quantization},
    booktitle = {Advances in Neural Information Processing Systems 15},
       editor = {S. Becker, S. Thrun and K. Obermayer},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2003,
}
@article{Roth-AlpermannetalJoN-10,
       author = {Roth-Alpermann, Claudia and Anjum, Farzana and Naumann, Robert and Brecht, Michael},
        title = {Cortical Organization in the Etruscan Shrew (Suncus etruscus)},
      journal = {Journal of Neurophysiology},
       volume = {104},
       number = {5},
         year = {2010},
        pages = {2389-2406},
     abstract = { Cortical organization in the Etruscan shrew is of comparative interest because of its small size and because the Etruscan shrew is an amazing tactile hunter. Here we investigated cortical organization in Etruscan shrews by electrophysiological mapping. We developed an anesthesia protocol for this very small mammal in which we combined massive application of local anesthesia, very slow induction of general anesthesia, and passive cooling. Under this anesthesia regime, we characterized auditory, visual, and somatosensory cortical responses. We found that large parts of shrew cortex respond to such stimuli. Of the responsive sites, a small fraction (\~14\%) responded to visual stimuli in a caudally located region. Another small fraction of sites (\~11\%) responded to auditory stimuli in a centrally located region. The majority of sites (\~75\%) responded to tactile stimuli. We identified two topographically organized somatosensory areas with small receptive fields referred to as putative primary somatosensory cortex and putative secondary somatosensory cortex. In a posterior-lateral region that partially overlaps with piriform cortex, we observed large somatosensory receptive fields and often polysensory responses. In an anterior-lateral region that partially overlaps with piriform cortex, we observed large unimodal somatosensory receptive fields. Our findings demonstrate a remarkable degree of tactile specialization in Etruscan shrew cortex. }
}
@inproceedings{RothandBlackCVPR-05,
       author = {Stefan Roth and Michael J. Black},
        title = {Fields of Experts: A Framework for Learning Image Priors with Applications},
    publisher = {IEEE},
    booktitle = {Proceedings {IEEE} Conference on Computer Vision and Pattern Recognition},
     location = {San Diego, California},
         year = 2005,
}
@article{RothandBlackIJCV-07,
       author = {Stefan Roth and Michael J. Black},
        title = {On the Spatial Statistics of Optical Flow},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 74,
        issue = 1,
         year = 2007,
        pages = {33-50},
}
@incollection{RothandvanRossumCMMN-09,
        title = {Modeling Synapses},
       author = {Arnd Roth and Mark C. W. van Rossum},
    booktitle = {Computational Modeling Methods for Neuroscientists},
    publisher = {The MIT Press},
         year = {2009}
}
@article{RothemundNATURE-06,
        title = {Folding {DNA} to create nanoscale shapes and patterns},
       author = {Rothemund, Paul W.},
      journal = {Nature},
       volume = 440,
        issue = 5,
         year = 2006,
        pages = {297-302},
}
@article{RothschildetalNATURE-NEUROSCIENCE-17,
       author = {Rothschild, Gideon and Eban, Elad and Frank, Loren M.},
        title = {A cortical-hippocampal-cortical loop of information processing during memory consolidation},
      journal = {Nature Neuroscience},
         year = {2017},
       volume = {20},
        issue = {2},
        pages = {251-259},
     abstract = {Hippocampal replay during sharp-wave ripple events (SWRs) is thought to drive memory consolidation in hippocampal and cortical circuits. Changes in neocortical activity can precede SWR events, but whether and how these changes influence the content of replay remains unknown. Here we show that during sleep there is a rapid cortical-hippocampal-cortical loop of information flow around the times of SWRs. We recorded neural activity in auditory cortex (AC) and hippocampus of rats as they learned a sound-guided task and during sleep. We found that patterned activation in AC precedes and predicts the subsequent content of hippocampal activity during SWRs, while hippocampal patterns during SWRs predict subsequent AC activity. Delivering sounds during sleep biased AC activity patterns, and sound-biased AC patterns predicted subsequent hippocampal activity. These findings suggest that activation of specific cortical representations during sleep influences the identity of the memories that are consolidated into long-term stores.},
}
@article{RottschyetalNEUROIMAGE-12,
       author = {Rottschy, C. and Langner, R. and Dogan, I. and Reetz, K. and Laird, A. R. and Schulz, J. B. and Fox, P. T. and Eickhoff, S. B.},
        title = {Modelling neural correlates of working memory: A coordinate-based meta-analysis},
      journal = {Neuroimage},
         year = {2012},
       volume = {60},
        issue = {1},
        pages = {830-846},
     abstract = {Working memory subsumes the capability to memorize, retrieve and utilize information for a limited period of time which is essential to many human behaviours. Moreover, impairments of working memory functions may be found in nearly all neurological and psychiatric diseases. To examine what brain regions are commonly and differently active during various working memory tasks, we performed a coordinate-based meta-analysis over 189 fMRI experiments on healthy subjects. The main effect yielded a widespread bilateral fronto-parietal network. Further meta-analyses revealed that several regions were sensitive to specific task components, e.g. Broca's region was selectively active during verbal tasks or ventral and dorsal premotor cortex were preferentially involved in memory for object identity and location, respectively. Moreover, the lateral prefrontal cortex showed a division in a rostral and a caudal part based on differential involvement in task-set and load effects. Nevertheless, a consistent but more restricted core network emerged from conjunctions across analyses of specific task designs and contrasts. This core network appears to comprise the quintessence of regions, which are necessary during working memory tasks. It may be argued that the core regions form a distributed executive network with potentially generalized functions for focusing on competing representations in the brain. The present study demonstrates that meta-analyses are a powerful tool to integrate the data of functional imaging studies on a (broader) psychological construct, probing the consistency across various paradigms as well as the differential effects of different experimental implementations.},
}
@article{RougierandOReillyCS-02,
       author = {Nicolas P. Rougier and Randall C. O'Reilly},
        title = {Learning Representations in a Gated Prefrontal Cortex Model of Dynamic Task Switching},
      journal = {Cognitive Science},
       volume = {26},
       number = {4},
        pages = {503-520},
         year = {2002},
     abstract = {We present a computational model of task switching based on dynamic context representations using a prefrontal cortex gating mechanism. Based solely on an external reward signal and task instructions, the model is able to search among available subtasks and stabilize itself on the correct context representation of the current subtask. Results are presented and demonstrate that in this framework activation-based memory leads to better performance than weight-based memory because it allows for rapid representation updating.},
}
@article {RougieretalPNAS-05,
       author = {Rougier, Nicolas P. and Noelle, David C. and Braver, Todd S. and Cohen, Jonathan D. and O'Reilly, Randall C.},
        title = {Prefrontal cortex and flexible cognitive control: Rules without symbols},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {102},
       number = {20},
        pages = {7338-7343},
         year = {2005},
     abstract = {Human cognitive control is uniquely flexible and has been shown to depend on prefrontal cortex (PFC). But exactly how the biological mechanisms of the PFC support flexible cognitive control remains a profound mystery. Existing theoretical models have posited powerful task-specific PFC representations, but not how these develop. We show how this can occur when a set of PFC-specific neural mechanisms interact with breadth of experience to self organize abstract rule-like PFC representations that support flexible generalization in novel tasks. The same model is shown to apply to benchmark PFC tasks (Stroop and Wisconsin card sorting), accurately simulating the behavior of neurologically intact and frontally damaged people.},
}
@inproceedings{RoweisNIPS-08,
     author = {Sam Roweis},
      title = {{EM} Algorithms for {PCA} and {SPCA}},
  booktitle = {Advances in Neural Information Processing Systems},
     volume = 10,
     editor = {Michael I. Jordan and Michael J. Kearns and Sara A. Solla},
       year = 1998,
}
@article{RoweisandGhahramaniNC-99,
     author = {Sam Roweis and Zoubin Ghahramani},
      title = {A Unifying Review of Linear {G}aussian Models},
    journal = {Neural Computation},
     volume = 11,
      issue = 2,
      pages = {305-345},
       year = 1999,
}
@article{RoweisandSaulSCIENCE-00,
       author = {Lawrence K. Saul and Sam T. Roweis},
        title = {Nonlinear Dimensionality Reduction by Locally Linear Embedding},
      journal = {Science},
       volume = {290},
        pages = {2323-2326},
         year = {2000},
     abstract = {Many areas of science depend on exploratory data analysis and visualization. The need to analyze large amounts of multivariate data raises the fundamental problem of dimensionality reduction: how to discover compact representations of high-dimensional data. Here, we introduce locally linear embedding (LLE), an unsupervised learning algorithm that computes low-dimensional, neighborhood-preserving embeddings of high-dimensional inputs. Unlike clustering methods for local dimensionality reduction, LLE maps its inputs into a single global coordinate system of lower dimensionality, and its optimizations do not involve local minima. By exploiting the local symmetries of linear reconstructions, LLE is able to learn the global structure of nonlinear manifolds, such as those generated by images of faces or documents of text.},
}
@inproceedings{RoweisetalNIPS-14,
       author = {Sam T. Roweis and  Lawrence K. Saul and Geoffrey E. Hinton},
        title = {Global Coordination of Local Linear Models},
    booktitle = {Advances in Neural Information Processing Systems 14},
         year = {2001},
        pages = {889-896},
}
@article{RoyIEEE-MULTIMEDIA-02,
       author = {Deb Roy},
      journal = {{IEEE Transactions on Multimedia}},
        title = {Grounded spoken language acquisition: experiments in word learning},
         year = {2003},
       volume = {5},
       number = {2},
        pages = {197-209},
     abstract = {Language is grounded in sensory-motor experience. Grounding connects concepts to the physical world enabling humans to acquire and use words and sentences in context. Currently most machines which process language are not grounded. Instead, semantic representations are abstract, pre-specified, and have meaning only when interpreted by humans. We are interested in developing computational systems which represent words, utterances, and underlying concepts in terms of sensory-motor experiences leading to richer levels of machine understanding. A key element of this work is the development of effective architectures for processing multisensory data. Inspired by theories of infant cognition, we present a computational model which learns words from untranscribed acoustic and video input. Channels of input derived from different sensors are integrated in an information-theoretic framework. Acquired words are represented in terms of associations between acoustic and visual sensory experience. The model has been implemented in a real-time robotic system which performs interactive language learning and understanding. Successful learning has also been demonstrated using infant-directed speech and images.},
}
@article{RoyandPentlandCS-02,
       author = {Roy, Deb K. and Pentland, Alex P.},
        title = {Learning words from sights and sounds: a computational model},
      journal = {Cognitive Science},
       volume = {26},
       number = {1},
         year = {2002},
        pages = {113-146},
     abstract = {This paper presents an implemented computational model of word acquisition which learns directly from raw multimodal sensory input. Set in an information theoretic framework, the model acquires a lexicon by finding and statistically modeling consistent cross-modal structure. The model has been implemented in a system using novel speech processing, computer vision, and machine learning algorithms. In evaluations the model successfully performed speech segmentation, word discovery and visual categorization from spontaneous infant-directed speech paired with video images of single objects. These results demonstrate the possibility of using state-of-the-art techniques from sensory pattern recognition and machine learning to implement cognitive models which can process raw sensor data without the need for human transcription or labeling.}
}
@article{RoyeretalNATURE-NEUROSCIENCE-12,
       author = {Royer, Sebastien and Zemelman, Boris V. and Losonczy, Attila and Kim, Jinhyun and Chance, Frances and Magee, Jeffrey C. and Buzsaki, Gyorgy},
        title = {Control of timing, rate and bursts of hippocampal place cells by dendritic and somatic inhibition},
      journal = {Nature Neuroscience},
         year = 2012,
    publisher = {Nature Publishing Group},
       volume = 15,
        issue = 5,
        pages = {769-775},
     abstract = {A consortium of inhibitory neurons control the firing patterns of pyramidal cells, but their specific roles in the behaving animal are largely unknown. We performed simultaneous physiological recordings and optogenetic silencing of either perisomatic (parvalbumin (PV) expressing) or dendrite-targeting (somatostatin (SOM) expressing) interneurons in hippocampal area CA1 of head-fixed mice actively moving a treadmill belt rich with visual-tactile stimuli. Silencing of either PV or SOM interneurons increased the firing rates of pyramidal cells selectively in their place fields, with PV and SOM interneurons having their largest effect during the rising and decaying parts of the place field, respectively. SOM interneuron silencing powerfully increased burst firing without altering the theta phase of spikes. In contrast, PV interneuron silencing had no effect on burst firing, but instead shifted the spikes' theta phase toward the trough of theta. These findings indicate that perisomatic and dendritic inhibition have distinct roles in controlling the rate, burst and timing of hippocampal pyramidal cells.}
}
@inproceedings{RozelletalICIP-07,
     author = {Rozell, C.J. and Johnson, D.H. and Baraniuk, R.G. and Olshausen, B.A.},
      title = {Locally competitive algorithms for sparse approximation},
  booktitle = {Proceedings of the International Conference on Image Processing (ICIP)},
       year = 2007,
    address = {San Antonio, TX},
      month = {September}
}
@article{RozelletalNC-08,
       author = {Rozell, C.J. and Johnson, D.H and Baraniuk, R.G. and Olshausen, B.A.},
        title = {Sparse coding via thresholding and local competition in neural circuits},
      journal = {Neural Computation},
         year = 2008,
       volume = 20,
        issue = 10,
        pages = {2526-2563},
     abstract = {While evidence indicates that neural systems may be employing sparse approximations to represent sensed stimuli, the mechanisms underlying this ability are not understood. We describe a locally competitive algorithm (LCA) that solves a collection of sparse coding principles minimizing a weighted combination of mean-squared error (MSE) and a coefficient cost function. LCAs are designed to be implemented in a dynamical system composed of many neuron-like elements operating in parallel. These algorithms use thresholding functions to induce local (usually one-way) inhibitory competitions between nodes to produce sparse representations. LCAs produce coefficients with sparsity levels comparable to the most popular centralized sparse coding algorithms while being readily suited for neural implementation. Additionally, LCA coefficients for video sequences demonstrate inertial properties that are both qualitatively and quantitatively more regular (i.e., smoother and more predictable) than the coefficients produced by greedy algorithms.},
}
@article{RubehnetalJNE-09,
    author = {Birthe Rubehn and Conrado Bosman and Robert Oostenveld and Pascal Fries and Thomas Stieglitz},
     title = {A {MEMS}-based flexible multichannel {ECoG}-electrode array},
   journal = {Journal of Neural Engineering},
    volume = 6,
     issue = 3,
      year = 2009,
     pages = 036003,
  abstract = {We present a micromachined 252-channel ECoG (electrocorticogram)-electrode array, which is made of a thin polyimide foil substrate enclosing sputtered platinum electrode sites and conductor paths. The array subtends an area of approximately 35 mm by 60 mm and is designed to cover large parts of a hemisphere of a macaque monkey's cortex. Eight omnetics connectors are directly soldered to the foil. This leads to a compact assembly size which enables a chronic implantation of the array and allows free movements of the animal between the recording sessions. The electrode sites are 1 mm in diameter and were characterized by electrochemical impedance spectroscopy. At 1 kHz, the electrode impedances vary between 1.5 kOmega and 5 kOmega. The yield of functioning electrodes in three assembled devices is 99.5\%. After implantation of a device with 100\% working electrodes, standard electrocorticographic signals can be obtained from every electrode. The response to visual stimuli can be measured with electrodes lying on the visual cortex. After an implantation time of 4.5 months, all electrodes are still working and no decline in signal quality could be observed.}
}
@book{Rubin95,
       author = {The Desktop Musician: Creating Music with Your Computer},
        title = {David M. Rubin},
    publisher = {Osborne McGraw-Hill},
      address = {Berkeley, California},
         year = 1995,
}
@inproceedings{RubleeetalICCV-11,
      title = {{ORB}: an efficient alternative to {SIFT} or {SURF}},
     author = {Ethan Rublee and Vincent Rabaud and Kurt Konolige and Gary Bradski},
  booktitle = {Proceedings of the International Conference on Computer Vision}, 
  publisher = {IEEE Computer Society},
       year = 2011,
}
@article{RudermanNETWORK-94,
       author = {D.L. Ruderman},
        title = {The statistics of natural images},
      journal = {Network: Computation in Neural Systems},
       volume = 5,
        issue = 4,
         year = 1994,
        pages = {517-548}
}
@article{RudermanVR-97,
       author = {D.L. Ruderman},
        title = {Origins of scaling in natural images},
      journal = {Vision Research},
       volume = 37,
        issue = 23,
         year = 1997,
        pages = {3385-3395}
}
@inproceedings{RudichFOCS-85,
       author = {Rudich, Steven},
        title = {Inferring the Structure of a {M}arkov Chain from its Output},
    booktitle = {Proceedings of the Twenty Sixth Annual Symposium on Foundations of Computer Science},
        pages = {321-326},
         year = 1985,
}
@article{RueckemannandBuffaloNATURE-2017,
       author = {Rueckemann, Jon W. and Buffalo, Elizabeth A.},
        title = {Auditory landscape on the cognitive map},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = {543},
         year = {2017},
        pages = {631},
}
@article{RumelhartandAbrahamsonCP-73,
       author = {David E. Rumelhart and Adele A. Abrahamson},
        title = {A model for analogical reasoning},
      journal = {Cognitive Psychology},
       volume = 5,
        issue = 1,
        pages = {1-28},
         year = 1973,
}
@book{RumelhartandMcClelland86,
        title = {Parallel Distributed Processing, Volume 1 - Explorations in the Microstructure of Cognition: Foundations},
       author = {Rumelhart, D. E. and McClelland, J. L.},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1986,
}
@book{RumelhartandMcClellandPDP_1-86,
        title = {Parallel Distributed Processing, Volume 1 - Explorations in the Microstructure of Cognition: Foundations},
       editor = {Rumelhart, D. E. and McClelland, J. L.},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1986,
}
@book{RumelhartandMcClellandPDP_2-87,
        title = {Parallel Distributed Processing, Volume 2 - Explorations in the Microstructure of Cognition: Psychological and Biological Models},
       editor = {Rumelhart, D. E. and McClelland, J. L.},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1986,
}
@article{RumelhartandZipserCS-85,
       author = {Rumelhart, David E. and Zipser, David},
        title = {Feature Discovery by Competitive Learning},
      journal = {Cognitive Science},
       volume = 9,
        issue = 1,
         year = 1985,
        pages = {75-112},
}
@incollection{RumelhartetalPDP-86a,
       author = {Rumelhart, D. E. and Hinton, G. E. and Williams, R. J.},
        title = {Learning Internal Representations by Error Propagation},
       editor = {Rumelhart, D. E. and McClelland, J. L.},
    booktitle = {Parallel Distributed Processing, Volume 1 - Explorations in the Microstructure of Cognition: Foundations},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1986,
        pages = {318-363},
}
@incollection{RumelhartetalPDP-86b,
       author = {Rumelhart, D. E. and Hinton, G. E. and McClelland, J. L.},
        title = {A General Framework for Parallel Distributed Processing},
    booktitle = {Parallel Distributed Processing, Volume 1 - Explorations in the Microstructure of Cognition: Foundations},
       editor = {Rumelhart, D. E. and McClelland, J. L.},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 1986,
        pages = {45-76},
}
@article{RunnelsetalMOLECULAR_IMAGING-06,
       author = {Runnels, Judith M. and Zamiri, Parisa and Spencer, Joel A. and Veilleux, Isreal and Wei, Xunbin and Bogdanov, Alexei and Lin, Charles P.},
        title = {Imaging Molecular Expression on Vascular Endothelial Cells by In Vivo Immunofluorescence Microscopy},
      journal = {Molecular Imaging},
         year = {2006},
       volume = {5},
        issue = {1},
        pages = {31-40},
     abstract = {Molecular expression on the vascular endothelium is critical in regulating the Interaction of circulating cells with the blood vessel wall. Leukocytes as well as circulating cancer cells gain entry into tissue by interacting with adhesion molecules on the endothelial cells (EC). Molecular targets on the EC arc increasingly being explored for tissue-specific delivery of therapeutic and Imaging agents. Here we use in vivo immunofluorescence microscopy to visualize the endothelial molecular expression in the vasculature of live animals. High-resolution images are obtained by optical sectioning through the intact skin using in vivo confocal and multiphoton microscopy after in situ labeling of EC surface markers with fluorescent antibodies. Other vascular beds such as the bone marrow and ocular blood vessels can be imaged with little or no tissue manipulation. Live imaging is particularly useful for following the dynamic expression of inducible molecules such as E-selectin during an inflammatory response.},
}
@article{RuoslahtiGLYCOBIOLOGY-96,
        title = {Brain extracellular matrix},
       author = {Erkki Ruoslahti},
      journal = {Glycobiology},
       volume = 6,
        issue = 5,
         year = 1996,
        pages = {489-492},
}
@article{RuschetalPLoS-07,
        title = {The {Sorcerer II} Global Ocean Sampling Expedition: Northwest Atlantic through Eastern Tropical Pacific},
       author = {Rusch, Douglas B and Halpern, Aaron L and Sutton, Granger and Heidelberg, Karla B and Williamson, Shannon and Yooseph, Shibu and Wu, Dongying and Eisen, Jonathan A and Hoffman, Jeff M and Remington, Karin and Beeson, Karen and Tran, Bao and Smith, Hamilton and Baden-Tillson, Holly and Stewart, Clare and Thorpe, Joyce and Freeman, Jason and Andrews-Pfannkoch, Cynthia and Venter, Joseph E and Li, Kelvin and Kravitz, Saul and Heidelberg, John F and Utterback, Terry and Rogers, Yu-Hui and Falcan, Luisa I and Souza, Valeria and Bonilla-Rosso, GermÃ¡n and Eguiarte, Luis E and Karl, David M and Sathyendranath, Shubha and Platt, Trevor and Bermingham, Eldredge and Gallardo, Victor and Tamayo-Castillo, Giselle and Ferrari, Michael R and Strausberg, Robert L and Nealson, Kenneth and Friedman, Robert and Frazier, Marvin and Venter, J. Craig},
      journal = {{PLoS} Biology},
    publisher = {Public Library of Science},
         year = 2007,
       volume = 5,
        issue = 3,
        pages = {e77},
}
@inproceedings{RushANNOTATED-TRANSFORMER-18,
        title = {The Annotated Transformer},
       author = {Rush, Alexander},
    booktitle = {Proceedings of Workshop for {NLP} Open Source Software ({NLP}-{OSS})},
      address = {Melbourne, Australia},
    publisher = {Association for Computational Linguistics},
         year = {2018},
        pages = {52--60},
     abstract = {A major goal of open-source NLP is to quickly and accurately reproduce the results of new work, in a manner that the community can easily use and modify. While most papers publish enough detail for replication, it still may be difficult to achieve good results in practice. This paper presents a worked exercise of paper reproduction with the goal of implementing the results of the recent Transformer model. The replication exercise aims at simple code structure that follows closely with the original work, while achieving an efficient usable system.},
}
@article{RusseletalIJCV-08,
       author = {Bryan Russell and Antonio Torralba and Kevin Murphy and William Freeman},
        title = {{LabelMe}: A Database and Web-Based Tool for Image Annotation},
      journal = {International Journal Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 77,
        issue = {1-3},
        pages = {157-173},
         year = 2008,
}
@misc{Russell,
       author = {Jerry Russell},
        title = {The structure of neocortex circuits},
          url = {http://www.regena.com/cortex.htm},
}
@book{Russell1967autobiography,
        title = {The Autobiography of Bertrand Russell, Volume III, 1944-1969},
       author = {Russell, Bertrand},
         year = {1967},
    publisher = {Little, Brown and Company}
}
@book{RussellandNorvig03,
       author = {Stuart Russell and Peter Norvig},
        title = {Artificial Intelligence: A Modern Approach},
    publisher = {Second edition, Prentice Hall},
      address = {Upper Saddle River, NJ},
         year = 2003,
}
@inproceedings{RussellandWefaldIJCAI-89,
       author = {Russell, Stuart J. and Wefald, Eric H.},
        title = {On Optimal Game-Tree Search using Rational Meta-Reasoning},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
         year = 1989,
        pages = {334-340},
}
@incollection{RussellandWefaldKR-89,
       author = {Russell, Stuart J. and Wefald, Eric H.},
        title = {Principles of Metareasoning},
       editor = {Brachman, Ronald J. and Levesque, Hector J. and Reiter, Raymond},
    booktitle = {Proceedings of the First International Conference on Principles of Knowledge Representation and Reasoning},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1989,
        pages = {400-411},
}
@inproceedings{RusselletalCVPR-06,
       author = {Bryan Russell and Alexei Efros and Josef Sivic and William Freeman and Andrew Zisserman},
        title = {Using Multiple Segmentations to Discover Objects and their Extent in Image Collections},
    booktitle = {Proceedings of the 2006 {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
       volume = 2,
        pages = {1605-1614},
         year = 2006,
}
@inproceedings{RusselletalIJCAI-95,
       author = {Russell, S. J. and Binder, J. and Koller, D. and Kanazawa, K.},
        title = {Local Learning in Probabilistic Networks with Hidden Variables},
    booktitle = {Proceedings {IJCAI} 14},
     location = {Montreal, Canada},
 organization = {IJCAII},
        pages = {1146-1152},
         year = 1995,
}
@incollection{RusselletalNIPS-07,
       author = {Bryan Russell and Antonio Torralba and Ce Liu and Rob Fergus and William Freeman},
        title = {Object Recognition by Scene Alignment},
    booktitle = {Advances in Neural Information Processing Systems},
      address = {Cambridge, MA},
    publisher = {MIT Press},
         year = 2007,
}
@techreport{RusselletatTR-94,
       author = {Russell, S. J., and Binder, J., and Koller, D.},
        title = {Adaptive Probabilistic Networks},
  institution = {University of California at Berkeley},
        issue = {UCB/CSD-94-824},
         year = 1994,
}
@article{RussinetalCoRR-19,
       author = {Jake Russin and Jason Jo and Randall C. O'Reilly and Yoshua Bengio},
        title = {Compositional generalization in a deep seq2seq model by separating syntax and semantics},
      journal = {CoRR},
       volume = {arXiv:1904.09708},
         year = {2019},
     abstract = {Standard methods in deep learning for natural language processing fail to capture the compositional structure of human language that allows for systematic generalization outside of the training distribution. However, human learners readily generalize in this way, e.g. by applying known grammatical rules to novel words. Inspired by work in neuroscience suggesting separate brain systems for syntactic and semantic processing, we implement a modification to standard approaches in neural machine translation, imposing an analogous separation. The novel model, which we call Syntactic Attention, substantially outperforms standard methods in deep learning on the SCAN dataset, a compositional generalization task, without any hand-engineered features or additional supervision. Our work suggests that separating syntactic from semantic learning may be a useful heuristic for capturing compositional structure.}
}
@inproceedings{RussinetalICLR-20,
       author = {Jacob Russin and Randy O'Reilly and Yoshua Bengio},
        title = {Deep learning needs a prefrontal cortex},
    booktitle = {Bridging AI and Cognitive Science},
         year = {2020},
     abstract = {Research seeking to build artificial systems capable of reproducing elements of human intelligence may benefit from a deeper consideration of the architecture and learning mechanisms of the human brain. In this brief review, we note a connection between many current challenges facing artificial intelligence and the functions of a particular brain area — the prefrontal cortex (PFC). This brain area is known to be involved in executive functions such as reasoning, rule-learning, deliberate or controlled processing, and abstract planning. Motivated by the hypothesis that these functions provide a form of out-of-distribution robustness currently not available in state-of-the-art AI systems, we elaborate on this connection and highlight some computational principles thought to be at work in PFC, with the goal of enhancing the synergy between neuroscience and machine learning.}
}
@article{RustetalNATURE-06,
        title = {How {MT} cells analyze the motion of visual patterns},
       author = {N. C. Rust and V. Mante and E. P. Simoncelli and J. A. Movshon},
      journal = {Nature Neuroscience},
       volume = 9,
        issue = 11,
         year = 2006,
        pages = {1421-1431},
}
@article{RusuetalCoRR-16,
       author = {Andrei A. Rusu and Neil C. Rabinowitz and Guillaume Desjardins and Hubert Soyer and James Kirkpatrick and Koray Kavukcuoglu and Razvan Pascanu and Raia Hadsell},
        title = {Progressive Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1606.04671},
         year = {2016},
     abstract = {Learning to solve complex sequences of tasks--while both leveraging transfer and avoiding catastrophic forgetting--remains a key obstacle to achieving human-level intelligence. The progressive networks approach represents a step forward in this direction: they are immune to forgetting and can leverage prior knowledge via lateral connections to previously learned features. We evaluate this architecture extensively on a wide variety of reinforcement learning tasks (Atari and 3D maze games), and show that it outperforms common baselines based on pretraining and finetuning. Using a novel sensitivity measure, we demonstrate that transfer occurs at both low-level sensory and high-level control layers of the learned policy.}
}
@incollection{SVM-Light,
       author = {T. Joachims},
        title = {Making large-scale {SVM} learning practical},
    booktitle = {Advances in Kernel Methods: Support Vector Machines},
       editor = {B. Sch\"{o}lkopf, C. Burges, A. Smola},
    publisher = {MIT Press, Cambridge, MA},
          url = {http://www-ai.cs.uni-dortmund.de/DOKUMENTE/joachims_99a.pdf},
         year = 1999,
}
@article{SaalfeldetalNATURE-12,
        title = {Elastic volume reconstruction from series of ultra-thin microscopy sections},
       author = {Saalfeld, Stephan and Fetter, Richard and Cardona, Albert and Tomancak, Pavel},
      journal = {Nature Methods},
       volume = 9,
        issue = 7,
    publisher = {Nature Publishing Group},
         year = 2012,
        pages = {717-720},
}
@book{Sacerdoti77,
       author = {Sacerdoti, Earl},
        title = {A Structure for Plans and Behavior},
    publisher = {American Elsevier},
      address = {New York},
         year = 1977,
      comment = {NOAH},
}
@article{SacerdotiAIJ-74,
       author = {Sacerdoti, Earl},
        title = {Planning in a Hierarchy of Abstraction Spaces},
      journal = AIJ,
       volume = 7,
         year = 1974,
        pages = {231-272},
}
@book{Sacks1998wifeforahat,
        title = {The Man Who Mistook His Wife For A Hat: And Other Clinical Tales},
       author = {Sacks, Oliver and Sacks, Oliver W.},
         year = {1998},
    publisher = {Simon \& Schuster}
}
@book{Sacks2010minsdeye,
        title = {The Mind's Eye},
       author = {Sacks, Oliver},
         year = {2010},
    publisher = {Knopf Doubleday Publishing Group}
}
@article{SadakaneetalCELL-15,
       author = {Sadakane, Osamu and Masamizu, Yoshito and Watakabe, Akiya and Terada, Shin-Ichiro and Ohtsuka, Masanari and Takaji, Masafumi and Mizukami, Hiroaki and Ozawa, Keiya and Kawasaki, Hiroshi and Matsuzaki, Masanori and Yamamori, Tetsuo},
        title = {Long-Term Two-Photon Calcium Imaging of Neuronal Populations with Subcellular Resolution in Adult Non-human Primates},
      journal = {Cell Reports},
       volume = {13},
        issue = {9},
         year = {2015},
        pages = {1989-1999},
     abstract = {Summary Two-photon imaging with genetically encoded calcium indicators (GECIs) enables long-term observation of neuronal activity in vivo. However, there are very few studies of GECIs in primates. Here, we report a method for long-term imaging of a GECI, GCaMP6f, expressed from adeno-associated virus vectors in cortical neurons of the adult common marmoset (Callithrix jacchus), a small New World primate. We used a tetracycline-inducible expression system to robustly amplify neuronal GCaMP6f expression and up- and downregulate it for more than 100 days. We succeeded in monitoring spontaneous activity not only from hundreds of neurons three-dimensionally distributed in layers 2 and 3 but also from single dendrites and axons in layer 1. Furthermore, we detected selective activities from somata, dendrites, and axons in the somatosensory cortex responding to specific tactile stimuli. Our results provide a way to investigate the organization and plasticity of cortical microcircuits at subcellular resolution in non-human primates.},
}
@inproceedings{SadeghietalCVPR-11,
       author = {Sadeghi, Mohammad Amin and Farhadi, Ali},
        title = {Recognition using Visual Phrases},
    booktitle = {{IEEE} Computer Vision and Pattern Recognition, Workshop on Biologically Consistent Vision},
         year = 2011,
}
@article{SadeketalNANO-10,
        title = {Wiring Nanoscale Biosensors with Piezoelectric Nanomechanical Resonators},
       author = {Sadek, A.S. and Karabalin, R.B. and Du, J. and Roukes, M.L. and Koch, C. and Masmanidis, S.C.},
      journal = {Nano Letters},
       volume = 10,
         year = 2010,
        pages = {1769-1773},
     abstract = {Nanoscale integrated circuits and sensors will require methods for unobtrusive interconnection with the macroscopic world to fully realize their potential. We report on a nanoelectromechanical system that may present a solution to the wiring problem by enabling information from multisite sensors to be multiplexed onto a single output line. The basis for this method is a mechanical Fourier transform mediated by piezoelectrically coupled nanoscale resonators. Our technique allows sensitive, linear, and real-time measurement of electrical potentials from conceivably any voltage-sensitive device. With this method, we demonstrate the direct transduction of neuronal action potentials from an extracellular microelectrode. This approach to wiring nanoscale devices could lead to minimally invasive implantable sensors with thousands of channels for in vivo neuronal recording, medical diagnostics, and electrochemical sensing.}
}
@article{SadrandSinhaCS-04,
    author = {Javid Sadr and Pawan Sinha},
     title = {Object recognition and Random Image Structure Evolution},
   journal = {Cognitive Science},
    volume = 28,
     issue = 2,
      year = 2004,
     pages = {259-287},
}
@article{Safiejko-MroczkaandBellJHC-96,
       author = {Safiejko-Mroczka, B. and Bell, P. B.},
        title = {Bifunctional protein cross-linking reagents improve labeling of cytoskeletal proteins for qualitative and quantitative fluorescence microscopy},
      journal = {Journal of Histochemistry \& Cytochemistry},
         year = {1996},
       volume = {44},
        issue = {6},
        pages = {641-656},
     abstract = {Because permeabilization of the cell membrane is necessary to label intracellular proteins with most fluorescent probes, it is important to optimize the preservation and labeling of the proteins under study. We used qualitative and quantitative fluorescence microscopy to evaluate the effects of six different bifunctional protein cross-linking reagents and several extraction conditions on the labeling of filamentous actin with phalloidin and the immunolabeling of tubulin and gelsolin. The labeling of cytoskeletal and associated proteins can be significantly enhanced by the appropriate combination of bifunctional protein cross-linking reagents and extraction conditions. However, the conditions that give the most intense labeling vary depending on the label used. The greatest intensity of labeling with either phalloidin or antibodies was obtained with the intermediate-length cross-linker DSP. The two-step procedure of cross-linking with DSP and extracting in Triton X-100 in microtubule-stabilizing buffer containing DSP gives maximal labeling with phalloidin. Maximal labeling of gelsolin and tubulin with antibodies is obtained by extracting DSP-cross-linked cells with Triton in Hank's saline containing DSP. Therefore, DSP reproducibly improves preservation of both soluble and filamentous proteins for quantitative and qualitative studies by fluorescence microscopy.}
}
@inproceedings{SageandBuxtonICPR-04,
        title = {Joint spatial and  temporal structure learning for task based control},
       author = {Sage, Kingsley H. and Buxton, Hilary},
    booktitle = {International Conference on Pattern Recognition (ICPR)},
       volume = 2,
        pages = {48-51},
         year = 2004,
}
@incollection{SageetalHCI-03,
        title = {Developing Content Sensitive {HMM} Gesture Recognition},
       author = {Sage, K. H. and Howell, A. J. and Buxton, H.},
    booktitle = {Gesture-based communication in Human Computer Interaction},
    publisher = {Springer LNAI},
       volume = 2915,
        pages = {277-287},
         year = 2003,
}
@incollection{SaglametalNIP-08,
     title = {A Retinal Circuit Model Accounting for Functions of Amacrine Cells},
    author = {Saglam, Murat and Hayashida, Yuki and Murayama, Nobuki},
 booktitle = {Neural Information Processing},
    editor = {Ishikawa, Masumi and Doya, Kenji and Miyamoto, Hiroyuki and Yamakawa, Takeshi},
 publisher = {Springer-Verlag},
   address = {Berlin, Heidelberg},
      year = 2008,
     pages = {1-6},
  abstract = {In previous experimental studies on vertebrates, high level processes of vision such as object segregation and spatio-temporal pattern adaptation were found to begin at retinal stage. In those visual functions, diverse subtypes of amacrine cells are believed to play essential roles by processing the excitatory and inhibitory signals laterally over a wide region on the retina to shape the ganglion cell responses. Previously, a simple "linear-nonlinear" model was proposed to explain a specific function of the retina, and could capture the spiking behavior of retinal output, although each class of the retinal neurons were largely omitted in it. Here, we present a spatio-temporal computational model based on the response function for each class of the retinal neurons and the anatomical intercellular connections. This model is not only capable of reproducing filtering properties of outer retina but also realizes high-order inner retinal functions such as object segregation mechanism via wide-field amacrine cells.},
}
@article{SahayetalNEURON-11,
        title = {Pattern Separation: A Common Function for New Neurons in Hippocampus and Olfactory Bulb},
      journal = {Neuron},
       volume = {70},
       number = {4},
        pages = {582-588},
         year = {2011},
       author = {Amar Sahay and Donald A. Wilson and Ren\'{e} Hen},
     abstract = {While adult-born neurons in the olfactory bulb (OB) and the dentate gyrus (DG) subregion of the hippocampus have fundamentally different properties, they may have more in common than meets the eye. Here, we propose that new granule cells in the OB and DG may function as modulators of principal neurons to influence pattern separation and that adult neurogenesis constitutes an adaptive mechanism to optimally encode contextual or olfactory information.}
}
@article{SahetalCSM-14,
       author = {Sah, M.P. and Hyongsuk Kim and Chua, L.O.},
      journal = {Circuits and Systems Magazine, IEEE},
        title = {Brains Are Made of Memristors},
         year = 2014,
       volume = 14,
        issue = 1,
        pages = {12-36},
     abstract = {This exposition shows that the potassium ion-channels and the sodium ion-channels that are distributed over the entire length of the axons of our neurons are in fact locally-active memristors. In particular, they exhibit all of the fingerprints of memristors, including the characteristic pinched hysteresis Lissajous figures in the voltage-current plane, whose loop areas shrink as the frequency of the periodic excitation signal increases. Moreover, the pinched hysteresis loops for the potassium ion-channel memristor, and the sodium ion-channel memristor, from the Hodgkin-Huxley axon circuit model are unique for each periodic excitation signal. An in-depth circuit-theoretic analysis and characterizations of these two classic biological memristors are presented via their small-signal memristive equivalent circuits, their frequency response, and their Nyquist plots. Just as the Hodgkin-Huxley circuit model has stood the test of time, its constituent potassium ion-channel and sodium ion-channel memristors are destined to be classic examples of locally-active memristors in future textbooks on circuit theory and bio-physics.},
}
@article{SahnietalCoRR-17,
       author = {Himanshu Sahni and Saurabh Kumar and Farhan Tejani and Charles L. Isbell},
        title = {Learning to Compose Skills},
      journal = {CoRR},
       volume = {arXiv:1711.11289},
         year = {2017},
     abstract = {We present a differentiable framework capable of learning a wide variety of compositions of simple policies that we call skills. By recursively composing skills with themselves, we can create hierarchies that display complex behavior. Skill networks are trained to generate skill-state embeddings that are provided as inputs to a trainable composition function, which in turn outputs a policy for the overall task. Our experiments on an environment consisting of multiple collect and evade tasks show that this architecture is able to quickly build complex skills from simpler ones. Furthermore, the learned composition function displays some transfer to unseen combinations of skills, allowing for zero-shot generalizations.}
}
@article{SakaietalCURRENT-BIOLOGY-11,
       author = {Sakai, T. and Mikami, A. and Tomonaga, M. and Matsui, M. and Suzuki, J. and Hamada, Y. and Tanaka, M. and Miyabe-Nishiwaki, T. and Makishima, H. and Nakatsukasa, M. and Matsuzawa, T. },
        title = {Differential prefrontal white matter development in chimpanzees and humans},
      journal = {Current Biology},
         year = {2011},
       volume = {21},
       number = {16},
        pages = {1397-1402},
     abstract = {A comparison of developmental patterns of white matter (WM) within the prefrontal region between humans and nonhuman primates is key to understanding human brain evolution. WM mediates complex cognitive processes and has reciprocal connections with posterior processing regions [1, 2]. Although the developmental pattern of prefrontal WM in macaques differs markedly from that in humans [3], this has not been explored in our closest evolutionary relative, the chimpanzee. The present longitudinal study of magnetic resonance imaging scans demonstrated that the prefrontal WM volume in chimpanzees was immature and had not reached the adult value during prepuberty, as observed in humans but not in macaques. However, the rate of prefrontal WM volume increase during infancy was slower in chimpanzees than in humans. These results suggest that a less mature and more protracted elaboration of neuronal connections in the prefrontal portion of the developing brain existed in the last common ancestor of chimpanzees and humans, and that this served to enhance the impact of postnatal experiences on neuronal connectivity. Furthermore, the rapid development of the human prefrontal WM during infancy may help the development of complex social interactions, as well as the acquisition of experience-dependent knowledge and skills to shape neuronal connectivity.}
}
@article{SakaietalPLoS-17,
       author = {Sakai, T. and Mikami, A. and Suzuki, J. and Miyabe-Nishiwaki, T. and Matsui, M. and Tomonaga, M. and Hamada, Y. and Matsuzawa, T. and Okano, H. and Oishi, K.},
        title = {{{D}evelopmental trajectory of the corpus callosum from infancy to the juvenile stage: {C}omparative {M}{R}{I} between chimpanzees and humans}},
      journal = {PLoS ONE},
         year = {2017},
       volume = {12},
       number = {6},
        pages = {e0179624},
     abstract = {How brains develop during early life is one of the most important topics in neuroscience because it underpins the neuronal functions that mature during this period. A comparison of the neurodevelopmental patterns among humans and nonhuman primates is essential to infer evolutional changes in neuroanatomy that account for higher-order brain functions, especially those specific to humans. The corpus callosum (CC) is the major white matter bundle that connects the cerebral hemispheres, and therefore, relates to a wide variety of neuronal functions. In humans, the CC area rapidly expands during infancy, followed by relatively slow changes. In chimpanzees, based on a cross-sectional study, slow changes in the CC area during the juvenile stage and later have also been reported. However, little is known about the developmental changes during infancy. A longitudinal study is also required to validate the previous cross-sectional observations about the chimpanzee CC. The present longitudinal study of magnetic resonance imaging scans demonstrates that the CC development in chimpanzees and humans is characterized by a rapid increase during infancy, followed by gradual increase during the juvenile stage. Several differences between the two species were also identified. First, there was a tendency toward a greater increase in the CC areas during infancy in humans. Second, there was a tendency toward a greater increase in the rostrum during the juvenile stage in chimpanzees. The rostral body is known to carry fibers between the bilateral prefrontal and premotor cortices, and is involved in behavior planning and control, verbal working memory, and number conception. The rostrum is known to carry fibers between the prefrontal cortices, and is involved in attention control. The interspecies differences in the developmental trajectories of the rostral body and the rostrum might be related to evolutional changes in the brain systems.}
}
@book{Sakamotoetal1986,
   author =    {Y. Sakamoto and M. Ishiguro and G. Kitagawa},
   title =     {Akaike Information Criterion Statistics},
   publisher = {D. Reidel Publishing Company},
   year =      1986,
}
@article{SaketalCoRR-14,
       author = {Hasim Sak and Andrew Senior and Fran\c{c}oise Beaufays},
        title = {Long Short-Term Memory Based Recurrent Neural Network Architectures for Large Vocabulary Speech Recognition},
      journal = {CoRR},
       volume = {arXiv:1402.1128},
         year = 2014,
}
@inproceedings{SaketalINTERSPEECH-14a,
        title = {Long Short-Term Memory Recurrent Neural Network Architectures for Large Scale Acoustic Modeling},
       author = {Hasim Sak and Andrew Senior and Francoise Beaufays},
    booktitle = {Proceedings of Interspeech},
         year = 2014,
          URL = {http://www.interspeech2014.org/},
        pages = {00-00},
      absract = {Long Short-Term Memory (LSTM) is a specific recurrent neural network (RNN) architecture that was designed to model temporal sequences and their long-range dependencies more accurately than conventional RNNs. In this paper, we explore LSTM RNN architectures for large scale acoustic modeling in speech recognition. We recently showed that LSTM RNNs are more effective than DNNs and conventional RNNs for acoustic modeling, considering moderately-sized models trained on a single machine. Here, we introduce the first distributed training of LSTM RNNs using asynchronous stochastic gradient descent optimization on a large cluster of machines. We show that a two-layer deep LSTM RNN where each LSTM layer has a linear recurrent projection layer can exceed state-of-the-art speech recognition performance. This architecture makes more effective use of model parameters than the others considered, converges quickly, and outperforms a deep feed forward neural network having an order of magnitude more parameters.}
}
@inproceedings{SaketalINTERSPEECH-14b,
        title = {Sequence Discriminative Distributed Training of Long Short-Term Memory Recurrent Neural Networks},
       author = {Hasim Sak and Oriol Vinyals and Georg Heigol and Andrew Senior and Erik McDermott and Rajat Monga and Mark Mao},
    booktitle = {Proceedings of Interspeech},
         year = 2014,
          URL = {http://www.interspeech2014.org/},
        pages = {00-00},
      absract = {We recently showed that Long Short-Term Memory (LSTM) recurrent neural networks (RNNs) outperform state-of-the-art deep neural networks (DNNs) for large scale acoustic modeling where the models were trained with the cross-entropy (CE) criterion. It has also been shown that sequence discriminative training of DNNs initially trained with the CE criterion gives significant improvements. In this paper, we investigate sequence discriminative training of LSTM RNNs in a large scale acoustic modeling task. We train the models in a distributed manner using asynchronous stochastic gradient descent optimization technique. We compare two sequence discriminative criteria - maximum mutual information and state-level minimum Bayes risk, and we investigate a number of variations of the basic training strategy to better understand issues raised by both the sequential model, and the objective function. We obtain significant gains over the CE trained LSTM RNN model using sequence discriminative training techniques.}
}
@article{SaketalNS-13,
       author = {Sak, J. and Grzybowski, A. and von Economo, C.},
        title = {{B}rain and aviation: on the 80th anniversary of {C}onstantin von {E}conomo's (1876-1931) death},
      journal = {Neurological Sciences},
         year = {2013},
       volume = {34},
       number = {3},
        pages = {387-391},
     abstract = {2011 marks the 80th anniversary of the death of Constantin Alexander von Economo who conducted advanced research on the cytoarchitectonics of the brain. This Austrian neurologist and the pioneer of aviation described encephalitis lethargica, discovered the spindle neurons, and postulated the existence of the sleep and wakefulness centre in the brain. What is more he realized two of the biggest dreams of humankind: conquering space and getting to know the secrets of the human brain.}
}
@inproceedings{SalakhutdinovandHintonAIS-07,
    author = {Ruslan Salakhutdinov and Geoffrey Hinton},
     title = {Learning a non-linear embedding by preserving class neighbourhood structure},
 booktitle = {Proceedings of {AI} and Statistics},
  location = {Puerto Rico},
      year = 2007,
}
@inproceedings{SalakhutdinovandHintonSIGIR-07,
    author = {Ruslan Salakhutdinov and Geoffrey Hinton},
     title = {Semantic Hashing},
 booktitle = {Proceedings of SIGIR Workshop on Information Retrieval and Applications of Graphical Models}, 
  location = {Amsterdam},
      year = 2007,
}
@inproceedings{SalakhutdinovandRoweisICML-03,
     author = {Ruslan Salakhutdinov and Sam T. Roweis},
      title = {Adaptive Overrelaxed Bound Optimization Methods},
  booktitle = {Proceedings of the 20th Annual International Conference on Machine Learning},
   location = {Washington, DC, USA},
  publisher = {AAAI Press},
       year = 2003,
      pages = {664-671},
}
@article{SalinasandSejnowskiJoN-00,
       author = {Salinas, Emilio and Sejnowski, Terrence J.},
        title = {Impact of Correlated Synaptic Input on Output Firing Rate and Variability in Simple Neuronal Models},
      journal = {The Journal of Neuroscience},
       volume = {20},
       number = {16},
        pages = {6193-6209},
         year = {2000},
     abstract = {Cortical neurons are typically driven by thousands of synaptic inputs. The arrival of a spike from one input may or may not be correlated with the arrival of other spikes from different inputs. How does this interdependence alter the probability that the postsynaptic neuron will fire? We constructed a simple random walk model in which the membrane potential of a target neuron fluctuates stochastically, driven by excitatory and inhibitory spikes arriving at random times. An analytic expression was derived for the mean output firing rate as a function of the firing rates and pairwise correlations of the inputs. This stochastic model made three quantitative predictions. (1) Correlations between pairs of excitatory or inhibitory inputs increase the fluctuations in synaptic drive, whereas correlations between excitatory-inhibitory pairs decrease them. (2) When excitation and inhibition are fully balanced (the mean net synaptic drive is zero), firing is caused by the fluctuations only. (3) In the balanced case, firing is irregular. These theoretical predictions were in excellent agreement with simulations of an integrate-and-fire neuron that included multiple conductances and received hundreds of synaptic inputs. The results show that, in the balanced regime, weak correlations caused by signals shared among inputs may have a multiplicative effect on the input-output rate curve of a postsynaptic neuron, i.e. they may regulate its gain; in the unbalanced regime, correlations may increase firing probability mainly around threshold, when output rate is low; and in all cases correlations are expected to increase the variability of the output spike train.},
}
@incollection{SalleeandOlshausenNIPS-02,
     author = {Phil Sallee and Bruno A. Olshausen},
      title = {Learning Sparse Multiscale Image Representations},
  booktitle = {Advances in Neural Information Processing Systems 15},
     editor = {S. Becker, S. Thrun and K. Obermayer},
  publisher = {MIT Press},
    address = {Cambridge, MA},
      pages = {1327-1334},
       year = 2003,
}
@article{SalterandStevensNATURE-MEDICINE-17,
       author = {Salter, Michael W. and Stevens, Beth},
        title = {Microglia emerge as central players in brain disease},
      journal = {Nature Medicine},
    publisher = {Nature Publishing Group},
       volume = {23},
         year = {2017},
        pages = {1018-1027},
     abstract = {There has been an explosion of new findings recently giving us insights into the involvement of microglia in central nervous system (CNS) disorders. A host of new molecular tools and mouse models of disease are increasingly implicating this enigmatic type of nervous system cell as a key player in conditions ranging from neurodevelopmental disorders such as autism to neurodegenerative disorders such as Alzheimer's disease and chronic pain. Contemporaneously, diverse roles are emerging for microglia in the healthy brain, from sculpting developing neuronal circuits to guiding learning-associated plasticity. Understanding the physiological functions of these cells is crucial to determining their roles in disease. Here we focus on recent developments in our rapidly expanding understanding of the function, as well as the dysfunction, of microglia in disorders of the CNS.},
}
@article{SalvucciandAndersonCS-01,
       author = {Salvucci, Dario D. and Anderson, John R.},
        title = {Integrating analogical mapping and general problem solving: the path-mapping theory},
      journal = {Cognitive Science},
       volume = {25},
       number = {1},
         year = {2001},
        pages = {67-110},
     abstract = {Abstract This article describes the path-mapping theory of how humans integrate analogical mapping and general problem solving. The theory posits that humans represent analogs with declarative roles, map analogs by lower-level retrieval of analogous role paths, and coordinate mappings with higher-level organizational knowledge. Implemented in the ACT-R cognitive architecture, the path-mapping theory enables models of analogical mapping behavior to incorporate and interface with other problem-solving knowledge. Path-mapping models thus can include task-specific skills such as encoding analogs or generating responses, and can make behavioral predictions at the level of real-world metrics such as latency or correctness. We show that the path-mapping theory can successfully account for the major phenomena addressed by previous theories of analogy. We also describe a path-mapping model that can account for subjects' incremental eye-movement and typing behavior in a story-mapping task. We discuss extensions and implications of this work to other areas of analogy and problem-solving research.},
}
@article {SalzetalJoN-16,
       author = {Salz, Daniel M. and Tiganj, Zoran and Khasnabish, Srijesa and Kohley, Annalyse and Sheehan, Daniel and Howard, Marc W. and Eichenbaum, Howard},
        title = {Time Cells in Hippocampal Area CA3},
      journal = {Journal of Neuroscience},
       volume = {36},
       number = {28},
         year = {2016},
        pages = {7476-7484},
    publisher = {Society for Neuroscience},
     abstract = {Studies on time cells in the hippocampus have so far focused on area CA1 in animals performing memory tasks. Some studies have suggested that temporal processing within the hippocampus may be exclusive to CA1 and CA2, but not CA3, and may occur only under strong demands for memory. Here we examined the temporal and spatial coding properties of CA3 and CA1 neurons in rats performing a maze task that demanded working memory and a control task with no explicit working memory demand. In the memory demanding task, CA3 cells exhibited robust temporal modulation similar to the pattern of time cell activity in CA1, and the same populations of cells also exhibited typical place coding patterns in the same task. Furthermore, the temporal and spatial coding patterns of CA1 and CA3 were equivalently robust when animals performed a simplified version of the task that made no demands on working memory. However, time and place coding did differ in that the resolution of temporal coding decreased over time within the delay interval, whereas the resolution of place coding was not systematically affected by distance along the track. These findings support the view that CA1 and CA3 both participate in encoding the temporal and spatial organization of ongoing experience. Significance Statement: Hippocampal "time cells" that fire at specific moments in a temporally structured memory task have so far been observed only in area CA1, and some studies have suggested that temporal coding within the hippocampus is exclusive to CA1. Here we describe time cells also in CA3, and time cells in both areas are observed even without working memory demands, similar to place cells in these areas. However, unlike equivalent spatial coding along a path, temporal coding is nonlinear, with greater temporal resolution earlier than later in temporally structured experiences. These observations reveal both similarities and differences in temporal and spatial coding within the hippocampus of importance to understanding how these features of memory are represented in the hippocampus.},
}
@misc{SamadaniandHeeger2013,
     title = {Discussions regarding the prospects for neurological assessment using eye-trackng technology},
    author = {Samadani, Uzma and Heeger, David},
      year = 2013,
howpublished = {Personal Communication},
}
@article{SamanidouetalCoRR-01,
       author = {E. Samanidou, E. Zschischang, D. Stauffer, T. Lux},
        title = {Microscopic Models of Financial Markets},
      journal = {CoRR},
       volume = {arXiv:cond-mat/0110354},
         year = {2001},
      comment = {Submitted to F. Schweitzer (ed.), Microscopic Models for Economic Dynamics, Lecture notes in physics, Springer, Berlin-Heidelberg 2002.},
     abstract = {This review deals with several microscopic models of financial markets which have been studied by economists and physicists over the last decade: Kim-Markowitz, Levy-Levy-Solomon, Cont-Bouchaud, Solomon-Weisbuch, Lux-Marchesi, Donangelo-Sneppen and Solomon-Levy-Huang. After an overview of simulation approaches in financial economics, we first give a summary of the Donangelo-Sneppen model of monetary exchange and compare it with related models in economics literature. Our selective review then outlines the main ingredients of some influential early models of multi-agent dynamics in financial markets (Kim-Markowitz, LevyLevy-Solomon). As will be seen, these contributions draw their inspiration from the complex appearance of investors' interactions in real-life markets. Their main aim is to reproduce (and, thereby, provide possible explanations) for the spectacular bubbles and crashes seen in certain historical episodes, but they lack (like almost all the work before 1998 or so) a perspective in terms of the universal statistical features of financial time series. In fact, awareness of a set of such regularities (power-law tails of the distribution of returns, temporal scaling of volatility) only gradually appeared over the nineties. With the more precise description of the formerly relatively vague characteristics (e.g. moving from the notion of fat tails to the more concrete one of a power-law with index around three), it became clear that financial markets dynamics give rise to some kind of universal scaling laws. Showing similarities with scaling laws for other systems with many interacting subunits, an exploration of financial markets as multi-agent systems appeared to be a natural consequence. This topic was pursued by quite a number of contributions appearing in both the physics and economics literature since the late nineties. From the wealth of different flavors of multi-agent models that have appeared by now, we discuss the Cont-Bouchaud, Solomon-Levy-Huang and Lux-Marchesi models. Open research questions are discussed in our concluding section.}
}
@article{SanandMaulikCoRR-17,
        title = {Neural network closures for nonlinear model order reduction},
       author = {Omer San, Romit Maulik},
      journal = {CoRR},
       volume = {arXiv:1705.08532},
          url = {https://arxiv.org/pdf/1705.08532.pdf},
     abstract = {Many reduced order models are neither robust with respect to the parameter changes nor cost-effective enough for handling the nonlinear dependence of complex dynamical systems. In this study, we put forth a robust machine learning framework for projection based reduced order modeling of such nonlinear and nonstationary systems. As a demonstration, we focus on a nonlinear advection-diffusion system given by the viscous Burgers equation, which is a prototype setting of more realistic fluid dynamics applications with the same quadratic nonlinearity. In our proposed methodology the effects of truncated modes are modeled using a singe layer feed-forward neural network architecture. The neural network architecture is trained by utilizing both the Bayesian regularization and extreme learning machine approaches, where the latter one is found to be computationally more efficient. A particular effort is devoted to the selection of basis functions considering the proper orthogonal decomposition and Fourier bases. It is shown that the proposed models yield significant improvements in the accuracy over the standard Galerkin projection models with a negligibly small computational overhead and provide reliable predictions with respect to the parameter changes.},
}
@article{SanchezetalCoRR-18,
       author = {Alvaro Sanchez{-}Gonzalez and Nicolas Heess and Jost Tobias Springenberg and Josh Merel and Martin A. Riedmiller and Raia Hadsell and Peter Battaglia},
        title = {Graph networks as learnable physics engines for inference and control},
      journal = {CoRR},
       volume = {arXiv:1806.01242},
         year = {2018},
     abstract = {Understanding and interacting with everyday physical scenes requires rich knowledge about the structure of the world, represented either implicitly in a value or policy function, or explicitly in a transition model. Here we introduce a new class of learnable models--based on graph networks--which implement an inductive bias for object- and relation-centric representations of complex, dynamical systems. Our results show that as a forward model, our approach supports accurate predictions from real and simulated data, and surprisingly strong and efficient generalization, across eight distinct physical systems which we varied parametrically and structurally. We also found that our inference model can perform system identification. Our models are also differentiable, and support online planning via gradient-based trajectory optimization, as well as offline policy optimization. Our framework offers new opportunities for harnessing and exploiting rich knowledge about the world, and takes a key step toward building machines with more human-like representations of the world.},
}
@article{SanderetalBOE-12,
     title = {Magnetoencephalography with a chip-scale atomic magnetometer}, 
    author = {T. H. Sander and J. Preusser and R. Mhaskar and J. Kitching and L. Trahms and S. Knappe}, 
   journal = {Biomedical Optics Express}, 
 publisher = {Optical Society of America},
    volume = 3, 
     issue = 5, 
      year = 2012,
     pages = {981-990}, 
  abstract = {We report on the measurement of somatosensory-evoked and spontaneous magnetoencephalography (MEG) signals with a chip-scale atomic magnetometer (CSAM) based on optical spectroscopy of alkali atoms. The uncooled, fiber-coupled CSAM has a sensitive volume of 0.77 mm3 inside a sensor head of volume 1 cm3 and enabled convenient handling, similar to an electroencephalography (EEG) electrode. When positioned over O1 of a healthy human subject, $\alpha$-oscillations were observed in the component of the magnetic field perpendicular to the scalp surface. Furthermore, by stimulation at the right wrist of the subject, somatosensory-evoked fields were measured with the sensors placed over C3. Higher noise levels of the CSAM were partly compensated by higher signal amplitudes due to the shorter distance between CSAM and scalp.},
}
@inproceedings{SandewallIJCAI-89,
       author = {Sandewall, Erik},
        title = {Filter Preferential Entailment for the Logic of Action in Almost Continuous Worlds},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
         year = 1989,
}
@incollection{SandewallKR-89,
       author = {Sandewall, Erik},
        title = {Combining Logic and Differential Equations for Describing Real-World Systems},
     crossref = {KR-89},
        pages = {412-420},
}
@incollection{SandewallandShoham94,
       author = {E.J. Sandewall and Y. Shoham},
        title = {Nonmonotonic Temporal Reasoning},
    booktitle = {Handbook of Logic in Artificial Intelligence and Logic Programming},
    publisher = {Elsevier},
       editor = {D. Gabbai},
         year = 1994,
}
@inproceedings{SandleretalNIPS-08,
     author = {Ted Sandler and John Blitzer and Partha Talukdar and Fernando Pereira},
      title = {Regularized Learning with Networks of Features},
  booktitle = {Neural Information Processing Systems},
  publisher = {MIT Press},
    address = {Cambridge, MA},
       year = 2009,
}
@article{SanesetalNEURON-10,
       author = {Sanes, Joshua R. and Zipursky, S. Lawrence},
        title = {Design Principles of Insect and Vertebrate Visual Systems},
      journal = {Neuron},
         year = {2010},
       volume = {66},
        issue = {1},
        pages = {15-36},
     abstract = {A century ago, Cajal noted striking similarities between the neural circuits that underlie vision in vertebrates and flies. Over the past few decades, structural and functional studies have provided strong support for Cajal's view. In parallel, genetic studies have revealed some common molecular mechanisms controlling development of vertebrate and fly visual systems and suggested that they share a common evolutionary origin. Here, we review these shared features, focusing on the first several layers - retina, optic tectum (superior colliculus) and lateral geniculate nucleus in vertebrates, and retina, lamina and medulla in fly. We argue that vertebrate and fly visual circuits utilize common design principles, and that taking advantage of this phylogenetic conservation will speed progress in elucidating both functional strategies and developmental mechanisms, as has already occurred in other areas of neurobiology ranging from electrical signaling and synaptic plasticity to neurogenesis and axon guidance.},
}
@article{SangandJacksonASENT-05,
       author = {Sang, Tzu-Kang and Jackson, George R.},
        title = {Drosophila Models of Neurodegenerative Disease},
      journal = {NeuroX},
    publisher = {The American Society for Experimental NeuroTherapeutics},
       volume = {2},
        issue = {3},
         year = {2005},
        pages = {438-446},
     abstract = {Over the last two decades, a number of mutations have been identified that give rise to neurodegenerative disorders, including familial forms of Alzheimer's disease, Parkinson's disease, and amyotrophic lateral sclerosis. Although in most cases sporadic cases vastly outnumber familial forms of such diseases, study of such inherited forms has the potential to provide powerful clues regarding the pathophysiological basis of neurodegeneration. One powerful approach to analyzing disease mechanisms is the development of transgenic animal models, most notably in the mouse. However, development and analysis of such models can be costly and time consuming. Development of improved transgenic technologies have contributed to the development of Drosophila models of a number of neurodegenerative disorders that have shown striking similarities to the human diseases. Moreover, genetic screens using such models have begun to unravel aspects of the pathophysiological basis of neurodegenerative disorders. Here, we provide a general overview of fly models pertinent to trinucleotide repeat expansion disorders, Alzheimer's, and Parkinson's diseases, and highlight key genetic modifiers that have been identified to date using such models.}
}
@article{SangandJacksonNEURORX-05,
       author = {Sang, Tzu-Kang and Jackson, George R},
        title = {Drosophila Models of Neurodegenerative Disease},
      journal = {NeuroRx},
    publisher = {The American Society for Experimental NeuroTherapeutics, Inc.},
       volume = {2},
        issue = {3},
         year = {2005},
        pages = {438-446},
     abstract = {Over the last two decades, a number of mutations have been identified that give rise to neurodegenerative disorders, including familial forms of Alzheimer's disease, Parkinson's disease, and amyotrophic lateral sclerosis. Although in most cases sporadic cases vastly outnumber familial forms of such diseases, study of such inherited forms has the potential to provide powerful clues regarding the pathophysiological basis of neurodegeneration. One powerful approach to analyzing disease mechanisms is the development of transgenic animal models, most notably in the mouse. However, development and analysis of such models can be costly and time consuming. Development of improved transgenic technologies have contributed to the development of Drosophila models of a number of neurodegenerative disorders that have shown striking similarities to the human diseases. Moreover, genetic screens using such models have begun to unravel aspects of the pathophysiological basis of neurodegenerative disorders. Here, we provide a general overview of fly models pertinent to trinucleotide repeat expansion disorders, Alzheimer's, and Parkinson's diseases, and highlight key genetic modifiers that have been identified to date using such models.}
}
@article{SangerNN-89,
        title = {Optimal unsupervised learning in a single-layer linear feedforward neural network},
       author = {Sanger, Terence D.},
      journal = {Neural Networks},
       volume = 2,
        issue = 6,
        pages = {459-473},
         year = 1989
}
@article{SantoroetalCUP-19,
        title = {Is coding a relevant metaphor for building AI? A commentary on "Is coding a relevant metaphor for the brain?", by Romain Brette},
       author = {Santoro, Adam and Hill, Felix and Barrett, David and Raposo, David and Botvinick, Matt and Lillicrap, Timothy},
       volume = {42},
      journal = {Behavioral and Brain Sciences},
    publisher = {Cambridge University Press (CUP)},
         year = {2019},
     abstract = {We agree with Lake and colleagues on their list of key ingredients for building humanlike intelligence, including the idea that model-based reasoning is essential. However, we favor an approach that centers on one additional ingredient: autonomy. In particular, we aim toward agents that can both build and exploit their own internal models, with minimal human hand-engineering. We believe an approach centered on autonomous learning has the greatest chance of success as we scale toward real-world complexity, tackling domains for which ready-made formal models are not available. Here we survey several important examples of the progress that has been made toward building autonomous agents with humanlike abilities, and highlight some outstanding challenges.}
}
@article{SantoroetalCoRR-21,
        title = {Symbolic Behaviour in Artificial Intelligence},
       author = {Adam Santoro and Andrew Lampinen and Kory Mathewson and Timothy Lillicrap and David Raposo},
       journAL = {CoRR},
         year = {2021},
       volume = {arXiv:2102.03406},
     abstract = {The ability to use symbols is the pinnacle of human intelligence, but has yet to be fully replicated in machines. Here we argue that the path towards symbolically fluent artificial intelligence (AI) begins with a reinterpretation of what symbols are, how they come to exist, and how a system behaves when it uses them. We begin by offering an interpretation of symbols as entities whose meaning is established by convention. But crucially, something is a symbol only for those who demonstrably and actively participate in this convention. We then outline how this interpretation thematically unifies the behavioural traits humans exhibit when they use symbols. This motivates our proposal that the field place a greater emphasis on symbolic behaviour rather than particular computational mechanisms inspired by more restrictive interpretations of symbols. Finally, we suggest that AI research explore social and cultural engagement as a tool to develop the cognitive machinery necessary for symbolic behaviour to emerge. This approach will allow for AI to interpret something as symbolic on its own rather than simply manipulate things that are only symbols to human onlookers, and thus will ultimately lead to AI with more human-like symbolic fluency.}
}
@inproceedings{SantoroetalICML-16,
        title = {Meta-Learning with Memory-Augmented Neural Networks},
       author = {Adam Santoro and Sergey Bartunov and Matthew Botvinick and Daan Wierstra and Timothy Lillicrap},
    booktitle = {Proceedings of The 33rd International Conference on Machine Learning},
        pages = {1842-1850},
         year = {2016},
       editor = {Maria Florina Balcan and Kilian Q. Weinberger},
       volume = {48},
       series = {Proceedings of Machine Learning Research},
      address = {New York, New York, USA},
    publisher = {PMLR},
     abstract = {Despite recent breakthroughs in the applications of deep neural networks, one setting that presents a persistent challenge is that of "one-shot learning." Traditional gradient-based networks require a lot of data to learn, often through extensive iterative training. When new data is encountered, the models must inefficiently relearn their parameters to adequately incorporate the new information without catastrophic interference. Architectures with augmented memory capacities, such as Neural Turing Machines (NTMs), offer the ability to quickly encode and retrieve new information, and hence can potentially obviate the downsides of conventional models. Here, we demonstrate the ability of a memory-augmented neural network to rapidly assimilate new data, and leverage this data to make accurate predictions after only a few samples. We also introduce a new method for accessing an external memory that focuses on memory content, unlike previous methods that additionally use memory location-based focusing mechanisms.}
}
@incollection{SantoroetalNIPS-17,
       author = {Santoro, Adam and Raposo, David and Barrett, David G and Malinowski, Mateusz and Pascanu, Razvan and Battaglia, Peter and Lillicrap, Timothy},
        title = {A simple neural network module for relational reasoning},
    booktitle = {Advances in Neural Information Processing Systems 30},
       editor = {I. Guyon and U. V. Luxburg and S. Bengio and H. Wallach and R. Fergus and S. Vishwanathan and R. Garnett},
    publisher = {Curran Associates, Inc.},
        pages = {4967-4976},
         year = {2017},
     abstract = {Relational reasoning is a central component of generally intelligent behavior, but has proven difficult for neural networks to learn. In this paper we describe how to use Relation Networks (RNs) as a simple plug-and-play module to solve problems that fundamentally hinge on relational reasoning. We tested RN-augmented networks on three tasks: visual question answering using a challenging dataset called CLEVR, on which we achieve state-of-the-art, super-human performance; text-based question answering using the bAbI suite of tasks; and complex reasoning about dynamical physical systems. Then, using a curated dataset called Sort-of-CLEVR we show that powerful convolutional networks do not have a general capacity to solve relational questions, but can gain this capacity when augmented with RNs. Thus, by simply augmenting convolutions, LSTMs, and MLPs with RNs, we can remove computational burden from network components that are not well-suited to handle relational reasoning, reduce overall network complexity, and gain a general ability to reason about the relations between entities and their properties.}
}
@phdthesis{SantosTR-92,
       author = {Eugene Santos Jr.},
        title = {A Linear Constraint Satisfaction Approach for Abductive Reasoning},
       school = {Department of Computer Science, Brown University},
         year = 1992,
}
@inproceedings{SantosUAI-91,
       author = {Eugene Santos Jr.},
        title = {On the Generation of Alternative Explanations with Implications for Belief Revision},
    booktitle = {Seventh Conference on Uncertainty in Artificial Intelligence},
         year = 1991,
}
@article{SaperetalNATURE-05,
       author = {Saper, Clifford B. and Scammell, Thomas E. and Lu, Jun},
      journal = {Nature},
        pages = {1257},
    publisher = {Nature Publishing Group},
        title = {Hypothalamic regulation of sleep and circadian rhythms},
       volume = {437},
         year = {2005},
     abstract = {A series of findings over the past decade has begun to identify the brain circuitry and neurotransmitters that regulate our daily cycles of sleep and wakefulness. The latter depends on a network of cell groups that activate the thalamus and the cerebral cortex. A key switch in the hypothalamus shuts off this arousal system during sleep. Other hypothalamic neurons stabilize the switch, and their absence results in inappropriate switching of behavioural states, such as occurs in narcolepsy. These findings explain how various drugs affect sleep and wakefulness, and provide the basis for a wide range of environmental influences to shape wake-sleep cycles into the optimal pattern for survival.}
}
@book{Sapolsky2004,
     title = {Why Zebras Don't Get Ulcers},
    author = {Robert Sapolsky},            
 publisher = {Henry Holt and Company},
   address = {New York, NY}, 
      year = 2004,
}
@book{Sapolsky2010,
     title = {Stress and Your Body}, 
    author = {Robert Sapolsky},
    series = {Great Courses}, 
 publisher = {Teaching Company},
   address = {Chantily, VA},
      year = 2010,
  abstract = {Contents: 

  Disc 1. Why don't zebras get ulcers? -- The nuts and bolts of the
  stress-response -- Stress and your heart -- Stress, metabolism, and
  liquidating your assets -- Stress, overeating, and your digestive
  tract -- Stress and growth - echoes from the womb.

  Disc 2. Stress, growth, and child development -- Stress and female
  reproduction -- Stress and male reproduction -- Stress and your
  immune system -- Stress and cancer -- Stress and pain .

  Disc 3. Stress, learning and memory -- Stress, judgment, and impulse
  control -- Stress, sleep, and lack of sleep -- Stress and aging --
  Understanding psychological stress -- Psychological modulators of
  stress.

  Disc 4. Stress and the biology of depression -- Stress and the
  psychology of depression -- Anxiety, hostility, repression, and
  reward -- Stress, health, and low social status -- Stress management
  - clues to success? -- Stress management - approaches and cautions.},
}
@book{Sapolsky2017,
        title = {Behave: The Biology of Humans at Our Best and Worst},
       author = {Sapolsky, Robert M.},
         year = {2017},
    publisher = {Penguin Publishing Group}
}
@book{SapolskyBEHAVE-17,
        title = {Behave: The Biology of Humans at Our Best and Worst},
       author = {Robert Sapolsky},
    publisher = {Penguin Publishing Group},
      address = {New York, NY},
         year = 2017,
}
@misc{SapolskyCAS-11,
       author = {Robert Sapolsky},
        title = {Are Humans Just Another Primate?},
        month = feb,
         year = {February, 2011},
 howpublished = {\url{http://fora.tv/2011/02/15/Robert_Sapolsky_Are_Humans_Just_Another_Primate}, {California Academy of Sciences}},
}
@inproceedings{SarachikRA-89,
       author = {Sarachik, Karen B.},
        title = {Characterizing an Indoor Environment with a Mobile Robot and Uncalibrated Stereo},
    booktitle = {Proceedings of the {IEEE} International Conference on Robotics and Automation},
         year = 1989,
}
@book{Saridis77,
       author = {Saridis, George N.},
        title = {Self-Organizing Controls of Stochastic Systems},
    publisher = {Marcel Dekker},
      address = {New York},
         year = 1977,
}
@inproceedings{SaridisSPIE-87,
       author = {Saridis, George N. and Valavanis, Kimon P.},
        title = {On the Theory of Intelligent Controls},
    booktitle = {Proceedings of the {SPIE} Conference on Advances in Intelligent
                   Robotic Systems},
 organization = {SPIE},
     location = {Cambridge, Massachusetts},
         year = 1987,
        pages = {488-495},
}
@article{SarntheinetalPNAS-98,
       author = {Sarnthein, J. and Petsche, H. and Rappelsberger, P. and Shaw, G. L. and von Stein, A.},
        title = {Synchronization between prefrontal and posterior association cortex during human working memory},
      journal = {Proceedings of the National Academy of Sciences of the United States of America},
    publisher = {The National Academy of Sciences},
       volume = {95},
        issue = {12},
         year = {1998},
        pages = {7092-7096},
     abstract = {We measured coherence between the electroencephalogram at different scalp sites while human subjects performed delayed response tasks. The tasks required the retention of either verbalizable strings of characters or abstract line drawings. In both types of tasks, a significant enhancement in coherence in the theta range (4-7 Hz) was found between prefrontal and posterior electrodes during 4-s retention intervals. During 6-s perception intervals, far fewer increases in theta coherence were found. Also in other frequency bands, coherence increased; however, the patterns of enhancement made a relevance for working memory processes seem unlikely. Our results suggest that working memory involves synchronization between prefrontal and posterior association cortex by phase-locked, low frequency (4-7 Hz) brain activity.},
}
@book{Sarpeshkar2010,
        title = {Ultra Low Power Bioelectronics: Fundamentals, Biomedical Applications, and Bio-inspired Systems},
       author = {Rahul Sarpeshkar},
         year = {2010},
    publisher = {Cambridge University Press},
}
@article{SarpeshkarNC-98,
       author = {Sarpeshkar, Rahul},
        title = {{A}nalog versus digital: extrapolating from electronics to neurobiology},
      journal = {Neural Computing},
       volume = 10,
        issue = 7,
         year = 1998,
        pages = {1601-1638},
}
@article{SasakietalCURRENT-OPINION-15,
       author = {Sasaki, T. and Leutgeb, S. and Leutgeb, J. K.},
        title = {Spatial and memory circuits in the medial entorhinal cortex},
      journal = {Current Opinion Neurobiology},
         year = {2015},
       volume = {32},
        pages = {16-23},
     abstract = {The large capacity of episodic memory is thought to be supported by the emergence of distinct hippocampal cell assemblies for unrelated memories, such that interference is minimized. In large-scale population recordings, the orthogonal nature of hippocampal representations across environments is evident in the complete reorganization of the firing locations of hippocampal place cells. Entorhinal grid cells provide inputs to the hippocampus, and their firing patterns shift relative to each other across different environments. Although this suggests that altered grid cell firing could generate distinct hippocampal population codes, it has recently been shown that new and distinct hippocampal place fields emerge while grid cell firing is compromised. We therefore propose that separate circuits within the medial entorhinal cortex are specialized for performing either spatial or memory-related computations.}
}
@book{Sastry89,
       author = {Sastry, Shankar and Bodson, Marc},
        title = {Adaptive Control: Stability, Convergence, and Robustness},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1989,
}
@article{SatiaandLaveOR-73,
       author = {Satia, J. K. and Lave, R. E.},
        title = {{M}arkovian Decision Processes with Uncertain Transition Probabilities},
      journal = {Operations Research},
       volume = 21,
         year = 1973,
        pages = {728-740},
}
@incollection{SatoandandTouharaCSIMFI-09,
       author = {Sato, K. and Touhara, K.},
       editor = {Korsching, Sigrun and Meyerhof, Wolfgang},
        title = {Insect Olfaction: Receptors, Signal Transduction, and Behavior},
    booktitle = {Chemosensory Systems in Mammals, Fishes, and Insects},
    publisher = {Springer Berlin Heidelberg},
      address = {Berlin, Heidelberg},
         year = {2009},
        pages = {203-220},
     abstract = {The insect olfactory system is a suitable model for exploring molecular function of odorant receptors, axonal projection of olfactory receptor neurons onto secondary neurons, and the neural circuit for odor perception. Recent progress in the study of insect olfaction revealed that the heteromeric insect olfactory receptor complex forms a cation nonselective ion channel directly gated by odor or pheromone ligands independent of known G-protein signaling pathways. Despite fundamental differences in transduction machineries between insects and vertebrates, the anatomical and functional features of insect odor-coding strategy are similar and thus justify any consideration of mammalian olfaction in the study of insects. The understanding of the molecular mechanism of insect olfaction will help in the development of insect repellents for controlling insect pest and vector populations for a wide range of pathogens.}
}
@article{SattaretalLANCET-10,
        title = {Statins and risk of incident diabetes: a collaborative meta-analysis of randomised statin trials},
       author = {Sattar, Naveed and Preiss, David and Murray, Heather M and Welsh, Paul and Buckley, Brendan M and 
              de Craen, Anton JM and Seshasai, Sreenivasa Rao Kondapally and McMurray, John J and Freeman, Dilys J 
              and Jukema, J Wouter and Macfarlane, Peter W and Packard, Chris J and Stott, David J and Westendorp, 
              Rudi G and Shepherd, James and Davis, Barry R and Pressel, Sara L and Marchioli, Roberto and Marfisi, 
              Rosa Maria and Maggioni, Aldo P and Tavazzi, Luigi and Tognoni, Gianni and Kjekshus, John and 
              Pedersen, Terje R and Cook, Thomas J and Gotto, Antonio M and Clearfield, Michael B and Downs, John 
              R and Nakamura, Haruo and Ohashi, Yasuo and Mizuno, Kyoichi and Ray, Kausik K and Ford, Ian},
      journal = {The Lancet},
       volume = 375,
        issue = 9716,
        pages = {735-742},
         year = 2010,
     abstract = {Trials of statin therapy have had conflicting findings on the risk of development of diabetes mellitus in patients given statins. We aimed to establish by a meta-analysis of published and unpublished data whether any relation exists between statin use and development of diabetes. We searched Medline, Embase, and the Cochrane Central Register of Controlled Trials from 1994 to 2009, for randomised controlled endpoint trials of statins. We included only trials with more than          year =  patients, with identical follow-up in both groups and duration of more than 1 year. We excluded trials of patients with organ transplants or who needed haemodialysis. We used theI2statistic to measure heterogeneity between trials and calculated risk estimates for incident diabetes with random-effect meta-analysis. We identified 13 statin trials with 91,140 participants, of whom 4278 (2226 assigned statins and 2052 assigned control treatment) developed diabetes during a mean of 4 years. Statin therapy was associated with a 9\% increased risk for incident diabetes, with little heterogeneity between trials. Meta-regression showed that risk of development of diabetes with statins was highest in trials with older participants, but neither baseline body-mass index nor change in LDL-cholesterol concentrations accounted for residual variation in risk. Treatment of 255 (95\% CI 150-852) patients with statins for 4 years resulted in one extra case of diabetes. Statin therapy is associated with a slightly increased risk of development of diabetes, but the risk is low both in absolute terms and when compared with the reduction in coronary events. Clinical practice in patients with moderate or high cardiovascular risk or existing cardiovascular disease should not change.}
}
@incollection{Sauer94,
       author = {Sauer, Tim},
        title = {Time Series Prediction by Using Delayed Coordinate Embedding},
    booktitle = {Time Series Prediction: Forecasting the Future and Understanding the Past},
       editor = {Weigend, Andreas S. and Gershenfeld, Neil A.},
    publisher = {Addison-Wesley},
         year = 1994,
}
@inproceedings{SaulandJordanNIPS-95,
       author = {Saul, L. K. and Jordan, M. I.},
        title = {Boltzmann Chains and Hidden {M}arkov Models},
       editor = {Touretzky, D. S. and Leen, T. K.},
    booktitle = {Advances in Neural Information Processing 7},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1995,
}
@inproceedings{SaulandJordanNIPS-96,
       author = {Saul, L. K. and Jordan, M. I.},
        title = {Exploiting Tractable Substructures in Intractable Networks},
       editor = {Touretzky, D. S. and Mozer, M. C. and Hasselmo, M.E.},
    booktitle = {Advances in Neural Information Processing 8},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1996,
}
@article{SaulandRoweisJMLR-03,
       author = {Saul, Lawrence K. and Roweis, Sam T.},
        title = {Think Globally, Fit Locally: Unsupervised Learning of Low Dimensional Manifolds},
      journal = {Journal Machine Learning Research},
       volume = {4},
         year = {2003},
        pages = {119-155},
     abstract = {The problem of dimensionality reduction arises in many fields of information processing, including machine learning, data compression, scientific visualization, pattern recognition, and neural computation. Here we describe locally linear embedding (LLE), an unsupervised learning algorithm that computes low dimensional, neighborhood preserving embeddings of high dimensional data. The data, assumed to be sampled from an underlying manifold, are mapped into a single global coordinate system of lower dimensionality. The mapping is derived from the symmetries of locally linear reconstructions, and the actual computation of the embedding reduces to a sparse eigenvalue problem. Notably, the optimizations in LLE---though capable of generating highly nonlinear embeddings---are simple to implement, and they do not involve local minima. In this paper, we describe the implementation of the algorithm in detail and discuss several extensions that enhance its performance. We present results of the algorithm applied to data sampled from known manifolds, as well as to collections of images of faces, lips, and handwritten digits. These examples are used to provide extensive illustrations of the algorithm's performance---both successes and failures---and to relate the algorithm to previous and ongoing work in nonlinear dimensionality reduction.},
}
@article{SausengandKlimeschNBR-08,
       author = {Sauseng, P. and Klimesch, W.},
        title = {What does phase information of oscillatory brain activity tell us about cognitive processes?},
      journal = {Neuroscience \& Biobehavioral Reviews},
       volume = 32,
        issue = 5,
         year = 2008,
       pages = {1001-1013},
     abstract = {The electroencephalogram (EEG) bears the possibility to investigate oscillatory processes in the human brain. In the animal brain it has been shown that the phase of cortical oscillations is related to the exact timing of neural activity. The potential role of oscillatory phase and phase synchronization for the explanation of cortical information processing has been largely underestimated in the human EEG until now. Here it is argued that EEG phase (synchronization) reflects the exact timing of communication between distant but functionally related neural populations, the exchange of information between global and local neuronal networks, and the sequential temporal activity of neural processes in response to incoming sensory stimuli. Three different kinds of phase synchronization are discussed: (i) phase coupling between brain sites, (ii) phase synchronization across frequencies, and (iii) phase-locking to external events. In this review recent work is presented demonstrating that EEG phase synchronization provides valuable information about the neural correlates of various cognitive processes, and that it leads to a better understanding of how memory and attention processes are interrelated.}
}
@book{Savage72,
       author = {Savage, Leonard J.},
        title = {The Foundations of Statistics, Second Edition},
    publisher = {Dover Publications},
      address = {New York},
         year = 1972,
}
@article{SavalletalNATURE-METHODS-15,
       author = {Savall, Joan and Ho, Eric Tatt Wei and Huang, Cheng and Maxey, Jessica R. and Schnitzer, Mark J.},
        title = {Dexterous robotic manipulation of alert adult Drosophila for high-content experimentation},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = {12},
        issue = {7},
         year = {2015},
        pages = {657-660},
     abstract = {We present a robot that enables high-content studies of alert adult Drosophila by combining operations including gentle picking; translations and rotations; characterizations of fly phenotypes and behaviors; microdissection; or release. To illustrate, we assessed fly morphology, tracked odor-evoked locomotion, sorted flies by sex, and dissected the cuticle to image neural activity. The robot's tireless capacity for precise manipulations enables a scalable platform for screening flies' complex attributes and behavioral patterns.}
}
@article{SawyerNARRATIVIUM-02,
       author = {Andy Sawyer},
        title = {Narrativium and Lies-to-Children: "Palatable Instruction" in "The Science of Discworld"},
       editor = {Brian Attebery},
      journal = {Journal of the Fantastic in the Arts},
    publisher = {International Association for the Fantastic in the Arts},
       volume = {13},
       number = {1 (49)},
         year = {2002},
        pages = {62-81},
     abstract = {This discusses not so much the science in science fiction as aspects of the science fiction in science: a fusion brought together with interesting effect in The Science of Discworld by Terry Pratchett, Jack Cohen, and Ian Stewart.}
}
@inproceedings{SaxeetalICML-11,
       author = {Andrew Saxe and Pang Wei Koh and Zhenghao Chen and Maneesh Bhand and Bipin Suresh and Andrew Ng},
        title = {On Random Weights and Unsupervised Feature Learning},
    booktitle = {Proceedings of the 28th International Conference on Machine Learning},
       editor = {Lise Getoor and Tobias Scheffer},
         year = 2011,
        pages = {1089-1096},
}
@inproceedings{SaxeetalNIPS-11,
        title = {Unsupervised learning models of primary cortical receptive fields and receptive field plasticity},
       author = {Saxe, A. and Bhand, M. and Mudur, R. and Suresh, B. and Ng, A.},
    booktitle = {Advances in Neural Information Processing Systems 24},
       editor = {J. Shawe-Taylor and R.S. Zemel and P. Bartlett and F. Pereira and K.Q. Weinberger},
    publisher = {MIT Press},
         year = 2011
}
@inproceedings{SaxenaetalICCV-07,
        title = {Learning 3-{D} Scene Structure from a Single Still Image},
       author = {Saxena, Ashutosh and Sun, Min and Ng, Andrew Y.},
    booktitle = {Proceedings of the 11th {IEEE} International Conference on Computer Vision},
        pages = {1-8},
         year = 2007,
}
@article{SaxenaetalIJCV-07,
       author = {Ashutosh Saxena and Sung Chung and Andrew Ng},
        title = {3-{D} Depth Reconstruction from a Single Still Image},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 76,
        issue = 1,
        pages = {53-69},
         year = 2007,
}
@inproceedings{SaxtonetalICLR-19,
        title = {Analysing Mathematical Reasoning Abilities of Neural Models},
       author = {David Saxton and Edward Grefenstette and Felix Hill and Pushmeet Kohli},
    booktitle = {International Conference on Learning Representations},
         year = {2019},
     abstract = {Mathematical reasoning-a core ability within human intelligence-presents some unique challenges as a domain: we do not come to understand and solve mathematical problems primarily on the back of experience and evidence, but on the basis of inferring, learning, and exploiting laws, axioms, and symbol manipulation rules. In this paper, we present a new challenge for the evaluation (and eventually the design) of neural architectures and similar system, developing a task suite of mathematics problems involving sequential questions and answers in a free-form textual input/output format. The structured nature of the mathematics domain, covering arithmetic, algebra, probability and calculus, enables the construction of training and test splits designed to clearly illuminate the capabilities and failure-modes of different architectures, as well as evaluate their ability to compose and relate knowledge and learned processes. Having described the data generation process and its potential future expansions, we conduct a comprehensive analysis of models from two broad classes of the most powerful sequence-to-sequence architectures and find notable differences in their ability to resolve mathematical problems and generalize their knowledge.}
}
@article{ScanzianiandHausserNATURE-09,
        title = {Electrophysiology in the age of light},
       author = {Scanziani, Massimo and Hausser, Michael},
      journal = {Nature},
        issue = 7266,
    publisher = {Nature Publishing Group},
       volume = 461,
         year = 2009,
        pages = {930-939},
     abstract = {Electrophysiology, the 'gold standard' for investigating neuronal signalling, is being challenged by a new generation of optical probes. Together with new forms of microscopy, these probes allow us to measure and control neuronal signals with spatial resolution and genetic specificity that already greatly surpass those of electrophysiology. We predict that the photon will progressively replace the electron for probing neuronal function, particularly for targeted stimulation and silencing of neuronal populations. Although electrophysiological characterization of channels, cells and neural circuits will remain necessary, new combinations of electrophysiology and imaging should lead to transformational discoveries in neuroscience.},
}
@article{ScellierandBengioCoRR-16a,
       author = {Benjamin Scellier and Yoshua Bengio},
        title = {Towards a Biologically Plausible Backprop},
      journal = {CoRR},
       volume = {arXiv:1602.05179v2},
         year = {2016},
     abstract = {Neuroscientists have long criticised deep learning algorithms as incompatible with current knowledge of neurobiology. We explore more biologically plausible versions of deep representation learning, focusing here mostly on unsupervised learning but developing a learning mechanism that could account for supervised, unsupervised and reinforcement learning. The starting point is that the basic learning rule believed to govern synaptic weight updates (Spike-TimingDependent Plasticity) arises out of a simple update rule that makes a lot of sense from a machine learning point of view and can be interpreted as gradient descent on some objective function so long as the neuronal dynamics push firing rates towards better values of the objective function (be it supervised, unsupervised, or rewarddriven). The second main idea is that this corresponds to a form of the variational EM algorithm, i.e., with approximate rather than exact posteriors, implemented by neural dynamics. Another contribution of this paper is that the gradients required for updating the hidden states in the above variational interpretation can be estimated using an approximation that only requires propagating activations forward and backward, with pairs of layers learning to form a denoising auto-encoder. Finally, we extend the theory about the probabilistic interpretation of auto-encoders to justify improved sampling schemes based on the generative interpretation of denoising auto-encoders, and we validate all these ideas on generative learning tasks.},
}
@article{ScellierandBengioCoRR-16b,
       author = {Benjamin Scellier and Yoshua Bengio},
        title = {Equilibrium Propagation: Bridging the Gap Between Energy-Based Models and Backpropagation},
      journal = {CoRR},
       volume = {arXiv:1602.05179v4},
         year = {2016},
     abstract = {We introduce Equilibrium Propagation (e-prop), a learning algorithm for energy-based models. This algorithm involves only one kind of neural computation both for the first phase (when the prediction is made) and the second phase (after the target is revealed) of training. Contrary to backpropagation in feedforward networks, there is no need for special computation in the second phase of our learning algorithm. Equilibrium Propagation combines features of Contrastive Hebbian Learning and Contrastive Divergence while solving the theoretical issues of both algorithms: the algorithm computes the exact gradient of a well defined objective function. Because the objective function is defined in terms of local perturbations, the second phase of e-prop corresponds to only nudging the first-phase fixed point towards a configuration that has lower cost value. In the case of a multi-layer supervised neural network, the output units are slightly nudged towards their target, and the perturbation introduced at the output layer propagates backward in the network. The theory developed in this paper shows that the signal 'back-propagated' during this second phase actually contains information about the error derivatives, which we use to implement a learning rule proved to perform gradient descent with respect to the objective function. Thus, this work makes it more plausible that a mechanism similar to backpropagation could be implemented by brains.}
}
@incollection{SchacteretalCN-09,
        title = {Constructive memory and the simulation of future events},
       author = {D.L. Schacter and Addis, D.R and R.L. Buckner},
    booktitle = {The Cognitive Neurosciences {IV}},
       editor = {M.S. Gazzaniga},
    publisher = {MIT Press},
 organization = {MIT Press},
      address = {Cambridge, MA},
         year = 2009,
        pages = {751-762},
}
@article{SchacteretalNEURON-12,
       author = {Daniel L. Schacter and Donna Rose Addis and Demis Hassabis and Victoria C. Martin and R. Nathan Spreng and Karl K. Szpunar},
        title = {The Future of Memory: Remembering, Imagining, and the Brain},
      journal = {Neuron},
       volume = 76,
         year = 2012,
        pages = {677-694},
     abstract = {During the past few years, there has been a dramatic increase in research examining the role of memory in imagination and future thinking. This work has revealed striking similarities between remembering the past and imagining or simulating the future, including the finding that a common brain network underlies both memory and imagination. Here, we discuss a number of key points that have emerged during recent years, focusing in particular on the importance of distinguishing between temporal and nontemporal factors in analyses of memory and imagination, the nature of differences between remembering the past and imagining the future, the identification of component processes that comprise the default network supporting memory-based simulations, and the finding that this network can couple flexibly with other networks to support complex goal-directed simulations. This growing area of research has broadened our conception of memory by highlighting the many ways in which memory supports adaptive functioning.}
}
@inproceedings{SchaferetalICANN-06,
       author = {Sch\"{a}fer, Anton Maximilian and Zimmermann, Hans Georg},
        title = {Recurrent Neural Networks Are Universal Approximators},
    booktitle = {Proceedings of the 16th International Conference on Artificial Neural Networks - Volume Part I},
    publisher = {Springer-Verlag},
     location = {Athens, Greece},
         year = {2006},
        pages = {632-640},
     abstract = {Neural networks represent a class of functions for the efficient identification and forecasting of dynamical systems. It has been shown that feedforward networks are able to approximate any (Borel-)measurable function on a compact domain. Recurrent neural networks (RNNs) have been developed for a better understanding and analysis of open dynamical systems. Compared to feedforward networks they have several advantages which have been discussed extensively in several papers and books. Still the question often arises if RNNs are able to map every open dynamical system, which would be desirable for a broad spectrum of applications. In this paper we give a proof for the universal approximation ability of RNNs in state space model form. The proof is based on the work of Hornik, Stinchcombe, and White about feedforward neural networks.}
}
@book{Schagrinetal85,
       author = {Schagrin, Morton L. and Rapaport, William J. and Dipert, Randall R.},
        title = {Logic: A Computer Approach},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1985,
}
@article{SchaleketalMM-11,
       author = {Schalek, R. and Kasthuri, N. and Hayworth, K. and Berger, D. and Tapia, J. and Morgan, J. and Turaga, S. and Fagerholm, E. and Seung, H. and Lichtman, J.},
        title = {Development of high-throughput, high-resolution {3D} reconstruction of large-volume biological tissue using automated tape collection ultramicrotomy and scanning electron microscopy},
      journal = {Microscopic Microanalysis},
       volume = 1752,
         year = 2011,
        pages = {966-967},
}
@inproceedings{SchalkwyketalASR-10,
        title = {{Google Search by Voice: A case study}},
       author = {Johan Schalkwyk and Doug Beeferman and Francoise Beaufays and Bill Byrne and Ciprian Chelba and Mike Cohen and Maryam Garret and Brian Strope},
    booktitle = {Advances in Speech Recognition: Mobile Environments, Call Centers and Clinics, Springer},
    publisher = {Springer},
         year = 2010,
        pages = {61-90}
}
@book{SchallerandSacks1991,
        title = {A Man Without Words},
       author = {Schaller, S. and Sacks, O.W.},
    publisher = {University of California Press},
         year = {1991},
}
@book{SchallerandSacks1995,
        title = {A Man Without Words},
       author = {Schaller, S. and Sacks, O. and Sacks, O.W.},
         year = {1995},
    publisher = {University of California Press}
}
@article{SchankCP-72,
        title = {Conceptual dependency: {A} theory of natural language understanding},
       author = {Roger C. Schank},
      journal = {Cognitive Psychology},
       volume = {3},
       number = {4},
         year = {1972},
        pages = {552-631},
}
@book{SchankandAbelson77,
       author = {Schank, Roger C. and Abelson, Robert P.},
        title = {Scripts, Plans, Goals and Understanding: An Inquiry into Human Knowledge Structures},
    publisher = {Lawrence Erlbaum},
      address = {Hillsdale, New Jersey},
         year = 1977,
}
@techreport{SchapireMSc,
       author = {Schapire, Robert E.},
        title = {Diversity-Based Inference of Finite Automata},
        issue = {Masters Thesis},
  institution = {{MIT} Laboratory for Computer Science},
         year = 1988,
}
@techreport{SchapirePhD,
       author = {Schapire, Robert E.},
        title = {The Design and Analysis of Efficient Learning Algorithms},
        issue = {MIT/LCS/TR-493},
  institution = {{MIT} Laboratory for Computer Science},
         year = 1991,
}
@article{SchapiroetalHIPPOCAMPUS-19,
       author = {Schapiro, Anna C. and Reid, Allison G. and Morgan, Alexandra and Manoach, Dara S. and Verfaellie, Mieke and Stickgold, Robert},
        title = {The hippocampus is necessary for the consolidation of a task that does not require the hippocampus for initial learning},
      journal = {Hippocampus},
       volume = {29},
       number = {11},
         year = {2019},
        pages = {1091-1100},
     abstract = {Abstract During sleep, the hippocampus plays an active role in consolidating memories that depend on it for initial encoding. There are hints in the literature that the hippocampus may have a broader influence, contributing to the consolidation of memories that may not initially require the area. We tested this possibility by evaluating learning and consolidation of the motor sequence task (MST) in hippocampal amnesics and demographically matched control participants. While the groups showed similar initial learning, only controls exhibited evidence of overnight consolidation. These results demonstrate that the hippocampus can be required for normal consolidation of a task without being required for its acquisition, suggesting that the area plays a broader role in coordinating memory consolidation than has previously been assumed.},
}
@article{SchapiroetalNATURE-COMMUNICATIONS-18,
        title = {Human hippocampal replay during rest prioritizes weakly learned information and predicts memory performance},
       author = {Schapiro, Anna C and McDevitt, Elizabeth A and Rogers, Timothy T and Mednick, Sara C and Norman, Kenneth A},
      journal = {Nature communications},
       volume = {9},
       number = {1},
        pages = {3920},
         year = {2018},
    publisher = {Nature Publishing Group}
}
@article{SchapiroetalPTRS-B-16,
       author = {Anna C. Schapiro  and Nicholas B. Turk-Browne  and Matthew M. Botvinick  and Kenneth A. Norman },
        title = {Complementary learning systems within the hippocampus: a neural network modelling approach to reconciling episodic memory with statistical learning},
      journal = {Philosophical Transactions of the Royal Society B: Biological Sciences},
       volume = {372},
       number = {1711},
        pages = {20160049},
         year = {2017},
     abstract = {A growing literature suggests that the hippocampus is critical for the rapid extraction of regularities from the environment. Although this fits with the known role of the hippocampus in rapid learning, it seems at odds with the idea that the hippocampus specializes in memorizing individual episodes. In particular, the Complementary Learning Systems theory argues that there is a computational trade-off between learning the specifics of individual experiences and regularities that hold across those experiences. We asked whether it is possible for the hippocampus to handle both statistical learning and memorization of individual episodes. We exposed a neural network model that instantiates known properties of hippocampal projections and subfields to sequences of items with temporal regularities. We found that the monosynaptic pathway-the pathway connecting entorhinal cortex directly to region CA1-was able to support statistical learning, while the trisynaptic pathway-connecting entorhinal cortex to CA1 through dentate gyrus and CA3-learned individual episodes, with apparent representations of regularities resulting from associative reactivation through recurrence. Thus, in paradigms involving rapid learning, the computational trade-off between learning episodes and regularities may be handled by separate anatomical pathways within the hippocampus itself. This article is part of the themed issue 'New frontiers for statistical learning in the cognitive sciences'. }
}
@article{ScharfmanandMyersFiNC-2012,
       author = {Scharfman, Helen and Myers, Catherine},
        title = {Hilar mossy cells of the dentate gyrus: A historical perspective},
      journal = {Frontiers in neural circuits},
         year = {2012},
        pages = {106},
       volume = {6},
     abstract = {The circuitry of the dentate gyrus (DG) of the hippocampus is unique compared to other hippocampal subfields because there are two glutamatergic principal cells instead of one: granule cells, which are the vast majority of the cells in the DG, and the so-called "mossy cells." The distinctive appearance of mossy cells, the extensive divergence of their axons, and their vulnerability to excitotoxicity relative to granule cells has led to a great deal of interest in mossy cells. Nevertheless, there is no consensus about the normal functions of mossy cells and the implications of their vulnerability. There even seems to be some ambiguity about exactly what mossy cells are. Here we review initial studies of mossy cells, characteristics that define them, and suggest a practical definition to allow investigators to distinguish mossy cells from other hilar neurons even if all morphological and physiological information is unavailable due to technical limitations of their experiments. In addition, hypotheses are discussed about the role of mossy cells in the DG network, reasons for their vulnerability and their implications for disease.}
}
@article{SchartneretalPLoS-15,
       author = {Schartner, M. and Seth, A. and Noirhomme, Q. and Boly, M. and Bruno, M. A. and Laureys, S. and Barrett, A.},
        title = {Complexity of multi-dimensional spontaneous {EEG} decreases during propofol induced general anaesthesia},
      journal = {PLoS ONE},
         year = {2015},
       volume = {10},
       number = {8},
        pages = {e0133532},
     abstract = {Emerging neural theories of consciousness suggest a correlation between a specific type of neural dynamical complexity and the level of consciousness: When awake and aware, causal interactions between brain regions are both integrated (all regions are to a certain extent connected) and differentiated (there is inhomogeneity and variety in the interactions). In support of this, recent work by Casali et al (2013) has shown that Lempel-Ziv complexity correlates strongly with conscious level, when computed on the EEG response to transcranial magnetic stimulation. Here we investigated complexity of spontaneous high-density EEG data during propofol-induced general anaesthesia. We consider three distinct measures: (i) Lempel-Ziv complexity, which is derived from how compressible the data are; (ii) amplitude coalition entropy, which measures the variability in the constitution of the set of active channels; and (iii) the novel synchrony coalition entropy (SCE), which measures the variability in the constitution of the set of synchronous channels. After some simulations on Kuramoto oscillator models which demonstrate that these measures capture distinct 'flavours' of complexity, we show that there is a robustly measurable decrease in the complexity of spontaneous EEG during general anaesthesia.},
}
@article{SchatzmannetalKER-06,
       author = {Jost Schatzmann and Karl Weilhammer and Matt Stuttle and Steve Young},
        title = {A Survey of Statistical User Simulation Techniques for Reinforcement-Learning of Dialogue Management Strategies},
      journal = {The Knowledge Engineering Review},
       volume = 21,
        issue = 2,
         year = 2006,
        pages = {97-126},
}
@article{SchauletalCoRR-15,
        title = {Prioritized Experience Replay},
       author = {Tom Schaul and John Quan and Ioannis Antonoglou and David Silver},
      journal = {CoRR},
       volume = {arXiv:1511.05952},
         year = {2015},
     abstract = {Experience replay lets online reinforcement learning agents remember and reuse experiences from the past. In prior work, experience transitions were uniformly sampled from a replay memory. However, this approach simply replays transitions at the same frequency that they were originally experienced, regardless of their significance. In this paper we develop a framework for prioritizing experience, so as to replay important transitions more frequently, and therefore learn more efficiently. We use prioritized experience replay in Deep Q-Networks (DQN), a reinforcement learning algorithm that achieved human-level performance across many Atari games. DQN with prioritized experience replay achieves a new state-of-the-art, outperforming DQN with uniform replay on 41 out of 49 game},
}
@inproceedings{SchefferDAC-12,
       author = {Scheffer, Louis K.},
        title = {Design Tools for Artificial Nervous Systems},
    booktitle = {Proceedings of the 49th Annual Design Automation Conference},
     location = {San Francisco, California},
    publisher = {ACM},
      address = {New York, NY, USA},
         year = {2012},
        pages = {717-722},
     abstract = {Electronic and biological systems both perform complex information processing, but they use very different techniques. Though electronics has the advantage in raw speed, biological systems have the edge in many other areas. They can be produced, and indeed self-reproduce, without expensive and finicky factories. They are tolerant of manufacturing defects, and learn and adapt for better performance. In many cases they can self-repair damage. These advantages suggest that biological systems might be useful in a wide variety of tasks involving information processing. So far, all attempts to use the nervous system of a living organism for information processing have involved selective breeding of existing organisms. This approach, largely independent of the details of internal operation, is used since we do not yet understand how neural systems work, nor exactly how they are constructed. However, as our knowledge increases, the day will come when we can envision useful nervous systems and design them based upon what we want them to do, as opposed to variations on what has been already built. We will then need tools, corresponding to our Electronic Design Automation tools, to help with the design. This paper is concerned with what such tools might look like.}
}
@article{ScheinostTP-13,
    author = {Scheinost, D. and Stoica, T. and Saksa, J. and Papademetris, X. and Constable, R.T. and Pittenger, C. and Hampson, M.},
     title = {Orbitofrontal cortex neurofeedback produces lasting changes in contamination anxiety and resting-state connectivity},
   journal = {Translational Psychiatry},
    volume = 3,
      year = 2013,
     pages = {e250},
}
@article{ScheireretalPAMI-14,
       author = {Scheirer, W. and Anthony, S. and Nakayama, K. and Cox, D.},
        title = {Perceptual Annotation: Measuring Human Performance to Improve Machine Vision},
      journal = {Transactions in Pattern Analysis and Machine Learning},
       volume = 36,
        issue = 8,
         year = 2014,
        pages = {1679-1686},
     abstract = {For many problems in computer vision, human learners are considerably better than machines. Humans possess highly accurate internal recognition and learning mechanisms that are not yet understood, and they frequently have access to more extensive training data through a lifetime of unbiased experience with the visual world. We propose to use visual psychophysics to directly leverage the abilities of human subjects to build better machine learning systems. First, we use an advanced online psychometric testing platform to make new kinds of annotation data available for learning. Second, we develop a technique for harnessing these new kinds of information-"perceptual annotations"-for support vector machines. A key intuition for this approach is that while it may remain infeasible to dramatically increase the amount of data and high-quality labels available for the training of a given system, measuring the exemplar-by-exemplar difficulty and pattern of errors of human annotators can provide important information for regularizing the solution of the system at hand. A case study for the problem face detection demonstrates that this approach yields state-of-the-art results on the challenging FDDB data set.}
}
@book{Schelling1978,
    author = {Thomas C. Schelling},
 publisher = {{W. W. Norton \& Company}},
     title = {Micromotives and Macrobehavior},
      year = 1978,
}
@article{SchermellehetalJCB-10,
     title = {A guide to super-resolution fluorescence microscopy},
    author = {Schermelleh, L. and Heintzmann, R. and Leonhardt, H.},
   journal = {Journal of Cell Biology},
    volume = 190, 
     issue = 2,
      year = 2010,
     pages = {165-175},
  abstract = {For centuries, cell biology has been based on light microscopy and at the same time been limited by its optical resolution. However, several new technologies have been developed recently that bypass this limit. These new super-resolution technologies are either based on tailored illumination, nonlinear fluorophore responses, or the precise localization of single molecules. Overall, these new approaches have created unprecedented new possibilities to investigate the structure and function of cells.}
}
@book{Schillacietal2017sensorimotor,
        title = {Re-Enacting Sensorimotor Experience for Cognition},
       author = {Schillaci, G. and Hafner, V.V. and Lara, B.},
    publisher = {Frontiers Media SA},
       series = {Frontiers Research Topics},
         year = {2017},
}
@article{SchippetalTiN-13,
    author = {Shipp, Stewart and Adams, Rick A. and Friston, Karl J.},
     title = {Reflections on agranular architecture: predictive coding in the motor cortex},
   journal = {Trends in Neurosciences},
    volume = {36},
     issue = {12},
      year = {2013},
     pages = {706-716},
  abstract = {Predictive coding explains the recursive hierarchical structure of cortical processes. Granular layer 4, which relays ascending cortical pathways, is absent from motor cortex. Perceptual inference results if ascending sensory data modify sensory predictions action, if spinal reflexes enact descending motor and/or proprioceptive predictions. Motor layer 4 regresses as motor predictions inherently require less modification. The agranular architecture of motor cortex lacks a functional interpretation. Here, we consider a predictive coding account of this unique feature based on asymmetries in hierarchical cortical connections. In sensory cortex, layer 4 (the granular layer) is the target of ascending pathways. We theorise that the operation of predictive coding in the motor system (a process termed active inference) provides a principled rationale for the apparent recession of the ascending pathway in motor cortex. The extension of this theory to interlaminar circuitry also accounts for a sub-class of mirror neuron in motor cortex whose activity is suppressed when observing an action explaining how predictive coding can gate hierarchical processing to switch between perception and action}
}
@article{SchkadeandKahnemanPS-98,
     title = {Does living in California make people happy? A focusing illusion in judgments of life satisfaction},
    author = {Schkade, D. and Kahneman, D.},
   journal = {Psychological Science},
    volume = 9,
     issue = 5,
      year = 1998,
     pages = {340-346}
}
@article{SchlaepferetalNPP-07,
     title = {Deep brain stimulation to reward circuitry alleviates anhedonia in refractory major depression},
    author = {Schlaepfer, Thomas E. and Cohen, Michael X. and Frick Caroline and Kosel, Markus and Brodesser, Daniela and Axmacher, Nikolai and Joe, Alexius Young Y. and Kreft, Martina and Lenartz, Doris and Sturm, Volker},
   journal = {Neuropsychopharmacology},
    volume = 33,
     issue = 2,
     pages = {368-377},
 publisher = {Nature Publishing Group},
      year = 2007,
  abstract = {Deep brain stimulation (DBS) to different sites allows interfering with dysfunctional network function implicated in major depression. Because a prominent clinical feature of depression is anhedonia --- the inability to experience pleasure from previously pleasurable activities --- and because there is clear evidence of dysfunctions of the reward system in depression, DBS to the nucleus accumbens might offer a new possibility to target depressive symptomatology in otherwise treatment-resistant depression. Three patients suffering from extremely resistant forms of depression, who did not respond to pharmacotherapy, psychotherapy, and electroconvulsive therapy, were implanted with bilateral DBS electrodes in the nucleus accumbens. Stimulation parameters were modified in a double-blind manner, and clinical ratings were assessed at each modification. Additionally, brain metabolism was assessed 1 week before and 1 week after stimulation onset. Clinical ratings improved in all three patients when the stimulator was on  and worsened in all three patients when the stimulator was turned off. Effects were observable immediately, and no side effects occurred in any of the patients. Using FDG-PET, significant changes in brain metabolism as a function of the stimulation in fronto-striatal networks were observed. No unwanted effects of DBS other than those directly related to the surgical procedure (eg pain at sites of implantation) were observed. Dysfunctions of the reward system --- in which the nucleus accumbens is a key structure  are implicated in the neurobiology of major depression and might be responsible for impaired reward processing, as evidenced by the symptom of anhedonia. These preliminary findings suggest that DBS to the nucleus accumbens might be a hypothesis-guided approach for refractory major depression.},
}
@article{SchlagandSchmidhuberICLR-18,
        title = {Gated Fast Weights for Associative Retrieval},
       author = {Imanol Schlag and J\"{u}rgen Schmidhuber},
      journal = {Submitted to International Conference on Learning Representations},
         year = {2018},
     abstract = {We improve previous end-to-end differentiable neural networks (NNs) with fast weight memories. A gate mechanism updates fast weights at every time step of a sequence through two separate outer-product-based matrices generated by slow parts of the net. The system is trained on a complex sequence to sequence variation of the Associative Retrieval Problem with roughly 70 times more temporal memory (i.e. time-varying variables) than similar-sized standard recurrent NNs (RNNs). In terms of accuracy and number of parameters, our architecture outperforms a variety of RNNs, including Long Short-Term Memory, Hypernetworks, and related fast weight architectures.}
}
@article{SchlichtingandPrestonCOiBS,
       author = {Schlichting, M. L. and Preston, A. R.},
        title = {Memory integration: neural mechanisms and implications for behavior},
      journal = {Current Opinion in Behavioral Science},
         year = {2015},
       volume = {1},
        pages = {1-8},
     abstract = {Everyday behaviors require a high degree of flexibility, in which prior knowledge is applied to inform behavior in new situations. Such flexibility is thought to be supported in part by memory integration, a process whereby related memories become interconnected in the brain through recruitment of overlapping neuronal populations. Recent advances in cognitive and behavioral neuroscience highlight the importance of a hippocampal-medial prefrontal circuit in memory integration. Emerging evidence suggests that abstracted representations in medial prefrontal cortex guide reactivation of related memories during new encoding events, thus promoting hippocampal integration of related experiences. Moreover, recent work indicates that integrated memories are called upon during novel situations to facilitate a host of behaviors, from spatial navigation to imagination.}
}
@article{SchlichtingandPrestonMLM-16,
       author = {Schlichting, M. L. and Preston, A. R.},
        title = {Hippocampal-medial prefrontal circuit supports memory updating during learning and post-encoding rest},
      journal = {Neurobiology Learning Memory},
       volume = {134 Pt A},
         year = {2016},
        pages = {91-106},
     abstract = {Learning occurs in the context of existing memories. Encountering new information that relates to prior knowledge may trigger integration, whereby established memories are updated to incorporate new content. Here, we provide a critical test of recent theories suggesting hippocampal (HPC) and medial prefrontal (MPFC) involvement in integration, both during and immediately following encoding. Human participants with established memories for a set of initial (AB) associations underwent fMRI scanning during passive rest and encoding of new related (BC) and unrelated (XY) pairs. We show that HPC-MPFC functional coupling during learning was more predictive of trial-by-trial memory for associations related to prior knowledge relative to unrelated associations. Moreover, the degree to which HPC-MPFC functional coupling was enhanced following overlapping encoding was related to memory integration behavior across participants. We observed a dissociation between anterior and posterior MPFC, with integration signatures during post-encoding rest specifically in the posterior subregion. These results highlight the persistence of integration signatures into post-encoding periods, indicating continued processing of interrelated memories during rest. We also interrogated the coherence of white matter tracts to assess the hypothesis that integration behavior would be related to the integrity of the underlying anatomical pathways. Consistent with our predictions, more coherent HPC-MPFC white matter structure was associated with better performance across participants. This HPC-MPFC circuit also interacted with content-sensitive visual cortex during learning and rest, consistent with reinstatement of prior knowledge to enable updating. These results show that the HPC-MPFC circuit supports on- and offline integration of new content into memory.}
}
@book{Schmahmann1997cerebellum,
        title = {The cerebellum and cognition},
       author = {Schmahmann, Jeremy D.},
       volume = {41},
         year = {1997},
    publisher = {Academic Press}
}
@article{SchmahmannNEUROSCIENCE-19,
       author = {Jeremy D. Schmahmann},
        title = {The cerebellum and cognition},
      journal = {Neuroscience Letters},
       volume = {688},
         year = {2019},
        pages = {62-75},
     abstract = {What the cerebellum does to sensorimotor and vestibular control, it also does to cognition, emotion, and autonomic function. This hypothesis is based on the theories of dysmetria of thought and the universal cerebellar transform, which hold that the cerebellum maintains behavior around a homeostatic baseline, automatically, without conscious awareness, informed by implicit learning, and performed according to context. Functional topography within the cerebellum facilitates the modulation of distributed networks subserving multiple different functions. The sensorimotor cerebellum is represented in the anterior lobe with a second representation in lobule VIII, and lesions of these areas lead to the cerebellar motor syndrome of ataxia, dysmetria, dysarthria and impaired oculomotor control. The cognitive / limbic cerebellum is in the cerebellar posterior lobe, with current evidence pointing to three separate topographic representations, the nature of which remain to be determined. Posterior lobe lesions result in the cerebellar cognitive affective syndrome (CCAS), the hallmark features of which include deficits in executive function, visual spatial processing, linguistic skills and regulation of affect. The affective dyscontrol manifests in autism spectrum and psychosis spectrum disorders, and disorders of emotional control, attentional control, and social skill set. This report presents an overview of the rapidly growing field of the clinical cognitive neuroscience of the cerebellum. It commences with a brief historical background, then discusses tract tracing experiments in animal models and functional imaging observations in humans that subserve the cerebellar contribution to neurological function. Structure function correlation studies following focal cerebellar lesions in adults and children permit a finer appreciation of the functional topography and nature of the cerebellar motor syndrome, cerebellar vestibular syndrome, and the third cornerstone of clinical ataxiology - the cerebellar cognitive affective syndrome. The ability to detect the CCAS in real time in clinical neurology with a brief and validated scale should make it possible to develop a deeper understanding of the clinical consequences of cerebellar lesions in a wide range of neurological and neuropsychiatric disorders with a link to the cerebellum.}
}
@article{SchmahmannNEUROSCIENCE-LETTERS-19,
        title = {The cerebellum and cognition},
       author = {Jeremy D. Schmahmann},
       volume = {68},
        pages = {62-75},
      journal = {Neuroscience Letters},
         year = {2019},
     abstract = {What the cerebellum does to sensorimotor and vestibular control, it also does to cognition, emotion, and autonomic function. This hypothesis is based on the theories of dysmetria of thought and the universal cerebellar transform, which hold that the cerebellum maintains behavior around a homeostatic baseline, automatically, without conscious awareness, informed by implicit learning, and performed according to context. Functional topography within the cerebellum facilitates the modulation of distributed networks subserving multiple different functions. The sensorimotor cerebellum is represented in the anterior lobe with a second representation in lobule VIII, and lesions of these areas lead to the cerebellar motor syndrome of ataxia, dysmetria, dysarthria and impaired oculomotor control. The cognitive / limbic cerebellum is in the cerebellar posterior lobe, with current evidence pointing to three separate topographic representations, the nature of which remain to be determined.Posterior lobe lesions result in the cerebellar cognitive affective syndrome (CCAS), the hallmark features ofwhich include deficits in executive function, visual spatial processing, linguistic skills and regulation of affect. The affective dyscontrol manifests in autism spectrum and psychosis spectrum disorders, and disorders ofemotional control, attentional control, and social skill set. This report presents an overview of the rapidly growingfield of the clinical cognitive neuroscience of the cerebellum. It commences with a brief historical background, then discusses tract tracing experiments in animal models and functional imaging observations in humans that subserve the cerebellar contribution to neurological function. Structure function correlation studies following focal cerebellar lesions in adults and children permit afiner appreciation of the functional topography and nature of the cerebellar motor syndrome, cerebellar vestibular syndrome, and the third cornerstone of clinical ataxiology - the cerebellar cognitive affective syndrome. The ability to detect the CCAS in real time inclinical neurology with a brief and validated scale should make it possible to develop a deeper understanding ofthe clinical consequences of cerebellar lesions in a wide range of neurological and neuropsychiatric disorderswith a link to the cerebellum.}
}
@article{SchmahmannetalARN-19,
       author = {Schmahmann, Jeremy D. and Guell, Xavier and Stoodley, Catherine J. and Halko, Mark A.},
        title = {The Theory and Neuroscience of Cerebellar Cognition},
      journal = {Annual Review of Neuroscience},
       volume = {42},
       number = {1},
        pages = {337-364},
         year = {2019},
     abstract = {Cerebellar neuroscience has undergone a paradigm shift. The theories of the universal cerebellar transform and dysmetria of thought and the principles of organization of cerebral cortical connections, together with neuroanatomical, brain imaging, and clinical observations, have recontextualized the cerebellum as a critical node in the distributed neural circuits subserving behavior. The framework for cerebellar cognition stems from the identification of three cognitive representations in the posterior lobe, which are interconnected with cerebral association areas and distinct from the primary and secondary cerebellar sensorimotor representations linked with the spinal cord and cerebral motor areas. Lesions of the anterior lobe primary sensorimotor representations produce dysmetria of movement, the cerebellar motor syndrome. Lesions of the posterior lobe cognitive-emotional cerebellum produce dysmetria of thought and emotion, the cerebellar cognitive affective/Schmahmann syndrome. The notion that the cerebellum modulates thought and emotion in the same way that it modulates motor control advances the understanding of the mechanisms of cognition and opens new therapeutic opportunities in behavioral neurology and neuropsychiatry.}
}
@techreport{SchmidandWysotzkiTR-95,
       author = {Schmid, Ute and Wysotzki, Fritz},
        title = {Program synthesis by learning and planning},
        issue = {Technical Report, Computer Science Department},
  institution = {Universit\"{a}t Berlin},
         year = {1995},
     abstract = {In artificial intelligence automatic programming hasbeen a topic of research since the Sixties. Its main concern is to model programming expertise with the aim of automating or supporting the process of software engineering. The applicability of fully automatisized synthesis systems is very restricted. Nevertheless, research in this area can lead to valuable models about reasoning and problem solving in general. The idea, that enriching program synthesis systems with a memory component may improve performance is not anew one. But the notion of combining program synthesis with learning has been scarcely exploited. In our research project we are elaboratingthis idea. There are two different approaches to learning which were investigated: learning by problem solving and learning by analogy. Furthermore, the idea of learning by problem solving can be applied not only to the domain of automatic programming. We investigated the acquisition of generalized rules from problem solving by state-space search as well as the acquisition of correct behavior in dynamic environments by reinforcement learning. We argue that learning of this kind can be also seen as program synthesis because the acquired rules can be looked upon as specific programs.}
}
@article{SchmidhuberCoRR-00,
       author = {J{\"{u}}rgen Schmidhuber},
        title = {One Big Net For Everything},
      journal = {CoRR},
       volume = {arXiv:1802.08864},
         year = {2018},
          url = {http://arxiv.org/abs/1802.08864},
     abstract = {The probability distribution P from which the history of our universe is sampled represents a theory of everything or TOE. We assume P is formally describable. Since most (uncountably many) distributions are not, this imposes a strong inductive bias. We show that P(x) is small for any universe x lacking a short description, and study the spectrum of TOEs spanned by two Ps, one reflecting the most compact constructive descriptions, the other the fastest way of computing everything. The former derives from generalizations of traditional computability, Solomonoff's algorithmic probability, Kolmogorov complexity, and objects more random than Chaitin's Omega, the latter from Levin's universal search and a natural resource-oriented postulate: the cumulative prior probability of all x incomputable within time t by this optimal algorithm should be 1/t. Between both Ps we find a universal cumulatively enumerable measure that dominates traditional enumerable measures; any such CEM must assign low probability to any universe lacking a short enumerating program. We derive P-specific consequences for evolving observers, inductive reasoning, quantum physics, philosophy, and the expected duration of our universe.},
}
@article{SchmidhuberCoRR-15,
       author = {J{\"{u}}rgen Schmidhuber},
        title = {On Learning to Think: Algorithmic Information Theory for Novel Combinations of Reinforcement Learning Controllers and Recurrent Neural World Models},
      journal = {CoRR},
       volume = {arXiv:1511.09249},
         year = {2015},
     abstract = {This paper addresses the general problem of reinforcement learning (RL) in partially observable environments. In 2013, our large RL recurrent neural networks (RNNs) learned from scratch to drive simulated cars from high-dimensional video input. However, real brains are more powerful in many ways. In particular, they learn a predictive model of their initially unknown environment, and somehow use it for abstract (e.g., hierarchical) planning and reasoning. Guided by algorithmic information theory, we describe RNN-based AIs (RNNAIs) designed to do the same. Such an RNNAI can be trained on never-ending sequences of tasks, some of them provided by the user, others invented by the RNNAI itself in a curious, playful fashion, to improve its RNN-based world model. Unlike our previous model-building RNN-based RL machines dating back to 1990, the RNNAI learns to actively query its model for abstract reasoning and planning and decision making, essentially "learning to think." The basic ideas of this report can be applied to many other cases where one RNN-like system exploits the algorithmic information content of another. They are taken from a grant proposal submitted in Fall 2014, and also explain concepts such as "mirror neurons." Experimental results will be described in separate papers.}
}
@techreport{SchmidhuberDNN-14,
        title = {Deep Learning in Neural Networks: An Overview},
       author = {J\"{u}rgen Schmidhuber},
         year = 2014,
          pdf = {arXiv:1404.7828},
  institution = {Technical Report IDSIA-03-14},
     abstract = {In recent years, deep neural networks (including recurrent ones) have won numerous contests in pattern recognition and machine learning. This historical survey compactly summarises relevant work, much of it from the previous millennium. Shallow and deep learners are distinguished by the depth of their credit assignment paths, which are chains of possibly learnable, causal links between actions and effects. I review deep supervised learning (also recapitulating the history of backpropagation), un- supervised learning, reinforcement learning \& evolutionary computation, and indirect search for short programs encoding deep and large networks.}
}
@article{SchmidhuberNC-92,
       author = {Schmidhuber, J\"{u}rgen},
        title = {Learning Complex, Extended Sequences Using the Principle of History Compression},
      journal = {Neural Compututation},
       volume = 4,
        issue = 2,
         year = 1992,
        pages = {234-242},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
      comment = {Previous neural network learning algorithms for sequence processing are computationally expensive and perform poorly when it comes to long time lags. This paper first introduces a simple principle for reducing the descriptions of event sequences without loss of information. A consequence of this principle is that only unexpected inputs can be relevant. This insight leads to the construction of neural architectures that learn to `divide and conquer' by recursively decomposing sequences. I describe two architectures. The first functions as a self-organizing multi-level hierarchy of recurrent networks. The second, involving only two recurrent networks, tries to collapse a multi-level predictor hierarchy into a single recurrent net. Experiments show that the system can require less computation per time step and many fewer training sequences than conventional training algorithms for recurrent nets.}
}
@article{Schmidt-HieberetalNATURE-04,
       author = {Schmidt-Hieber, C. and Jonas, P. and Bischofberger, J.},
        title = {Enhanced synaptic plasticity in newly generated granule cells of the adult hippocampus},
      journal = {Nature},
       volume = {429},
       number = {6988},
         year = {2004},
        pages = {184-187},
     abstract = {Neural stem cells in various regions of the vertebrate brain continuously generate neurons throughout life. In the mammalian hippocampus, a region important for spatial and episodic memory, thousands of new granule cells are produced per day, with the exact number depending on environmental conditions and physical exercise. The survival of these neurons is improved by learning and conversely learning may be promoted by neurogenesis. Although it has been suggested that newly generated neurons may have specific properties to facilitate learning, the cellular and synaptic mechanisms of plasticity in these neurons are largely unknown. Here we show that young granule cells in the adult hippocampus differ substantially from mature granule cells in both active and passive membrane properties. In young neurons, T-type Ca2+ channels can generate isolated Ca2+ spikes and boost fast Na+ action potentials, contributing to the induction of synaptic plasticity. Associative long-term potentiation can be induced more easily in young neurons than in mature neurons under identical conditions. Thus, newly generated neurons express unique mechanisms to facilitate synaptic plasticity, which may be important for the formation of new memories.}
}
@article{Schmidt-Pruzanetal85,
   author = {J. Schmidt-Pruzan and E. Shamir and E. Upfal},
    title = {Random hypergraph coloring algorithms and the weak chromatic number},
  journal = {Journal of Graph Theory},
   volume = 8,
    pages = {347-362},
     year = 1985,
}
@article{SchmidtandLipsonSCIENCE-09, 
      title = {Distilling free-form natural laws from experimental data}, 
     author = {Schmidt, Michael and Lipson, Hod}, 
    journal = {Science}, 
     volume = 324, 
      issue = 5923, 
  publisher = {AAAS}, 
       year = 2009, 
      pages = {81-85},
}
@article{SchmidtetalBBB-12,
       author = {Brandy Schmidt and Diano F. Marrone and Etan J. Markus},
        title = {Disambiguating the similar: {T}he dentate gyrus and pattern separation},
      journal = {Behavioural Brain Research},
       volume = {226},
       number = {1},
         year = {2012},
        pages = {56-65},
     abstract = {The human hippocampus supports the formation of episodic memory without confusing new memories with old ones. To accomplish this, the brain must disambiguate memories (i.e., accentuate the differences between experiences). There is convergent evidence linking pattern separation to the dentate gyrus. Damage to the dentate gyrus reduces an organism's ability to differentiate between similar objects. The dentate gyrus has tenfold more principle cells than its cortical input, allowing for a divergence in information flow. Dentate gyrus granule neurons also show a very different pattern of representing the environment than "classic" place cells in CA1 and CA3, or grid cells in the entorhinal cortex. More recently immediate early genes have been used to "timestamp" activity of individual cells throughout the dentate gyrus. These data indicate that the dentate gyrus robustly differentiates similar situations. The degree of differentiation is non-linear, with even small changes in input inducing a near maximal response in the dentate. Furthermore this differentiation occurs throughout the dentate gyrus longitudinal (dorsal-ventral) axis. Conversely, the data point to a divergence in information processing between the dentate gyrus suprapyramidal and infrapyramidal blades possibly related to differences in organization within these regions. The accumulated evidence from different approaches converges to support a role for the dentate gyrus in pattern separation. There are however inconsistencies that may require incorporation of neurogenesis and hippocampal microcircuits into the currents models. They also suggest different roles for the dentate gyrus suprapyramidal and infrapyramidal blades, and the responsiveness of CA3 to dentate input.}
}
@inproceedings{SchmidtetalECML-07,
     author = {Mark Schmidt and Glenn Fung and R\'{o}mer Rosales},
      title = {Fast Optimization Methods for {L1} Regularization: A Comparative Study and Two New Approaches},
  booktitle = {Proceedings of the Eighteenth European Conference on Machine Learning},
       year = 2007,
      pages = {286-297},
}
@article{SchmuckandBazantCoRR-12,
       author = {Markus Schmuck and Martin Z. Bazant},
        title = {Homogenization of the {P}oisson-{N}ernst-{P}lanck Equations for Ion Transport in Charged Porous Media},
      journal = {CoRR},
       volume = {arXiv:1202.1916},
         year = {2012},
     abstract = {Effective Poisson-Nernst-Planck (PNP) equations are derived for macroscopic ion transport in charged porous media under periodic fluid flow by an asymptotic multi-scale expansion with drift. The microscopic setting is a two-component periodic composite consisting of a dilute electrolyte continuum (described by standard PNP equations) and a continuous dielectric matrix, which is impermeable to the ions and carries a given surface charge. Four new features arise in the upscaled equations: (i) the effective ionic diffusivities and mobilities become tensors, related to the microstructure; (ii) the effective permittivity is also a tensor, depending on the electrolyte/matrix permittivity ratio and the ratio of the Debye screening length to the macroscopic length of the porous medium; (iii) the microscopic fluidic convection is replaced by a diffusion-dispersion correction in the effective diffusion tensor; and (iv) the surface charge per volume appears as a continuous "background charge density", as in classical membrane models. The coefficient tensors in the upscaled PNP equations can be calculated from periodic reference cell problems. For an insulating solid matrix, all gradients are corrected by the same tensor, and the Einstein relation holds at the macroscopic scale, which is not generally the case for a polarizable matrix, unless the permittivity and electric field are suitably defined. In the limit of thin double layers, Poisson's equation is replaced by macroscopic electroneutrality (balancing ionic and surface charges). The general form of the macroscopic PNP equations may also hold for concentrated solution theories, based on the local-density and mean-field approximations. These results have broad applicability to ion transport in porous electrodes, separators, membranes, ion-exchange resins, soils, porous rocks, and biological tissues.}
}
@article {Schneider-MizelletalBIORXIV-15,
       author = {Schneider-Mizell, Casey M and Gerhard, Stephan and Longair, Mark and Kazimiers, Tom and Li, Feng and Zwart, Maarten F. and Champion, Andrew and Midgley, Frank and Fetter, Richard and Saalfeld, Stephan and Cardona, Albert},
        title = {Quantitative neuroanatomy for connectomics in Drosophila},
      journal = {bioRxiv},
         year = 2015,
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {Large-scale neuronal circuit mapping using electron microscopy demands laborious proofreading by humans who resolve local ambiguities with larger contextual cues or by reconciling multiple independent reconstructions. We developed a new method that empowers expert neuroanatomists to apply quantitative arbor and network context to proofread and reconstruct neurons and circuits. We implemented our method in the web application CATMAID, supporting a group of collaborators to concurrently reconstruct neurons in the same circuit. We measured the neuroanatomical underpinnings of circuit connectivity in Drosophila neurons. We found that across life stages and cell types, synaptic inputs were preferentially located on spine-like microtubule-free branches, "twigs", while synaptic outputs were typically on microtubule-containing "backbone". The differential size and tortuosity of small twigs and rigid backbones was reflected in reconstruction errors, with nearly all errors being omission or truncation of twigs. The combination of redundant twig connectivity and low backbone error rates allows robust mapping of Drosophila circuits without time-consuming independent reconstructions. As a demonstration, we mapped a large sensorimotor circuit in the larva. We found anatomical pathways for proprioceptive feedback into motor circuits and applied novel methods of representing neuroanatomical compartments to describe their detailed structure. Our work suggests avenues for incorporating neuroanatomy into machine-learning approaches to connectomics and reveals the largely unknown circuitry of larval locomotion.},
}
@article{SchneiderandShiffrinPR-77,
       author = {Schneider, Walter and Shiffrin, Richard M.},
        title = {Controlled and automatic human information processing: I. Detection, search, and attention.},
      journal = {Psychological Review},
       volume = {84},
       number = {1},
         year = {1977},
        pages = {1-66},
     abstract = {A 2-process theory of human information processing is proposed and applied to detection, search, and attention phenomena. Automatic processing is activation of a learned sequence of elements in long-term memory that is initiated by appropriate inputs and then proceeds automatically--without S control, without stressing the capacity limitations of the system, and without necessarily demanding attention. Controlled processing is a temporary activation of a sequence of elements that can be set up quickly and easily but requires attention, is capacity-limited (usually serial in nature), and is controlled by the S. A series of studies, with approximately 8 Ss, using both reaction time and accuracy measures is presented, which traces these concepts in the form of automatic detection and controlled search through the areas of detection, search, and attention. Results in these areas are shown to arise from common mechanisms. Automatic detection is shown to develop following consistent mapping of stimuli to responses over trials. Controlled search was utilized in varied-mapping paradigms, and in the present studies, it took the form of serial, terminating search.}
}
@incollection{SchneiderandVelmansBCC2-17,
       author = {Susan Schneider and Max Velmans},
        title = {Introduction},
    booktitle = {The Blackwell Companion to Consciousness, 2nd Edition},
       editor = {Susan Schneider and Max Velmans},
    publisher = {Wiley-Blackwell},
         year = {2017},
}
@article{SchneidmanetalNATURE-06,
       author = {Schneidman, Elad and Berry, Michael J. and Segev, Ronen and Bialek, William},
        title = {Weak pairwise correlations imply strongly correlated network states in a neural population},
      journal = {Nature},
       volume = 440,
        issue = 7087,
         year = 2006,
        pages = {1007-1012},
}
@inproceedings{SchneidmanetalNIPS-02,
       author = {Elad Schneidman and William Bialek and Michael J. Berry II},
        title = {An Information Theoretic Approach to the Functional Classification of Neurons},
    booktitle = {Advances in Neural Information Processing Systems 15},
       editor = {Suzanna Becker and Sebastian Thrun and Klaus Obermayer},
         year = 2002,
        pages = {197-204},
}
@book{ScholkopfandSmolaKERNELS-01,
       author = {Sch\"{o}lkopf, Bernhard and Smola, Alexander J.},
        title = {Learning with Kernels: Support Vector Machines, Regularization, Optimization, and Beyond},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
         year = 2001,
}
@inproceedings{SchoppersIJCAI-87,
       author = {Schoppers, Marcel J.},
        title = {Universal Plans for Reactive Robots in Unpredictable Environments},
    booktitle = {Proceedings {IJCAI} 10},
     location = {Milan, Italy},
 organization = {IJCAII},
         year = 1987,
        pages = {1039-1046},
}
@book{SchorrandRappaportIAAI-89,
       editor = {Schorr, Herbert and Rappaport, Alain},
        title = {Innovative Applications of Artificial Intelligence},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1989,
}
@book{Schrijver86,
       author = {Alexander Schrijver},
        title = {Theory of linear and integer programming},
    publisher = {Wiley-Interscience},
         year = 1986,
}
@article{SchrodeletalNATURE-METHODS-13,
       author = {Schr\"{o}del, Tina and Prevedel, Robert and Aumayr, Karin and Zimmer, Manuel and Vaziri, Alipasha},
        title = {Brain-wide {3D} imaging of neuronal activity in {\it{Caenorhabditis elegans}} with sculpted light},
      journal = {Nature Methods},
         year = 2013,
    publisher = {Nature Publishing Group},
       volume = 10,
        issue = 10,
        pages = {1013-1020},
     abstract = {Recent efforts in neuroscience research have been aimed at obtaining detailed anatomical neuronal wiring maps as well as information on how neurons in these networks engage in dynamic activities. Although the entire connectivity map of the nervous system of Caenorhabditis elegans has been known for more than 25 years, this knowledge has not been sufficient to predict all functional connections underlying behavior. To approach this goal, we developed a two-photon technique for brain-wide calcium imaging in C. elegans, using wide-field temporal focusing (WF-TeFo). Pivotal to our results was the use of a nuclear-localized, genetically encoded calcium indicator, NLS-GCaMP5K, that permits unambiguous discrimination of individual neurons within the densely packed head ganglia of C. elegans. We demonstrate near-simultaneous recording of activity of up to 70\% of all head neurons. In combination with a lab-on-a-chip device for stimulus delivery, this method provides an enabling platform for establishing functional maps of neuronal networks.},
}
@incollection{SchubertandHwangEPILOG-00,
        title = {Logic Meets Little Red Riding Hood: A Comprehensive, Natural Representation for Language Understanding},
       author = {Lenhart K. Schubert and Chung Hee Hwang},
    booktitle = {Natural Language Processing and Knowledge Representation: Language for Knowledge and Knowledge for Language},
      address = {Menlo Park, CA, and Cambridge, MA},
       editor = {Lucja Iwanska and Stuart C. Shapiro},
    publisher = {MIT/AAAI Press},
         year = 2000,
        pages = {111-174},
}
@inproceedings{SchubertetalCSK-10,
       author = {Lenhart K. Schubert and Benjamin Van Durme and Marzieh Bazrafshan},
        title = {Entailment Inference in a Natural Logic-like General Reasoner},
    booktitle = {Commonsense Knowledge, Papers from the 2010 {AAAI} Fall Symposium},
         year = 2010,
}
@article{SchuckandNivDOI-18,
        title = {Sequential replay of non-spatial task states in the human hippocampus. bioRxiv},
       author = {Schuck, NW and Niv, Y},
      journal = {DOI},
       volume = {10},
        pages = {315978},
         year = {2018}
}
@article{SchuhandDrurySAPMP-97,
       author = {Schuh, Lori and Drury, Ivo},
        title = {Intraoperative electrocorticography and direct cortical electrical stimulation},
      journal = {Seminars in Anesthesia, Perioperative Medicine and Pain},
       volume = 16,
        issue = 1,
         year = 1997,
        pages = {46-55},
     abstract = {First performed by Victor Horsley in 1886, the surgical management of epilepsy has increased substantially over the past 20 years. The goals of surgery are to render a patient seizure-free without additional neurological deficits. ECoG provides information on the location of abnormal spiking cortex in lesional and nonlesional epilepsy surgery and may aid in determining margins of resection in epilepsy surgery. DCES provides information on the location of eloquent areas of cortex that should not be removed during epilepsy surgery. Conscious sedation and general anesthetic regimens are available, which provide maximal comfort for a patient without significant alteration of ECoG or alteration in ability to cooperate with functional mapping procedures when needed.},
}
@inproceedings{SchuldtetalICPR-04,
     author = {Christian Schuldt and Ivan Laptev and Barbara Caputo},
      title = {Recognizing Human Actions: A Local {SVM} Approach},
  booktitle = {Proceedings of the International Conference on Pattern Recognition},
  publisher = {IEEE Computer Society},
       year = 2004,
}
@article{SchulleretalNATURE-10,
        title = {Plasmonics for extreme light concentration and manipulation},
       author = {Schuller, Jon A. and Barnard, Edward S. and Cai, Wenshan and Jun, Young Chul and  White, Justin S. and Brongersma, Mark L.},
      journal = {Nature Materials},
       volume = 9,
        issue = 3,
         year = 2010,
        pages = {193-204},
}
@article{SchusterandPaliwalTSP-97,
       author = {Schuster, Michael and Paliwal, Kuldip K.},
      journal = {Signal Processing, IEEE Transactions on},
        title = {Bidirectional recurrent neural networks},
         year = {1997},
       volume = {45},
        issue = {11},
        pages = {2673-2681},
     abstract = {In the first part of this paper, a regular recurrent neural network (RNN) is extended to a bidirectional recurrent neural network (BRNN). The BRNN can be trained without the limitation of using input information just up to a preset future frame. This is accomplished by training it simultaneously in positive and negative time direction. Structure and training procedure of the proposed network are explained. In regression and classification experiments on artificial data, the proposed structure gives better results than other approaches. For real data, classification experiments for phonemes from the TIMIT database show the same tendency. In the second part of this paper, it is shown how the proposed bidirectional structure can be easily modified to allow efficient estimation of the conditional posterior probability of complete symbol sequences without making any explicit assumption about the shape of the distribution. For this part, experiments on real data are reported},
}
@article{SchusteretalCURRENT-BIOLOGY-02,
       author = {Stefan Schuster and Roland Strauss and Karl G G\"{o}tz},
        title = {Virtual-Reality Techniques Resolve the Visual Cues Used by Fruit Flies to Evaluate Object Distances},
      journal = {Current Biology},
       volume = {12},
       number = {18},
         year = {2002},
        pages = {1591-1594},
     abstract = {Insects can estimate distance or time-to-contact of surrounding objects from locomotion-induced changes in their retinal position and/or size [1-8]. Freely walking fruit flies (Drosophila melanogaster) use the received mixture of different distance cues to select the nearest objects for subsequent visits [9, 10]. Conventional methods of behavioral analysis fail to elucidate the underlying data extraction. Here we demonstrate first comprehensive solutions of this problem by substituting virtual for real objects; a tracker-controlled 360° panorama converts a fruit fly's changing coordinates into object illusions that require the perception of specific cues to appear at preselected distances up to infinity. An application reveals the following: (1) en-route sampling of retinal-image changes accounts for distance discrimination within a surprising range of at least 8-80 body lengths (20-200 mm). Stereopsis and peering are not involved. (2) Distance from image translation in the expected direction (motion parallax) outweighs distance from image expansion, which accounts for impact-avoiding flight reactions to looming objects. (3) The ability to discriminate distances is robust to artificially delayed updating of image translation. Fruit flies appear to interrelate self-motion and its visual feedback within a surprisingly long time window of about 2 s. The comparative distance inspection practiced in the small fruit fly deserves utilization in self-moving robots.}
}
@article{SchwartzandSharir,
       author = {Schwartz, J. T. and Sharir, M.},
        title = {On the Pianos Movers' Problem: {I}},
      journal = {Communications on Pure and Applied Mathematics},
       volume = 36,
         year = 1983,
        pages = {345-398},
}
@book{Schwartzetal87,
       author = {Schwartz, J. T. and Sharir, M. and Hopcroft, J.},
        title = {Planning, Geometry, and Complexity of Robot Motion},
    publisher = {Ablex},
      address = {Norwood, New Jersey},
         year = 1987,
}
@article{SchwartzetalPLoS-13,
        title = {Personality, Gender, and Age in the Language of Social Media: The Open-Vocabulary Approach},
       author = {H. Andrew Schwartz and Johannes C. Eichstaedt and Margaret L. Kern and Lukasz Dziurzynski and Stephanie M. Ramones and Megha Agrawali and Achal Shah and Michal Kosinski and David Stillwell and Martin E. P. Seligman and Lyle H. Ungar},
      journal = {Public Library of Science},
       volume = 8,
        issue = 9,
         year = 2013,
        pages = {e73791},
}
@article{SchweighoferetalFIN-13,
       author = {Schweighofer, Nicolas and Lang, Eric J. and Kawato, Mitsuo},
        title = {Role of the Olivo-cerebellar Complex in Motor Learning and Control},
      journal = {Frontiers in Neural Circuits},
       volume = 7,
        issue = 94,
         year = 2013,
     abstract = {How is the cerebellum capable of efficient motor learning and control despite very low firing of the inferior olive inputs, which are postulated to carry errors needed for learning and contribute to on-line motor control? Inferior olive neurons form the largest electrically coupled network in the adult human brain. Here, we discuss how intermediate coupling strengths can lead to chaotic resonance and increase information transmission of the error signal despite the very low inferior olive firing rate. This increased information transmission can then lead to more efficient learning than with weak or strong coupling. In addition, we argue that a dynamic modulation of inferior olive electrical coupling via the Purkinje cell-deep cerebellar neurons --- inferior olive triangle could speed up learning and improve on-line control. Initially strong coupling would allow transmission of large errors to multiple functionally related Purkinje cells, resulting in fast but coarse learning as well as significant effects on deep cerebellar nucleus and on-line motor control. In the late phase of learning decreased coupling would allow de-synchronized IO firing, allowing high-fidelity transmission of error, resulting in slower but fine learning, and little on-line motor control effects.}
}
@article{SchweighoferetalJoN-11,
       author = {Schweighofer, N.  and Lee, J. Y.  and Goh, H. T. and Choi, Y.  and Kim, S. S. and Stewart, J. C. and Lewthwaite, R. and Winstein, C. J.},
        title = {Mechanisms of the contextual interference effect in individuals poststroke},
      journal = {Journal of Neurophysiollgy},
         year = {2011},
       volume = {106},
       number = {5},
        pages = {2632-2641},
     abstract = {Although intermixing different motor learning tasks via random schedules enhances long-term retention compared with "blocked" schedules, the mechanism underlying this contextual interference effect has been unclear. Furthermore, previous studies have reported inconclusive results in individuals poststroke. We instructed participants to learn to produce three grip force patterns in either random or blocked schedules and measured the contextual interference effect by long-term forgetting: the change in performance between immediate and 24-h posttests. Nondisabled participants exhibited the contextual interference effect: no forgetting in the random condition but forgetting in the blocked condition. Participants at least 3 mo poststroke exhibited no forgetting in the random condition but marginal forgetting in the blocked condition. However, in participants poststroke, the integrity of visuospatial working memory modulated long-term retention after blocked schedule training: participants with poor visuospatial working memory exhibited little forgetting at 24 h. These counterintuitive results were predicted by a computational model of motor memory that contains a common fast process and multiple slow processes, which are competitively updated by motor errors. In blocked schedules, the fast process quickly improved performance, therefore reducing error-driven update of the slow processes and thus poor long-term retention. In random schedules, interferences in the fast process led to slower change in performance, therefore increasing error-driven update of slow processes and thus good long-term retention. Increased forgetting rates in the fast process, as would be expected in individuals with visuospatial working memory deficits, led to small updates of the fast process during blocked schedules and thus better long-term retention.},
}
@techreport{SchweinhartPhD-15,
       author = {Benjamin Schweinhart},
        title = {Statistical Topology of Embedded Graphs},
  institution = {Princeton University},
         year = 2015,
}
@incollection{Schweitzer84,
       author = {Schweitzer, Paul J.},
        title = {Aggregation Methods for Large {M}arkov Chains},
       editor = {Iazola, G. and Coutois, P. J. and Hordijk, A.},
    booktitle = {Mathemaical Computer Performance and Reliability},
    publisher = {Elsevier},
      address = {Amsterdam, Holland},
         year = 1984,
        pages = {275-302},
}
@article{SchweitzeretalOR-85,
       author = {Schweitzer, Paul J. and Puterman, Martin L. and Kindle, Kyle W.},
        title = {Iterative Aggregation-Disaggregation Procedures for Discounted Semi-{M}arkov Reward Processes},
      journal = {Operations Research},
       volume = 33,
        issue = 3,
         year = 1985,
        pages = {589-605},
}
@article{SchwemmeretalJoN-15,
       author = {Schwemmer, Michael A. and Fairhall, Adrienne L. and Den\'{e}ve, Sophie and Shea-Brown, Eric T.},
        title = {Constructing Precisely Computing Networks with Biophysical Spiking Neurons},
      journal = {The Journal of Neuroscience},
       volume = {35},
       number = {28},
         year = {2015},
        pages = {10112-10134}, 
    abstract = {While spike timing has been shown to carry detailed stimulus information at the sensory periphery, its possible role in network computation is less clear. Most models of computation by neural networks are based on population firing rates. In equivalent spiking implementations, firing is assumed to be random such that averaging across populations of neurons recovers the rate-based approach. Recently, however, Den\'{e}ve and colleagues have suggested that the spiking behavior of neurons may be fundamental to how neuronal networks compute, with precise spike timing determined by each neuron's contribution to producing the desired output (Boerlin and Deneve, 2011; Boerlin et al., 2013). By postulating that each neuron fires to reduce the error in the network's output, it was demonstrated that linear computations can be performed by networks of integrate-and-fire neurons that communicate through instantaneous synapses. This left open, however, the possibility that realistic networks, with conductance-based neurons with subthreshold nonlinearity and the slower timescales of biophysical synapses, may not fit into this framework. Here, we show how the spike-based approach can be extended to biophysically plausible networks. We then show that our network reproduces a number of key features of cortical networks including irregular and Poisson-like spike times and a tight balance between excitation and inhibition. Lastly, we discuss how the behavior of our model scales with network size or with the number of neurons "recorded" from a larger computing network. These results significantly increase the biological plausibility of the spike-based approach to network computation. SIGNIFICANCE STATEMENT We derive a network of neurons with standard spike-generating currents and synapses with realistic timescales that computes based upon the principle that the precise timing of each spike is important for the computation. We then show that our network reproduces a number of key features of cortical networks including irregular, Poisson-like spike times, and a tight balance between excitation and inhibition. These results significantly increase the biological plausibility of the spike-based approach to network computation, and uncover how several components of biological networks may work together to efficiently carry out computation.},
}
@book{Searle1984mindsandbrains,
        title = {Minds, Brains, and Science},
       author = {Searle, John},
    publisher = {Harvard University Press},
      address = {Cambridge, MA},
         year = {1984},
}
@book{Searle83,
       author = {Searle, J. R.},
        title = {Intentionality: An Essay in the Philosophy of Mind},
    publisher = {Cambridge University Press},
         year = {1983},
}
@article{SearleBBS-80,
       author = {Searle, John},
        title = {Minds, Brains, and Programs},
      journal = {Behavioral and Brain Sciences},
       volume = {3},
         year = {1980},
        pages = {417-424},
}
@article{SebanzandKnoblichAPS-21,
       author = {Natalie Sebanz and G{\"{u}}nther Knoblich},
        title = {Progress in Joint-Action Research},
      journal = {Current Directions in Psychological Science},
       volume = {6},
       number = {1},
        pages = {0963721420984425},
         year = {2021},
     abstract = {Humans have a striking ability to coordinate their actions with each other to achieve joint goals. The tight interpersonal coordination that characterizes joint actions is achieved through processes that help with preparing for joint action as well as processes that are active while joint actions are being performed. To prepare for joint action, partners form representations of each other's actions and tasks and the relation between them. This enables them to predict each other's upcoming actions, which, in turn, facilitates coordination. While performing joint actions, partners' coordination is maintained by (a) monitoring whether individual and joint outcomes correspond to what was planned, (b) predicting partners' action parameters on the basis of familiarity with their individual actions, (c) communicating task-relevant information unknown to partners in an action-based fashion, and (d) relying on coupling of predictions through dense perceptual-information flow between coactors. The next challenge for the field of joint action is to generate an integrated perspective that links coordination mechanisms to normative, evolutionary, and communicative frameworks.}
}
@inproceedings{SeeetalACL-17,
       author = {Abigail See and Peter J. Liu and Christopher D. Manning},
        title = {Get To The Point: Summarization with Pointer-Generator Networks},
    booktitle = {Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics},
       volume = {arXiv:1704.04368},
         year = {2017},
          url = {https://github.com/abisee/pointer-generator},
     abstract = {Neural sequence-to-sequence models have provided a viable new approach for abstractive text summarization (meaning they are not restricted to simply selecting and rearranging passages from the original text). However, these models have two shortcomings: they are liable to reproduce factual details inaccurately, and they tend to repeat themselves. In this work we propose a novel architecture that augments the standard sequence-to-sequence attentional model in two orthogonal ways. First, we use a hybrid pointer-generator network that can copy words from the source text via pointing, which aids accurate reproduction of information, while retaining the ability to produce novel words through the generator. Second, we use coverage to keep track of what has been summarized, which discourages repetition. We apply our model to the CNN / Daily Mail summarization task, outperforming the current abstractive state-of-the-art by at least 2 ROUGE points.}
}
@article{SeeligandJayaramanNATURE-15,
       author = {Seelig, Johannes D. and Jayaraman, Vivek},
        title = {Neural dynamics for landmark orientation and angular path integration},
      journal = {Nature},
         year = {2015},
    publisher = {Nature Publishing Group},
       volume = {521},
        issue = {7551},
        pages = {186-191},
     abstract = {Many animals navigate using a combination of visual landmarks and path integration. In mammalian brains, head direction cells integrate these two streams of information by representing an animal's heading relative to landmarks, yet maintaining their directional tuning in darkness based on self-motion cues. Here we use two-photon calcium imaging in head-fixed Drosophila melanogaster walking on a ball in a virtual reality arena to demonstrate that landmark-based orientation and angular path integration are combined in the population responses of neurons whose dendrites tile the ellipsoid body, a toroidal structure in the centre of the fly brain. The neural population encodes the fly's azimuth relative to its environment, tracking visual landmarks when available and relying on self-motion cues in darkness. When both visual and self-motion cues are absent, a representation of the animal's orientation is maintained in this network through persistent activity, a potential substrate for short-term memory. Several features of the population dynamics of these neurons and their circular anatomical arrangement are suggestive of ring attractors, network structures that have been proposed to support the function of navigational brain circuits.},
}
@article{SeeligetalNATURE-METHODS-10,
       author = {Seelig, Johannes D. and Chiappe, M. Eugenia and Lott, Gus K. and Dutta, Anirban and Osborne, Jason E. and Reiser, Michael B. and Jayaraman, Vivek},
        title = {Two-photon calcium imaging from motion-sensitive neurons in head-fixed Drosophila during optomotor walking behavior},
      journal = {Nature methods},
    publisher = {Nature Publishing Group},
       volume = {7},
        issue = {7},
         year = {2010},
        pages = {535-540},
     abstract = {Drosophila melanogster is a model organism rich in genetic tools to manipulate and identify neural circuits involved in specific behaviors. Here we present a novel technique for two-photon calcium imaging in the central brain of head-fixed Drosophila walking on an air-supported ball. The ball's motion is tracked at high resolution and can be treated as a proxy for the fly's own movements. We used the genetically encoded calcium sensor, GCaMP3.0, to record from important elements of the motion-processing pathway, the horizontal-system (HS) lobula plate tangential cells (LPTCs) in the fly optic lobe. We presented motion stimuli to the tethered fly and found that calcium transients in HS-neurons correlated with robust optomotor behavior during walking. Our technique allows an entirely new set of questions to be addressed by monitoring behavior and physiology in identified neurons in a powerful genetic model organism with an extensive repertoire of walking behaviors.}
}
@article{SefahetalCell-SELEX_PROTOCOL-10,
       author = {Sefah, Kwame and Shangguan, Dihua and Xiong, Xiangling and O'Donoghue, Meghan B. and Tan, Weihong},
        title = {{Development of DNA aptamers using Cell-SELEX}},
      journal = {Nature Protocols},
    publisher = {Nature Publishing Group},
       volume = 5,
        issue = 6,
         year = 2010,
        pages = {1169-1185},
     abstract = {In the past two decades, high-affinity nucleic acid aptamers have been developed for a wide variety of pure molecules and complex systems such as live cells. Conceptually, aptamers are developed by an evolutionary process, whereby, as selection progresses, sequences with a certain conformation capable of binding to the target of interest emerge and dominate the pool. This protocol, cell-SELEX (systematic evolution of ligands by exponential enrichment), is a method that can generate DNA aptamers that can bind specifically to a cell type of interest. Commonly, a cancer cell line is used as the target to generate aptamers that can differentiate that cell type from other cancers or normal cells. A single-stranded DNA (ssDNA) library pool is incubated with the target cells. Nonbinding sequences are washed off and bound sequences are recovered from the cells by heating cell-DNA complexes at 95 \hmdegree{}, followed by centrifugation. The recovered pool is incubated with the control cell line to filter out the sequences that bind to common molecules on both the target and the control, leading to the enrichment of specific binders to the target. Binding sequences are amplified by PCR using fluorescein isothiocyanate-labeled sense and biotin-labeled antisense primers. This is followed by removal of antisense strands to generate an ssDNA pool for subsequent rounds of selection. The enrichment of the selected pools is monitored by flow cytometry binding assays, with selected pools having increased fluorescence compared with the unselected DNA library. The procedure, from design of oligonucleotides to enrichment of the selected pools, takes \approx{} 3 months.}
}
@book{Segaller98,
      author = {Stephen Segaller},
       title = {Nerds 2.0.1: A Brief History of the Internet},
   publisher = {TV Books},
     address = {New York, NY},
        year = 1998,
}
@article{SegerandMillerARN-10,
       author = {Seger, Carol A. and Miller, Earl K.},
        title = {Category Learning in the Brain},
      journal = {Annual Review of Neuroscience},
       volume = 33,
        issue = 1,
         year = 2010,
        pages = {203-219},
     abstract = {The ability to group items and events into functional categories is a fundamental characteristic of sophisticated thought. It is subserved by plasticity in many neural systems, including neocortical regions (sensory, prefrontal, parietal, and motor cortex), the medial temporal lobe, the basal ganglia, and midbrain dopaminergic systems. These systems interact during category learning. Corticostriatal loops may mediate recursive, bootstrapping interactions between fast reward-gated plasticity in the basal ganglia and slow reward-shaded plasticity in the cortex. This can provide a balance between acquisition of details of experiences and generalization across them. Interactions between the corticostriatal loops can integrate perceptual, response, and feedback-related aspects of the task and mediate the shift from novice to skilled performance. The basal ganglia and medial temporal lobe interact competitively or cooperatively, depending on the demands of the learning task.}
}
@article{SegevJM-06,
       author = {Segev, I. and Rall, W.},
        title = {What do dendrites and their synapses tell the neuron?},
      journal = {Journal of Neurophysiology},
       volume = {95},
       number = {3},
         year = {2006},
        pages = {1295-1297},
     abstract = {This essay looks at the historical significance of four APS classic papers that are freely available online: Rall W. Distinguishing theoretical synaptic potentials computed for different soma-dendritic distributions of synaptic input. J Neurophysiol 30: 1138-1168, 1967 (http://jn.physiology.org/cgi/reprint/30/5/1138). Rall W, Burke RE, Smith TG, Nelson PG, and Frank K. Dendritic location of synapses and possible mechanisms for the monosynaptic EPSP in motoneurons. J Neurophysiol 30: 1169-1193, 1967 (http://jn.physiology.org/cgi/reprint/30/5/1169). Rall W and Shepherd GM. Theoretical reconstruction of field potentials and dendrodendritic synaptic interactions in olfactory bulb. J Neurophysiol 31: 884-915, 1968 (http://jn.physiology.org/cgi/reprint/31/6/884). Segev I and Rall W. Computational study of an excitable dendritic spine. J Neurophysiol 60: 499-523, 1988 (http://jn.physiology.org/cgi/reprint/60/2/499).}
}
@article{SeileretalSIGGRAPH-08,
     author = {Seiler, Larry and Carmean, Doug and Sprangle, Eric and Forsyth, Tom and 
              Abrash, Michael and Dubey, Pradeep and Junkins, Stephen and Lake, Adam and 
              Sugerman, Jeremy and Cavin, Robert and Espasa, Roger and Grochowski, Ed and 
              Juan, Toni and Hanrahan, Pat},
      title = {Larrabee: a many-core x86 architecture for visual computing},
    journal = {ACM Transactions on Graphics},
  publisher = {ACM},
    address = {New York, NY, USA},
     volume = 27,
      issue = 3,
       year = 2008,
      pages = {1-15},
}
@inproceedings{SeimetalWHC-15,
       author = {Caitlyn E. Seim and Tanya Estes and Thad Starner},
        title = {Towards Passive Haptic Learning of piano songs},
    booktitle = {2015 {IEEE} World Haptics Conference, {WHC} 2015},
     location = {Evanston, IL, USA},
        pages = {445-450},
         year = {2015},
     abstract = {Passive Haptic Learning (PHL) enables users to acquire motor skills by receiving tactile stimulation while no perceived attention is given to learning. Initial work used gloves with embedded vibration motors to passively teach users how to play simple, one-handed, one-note-at-a-time piano melodies. In an effort to create a practical system for learning full piano pieces, we have developed a method of passively teaching two-handed chorded skills, initially focusing on Braille typing. Here, we extend this effort to piano and show that passive stimulation is more effective at teaching piano pieces when presented on both hands simultaneously as opposed to training the left hand and then the right, as is common in many active teaching methods. We also demonstrate that accompanying audio is not needed for passive learning of piano melodies, which allows mobile PHL gloves to be used in more everyday situations.}
}
@article{SeirikietalNEURON-17,
       author = {Seiriki, Kaoru and Kasai, Atsushi and Hashimoto, Takeshi and Schulze, Wiebke and Niu, Misaki and Yamaguchi, Shun and Nakazawa, Takanobu and Inoue, Ken-ichi and Uezono, Shiori and Takada, Masahiko and Naka, Yuichiro and Igarashi, Hisato and Tanuma, Masato and Waschek, James A. and Ago, Yukio and Tanaka, Kenji F. and Hayata-Takano, Atsuko and Nagayasu, Kazuki and Shintani, Norihito and Hashimoto, Ryota and Kunii, Yasuto and Hino, Mizuki and Matsumoto, Junya and Yabe, Hirooki and Nagai, Takeharu and Fujita, Katsumasa and Matsuda, Toshio and Takuma, Kazuhiro and Baba, Akemichi and Hashimoto, Hitoshi},
        title = {High-Speed and Scalable Whole-Brain Imaging in Rodents and Primates},
      journal = {Neuron},
    publisher = {Elsevier},
       volume = {94},
        issue = {6},
         year = {2017},
        pages = {1085-1100},
     abstract = {Subcellular resolution imaging of the whole brain and subsequent image analysis are prerequisites for understanding anatomical and functional brain networks. Here, we have developed a very high-speed serial-sectioning imaging system named FAST (block-face serial microscopy tomography), which acquires high-resolution images of a whole mouse brain in a speed range comparable to that of light-sheet fluorescence microscopy. FAST enables complete visualization of the brain at a resolution sufficient to resolve all cells and their subcellular structures. FAST renders unbiased quantitative group comparisons of normal and disease model brain cells for the whole brain at a high spatial resolution. Furthermore, FAST is highly scalable to non-human primate brains and human postmortem brain tissues, and can visualize neuronal projections in a whole adult marmoset brain. Thus, FAST provides new opportunities for global approaches that will allow for a better understanding of brain systems in multiple animal models and in human diseases.},
}
@article{SejnowskiandRosenbergCOMPLEX-87,
   author =  {Sejnowski, T. J. and Rosenberg, C. R.},
   title =   {Parallel Networks that Learn to Pronounce English Text},
   journal = {Complex Systems},
   volume =  1,
   page =    {145-168},
   year =    1987,
}
@mastersthesis{SeldinMSc2001,
    author = {Yevgeny Seldin},
     title = {On Unsupervised Learning of Mixtures of {M}arkovian Sources},
    school = {The Hebrew University of Jerusalem},
      year = 2001,
}
@inproceedings{SeldinetalICML-01,
    author = {Yevgeny Seldin and Gill Bejerano and Naftali Tishby},
     title = {Unsupervised Sequence Segmentation by a Mixture of Switching Variable Memory {M}arkov Sources},
 booktitle = {Proceedings of the Eighteenth International Conference on Machine Learning (ICML 2001)},
     pages = {513-520},
      year = 2001,
}
@inproceedings{SeldinetalICSS-01,
    author = {Yevgeny Seldin and Gill Bejerano and Naftali Tishby},
     title = {Unsupervised Segmentation and Classification of Mixtures of {M}arkovian Sources},
 booktitle = {Proceedings of the 33rd Symposium on the Interface of Computing Science and Statistics},
      year = 2001,
}
@inproceedings{SelfridgeSMTP-59,
     author = {Oliver G. Selfridge},
      title = {Pandemonium: A paradigm for learning},
     editor = {D. V. Blake and A. M. Uttley},
  booktitle = {Proceedings of the Symposium on Mechanisation of Thought Processes}, 
      pages = {511-529}, 
   location = {London}, 
       year = 1959
}
@book{Seligman2011,
        title = {Authentic Happiness},
       author = {Seligman, M.},
         year = 2011,
    publisher = {Random House Australia}
}
@inproceedings{SelleeandOlshausenNIPS-04,
    author = {Phil Sallee and Bruno A. Olshausen},
     title = {Learning Sparse Multiscale Image Representations},
 booktitle = {Advances in Neural Information Processing Systems},
    volume = 16,
     pages = {1327-1334},
      year = 2004,
}
@misc{SellikSURVEY-19,
        title = {Natural Language Processing Techniques for Code Generation},
       author = {Hendrig Sellik},
         year = 2019,
 howpublished = {Delft University of Technology Technical Report},
          url = {https://hsellik.github.io/},
     abstract = {Software development is difficult and requires knowledge on many different levels such as understanding programming algorithms, languages, and frameworks. In addition, before code is being worked on, the system requirements and functionality are first discussed in natural language, after which it is sometimes visualized for the developers in a more formal language such as Unified Modeling Language. Recently, researchers have tried to close the gap between natural language description of the system and the actual implementation in code using natural language processing techniques. The techniques from NLP have also proven to be useful at generating code snippets while developers work on source code. This literature survey aims to present an overview of the field of code generation using Natural Language Processing techniques.},
}
@inproceedings{SelmanetalAAAI-92,
    author = {Selman, Bart and Levesque, Hector and Mitchell, David},
     title = {A New Method for Solving Hard Satisfiability Problems},
 booktitle = {Proceedings AAAI-92},
  location = {San Jose, California},
organization = {AAAI},
     pages = {440-446},
      year = 1992,
}
@inproceedings{SelmanetalAAAI-96,
       author = {Bart Selman and Rodney A. Brooks and Thomas Dean and Eric Horvitz and Tom M. Mitchell and Nils J. Nilsson},
        title = {Challenge Problems for Artificial Intelligence (Panel Statements)},
    publisher = {MIT Press},
      address = {Cambridge, Massachusetts},
    booktitle = {Proceedings of {AAAI}-96},
 organization = {AAAI},
        pages = {1340-1345},
         year = 1996,
}
@book{SeltzerMEG-08,
     title = {Developments in Alkali-metal Atomic Magnetometry},
    author = {Seltzer, S.J. and Princeton University},
      year = 2008,
 publisher = {Princeton University}
}
@article{SemendeferietalAJPA-01,
     title = {Prefrontal Cortex in Humans and Apes: A comparative study of area 10},
    author = {Semendeferi, K. and Armstrong, E. and Schleicher, A. and Zilles, K. and Hoesen, G. W. Van},
   journal = {American Journal of Physical Anthropology},
    volume = 114,
     issue = 3,
     pages = {224-241},
      year = 2001,
}
@article{SeminowiczetalNI-04,
     title = {Limbic-frontal circuitry in major depression: a path modeling metanalysis},
    author = {D.A. Seminowicz and H.S. Mayberg and A.R. McIntosh and K. Goldapple and S. Kennedy and Z. Segal and S. Rafi-Tari},
   journal = {NeuroImage},
    volume = 22,
     issue = 1,
      year = 2004,
     pages = {409-418},
}
@article{SeoandMilanfarPAMI-10,
    author = {Hae Jong Seo and Peyman Milanfar},
     title = {Training-Free, Generic Object Detection Using Locally Adaptive Regression Kernels},
   journal = {IEEE Transactions Pattern Analysis and Machine Intelligence},
 publisher = {IEEE Computer Society},
    volume = 32,
     issue = 9,
      year = 2010,
     pages = {1688-1704},
}
@inproceedings{SeoetalACM-17,
       author = {Seo, Sungyong and Huang, Jing and Yang, Hao and Liu, Yan},
        title = {Interpretable Convolutional Neural Networks with Dual Local and Global Attention for Review Rating Prediction},
    booktitle = {Proceedings of the Eleventh ACM Conference on Recommender Systems},
    publisher = {Association for Computing Machinery},
      address = {New York, NY, USA},
         year = {2017},
        pages = {297-305},
     abstract = {Recently, many e-commerce websites have encouraged their users to rate shopping items and write review texts. This review information has been very useful for understanding user preferences and item properties, as well as enhancing the capability to make personalized recommendations of these websites. In this paper, we propose to model user preferences and item properties using convolutional neural networks (CNNs) with dual local and global attention, motivated by the superiority of CNNs to extract complex features. By using aggregated review texts from a user and aggregated review text for an item, our model can learn the unique features (embedding) of each user and each item. These features are then used to predict ratings. We train these user and item networks jointly which enable the interaction between users and items in a similar way as matrix factorization. The local attention provides us insight on a user's preferences or an item's properties. The global attention helps CNNs focus on the semantic meaning of the whole review text. Thus, the combined local and global attentions enable an interpretable and better-learned representation of users and items. We validate the proposed models by testing on popular review datasets in Yelp and Amazon and compare the results with matrix factorization (MF), the hidden factor and topical (HFT) model, and the recently proposed convolutional matrix factorization (ConvMF+). Our proposed CNNs with dual attention model outperforms HFT and ConvMF+ in terms of mean square errors (MSE). In addition, we compare the user/item embeddings learned from these models for classification and recommendation. These results also confirm the superior quality of user/item embeddings learned from our model.},
}
@article{SeoetalARXIV-13,
        title = {Neural Dust: An Ultrasonic, Low Power Solution for Chronic Brain-Machine Interfaces},
       author = {Dongjin Seo and Jose M. Carmena and Jan M. Rabaey and Elad Alon and and Michel M. Maharbiz},
         year = 2013,
      journal = {CoRR},
       volume = {arXiv:1307.2196},
          url = {http://arxiv.org/abs/1307.2196},
}
@article{SeoetalNEURON-16,
       author = {Seo, Dongjin and Neely, Ryan M. and Shen, Konlin and Singhal, Utkarsh and Alon, Elad and Rabaey, Jan M. and Carmena, Jose M. and Maharbiz, Michel M.},
        title = {Wireless Recording in the Peripheral Nervous System with Ultrasonic Neural Dust},
      journal = {Neuron},
         year = {2016},
    publisher = {Elsevier},
       volume = {91},
        issue = {3},
        pages = {529-539},
     abstract = {The emerging field of bioelectronic medicine seeks methods for deciphering and modulating electrophysiological activity in the body to attain therapeutic effects at target organs. Current approaches to interfacing with peripheral nerves and muscles rely heavily on wires, creating problems for chronic use, while emerging wireless approaches lack the size scalability necessary to interrogate small-diameter nerves. Furthermore, conventional electrode-based technologies lack the capability to record from nerves with high spatial resolution or to record independently from many discrete sites within a nerve bundle. Here, we demonstrate neural dust, a wireless and scalable ultrasonic backscatter system for powering and communicating with implanted bioelectronics. We show that ultrasound is effective at delivering power to mm-scale devices in tissue; likewise, passive, battery-less communication using backscatter enables high-fidelity transmission of electromyogram (EMG) and electroneurogram (ENG) signals from anesthetized rats. These results highlight the potential for an ultrasound-based neural interface system for advancing future bioelectronics-based therapies.},
}
@article{SeoetalNEUROSCIENCE-METHODS-15,
       author = {Dongjin Seo and Jose M. Carmena and Jan M. Rabaey and Michel M. Maharbiz and Elad Alon},
        title = {Model validation of untethered, ultrasonic neural dust motes for cortical recording},
      journal = {Journal of Neuroscience Methods},
       volume = 244,
         year = 2015,
        pages = {114-122},
     abstract = {A major hurdle in brain-machine interfaces (BMI) is the lack of an implantable neural interface system that remains viable for a substantial fraction of the user's lifetime. Recently, sub-mm implantable, wireless electromagnetic (EM) neural interfaces have been demonstrated in an effort to extend system longevity. However, \{EM\} systems do not scale down in size well due to the severe inefficiency of coupling radio-waves at those scales within tissue. This paper explores fundamental system design trade-offs as well as size, power, and bandwidth scaling limits of neural recording systems built from low-power electronics coupled with ultrasonic power delivery and backscatter communication. Such systems will require two fundamental technology innovations: (1) 10-100 \hmmu{}m scale, free-floating, independent sensor nodes, or neural dust, that detect and report local extracellular electrophysiological data via ultrasonic backscattering and (2) a sub-cranial ultrasonic interrogator that establishes power and communication links with the neural dust. We provide experimental verification that the predicted scaling effects follow theory; (127 \hmmu{}m)3 neural dust motes immersed in water 3 cm from the interrogator couple with 0.002064\% power transfer efficiency and 0.04246 ppm backscatter, resulting in a maximum received power of \~0.5 \hmmu{}W with \~1 nW of change in backscatter power with neural activity. The high efficiency of ultrasonic transmission can enable the scaling of the sensing nodes down to 10s of micrometer. We conclude with a brief discussion of the application of neural dust for both central and peripheral nervous system recordings, and perspectives on future research directions.}
}
@article{SepulcreetalPLoS-10,
      title = {The Organization of Local and Distant Functional Connectivity in the Human Brain},
     author = {Sepulcre, J. and Liu, H. and Talukdar, T. and Martincorena, I. and Yeo, B.T.T.},
    journal = {{PLoS} Computational Biology},
  publisher = {Public Library of Science},
     volume = 6, 
      issue = 6,
       year = 2010,
}
@inproceedings{SerbanetalAAAI-16,
       author = {Serban, Iulian V. and Sordoni, Alessandro and Bengio, Yoshua and Courville, Aaron and Pineau, Joelle},
        title = {Building End-to-end Dialogue Systems Using Generative Hierarchical Neural Network Models},
    booktitle = {Proceedings of the Thirtieth AAAI Conference on Artificial Intelligence},
    publisher = {AAAI Press},
     location = {Phoenix, Arizona},
         year = {2016},
        pages = {3776-3783},
     abstract = {We investigate the task of building open domain, conversational dialogue systems based on large dialogue corpora using generative models. Generative models produce system responses that are autonomously generated word-by-word, opening up the possibility for realistic, flexible interactions. In support of this goal, we extend the recently proposed hierarchical recurrent encoder-decoder neural network to the dialogue domain, and demonstrate that this model is competitive with state-of-the-art neural language models and backoff n-gram models. We investigate the limitations of this and similar approaches, and show how its performance can be improved by bootstrapping the learning from a larger question-answer pair corpus and from pretrained word embeddings.}
}
@article{SerbanetalCoRR-15,
       author = {Iulian Vlad Serban and Alessandro Sordoni and Yoshua Bengio and Aaron C. Courville and Joelle Pineau},
        title = {Hierarchical Neural Network Generative Models for Movie Dialogues},
      journal = {CoRR},
       volume = {arXiv:1507.04808},
         year = {2015},
     abstract = {We investigate the task of building open domain, conversational dialogue systems based on large dialogue corpora using generative models. Generative models produce system responses that are autonomously generated word-by-word, opening up the possibility for realistic, flexible interactions. In support of this goal, we extend the recently proposed hierarchical recurrent encoder-decoder neural network to the dialogue domain, and demonstrate that this model is competitive with state-of-the-art neural language models and back-off n-gram models. We investigate the limitations of this and similar approaches, and show how its performance can be improved by bootstrapping the learning from a larger question-answer pair corpus and from pretrained word embeddings.}
}
@inproceedings{SerenoCSI-91,
        title = {{Language and the Primate Brain}},
       author = {Martin I. Sereno},
    booktitle = {{Proceedings Thirteenth Annual Conference of the Cognitive Science Society}},
    publisher = {Lawrence Erlbaum Associates},
      address = {Hillsdale, NJ},
         year = {1991},
        pages = {79-84},
     abstract = {New data on the large number of modality-specific areas in the post-central cortex of several non-human primates, and recent anatomical and functional studies of the human brain suggest that very little of the cortex consists of poly-modal 'association' areas. These observations are used to reinterpret psychological and neuropsychological data on language comprehension in normal and brain-damaged humans. I argue that language comprehension in sighted people might best be thought of as a kind of code-directed scene comprehension that draws heavily upon specifically visual, and probably largely prelinguistic processing constraints. The key processes of word-recognition and the assembly of visual word meaning patterns into interacting chains, however, may be mediated in part by species-specific activity patterns in secondary auditory cortex similar to those generated by uninterpreted speech-sound sequences.}
}
@article{SerenoandLehkyFCR-10,
      title = {Population coding of visual space: comparison of spatial representations in the dorsal and ventral pathways},   
     author = {Anne B. Sereno and Sidney R. Lehky},
    journal = {Frontiers in Computational Neuroscience},    
     volume = 4, 
      issue = 0,
      pages = 159,
       year = 2010,
}
@incollection{SeressHDCN-01,
        title = {Seress, L.},
       author = {Morphological changes of the human hippocampal formation from midgestation to early childhood},
    booktitle = {Handbook of developmental cognitive neuroscience},
       editor = {Nelson, C.A. and Luciana, M.},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = {2001},
        pages = {45-58},
}
@article{Serrano-GotarredonaetalFiN-13,
       author = {Serrano-Gotarredona, T. and Masquelier, T. and Prodromakis, T. and Indiveri, G. and Linares-Barranco, B.},
        title = {{STDP} and {STDP} variations with memristors for spiking neuromorphic learning systems},
      journal = {Frontiers in Neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = 7,
        pages = 2,
         year = 2013,
     abstract = {In this paper we review several ways of realizing asynchronous Spike-Timing-Dependent-Plasticity (STDP) using memristors as synapses. Our focus is on how to use individual memristors to implement synaptic weight multiplications, in a way such that it is not necessary to (a) introduce global synchronization and (b) to separate memristor learning phases from memristor performing phases. In the approaches described, neurons fire spikes asynchronously when they wish and memristive synapses perform computation and learn at their own pace, as it happens in biological neural systems. We distinguish between two different memristor physics, depending on whether they respond to the original "moving wall" or to the "filament creation and annihilation" models. Independent of the memristor physics, we discuss two different types of STDP rules that can be implemented with memristors: either the pure timing-based rule that takes into account the arrival time of the spikes from the pre- and the post-synaptic neurons, or a hybrid rule that takes into account only the timing of pre-synaptic spikes and the membrane potential and other state variables of the post-synaptic neuron. We show how to implement these rules in cross-bar architectures that comprise massive arrays of memristors, and we discuss applications for artificial vision.},
}
@techreport{SerreetalCBCL-05,
 author = {T. Serre and M. Kouh and C. Cadieu and U. Knoblich and
           G. Kreiman and T. Poggio},
  title = {A Theory of Object Recognition: Computations and Circuits
           in the Feedforward Path of the Ventral Stream in Primate
           Visual Cortex},
institution = {{MIT} Center for Biological \& Computational Learning},
  issue = {{AI} Memo 2005-036 ({CBCL} Memo 259)},
  month = {December},
   year = 2005,
}
@article{SerreetalPAMI-07,
    author = {T. Serre and L. Wolf and S. Bileschi and M. Riesenhuber and T. Poggio}, 
     title = {Object Recognition with Cortex-like Mechanisms},
   journal = {{IEEE} Transactions on Pattern Analysis and Machine Intelligence},
    volume = 29,
     issue = 3,
     pages = {411-426},
      year = 2007,
}
@article{SethetalCEREBRAL-CORTEX-04,
        title = {Visual binding through reentrant connectivity and dynamic synchronization in a brain-based device.},
       author = {Anil K. Seth and Jeffrey L. McKinstry and Gerald M. Edelman and Jeffrey L. Krichmar},
      journal = {Cerebral cortex},
       volume = {14 11},
         year = {2004},
        pages = {1185-99}
}
@book{Seung2012,
     title = {Connectome: How the Brain's Wiring Makes Us Who We Are},
    author = {Sebastian Seung},
 publisher = {Houghton Mifflin Harcourt},
   address = {Boston},
      year = 2012,
}
@article{SeungNATURE-11,
       author = {Seung, H. Sebastian},
        title = {Neuroscience: Towards functional connectomics},
      journal = {Nature},
         year = {2011},
    publisher = {Nature Publishing Group},
       volume = {471},
        issue = {7337},
        pages = {170-172},
     abstract = {To understand the brain, the thousands of synaptic connections made by each of billions of neurons should be mapped and related to neuronal function. First steps towards this formidable goal are now reported.}
}
@article{SeungNEURON-03,
        title = {Learning in Spiking Neural Networks by Reinforcement of Stochastic Synaptic Transmission},
      journal = {Neuron},
       volume = 40,
        issue = 6,
        pages = {1063-1073},
         year = 2003,
       author = {H. Sebastian Seung},
     abstract = {It is well-known that chemical synaptic transmission is an unreliable process, but the function of such unreliability remains unclear. Here I consider the hypothesis that the randomness of synaptic transmission is harnessed by the brain for learning, in analogy to the way that genetic mutation is utilized by Darwinian evolution. This is possible if synapses are "hedonistic," responding to a global reward signal by increasing their probabilities of vesicle release or failure, depending on which action immediately preceded reward. Hedonistic synapses learn by computing a stochastic approximation to the gradient of the average reward. They are compatible with synaptic dynamics such as short-term facilitation and depression and with the intricacies of dendritic integration and action potential generation. A network of hedonistic synapses can be trained to perform a desired computation by administering reward appropriately, as illustrated here through numerical simulations of integrate-and-fire model neurons.}
}
@article{SeungNEURON-09,
     title = {Reading the Book of Memory: Sparse Sampling versus Dense Mapping of Connectomes},
    author = {H. Sebastian Seung},
   journal = {Neuron},
    volume = 62,
     issue = 1,
      year = 2009,
     pages = {17-29},
}
@article{SeungandZungCoRR-17,
        title = {A correlation game for unsupervised learning yields computational interpretations of Hebbian excitation, anti-Hebbian inhibition, and synapse elimination},
       author = {Sebastian Seung and Jonathan Zung},
      journal = {CoRR},
       volume = {arXiv:1704.00646},
         year = 2017,
     abstract = {Much has been learned about plasticity of biological synapses from empirical studies. Hebbian plasticity is driven by correlated activity of presynaptic and postsynaptic neurons. Synapses that converge onto the same neuron often behave as if they compete for a fixed resource; some survive the competition while others are eliminated. To provide computational interpretations of these aspects of synaptic plasticity, we formulate unsupervised learning as a zero-sum game between Hebbian excitation and anti-Hebbian inhibition in a neural network model. The game formalizes the intuition that Hebbian excitation tries to maximize correlations of neurons with their inputs, while anti-Hebbian inhibition tries to decorrelate neurons from each other. We further include a model of synaptic competition, which enables a neuron to eliminate all connections except those from its most strongly correlated inputs. Through empirical studies, we show that this facilitates the learning of sensory features that resemble parts of objects.}
}
@inproceedings{SeyedhosseinietalMICCAI-11,
       author = {Seyedhosseini, Mojtaba and Kumar, Ritwik and Jurrus, Elizabeth and Giuly, Rick and Ellisman, Mark and Pfister, Hanspeter and Tasdizen, Tolga},
        title = {Detection of Neuron Membranes in Electron Microscopy Images Using Multi-scale Context and Radon-like Features},
    booktitle = {International Conference on Medical Image Computing and Computer-assisted Intervention},
    publisher = {Springer-Verlag},
         year = 2011,
     location = {Toronto, Canada},
        pages = {670-677},
     abstract = {Automated neural circuit reconstruction through electron microscopy (EM) images is a challenging problem. In this paper, we present a novel method that exploits multi-scale contextual information together with Radon-like features (RLF) to learn a series of discriminative models. The main idea is to build a framework which is capable of extracting information about cell membranes from a large contextual area of an EM image in a computationally efficient way. Toward this goal, we extract RLF that can be computed efficiently from the input image and generate a scale-space representation of the context images that are obtained at the output of each discriminative model in the series. Compared to a single-scale model, the use of a multi-scale representation of the context image gives the subsequent classifiers access to a larger contextual area in an effective way. Our strategy is general and independent of the classifier and has the potential to be used in any context based framework. We demonstrate that our method outperforms the state-of-the-art algorithms in detection of neuron membranes in EM images.}
}
@article{ShababoetalNIPS-12,
       author = {Ben Shababo and Kui Tang and Frank Wood},
        title = {Inferring Direct and Indirect Functional Connectivity Between Neurons From Multiple Neural Spike Train Data},
      journal = {Unknown Journal Attribution},
         year = {2012},
     abstract = {Our project aims to model the functional connectivity of neural microcircuits. On this scale, we are concerned with how the activity of each individual neuron relates to other nearby neurons in the population. Though several models and methods have been implemented to infer neural microcircuit connectivity, these fail to capture unobserved influences on the microcircuit. In this paper, we address these hidden influences on the microcircuit by developing a model which takes into account functional connectivity between observed neurons over more than one time step. We then test this model by simulating a large population of neurons but only observing a subpopulation which allows us to compare our inferred indirect con-nectivity with the known direct connectivity of the total population. With a better understanding of the functional patterns of neural activity at the cellular level, we can begin to decode the building blocks of neural computation.}
}
@inproceedings{ShachterAAAI-90,
       author = {Shachter, Ross D. and D'Ambrosio, Bruce and Del Favero, Brendan A.},
        title = {Symbolic probabilistic inference in belief networks},
        pages = {126-131},
    booktitle = {Proceedings AAAI-90},
     location = {Boston, Massachusetts},
 organization = {AAAI},
         year = 1990,
}
@article{ShachterOR-86,
       author = {Ross D. Shachter},
        title = {Evaluating Influence Diagrams},
      journal = {Operations Research},
       volume = 34,
        issue = 6,
         year = 1986,
        pages = {871-882},
}
@article{ShachterOR-88,
       author = {Shachter, Ross D.},
        title = {Probabilistic Inference and Influence Diagrams},
      journal = {Operations Research},
       volume = 36,
         year = 1988,
        pages = {589-604},
}
@inproceedings{ShachterPeotUAI-92,
       author = {Shachter, Ross and Mark Peot},
        title = {Decision Making Using Probabilistic Inference Models},
    booktitle = {Eighth Conference on Uncertainty in Artificial Intelligence},
     location = {Los Angeles, California},
         year = 1992,
}
@inproceedings{ShachterUAI-89,
       author = {Ross D. Shachter and Mark A. Peot},
        title = {Simulation approaches to general probabilistic inference on belief networks},
    booktitle = UAI-89,
         year = 1989,
        pages = {311-318},
}
@article{ShadidetalPC-13,
       author = {John Shadid and Scott Hutchinson and Gary Hennigan and Harry Moffat and Karen Devine and A. G. Salinger},
        title = {Efficient Parallel Computation of Unstructured Finite Element Reacting Flow Solutions},
      journal = {Parallel Computing },
       volume = {23},
         year = {1997},
        pages = {1307-1325},
     abstract = {A parallel unstructured finite element (FE) reacting flow solver designed for message passing MIMD computers is described. This implementation employs automated partitioning algorithms for load balancing unstructured grids, a distributed sparse matrix representation of the global FE equations, and parallel Krylov subspace iterative solvers. In this paper, a number of issues related to the efficient implementation of parallel unstructured mesh applications are presented. These issues include the differences between structured and unstructured mesh parallel applications, major communication kernels for unstructured Krylov iterative solvers, automatic mesh partitioning algorithms, and the influence of mesh partitioning metrics and single-node CPU performance on parallel performance. Results are presented for example FE heat transfer, fluid flow and full reacting flow applications on a 1024 processor nCUBE 2 hypercube and a 1904 processor Intel Paragon. Results indicate that very high computational rates and high scaled efficiencies can be achieved for large problems despite the use of sparse matrix data structures and the required unstructured data communication.}
}
@article{ShaetalNATURE-00,
       author = {Mi, Sha and Lee, Xinhua and Li, Xiang-ping and Veldman, Geertruida M. and Finnerty, Heather and Racie, Lisa and LaVallie, Edward and Tang, Xiang-Yang and Edouard, Philippe and Howes, Steve and Keith Jr, James C. and McCoy, John M.},
        title = {Syncytin is a captive retroviral envelope protein involved in human placental morphogenesis},
      journal = {Nature},
    publisher = {Macmillan Magazines Ltd},
       volume = {403},
         year = {2000},
        pages = {785-789},
     abstract = {Many mammalian viruses have acquired genes from their hosts during their evolution. The rationale for these acquisitions is usually quite clear: the captured genes are subverted to provide a selective advantage to the virus. Here we describe the opposite situation, where a viral gene has been sequestered to serve an important function in the physiology of a mammalian host. This gene, encoding a protein that we have called syncytin, is the envelope gene of a recently identified human endogenous defective retrovirus, HERV-W2. We find that the major sites of syncytin expression are placental syncytiotrophoblasts, multinucleated cells that originate from fetal trophoblasts. We show that expression of recombinant syncytin in a wide variety of cell types induces the formation of giant syncytia, and that fusion of a human trophoblastic cell line expressing endogenous syncytin can be inhibited by an anti-syncytin antiserum. Our data indicate that syncytin may mediate placental cytotrophoblast fusion in vivo, and thus may be important in human placental morphogenesis.}
}
@book{Shafer75,
   author =    {Shafer, Glenn},
   title =     {Mathematical Theory of Evidence},
   publisher = {Princeton University Press},
   year =      1975,
}
@incollection{ShaferUAI-88,
   author =   {Shafer, Glenn},
   title =    {Probability Judgement in Artificial Intelligence},
   crossref = UAI-88,
}
@book{ShaferandPearl90,
   editor =    {Shafer, Glenn and Pearl, Judea},
   title =     {Readings in Uncertain Reasoning},
   publisher = {Morgan Kaufmann},
   address =   {San Francisco, California},
   year =      1990,
}
@book{ShafferandKipp2010development,
        title = {Developmental Psychology: Childhood \& Adolescence: Childhood and Adolescence},
       author = {Shaffer, D.D.R. and Kipp, K.},
         year = {2010},
    publisher = {Wadsworth Cengage Learning}
}
@article{ShahandWakaiPMB-13,
       author = {Shah, V. K. and Wakai, R. T.},
        title = {A compact, high performance atomic magnetometer for biomedical applications},
      journal = {Physics in Medicine Biology},
         year = 2013,
       volume = 58,
        issue = 22,
        pages = {8153-8161},
}
@article{ShaietalPLoS-15,
       author = {Shai, A. S. and Anastassiou, C. A. and Larkum, M. E. and Koch, C.},
        title = {Physiology of {L}ayer 5 Pyramidal Neurons in Mouse Primary Visual Cortex: Coincidence Detection through Bursting},
      journal = {{PLoS} Computational Biology},
         year = 2015,
       volume = 11,
        issue = 3,
     abstract = {L5 pyramidal neurons are the only neocortical cell type with dendrites reaching all six layers of cortex, casting them as one of the main integrators in the cortical column. What is the na- ture and mode of computation performed in mouse primary visual cortex (V1) given the physiology of L5 pyramidal neurons? First, we experimentally establish active properties of the dendrites of L5 pyramidal neurons of mouse V1 using patch-clamp recordings. Using a detailed multi-compartmental model, we show this physiological setup to be well suited for coincidence detection between basal and apical tuft inputs by controlling the frequency of spike output. We further show how direct inhibition of calcium channels in the dendrites modulates such coincidence detection. To establish the singe-cell computation that this bio- physics supports, we show that the combination of frequency-modulation of somatic output by tuft input and (simulated) calcium-channel blockage functionally acts as a composite sigmoidal function. Finally, we explore how this computation provides a mechanism whereby dendritic spiking contributes to orientation tuning in pyramidal neurons.},
}
@phdthesis{ShakhnarovichPhD,
   author = {Greg Shakhnarovich},
    title = {Learning Task-Specific Similarity},
  comment = {Advisor: Trevor Darrell},
   school = {MIT},
     year = 2006,
}
@inproceedings{Shalev-ShwartzandTewariICML-09,
    author = {Shai Shalev-Shwartz and Ambuj Tewari},
     title = {Stochastic Methods for {$L_1$} Regularized Loss Minimization},
 booktitle = {Proceedings of the 26th International Conference on Machine Learning},
     pages = {929-936},
      year = 2009,
    editor = {L\'{e}on Bottou and Michael Littman},
   address = {Montreal},
     month = {June},
}
@inproceedings{Shalev-ShwartzetalICML-07,
    author = {Shalev-Shwartz, Shai and Singer, Yoram and Srebro, Nathan},
     title = {Pegasos: Primal Estimated sub-GrAdient SOlver for SVM},
 booktitle = {Proceedings of the 24th International Conference on Machine Learning},
      year = 2007,
     pages = {807-814},
  location = {Corvalis, Oregon},
 publisher = {ACM},
   address = {New York, NY, USA},
}
@techreport{ShalizietalSFI-02,
       author = {Cosma Rohilla Shalizi and Kristina Lisa Shalizi and James P. Crutchfield},
        title = {Pattern Discovery in Time Series, Part I: Theory, Algorithm, Analysis, and Convergence},
  institution = {Santa Fe Institute},
        issue = {SFI WORKING PAPER: 2002-10-060},
         year = 2015,
}
@article{ShamirandSompolinskyNC-04,
       author = {Shamir, M. and Sompolinsky, H.},
        title = {Nonlinear population codes},
      journal = {Neural Computation},
       volume = {16},
       number = {6},
         year = {2004},
        pages = {1105-1136},
     abstract = {Theoretical and experimental studies of distributed neuronal representations of sensory and behavioral variables usually assume that the tuning of the mean firing rates is the main source of information. However, recent theoretical studies have investigated the effect of cross-correlations in the trial-to-trial fluctuations of the neuronal responses on the accuracy of the representation. Assuming that only the first-order statistics of the neuronal responses are tuned to the stimulus, these studies have shown that in the presence of correlations, similar to those observed experimentally in cortical ensembles of neurons, the amount of information in the population is limited, yielding nonzero error levels even in the limit of infinitely large populations of neurons. In this letter, we study correlated neuronal populations whose higher-order statistics, and in particular response variances, are also modulated by the stimulus. We ask two questions: Does the correlated noise limit the accuracy of the neuronal representation of the stimulus? and, How can a biological mechanism extract most of the information embedded in the higher-order statistics of the neuronal responses? Specifically, we address these questions in the context of a population of neurons coding an angular variable. We show that the information embedded in the variances grows linearly with the population size despite the presence of strong correlated noise. This information cannot be extracted by linear readout schemes, including the linear population vector. Instead, we propose a bilinear readout scheme that involves spatial decorrelation, quadratic nonlinearity, and population vector summation. We show that this nonlinear population vector scheme yields accurate estimates of stimulus parameters, with an efficiency that grows linearly with the population size. This code can be implemented using biologically plausible neurons.}
}
@article{ShamirandUpfal81,
       author = {E. Shamir and E. Upfal},
        title = {One factor in random graphs},
      journal = {Israel Journal of Math},
       volume = 39,
        pages = {296-302},
         year = 1981,
}
@article{ShamirandUpfal82,
       author = {E. Shamir and E. Upfal},
        title = {One-factor in random graphs based on vertex choice},
      journal = {Discrete Math},
       volume = 41,
        pages = {281-286},
         year = 1982,
}
@article{ShamirandUpfal84,
       author = {E. Shamir and E. Upfal},
        title = {Large regular factors in random graphs},
      journal = {Annals of Discrete Math},
       volume = 20,
        pages = {271-282},
         year = 1984,
}
@book{Shampine75,
       author = {Shampine, L. F. and Gordon, M. K.},
        title = {Computer Solution of Ordinary Differential Equations},
    publisher = {W. H. Freeman and Company},
     location = {San Francisco, California},
         year = 1975,
}
@article{ShanahanPTRS_B-12,
       author = {Shanahan, Murray},
        title = {The brain's connective core and its role in animal cognition},
      journal = {Philosophical Transactions of the Royal Society B: Biological Sciences},
       volume = {367},
       number = {1603},
         year = {2012},
        pages = {2704-2714},
     abstract = {This paper addresses the question of how the brain of an animal achieves cognitive integration - that is to say how it manages to bring its fullest resources to bear on an ongoing situation. To fully exploit its cognitive resources, whether inherited or acquired through experience, it must be possible for unanticipated coalitions of brain processes to form. This facilitates the novel recombination of the elements of an existing behavioural repertoire, and thereby enables innovation. But in a system comprising massively many anatomically distributed assemblies of neurons, it is far from clear how such open-ended coalition formation is possible. The present paper draws on contemporary findings in brain connectivity and neurodynamics, as well as the literature of artificial intelligence, to outline a possible answer in terms of the brain's most richly connected and topologically central structures, its so-called connective core.}
}
@article{ShanetalELIFE-17,
       author = {Xu, C. Shan and Hayworth, Kenneth J. and Lu, Zhiyuan and Grob, Patricia and Hassan, Ahmed M. and Garc\'{i}a-Cerd\'{a}n, Jos\'{e} G. and Niyogi, Krishna K. and Nogales, Eva and Weinberg, Richard J and Hess, Harald F},
       editor = {Nathans, Jeremy},
        title = {Enhanced {FIB-SEM} systems for large-volume {3D} imaging},
      journal = {eLife},
    publisher = {eLife Sciences Publications, Ltd},
       volume = {6},
         year = {2017},
        pages = {e25916},
     abstract = {Focused Ion Beam Scanning Electron Microscopy (FIB-SEM) can automatically generate 3D images with superior z-axis resolution, yielding data that needs minimal image registration and related post-processing. Obstacles blocking wider adoption of FIB-SEM include slow imaging speed and lack of long-term system stability, which caps the maximum possible acquisition volume. Here we present techniques that accelerate image acquisition while greatly improving FIB-SEM reliability, allowing the system to operate for months and generating continuously imaged volumes > 106 µm3. These volumes are large enough for connectomics, where the excellent z resolution can help in tracing of small neuronal processes and accelerate the tedious and time-consuming human proofreading effort. Even higher resolution can be achieved on smaller volumes. We present example data sets from mammalian neural tissue, Drosophila brain, and Chlamydomonas reinhardtii to illustrate the power of this novel high-resolution technique to address questions in both connectomics and cell biology.},
}
@article{ShangCoRR-19,
       author = {Junyuan Shang and Tengfei Ma and Cao Xiao and Jimeng Sun},
        title = {Pre-training of Graph Augmented Transformers for Medication Recommendation},
      journal = {CoRR},
       volume = {arXiv:1906.00346},
         year = {2019},
     abstract = {Medication recommendation is an important healthcare application. It is commonly formulated as a temporal prediction task. Hence, most existing works only utilize longitudinal electronic health records (EHRs) from a small number of patients with multiple visits ignoring a large number of patients with a single visit (selection bias). Moreover, important hierarchical knowledge such as diagnosis hierarchy is not leveraged in the representation learning process. To address these challenges, we propose G-BERT, a new model to combine the power of Graph Neural Networks (GNNs) and BERT (Bidirectional Encoder Representations from Transformers) for medical code representation and medication recommendation. We use GNNs to represent the internal hierarchical structures of medical codes. Then we integrate the GNN representation into a transformer-based visit encoder and pre-train it on EHR data from patients only with a single visit. The pre-trained visit encoder and representation are then fine-tuned for downstream predictive tasks on longitudinal EHRs from patients with multiple visits. G-BERT is the first to bring the language model pre-training schema into the healthcare domain and it achieved state-of-the-art performance on the medication recommendation task.}
}
@inproceedings{ShangetalKDD-18,
       author = {Jingbo Shang and Q. Zhu and J. Shen and X. Wang and X. Gu and L. Kaplan and T. Harratty and J. Han},
        title = {AutoNet: Automated Network Construction and Exploration System from Domain-Specific Corpora},
    booktitle = {International Conference on Knowledge Discovery \& Data Mining},
         year = {2018},
        pages = {144-152},
    publisher = {Society for Industrial and Applied Mathematics},
     abstract = {As a collaborative project funded by US Army Research Lab, our goal is to turn massive unstructured text data into structured heterogeneous information networks (HINs), on which actionable knowledge can be further uncovered flexibly and effectively based on user's instructions. Taking advantage of open knowledge bases, we develop an end-to-end, data-driven system, AutoNet, with no additional human curation and annotation. AutoNet constructs a large-scale HIN from massive (user-provided) domain-specific text corpora (e.g., scientific papers) using our innovative phrase mining, entity typing, and relation extraction methods, and saves these models for later usage. After that, AutoNet supports two real-time functions: (1) discovery: given a few user-provided documents, AutoNet will construct a new HIN on the fly and highlight those new nodes (i.e., entities) and/or edges (i.e., relations), which are not in the pre-stored network; and (2) exploration: given some user-provided keywords, AutoNet will retrieve a related subnetwork from the large pre-stored HIN. We further design effective visualization tools for both functions.},
}
@article{Shannon48,
       author = {Claude Shannon},
        title = {A mathematical theory of communication},
      journal = {Bell System Technical Journal},
       volume = 27,
        issue = {July and October},
        pages = {379-423 and 623-656},
         year = 1948,
}
@article{Shannon49,
       author = {Shannon, C. E.},
        title = {Communication in the Presence of Noise},
      journal = {Proceedings of the IRE},
       volume = 37,
        pages = {10-21},
         year = 1949,
}
@book{ShannonandWeaver49,
       author = {C. E. Shannon and W. W. Weaver},
        title = {The Mathematical Theory of Communication},
    publisher = {University of Illinois Press},
      address = {Urbana, IL},
         year = 1949,
}
@article{ShaoetalNATURE-12,
        title = {Protein typing of circulating microvesicles allows real-time monitoring of glioblastoma therapy},
       author = {Shao, Huilin and Chung, Jaehoon and Balaj, Leonora and Charest, Alain and Bigner, Darell D. and Carter, Bob S. and Hochberg, Fred H. and Breakefield, Xandra O. and Weissleder, Ralph and Lee, Hakho},
      journal = {Nature Medicine},
    publisher = {Nature Publishing Group},
       volume = 18,
         year = 2012,
        pages = {1835-1840},
}
@article{ShaoetalTBE-14,
       author = {H. C. Shao and C. C. Wu and G. Y. Chen and H. M. Chang and A. S. Chiang and Y. C. Chen},
        title = {Developing a Stereotypical Drosophila Brain Atlas},
      journal = {{IEEE Transactions on Biomedical Engineering}},
       volume = {61},
       number = {12},
         year = {2014},
        pages = {2848-2858},
     abstract = {Brain research requires a standardized brain atlas to describe both the variance and invariance in brain anatomy and neuron connectivity. In this study, we propose a system to construct a standardized 3D Drosophila brain atlas by integrating labeled images from different preparations. The 3D fly brain atlas consists of standardized anatomical global and local reference models, e.g., the inner and external brain surfaces and the mushroom body. The averaged global and local reference models are generated by the model averaging procedure, and then the standard Drosophila brain atlas can be compiled by transferring the averaged neuropil models into the averaged brain surface models. The main contribution and novelty of our study is to determine the average 3D brain shape based on the isosurface suggested by the zero-crossings of a 3D accumulative signed distance map. Consequently, in contrast with previous approaches that also aim to construct a stereotypical brain model based on the probability map and a user-specified probability threshold, our method is more robust and thus capable to yield more objective and accurate results. Moreover, the obtained 3D average shape is useful for defining brain coordinate systems and will be able to provide boundary conditions for volume registration methods in the future. This method is distinguishable from those focusing on 2D + Z image volumes because its pipeline is designed to process 3D mesh surface models of Drosophila brains.},
}
@article{ShapiroARM-98,
       author = {Shapiro, J. A.},
        title = {Thinking about bacterial populations as multicellular organisms},
      journal = {Annual Review Microbiology},
         year = {1998},
       volume = {52},
        pages = {81-104},
     abstract = {It has been a decade since multicellularity was proposed as a general bacterial trait. Intercellular communication and multicellular coordination are now known to be widespread among prokaryotes and to affect multiple phenotypes. Many different classes of signaling molecules have been identified in both Gram-positive and Gram-negative species. Bacteria have sophisticated signal transduction networks for integrating intercellular signals with other information to make decisions about gene expression and cellular differentiation. Coordinated multicellular behavior can be observed in a variety of situations, including development of E. coli and B. subtilis colonies, swarming by Proteus and Serratia, and spatially organized interspecific metabolic cooperation in anaerobic bioreactor granules. Bacteria benefit from multicellular cooperation by using cellular division of labor, accessing resources that cannot effectively be utilized by single cells, collectively defending against antagonists, and optimizing population survival by differentiating into distinct cell types.}
}
@article{ShapiroandWilkBIOMETRIKA-65,
       author = {Shapiro, S. S. and Wilk, M. B.},
        title = {An Analysis of Variance Test for Normality},
      journal = {Biometrika},
       volume = 52,
         year = 1965,
        pages = {591-612},
}
@misc{SharafPOINTER-NETWORKS-17,
        title = {Inroduction to Pointer Networks - Medium, June 2, 2017},
       author = {Amr Sharaf},
 howpublished = {\url{https://medium.com/@sharaf/a-paper-a-day-11-pointer-networks-59f7af1a611c}},
         year = {2017},
}
@article{SharmaetalNATURE-NEUROSCIENCE-15,
       author = {Sharma, Kirti and Schmitt, Sebastian and Bergner, Caroline G. and Tyanova, Stefka and Kannaiyan, Nirmal and Manrique-Hoyos, Natalia and Kongi, Karina and Cantuti, Ludovico and Hanisch, Uwe-Karsten and Philips, Mari-Anne and Rossner, Moritz J. and Mann, Matthias and Simons, Mikael},
        title = {Cell type- and brain region-resolved mouse brain proteome},
      journal = {Nature Neuroscience},
         year = 2015,
    publisher = {Nature Publishing Group},
     abstract = {Brain transcriptome and connectome maps are being generated, but an equivalent effort on the proteome is currently lacking. We performed high-resolution mass spectrometry-based proteomics for in-depth analysis of the mouse brain and its major brain regions and cell types. Comparisons of the 12,934 identified proteins in oligodendrocytes, astrocytes, microglia and cortical neurons with deep sequencing data of the transcriptome indicated deep coverage of the proteome. Cell type-specific proteins defined as tenfold more abundant than average expression represented about a tenth of the proteome, with an overrepresentation of cell surface proteins. To demonstrate the utility of our resource, we focused on this class of proteins and identified Lsamp, an adhesion molecule of the IgLON family, as a negative regulator of myelination. Our findings provide a framework for a system-level understanding of cell-type diversity in the CNS and serves as a rich resource for analyses of brain development and function.},
}
@book{ShashaandLazere95,
     author = {Dennis Elliot Shasha and Cathy A. Lazere},
      title = {Out of Their Minds: The Lives and Discoveries of 15 Great Computer Scientists},
  publisher = {Copernicus Books},
    comment = {Copernicus is an imprint of Springer-Verlag},
    address = {New York, NY},
       year = 1995,
}
@inproceedings{ShatkayandKaelblingIJCAI-97,
       author = {Shatkay, Hagit and Kaelbling, Leslie},
        title = {Learning Topological Maps with Weak Local Odometric Information},
    booktitle = {Proceedings {IJCAI} 15},
     location = {Nagoya, Japan},
 organization = {IJCAII},
         year = 1997,
}
@book{ShavlikandDietterich90,
       editor = {Shavlik, Jude W. and Dietterich, Thomas G.},
        title = {Readings in Machine Learning},
    booktitle = {Readings in Machine Learning},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1990,
}
@book{Shawe-TaylorandCristianini04,
        title = {Kernel Methods for Pattern Analysis},
       author = {John Shawe-Taylor and Nello Cristianini},
    publisher = {Cambridge University Press},
      address = {New York, NY},
         year = 2004,
}
@book{Shawe-TaylorandCristianiniKERNELS-04,
       author = {Shawe-Taylor, John and Cristianini, Nello},
        title = {Kernel Methods for Pattern Analysis},
    publisher = {Cambridge University Press},
         year = 2004,
}
@book{Shawe-TaylorandCristianiniSVMS-00,
        title = {An Introduction to Support Vector Machines},
    publisher = {Cambridge University Press},
       author = {Cristianini, Nello and Shawe-Taylor, John},
         year = 2000,
}
@article{ShazeeretalCoRR-17,
       author = {Noam Shazeer and Azalia Mirhoseini and Krzysztof Maziarz and Andy Davis and Quoc V. Le and Geoffrey E. Hinton and Jeff Dean},
        title = {Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer},
      journal = {CoRR},
       volume = {arXiv:1701.06538},
         year = {2017},
          url = {http://arxiv.org/abs/1701.06538},
     abstract = {The capacity of a neural network to absorb information is limited by its number of parameters. Conditional computation, where parts of the network are active on a per-example basis, has been proposed in theory as a way of dramatically increasing model capacity without a proportional increase in computation. In practice, however, there are significant algorithmic and performance challenges. In this work, we address these challenges and finally realize the promise of conditional computation, achieving greater than          year = x improvements in model capacity with only minor losses in computational efficiency on modern GPU clusters. We introduce a Sparsely-Gated Mixture-of-Experts layer (MoE), consisting of up to thousands of feed-forward sub-networks. A trainable gating network determines a sparse combination of these experts to use for each example. We apply the MoE to the tasks of language modeling and machine translation, where model capacity is critical for absorbing the vast quantities of knowledge available in the training corpora. We present model architectures in which a MoE with up to 137 billion parameters is applied convolutionally between stacked LSTM layers. On large language modeling and machine translation benchmarks, these models achieve significantly better results than state-of-the-art at lower computational cost.}
}
@inproceedings{ShazeeretalICLR-17,
       author = {Noam Shazeer and Azalia Mirhoseini and Krzysztof Maziarz and Andy Davis and Quoc V. Le and Geoffrey E. Hinton and Jeff Dean},
        title = {Outrageously Large Neural Networks: The Sparsely-Gated Mixture-of-Experts Layer},
    booktitle = {International Conference on Learning Representations},
         year = {2017},
     abstract = {The capacity of a neural network to absorb information is limited by its number of parameters. Conditional computation, where parts of the network are active on a per-example basis, has been proposed in theory as a way of dramatically increasing model capacity without a proportional increase in computation. In practice, however, there are significant algorithmic and performance challenges. In this work, we address these challenges and finally realize the promise of conditional computation, achieving greater than one thousand times improvements in model capacity with only minor losses in computational efficiency on modern GPU clusters. We introduce a Sparsely-Gated Mixture-of-Experts layer (MoE), consisting of up to thousands of feed-forward sub-networks. A trainable gating network determines a sparse combination of these experts to use for each example. We apply the MoE to the tasks of language modeling and machine translation, where model capacity is critical for absorbing the vast quantities of knowledge available in the training corpora. We present model architectures in which a MoE with up to 137 billion parameters is applied convolutionally between stacked LSTM layers. On large language modeling and machine translation benchmarks, these models achieve significantly better results than state-of-the-art at lower computational cost.}
}
@inproceedings{ShechtmanandIraniCVPR-07,
       author = {Eli Shechtman and Michal Irani},
        title = {Matching Local Self-Similarities across Images and Videos},
    booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
          url = {www.wisdom.weizmann.ac.il/~vision/SelfSimilarities.html},
         year = 2007,
        pages = {1-8},
}
@article{ShelhameretalPAMI-16,
        title = {Fully Convolutional Networks for Semantic Segmentation},
       author = {Evan Shelhamer and Jonathan Long and Trevor Darrell},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
         year = 2016,
     abstract = {Convolutional networks are powerful visual models that yield hierarchies of features. We show that convolutional networks by themselves, trained end-to-end, pixels-to-pixels, improve on the previous best result in semantic segmentation. Our key insight is to build "fully convolutional" networks that take input of arbitrary size and produce correspondingly-sized output with efficient inference and learning. We define and detail the space of fully convolutional networks, explain their application to spatially dense prediction tasks, and draw connections to prior models. We adapt contemporary classification networks (AlexNet, the VGG net, and GoogLeNet) into fully convolutional networks and transfer their learned representations by fine-tuning to the segmentation task. We then define a skip architecture that combines semantic information from a deep, coarse layer with appearance information from a shallow, fine layer to produce accurate and detailed segmentations. Our fully convolutional network achieves improved segmentation of PASCAL VOC (30\% relative improvement to 67.2\% mean IU on 2012), NYUDv2, SIFT Flow, and PASCAL-Context, while inference takes one tenth of a second for a typical image.}
}
@article{ShenetalETS-09,
       author = {Liping Shen and Minjuan Wang and Ruimin Shen},
        title = {Affective e-Learning: Using "Emotional" Data to Improve Learning in Pervasive Learning Environment},
      journal = {Educational Technology \& Society},
       volume = 12,
        issue = 2,
         year = 2009,
        pages = {176-189},
     abstract = {Using emotion detection technologies from biophysical signals, this study explored how emotion evolves during learning process and how emotion feedback could be used to improve learning experiences. This article also described a cutting-edge pervasive e-Learning platform used in a Shanghai online college and proposed an affective e-Learning model, which combined learners' emotions with the Shanghai e-Learning platform. The study was guided by Russell's circumplex model of affect and Kort's learning spiral model. The results about emotion recognition from physiological signals achieved a best-case accuracy (86.3\%) for four types of learning emotions. And results from emotion revolution study showed that engagement and confusion were the most important and frequently occurred emotions in learning, which is consistent with the findings from AutoTutor project. No evidence from this study validated Kort's learning spiral model.},
}
@article{ShenoyetalARN-13,
       author = {Krishna V. Shenoy and Maneesh Sahani and Mark M. Churchland},
        title = {Cortical Control of Arm Movements: A Dynamical Systems Perspective},
      journal = {Annual Review of Neuroscience},
       volume = {36},
       number = {1},
        pages = {337-359},
         year = {2013},
     abstract = {Our ability to move is central to everyday life. Investigating the neural control of movement in general, and the cortical control of volitional arm movements in particular, has been a major research focus in recent decades. Studies have involved primarily either attempts to account for single-neuron responses in terms of tuning for movement parameters or attempts to decode movement parameters from populations of tuned neurons. Even though this focus on encoding and decoding has led to many seminal advances, it has not produced an agreed-upon conceptual framework. Interest in understanding the underlying neural dynamics has recently increased, leading to questions such as how does the current population response determine the future population response, and to what purpose? We review how a dynamical systems perspective may help us understand why neural activity evolves the way it does, how neural activity relates to movement parameters, and how a unified conceptual framework may result.}
}
@article{ShepherdANYAS-09,
       author = {Shepherd, Gordon M.},
        title = {Dendrodendritic synapses: past, present and future},
      journal = {Annals of the New York Academy of Science},
         year = {2009},
       volume = {1170},
        pages = {1749-6632},
     abstract = {Synapses between dendrites are at the core of mechanisms for processing odor stimuli, as well as for processing in many other brain systems. A perspective on the development of our understanding of these mechanisms may therefore be of interest. Studies of the olfactory bulb leading to the discovery of dendrodendritic synapses began in 1959. They involved a multidisciplinary approach that included Golgi cell morphology, electrophysiology, a microcircuit wiring diagram, membrane biophysics, theory of field potentials, cable theory, dendritic electrotonus theory, computational models of mitral and granule cells, prediction by the models of dendrodendritic synaptic interactions, confirmation with electron microscopy using single sections and serial sections, and final integration in the reports of feedback and lateral inhibitory interactions in 1966 and 1968. Following the discovery of glomerular odor maps in the 1970s, the functional significance of the dendrodendritic inhibition in processing the maps has been increasingly documented. Recent experimental and computational studies are revealing how these synapses are organized into distributed systems for processing the odor maps. Future studies need to situate dendrodendritic mechanisms in these distributed systems and correlate them with the tight functional loops between olfactory bulb and olfactory cortex. Studies in awake behaving animals will be increasingly important. The relations of dendritic mechanisms to perception, memory, and the pathogenesis of disorders such as Alzheimer's will be rich fields for study. Dendrites and their synapses should continue to provide ideal models for the study of basic mechanisms of cortical integration and the neural basis of smell.},
}
@book{ShepherdandGrillner2018,
        title = {Handbook of Brain Microcircuits},
       author = {Shepherd, G.M. and Grillner, S.},
         year = {2018},
    publisher = {Oxford University Press}
}
@book{Shermer2011,
       author = {Michael Shermer},
        title = {The Believing Brain: From Ghosts and Gods to Politics and Conspiracies --- How We Construct Beliefs and Reinforce Them as Truths},
    publisher = {Henry Holt and Company},
      address = {New York, NY},
         year = 2011,
      comment = {\urlh{http://en.wikipedia.org/wiki/Michael_Shermer}{Michael Shermer} is 
               the founder of The Skeptics Society, and Editor in Chief of its magazine 
               {\it{Skeptic}}. His descriptions of {\it{paternicity}} --- our tendency 
               to infer causal associations on the basis of a single example --- and 
               {\it{agenticity}} --- our otherwise evolutionarily invaluable theory of 
               mind run amok which encourages us to impute mental states to just about
               any random collection of matter or wacky fabrication --- include lots of 
               background research.},
}
@book{Sherrington1906,
        title = {The integrative action of the nervous system},
       author = {Sherrington, Charles},
    publisher = {Yale University Press},
      address = {New Haven, CT},
         year = 1906,
}
@article{SherwoodetalJoA-08,
       author = {Sherwood, Chet C. and Subiaul, Francys and Zawidzki, Tadeusz W.},
        title = {A natural history of the human mind: tracing evolutionary changes in brain and cognition},
      journal = {Journal of Anatomy},
         year = {2008},
    publisher = {Blackwell Science Inc},
       volume = {212},
        issue = {4},
        pages = {426-454},
     abstract = {Since the last common ancestor shared by modern humans, chimpanzees and bonobos, the lineage leading to Homo sapiens has undergone a substantial change in brain size and organization. As a result, modern humans display striking differences from the living apes in the realm of cognition and linguistic expression. In this article, we review the evolutionary changes that occurred in the descent of Homo sapiens by reconstructing the neural and cognitive traits that would have characterized the last common ancestor and comparing these with the modern human condition. The last common ancestor can be reconstructed to have had a brain of approximately 300-400 g that displayed several unique phylogenetic specializations of development, anatomical organization, and biochemical function. These neuroanatomical substrates contributed to the enhancement of behavioral flexibility and social cognition. With this evolutionary history as precursor, the modern human mind may be conceived as a mosaic of traits inherited from a common ancestry with our close relatives, along with the addition of evolutionary specializations within particular domains. These modern human-specific cognitive and linguistic adaptations appear to be correlated with enlargement of the neocortex and related structures. Accompanying this general neocortical expansion, certain higher-order unimodal and multimodal cortical areas have grown disproportionately relative to primary cortical areas. Anatomical and molecular changes have also been identified that might relate to the greater metabolic demand and enhanced synaptic plasticity of modern human brain's. Finally, the unique brain growth trajectory of modern humans has made a significant contribution to our species' cognitive and linguistic abilities.},
}
@article{ShevchenkoIVC-91,
       author = {Mikhail Shevchenko and David Windridge and Josef Kittler},
        title = {A linear-complexity reparameterisation strategy for the hierarchical bootstrapping of capabilities within perception-action architectures},
      journal = {Image and Vision Computing},
       volume = {27},
       number = {11},
        pages = {1702-1714},
         year = {2009},
     abstract = {Perception-action (PA) architectures are capable of solving a number of problems associated with artificial cognition, in particular, difficulties concerned with framing and symbol grounding. Existing PA algorithms tend to be 'horizontal' in the sense that learners maintain their prior percept-motor competences unchanged throughout learning. We here present a methodology for simultaneous 'horizontal' and 'vertical' perception-action learning in which there additionally exists the capability for incremental accumulation of novel percept-motor competences in a hierarchical fashion. The proposed learning mechanism commences with a set of primitive 'innate' capabilities and progressively modifies itself via recursive generalising of parametric spaces within the linked perceptual and motor domains so as to represent environmental affordances in maximally-compact manner. Efficient reparameterising of the percept domain is here accomplished by the exploratory elimination of dimensional redundancy and environmental context. Experimental results demonstrate that this approach exhibits an approximately linear increase in computational requirements when learning in a typical unconstrained environment, as compared with at least polynomially-increasing requirements for a classical perception-action system.}
}
@phdthesis{ShewchukPhD,
       author = {Shewchuk, John},
        title = {Reinforcement Learning in Dynamical Systems With High Input and Output Dimensionality},
       school = {Brown University},
         year = {Forthcoming},
      address = {Providence, RI},
}
@inproceedings{ShewchukandDeanIC-90,
       author = {Shewchuk, John and Dean, Thomas},
        title = {Towards Learning Time-Varying Functions With High Input Dimensionality},
    booktitle = {Proceedings of the 5th {IEEE} International Symposium on Intelligent Control},
     location = {Philadelphia, Pennsylvania},
 organization = {IEEE},
        pages = {383-388},
         year = 1990,
}
@techreport{ShewchukandViolaTR-89,
       author = {Shewchuk, John and Viola, Paul},
        title = {Implementing a Learning System for Subsumption Architectures},
        issue = {CS-89-17},
  institution = {Brown University Department of Computer Science},
         year = 1989,
}
@article{ShiandMalikPAMI-00,
       author = {Jianbo Shi and Jitendra Malik},
        title = {Normalized Cuts and Image Segmentation},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
       volume = 22,
        issue = 8,
         year = 2000,
        pages = {888-905},
}
@inproceedings{ShiandTomasiCVPR-94,
       author = {Jianbo Shi and Carlo Tomasi},
        title = {Good Features to Track},
    booktitle = {Proceedings of the 1994 {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
      address = {Seattle},
        pages = {593-600},
         year = 1994
}
@article{ShietalCoRR-16,
       author = {Wenzhe Shi and Jose Caballero and Ferenc Husz\'{a}r and Johannes Totz and Andrew P. Aitken and Rob Bishop and Daniel Rueckert and Zehan Wang},
        title = {Real-Time Single Image and Video Super-Resolution Using an Efficient Sub-Pixel Convolutional Neural Network},
      journal = {CoRR},
       volume = {arXiv:1609.05158},
         year = {2016},
      comment = {http://www.inference.vc/holiday-special-deriving-the-subpixel-cnn-from-first-principles/},
     abstract = {Recently, several models based on deep neural networks have achieved great success in terms of both reconstruction accuracy and computational performance for single image super-resolution. In these methods, the low resolution (LR) input image is upscaled to the high resolution (HR) space using a single filter, commonly bicubic interpolation, before reconstruction. This means that the super-resolution (SR) operation is performed in HR space. We demonstrate that this is sub-optimal and adds computational complexity. In this paper, we present the first convolutional neural network (CNN) capable of real-time SR of 1080p videos on a single K2 GPU. To achieve this, we propose a novel CNN architecture where the feature maps are extracted in the LR space. In addition, we introduce an efficient sub-pixel convolution layer which learns an array of upscaling filters to upscale the final LR feature maps into the HR output. By doing so, we effectively replace the handcrafted bicubic filter in the SR pipeline with more complex upscaling filters specifically trained for each feature map, whilst also reducing the computational complexity of the overall SR operation. We evaluate the proposed approach using images and videos from publicly available datasets and show that it performs significantly better (+0.15dB on Images and +0.39dB on Videos) and is an order of magnitude faster than previous CNN-based methods.},
}
@incollection{ShietalIAR-12,
        title = {Biologically Motivated Local Contextual Modulation Improves Low-Level Visual Feature Representations},
       author = {Shi, Xun and Bruce, Neil D.B. and Tsotsos, John K.},
    booktitle = {Image Analysis and Recognition},
       volume = 7324,
       series = {Lecture Notes in Computer Science},
       editor = {Campilho, Aurelio and Kamel, Mohamed},
    publisher = {Springer Berlin Heidelberg},
         year = 2012,
        pages = {79-88},
}
@article{ShietalNATURE-13,
       author = {Shi, Jian and Ha, Sieu D. and Zhou, You and Schoofs, Frank and Ramanathan, Shriram},
        title = {A correlated nickelate synaptic transistor},
      journal = {Nature Communications},
         year = {2013},
    publisher = {Nature Publishing Group},
       volume = {4},
     abstract = {Inspired by biological neural systems, neuromorphic devices may open up new computing paradigms to explore cognition, learning and limits of parallel computation. Here we report the demonstration of a synaptic transistor with SmNiO$^3$, a correlated electron system with insulator-metal transition temperature at 130\hmdegree{}C in bulk form. Non-volatile resistance and synaptic multilevel analogue states are demonstrated by control over composition in ionic liquid-gated devices on silicon platforms. The extent of the resistance modulation can be dramatically controlled by the film microstructure. By simulating the time difference between postneuron and preneuron spikes as the input parameter of a gate bias voltage pulse, synaptic spike-timing-dependent plasticity learning behaviour is realized. The extreme sensitivity of electrical properties to defects in correlated oxides may make them a particularly suitable class of materials to realize artificial biological circuits that can be operated at and above room temperature and seamlessly integrated into conventional electronic circuits.},
}
@article{ShinetalICLR-18,
        title = {Differentiable Neural Network Architecture Search},
       author = {Richard Shin and Charles Packer and Dawn Song},
      journal = {International Conference on Learning Representations},
         year = {2018},
abstract = {The successes of deep learning in recent years has been fueled by the development of innovative new neural network architectures. However, the design of a neural network architecture remains a difficult problem, requiring significant human expertise as well as computational resources. In this paper, we propose a method for transforming a discrete neural network architecture space into a continuous and differentiable form, which enables the use of standard gradient-based optimization techniques for this problem, and allows us to learn the architecture and the parameters simultaneously. We evaluate our methods on the Udacity steering angle prediction dataset, and show that our method can discover architectures with similar or better predictive accuracy but significantly fewer parameters and smaller computational cost.}
}
@inproceedings{ShinetalICLR-18b,
        title = {Towards Specification-Directed Program Repair},
       author = {Richard Shin and Illia Polosukhin and Dawn Song},
    booktitle = {Submitted to International Conference on Learning Representations},
         year = {2018},
     abstract = {Several recent papers have developed neural network program synthesizers by using supervised learning over large sets of randomly generated programs and specifications. In this paper, we investigate the feasibility of this approach for program repair: given a specification and a candidate program assumed similar to a correct program for the specification, synthesize a program which meets the specification.  Working in the Karel domain with a dataset of synthetically generated candidates, we develop models that can make effective use of the extra information in candidate programs, achieving 40\% error reduction compared to a baseline program synthesis model that only receives the specification and not a candidate program.}
}
@article{ShipmanetalSCIENCE-16,
       author = {Shipman, Seth L. and Nivala, Jeff and Macklis, Jeffrey D. and Church, George M.},
        title = {Molecular recordings by directed {CRISPR} spacer acquisition},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {353},
       number = {6298},
         year = {2016},
     abstract = {Although recent advances in DNA synthesis and sequencing technologies have made practical the writing and readout of arbitrary data in the form of synthetic DNA, still lacking are the robust tools necessary to generate a dynamic record of such information within the genomes of living cells. An in vivo system, built out of biological parts with large storage capacity, would enable the recording of defined biological events into stable genetic memory and facilitate the tracking of long molecular and cellular histories.
        The CRISPR (clustered regularly interspaced short palindromic repeats)-Cas system is a prokaryotic type of immunological memory. Foreign DNA sequences originating from viral infections are stored within genome-based arrays in the form of short sequences-called spacers-that confer sequence-specific resistance to the invading nucleic acids. These arrays not only preserve the spacer sequences but also record the order in which the sequences are acquired, generating a temporal record of acquisition events. We harnessed this system to record arbitrary DNA sequences into a genomic CRISPR array in the form of spacers acquired from synthetic oligonucleotides electroporated into a population of cells overexpressing the CRISPR adaptation proteins Cas1 and Cas2. This enabled the recording of defined molecular events into a stable genomic locus over time and the storage of arbitrary information across a population of cells.RESULTSWe show that the Cas1-Cas2 complex can be used in vivo to integrate synthetic DNA of a defined sequence into the Escherichia coli genome. We used this feature to examine the type I-E CRISPR-Cas spacer acquisition process and optimized the synthetic spacer design to achieve higher acquisition efficiency and specific integration orientation through the addition of an AAG protospacer adjacent motif (PAM). We then generated stable genomic recordings of multiple molecular events by electroporating sets of oligonucleotides over several days. These molecular records were read out with high-throughput sequencing and then decoded with a program that identified and faithfully reconstructed the temporal event order.Last, we used directed evolution to generate many Cas1-Cas2 mutants with modified PAM specificity (PAMNC). By modulating expression of these mutant and wild-type Cas1-Cas2 complexes, we could dynamically control the orientation of spacer integration. This enabled us to record acquisition events in multiple modes. That is, information was encoded in both the temporal order of the spacers and the orientation in which they were integrated.
        Our results establish a recording system that uses the nucleotide content, temporal ordering, and orientation of defined DNA sequences within a CRISPR array in order to encode arbitrary information within the genomes of a population of cells. Because information can be encoded in spacer nucleotide space (up to two bits per base) and in alternate modes, the system has the potential to record and permanently store higher capacities of information than any other synthetic biological system to date. This lays the foundation for an in vivo recording device that could be coupled with diverse molecular phenomena and used for applications that require tracing of long molecular histories. We also demonstrate that delivery of synthetic DNA substrates to a CRISPR-Cas adaptation system in vivo is a practical method to probe and adapt the system.Two modes of encoding information into the CRISPR locus.(A) Oligonucleotides containing an AAG PAM and 32 variable bases were electroporated into cells overexpressing Cas1-Cas2 and inserted into the genomic CRISPR array. Delivery of oligos with distinct sequence over time generates a molecular record. (B) Cas1-Cas2 mutants identified through directed evolution alter the orientation of acquisition. Varying expression ratios of wild-type and mutant Cas1-Cas2 over time generates a record encoded in spacer orientation.The ability to write a stable record of identified molecular events into a specific genomic locus would enable the examination of long cellular histories and have many applications, ranging from developmental biology to synthetic devices. We show that the type I-E CRISPR (clustered regularly interspaced short palindromic repeats){\textendash}Cas system of Escherichia coli can mediate acquisition of defined pieces of synthetic DNA. We harnessed this feature to generate records of specific DNA sequences into a population of bacterial genomes. We then applied directed evolution so as to alter the recognition of a protospacer adjacent motif by the Cas1-Cas2 complex, which enabled recording in two modes simultaneously. We used this system to reveal aspects of spacer acquisition, fundamental to the CRISPR-Cas adaptation process. These results lay the foundations of a multimodal intracellular recording device.},
}
@article{ShiptonetalPNAS-14,
       author = {Shipton, Olivia A. and El-Gaby, Mohamady and Apergis-Schoute, John and Deisseroth, Karl and Bannerman, David M. and Paulsen, Ole and Kohl, Michael M.},
        title = {Left-right dissociation of hippocampal memory processes in mice},
      journal = {Proceedings of the National Academy Science USA},
    publisher = {National Academy of Sciences},
       volume = {111},
        issue = {42},
         year = {2014},
        pages = {15238-15243},
     abstract = {Left-right asymmetries have likely evolved to make optimal use of bilaterian nervous systems; however, little is known about the synaptic and circuit mechanisms that support divergence of function between equivalent structures in each hemisphere. Here we examined whether lateralized hippocampal memory processing is present in mice, where hemispheric asymmetry at the CA3-CA1 pyramidal neuron synapse has recently been demonstrated, with different spine morphology, glutamate receptor content, and synaptic plasticity, depending on whether afferents originate in the left or right CA3. To address this question, we used optogenetics to acutely silence CA3 pyramidal neurons in either the left or right dorsal hippocampus while mice performed hippocampus-dependent memory tasks. We found that unilateral silencing of either the left or right CA3 was sufficient to impair short-term memory. However, a striking asymmetry emerged in long-term memory, wherein only left CA3 silencing impaired performance on an associative spatial long-term memory task, whereas right CA3 silencing had no effect. To explore whether synaptic properties intrinsic to the hippocampus might contribute to this left-right behavioral asymmetry, we investigated the expression of hippocampal long-term potentiation. Following the induction of long-term potentiation by high-frequency electrical stimulation, synapses between CA3 and CA1 pyramidal neurons were strengthened only when presynaptic input originated in the left CA3, confirming an asymmetry in synaptic properties. The dissociation of hippocampal long-term memory function between hemispheres suggests that memory is routed via distinct left-right pathways within the mouse hippocampus, and provides a promising approach to help elucidate the synaptic basis of long-term memory.},
}
@article{ShirakawaetalJCS-77,
       author = {Shirakawa, Hideki and Louis, Edwin J. and MacDiarmid, Alan G. and Chiang, Chwan K. and Heeger, Alan J.},
        title = {Synthesis of electrically conducting organic polymers: halogen derivatives of polyacetylene, (CH)},
      journal = {Journal of the Chemical Society, Chemical Communications},
         year = 1977,
        issue = 16,
        pages = {578-580},
    publisher = {The Royal Society of Chemistry},
     abstract = {When silvery films of the semiconducting polymer, 'polyacetylene', (CH), are exposed to chlorine, bromine, or iodine vapour, uptake of halogen occurs, and the conductivity increases markedly (over seven orders of magnitude in the case of iodine) to give, depending on the extent of halogenation, silvery or silvery-black films, some of which have a remarkably high conductivity at room temperature.}
}
@article{ShiroguchiandKinositaSCIENCE-07,
        title = {Myosin {V} Walks by Lever Action and Brownian Motion},
       author = {Shiroguchi, Katsuyuki and Kinosita, Kazuhiko},
      journal = {Science},
       volume = 316,
        issue = 5828,
         year = 2007,
        pages = {1208-1212},
     abstract = {Myosin V is a molecular motor that moves cargo along actin filaments. Its two heads, each attached to a long and relatively stiff neck, move alternately forward in a "hand-over-hand" fashion. To observe under a microscope how the necks move, we attached a micrometer-sized rod to one of the necks. The leading neck swings unidirectionally forward, whereas the trailing neck, once lifted, undergoes extensive Brownian rotation in all directions before landing on a site ahead of the leading head. The neck-neck joint is essentially free, and the neck motion supports a mechanism where the active swing of the leading neck biases the random motion of the lifted head to let it eventually land on a forward site.},
}
@inproceedings{ShivandQuirkNIPS-19,
       author = {Shiv, Vighnesh Leonardo and Quirk, Chris},
        title = {Novel positional encodings to enable tree-based transformers},
    booktitle = {NeurIPS 2019},
         year = {2019},
     abstract = {Neural models optimized for tree-based problems are of great value in tasks like SQL query extraction and program synthesis. On sequence-structured data, transformers have been shown to learn relationships across arbitrary pairs of positions more reliably than recurrent models. Motivated by this property, we propose a method to extend transformers to tree-structured data, enabling sequence-to-tree, tree-to-sequence, and tree-to-tree mappings. Our approach abstracts the transformer's sinusoidal positional encodings, allowing us to instead use a novel positional encoding scheme to represent node positions within trees. We evaluated our model in tree-to-tree program translation and sequence-to-tree semantic parsing settings, achieving superior performance over both sequence-to-sequence transformers and state-of-the-art tree-based LSTMs on several datasets. In particular, our results include a 22\% absolute increase in accuracy on a JavaScript to CoffeeScript translation dataset.}
}
@article{ShlensetalNEUROSCIENCE-06,
        title = {The structure of multi-neuron firing patterns in primate retina},
       author = {Shlens, J. and Field, G.D. and Gauthier, J.L. and Grivich, M.I. and Petrusca, D. and Sher, A. and Litke, A.M. and Chichilnisky, E.J.},
      journal = {Journal of Neuroscience},
       volume = 26,
        issue = 32,
        pages = {8254-66},
         year = 2006,
     abstract = {Current understanding of many neural circuits is limited by our ability to explore the vast number of potential interactions between different cells. We present a new approach that dramatically reduces the complexity of this problem. Large-scale multi-electrode recordings were used to measure electrical activity in nearly complete, regularly spaced mosaics of several hundred ON and OFF parasol retinal ganglion cells in macaque monkey retina. Parasol cells exhibited substantial pairwise correlations, as has been observed in other species, indicating functional connectivity. However, pairwise measurements alone are insufficient to determine the prevalence of multi-neuron firing patterns, which would be predicted from widely diverging common inputs and have been hypothesized to convey distinct visual messages to the brain. The number of possible multi-neuron firing patterns is far too large to study exhaustively, but this problem may be circumvented if two simple rules of connectivity can be established: (1) multi-cell firing patterns arise from multiple pairwise interactions, and (2) interactions are limited to adjacent cells in the mosaic. Using maximum entropy methods from statistical mechanics, we show that pairwise and adjacent interactions accurately accounted for the structure and prevalence of multi-neuron firing patterns, explaining approximately 98\% of the departures from statistical independence in parasol cells and approximately 99\% of the departures that were reproducible in repeated measurements. This approach provides a way to define limits on the complexity of network interactions and thus may be relevant for probing the function of many neural circuits.}
}
@article{ShlensetalNEUROSIENCE-09,
        title = {The structure of large-scale synchronized firing in primate retina},
       author = {Shlens, J. and Field, G.D. and Gauthier, J.L. and Greschner, M. and Sher, A. and Litke, A.M. and Chichilnisky, E.J.},
      journal = {Journal of Neuroscience},
       volume = 29,
        issue = 15,
        pages = {5022-31},
         year = 2009,
     abstract = {Synchronized firing among neurons has been proposed to constitute an elementary aspect of the neural code in sensory and motor systems. However, it remains unclear how synchronized firing affects the large-scale patterns of activity and redundancy of visual signals in a complete population of neurons. We recorded simultaneously from hundreds of retinal ganglion cells in primate retina, and examined synchronized firing in completely sampled populations of approximately 50-100 ON-parasol cells, which form a major projection to the magnocellular layers of the lateral geniculate nucleus. Synchronized firing in pairs of cells was a subset of a much larger pattern of activity that exhibited local, isotropic spatial properties. However, a simple model based solely on interactions between adjacent cells reproduced 99\% of the spatial structure and scale of synchronized firing. No more than 20\% of the variability in firing of an individual cell was predictable from the activity of its neighbors. These results held both for spontaneous firing and in the presence of independent visual modulation of the firing of each cell. In sum, large-scale synchronized firing in the entire population of ON-parasol cells appears to reflect simple neighbor interactions, rather than a unique visual signal or a highly redundant coding scheme.}
}
@misc{ShlizermanandBabbCOSYNE-14,
        title = {Inference of Functional Sub-Circuits in the Dynamical Connectome of the {\it{Caenorhabditis elegans}} Worm},
       author = {Eli Shlizerman1 and S. Grant Babb},
         year = 2014,
 howpublished = {Proposal submitted to Google Research Awards},
}
@article{ShmueliSTATISTICAL-SCIENCE-10,
       author = {Galit Shmueli},
        title = {To explain or predict?},
      journal = {Statistical Science},
    publisher = {Institute of Mathematical Statistics},
       volume = {25},
        issue = {3},
         year = {2010},
        pages = {289-310},
}
@book{Shoham94,
       author = {Shoham, Yoav},
        title = {Artificial Intelligence Techniques in Prolog},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1994,
}
@inproceedings{ShohamAAAI-86,
       author = {Shoham, Yoav},
        title = {Chronological Ignorance: Time, Nonmonotonicity, Necessity, and Causal Theories},
    booktitle = {Proceedings AAAI-86},
 organization = {AAAI},
     location = {Philadelphia, Pennsylvania},
         year = 1986,
        pages = {389-393},
}
@inproceedings{ShohamIJCAI-89,
       author = {Shoham, Yoav},
        title = {Time for Action: On the Relation Between Time, Knowledge, and Action},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
         year = 1989,
        pages = {954-959},
}
@book{ShohamPhD,
       author = {Shoham, Yoav},
        title = {Reasoning About Change: Time and Causation from the Standpoint of Artificial Intelligence},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1988,
}
@techreport{ShohamTR-86,
       author = {Shoham, Yoav},
        title = {Reasoning About Change: Time and Causation from the Standpoint of Artificial Intelligence},
        issue = 507,
  institution = {Yale University Computer Science Department},
         year = 1986,
}
@techreport{ShohamTR-89,
       author = {Shoham, Yoav},
        title = {Agent-Oriented Programming},
        issue = {Working Paper},
  institution = {Stanford Knowledge Systems Laboratory},
         year = 1989,
}
@inproceedings{ShohamandDeanCS-85,
       author = {Shoham, Yoav and Thomas Dean},
        title = {Temporal Notation and Causal Terminology},
    booktitle = {Proceedings 7th Annual Conference of the Cognitive Science Society},
 organization = {Cognitive Science Society},
     location = {Irvine, California},
        pages = {90-99},
         year = 1985,
}
@inproceedings{ShohamandMosesIJCAI-89,
       author = {Shoham, Yoav and Moses, Yoram},
        title = {Belief as Defeasible Knowledge},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
         year = 1989,
        pages = {1168-1173},
}
@article{ShohamyandDawTiBS-15,
       author = {Daphna Shohamy and Nathaniel D. Daw},
        title = {Integrating memories to guide decisions},
      journal = {Current Opinion in Behavioral Sciences},
       volume = {5},
         year = {2015},
        pages = {85-90},
     abstract = {Adaptive decisions are guided by past experience. Yet, decisions are often made between alternatives that have not been directly experienced before, requiring the integration of memories across multiple past events. We review emerging findings supporting at least two seemingly distinct mechanisms for how the brain draws on memories in the service of choice. Prospective integration is triggered when a new decision is faced, allowing inferences to be drawn on the spot. A complementary retrospective mechanism integrates existing memories into a network of related experiences before a decision is actually faced. We discuss evidence supporting each of these mechanisms and the implications for understanding the role of memory in decision-making.}
}
@book{Shortliffe76,
       author = {Shortliffe, E. H.},
        title = {Computer-Based Medical Consultations: {MYCIN}},
    publisher = {American Elsevier},
      address = {New York},
         year = 1976,
}
@inproceedings{ShortreedandMeilaUAI-05,
    author = {Susan Shortreed and Marina Meil\u{a}},
     title = {Unsupervised spectral learning},
 booktitle = {Proceedings of the 21th Annual Conference on Uncertainty in Artificial Intelligence},
 publisher = {AUAI Press},
   address = {Arlington, Virginia},
     pages = {534-54},
      year = 2005,
}
@inproceedings{ShottonetalCVPR-08,
     author = {J. Shotton and M. Johnson and R. Cipolla},
      title = {Semantic Texton Forests for Image Categorization and Segmentation},
  booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
       year = 2008,
}
@inproceedings{ShottonetalCVPR-11,
      title = {Real-Time Human Pose Recognition in Parts from Single Depth Images},
     author = {Jamie Shotton and Andrew Fitzgibbon and Mat Cook and Toby Sharp and 
               Mark Finocchio and Richard Moore and Alex Kipman and Andrew Blake},
  booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
       year = 2011,
}
@inproceedings{ShottonetalICCV-05,
     author = {Shotton, Jamie and Blake, Andrew and Cipolla, Roberto},
      title = {Contour-Based Learning for Object Detection},
  booktitle = {Proceedings of the 10th {IEEE} International Conference on Computer Vision},
      pages = {503-510},
       year = 2005,
}
@article{ShottonetalPAMI-07,
     author = {Shotton, Jamie. and Blake, Andrew and Cipolla, Roberto},
      title = {Multi-Scale Categorical Object Recognition Using Contour Fragments},
    journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
       year = 2007,
}
@article{ShrivastavaetalCoRR-16,
       author = {Abhinav Shrivastava and Rahul Sukthankar and Jitendra Malik and Abhinav Gupta},
        title = {Beyond Skip Connections: Top-Down Modulation for Object Detection},
      journal = {CoRR},
       volume = {arXiv:1612.06851},
         year = {2016},
     abstract = {In recent years, we have seen tremendous progress in the field of object detection. Most of the recent improvements have been achieved by targeting deeper feedforward networks. However, many hard object categories, such as bottle and remote, require representation of fine details and not coarse, semantic representations. But most of these fine details are lost in the early convolutional layers. What we need is a way to incorporate finer details from lower layers into the detection architecture. Skip connections have been proposed to combine high-level and lowlevel features, but we argue that selecting the right features from low-level requires top-down contextual information. Inspired by the human visual pathway, in this paper we propose top-down modulations as a way to incorporate fine details into the detection framework. Our approach supplements the standard bottom-up, feedforward ConvNet with a top-down modulation (TDM) network, connected using lateral connections. These connections are responsible for the modulation of lower layer filters, and the top-down network handles the selection and integration of features. The proposed architecture provides a significant boost on the COCO benchmark for VGG16, ResNet101, and InceptionResNet-v2 architectures. Preliminary experiments using InceptionResNet-v2 achieve 36.8 AP, which is the best performance to-date on the COCO benchmark using a single-model without any bells and whistles (e.g., multi-scale, iterative box refinement, etc.).}
}
@article{ShuetalPLoS-BIOLOGY-11,
       author = {Shu, Xiaokun and Lev-Ram, Varda and Deerinck, Thomas J. and Qi, Yingchuan and Ramko, Ericka B. and Davidson, Michael W. and Jin, Yishi and Ellisman, Mark H. and Tsien, Roger Y.},
      journal = {{PLoS} Biology},
    publisher = {Public Library of Science},
        title = {A Genetically Encoded Tag for Correlated Light and Electron Microscopy of Intact Cells, Tissues, and Organisms},
         year = 2011,
       volume = 9,
        issue = 4,
        pages = {e1001041},
     abstract = {Electron microscopy (EM) once revolutionized cell biology by revealing subcellular anatomy at resolutions of tens of nanometers, well below the diffraction limit of light microscopy. Over the past two decades, light microscopy has been revitalized by the development of spontaneously fluorescent proteins, which allow nearly any protein of interest to be specifically tagged by genetic fusion. EM has lacked comparable genetic tags that are generally applicable. Here, we introduce "miniSOG", a small (106-residue) fluorescent flavoprotein that efficiently generates singlet oxygen when illuminated by blue light. In fixed tissue, photogenerated singlet oxygen locally polymerizes diaminobenzidine into a precipitate that is stainable with osmium and therefore can be readily imaged at high resolution by EM. Thus miniSOG is a versatile label for correlated light and electron microscopy of genetically tagged proteins in cells, tissues, and organisms including intact nematodes and mice. As a demonstration of miniSOG's capabilities, controversies about the localization of synaptic cell adhesion molecules are resolved by EM of miniSOG fusions in neuronal culture and intact mouse brain.}
}
@article{ShulakeretalNATURE-13,
    author = {Shulaker, Max M. and Hills, Gage and Patil, Nishant and Wei, Hai and Chen, Hong-Yu and Wong, H.-S. Philip and Mitra, Subhasish},
     title = {Carbon nanotube computer},
   journal = {Nature},
    volume = 501,
     issue = 7468,
      year = 2013,
     pages = {526-530},
}
@article{ShultzetalNATURE-11,
     title = {Stepwise evolution of stable sociality in primates},
    author = {Shultz, Susanne and Opie, Christopher and Atkinson, Quentin D.},
   journal = {Nature},
    volume = 479,
     issue = 7372,
 publisher = {Nature Publishing Group},
      year = 2011,
     pages = {219-222},
}
@article{ShurrabetalERMD-14,
       author = {Shurrab, M. and Schilling, R. and Gang, E. and Khan, E. M. and Crystal, E.},
        title = {Robotics in invasive cardiac electrophysiology},
      journal = {Expert Review Medical Devices},
         year = 2014,
       volume = 11,
        issue = 4,
        pages = {375-381},
}
@article{ShutohetalNEUROSCIENCE-06,
       author = {Shutoh, F.  and Ohki, M.  and Kitazawa, H. and Itohara, S. and Nagao, S.},
        title = {Memory trace of motor learning shifts transsynaptically from cerebellar cortex to nuclei for consolidation},
      journal = {Neuroscience},
       volume = {139},
       number = {2},
         year = {2006},
        pages = {767-777},
     abstract = {Adaptation of ocular reflexes is a prototype of motor learning. While the cerebellum is acknowledged as the critical site for motor learning, the functional differences between the cerebellar cortex and nuclei in motor memory formation are not precisely known. Two different views are proposed: one that the memory is formed within the cerebellar flocculus, and the other that the memory is formed within vestibular nuclei. Here we developed a new paradigm of long-term adaptation of mouse horizontal optokinetic response eye movements and examined the location of its memory trace. We also tested the role of flocculus and inferior olive in long-term adaptation by chronic lesion experiments. Reversible bilateral flocculus shutdown with local application of 0.5 microl-5\% lidocaine extinguished the memory trace of day-long adaptation, while it very little affected the memory trace of week-long adaptation. The responsiveness of vestibular nuclei after week-long adaptation was examined by measuring the extracellular field responses to the electrical stimulation of vestibular nerve under trichloroacetaldehyde anesthesia. The amplitudes and slopes of evoked monosynaptic field response (N1) of week-long adapted mice were enhanced around the medial vestibular nucleus compared with those of control mice. Chronic flocculus or inferior olive lesions abolished both day and week-long adaptations. These results suggest that the functional memory trace of short-term adaptation is formed initially within the cerebellar cortex, and later transferred to vestibular nuclei to be consolidated to a long-term memory. Both day and week-long adaptations were markedly depressed when neural nitric oxide was pharmacologically blocked locally and when neuronal nitric oxide synthase was ablated by gene knockout, suggesting that cerebellar long-term depression underlies both acquisition and consolidation of motor memory.}
}
@article{ShweetalMIM-91,
    author = {M.A. Shwe and B. Middleton and D.E. Heckerman and M. Henrion and E.J. Horvitz and H.P. Lehmann and G.F. Cooper},
     title = {Probabilistic diagnosis using a reformulation of the {INTERNIST-1/QMR} knowledge base {I}: The Probabilistic Model and Inference Algorithms},
   journal = {Methods of Information in Medicine},
    volume = 30,
     issue = 4,
     month = {October},
     pages = {241-255},
      year = 1991
}
@article{SidiropoulouetalEMBO-06,
       author = {Sidiropoulou, Kyriaki and Pissadaki, Eleftheria Kyriaki and Poirazi, Panayiota},
        title = {Inside the brain of a neuron},
      journal = {EMBO Reports},
         year = {2006},
       volume = {7},
        issue = {9},
        pages = {886-892},
     abstract = {For many decades, neurons were considered to be the elementary computational units of the brain and were assumed to summate incoming signals and elicit action potentials only in response to suprathreshold stimuli. Although modelling studies predicted that single neurons constitute a much more powerful computational entity, able to perform an array of nonlinear calculations, this possibility was not explored experimentally until the discovery of active mechanisms in the dendrites of most neuron types. Here, we review several modelling studies that have addressed information processing in single neurons, starting with those characterizing the arithmetic of different dendritic components, to those tackling neuronal integration at the cell body and, finally, those analysing the computational abilities of the axon. We present modelling predictions along with supporting experimental data in an effort to highlight the significant contribution of modelling work to enhancing our understanding of single-neuron arithmetic.},
}
@inproceedings{SieberandKrennACL-10,
       author = {Gregor Sieber and Brigitte Krenn},
        title = {Episodic Memory for Companion Dialogue},
    booktitle = {Proceedings of the 2010 Workshop on Companionable Dialogue Systems},
    publisher = {Association for Computational Linguistics},
        pages = {2010},
         year = {1-6},
     abstract = {We present an episodic memory component for enhancing the dialogue of artificial companions with the capability to refer to, take up and comment on past interactions with the user, and to take into account in the dialogue long-term user preferences and interests. The proposed episodic memory is based on RDF representations of the agent's experiences and is linked to the agent's semantic memory containing the agent's knowledge base of ontological data and information about the user's interests.}
}
@misc{SieberandKrennCOGSYS-10,
       author = {Gregor Sieber and Brigitte Krenn},
        title = {Enhancing Companion Dialogue with Episodic Memory},
 howpublished = {Poster at Cognitive Systems 2010},
         year = {2010}
}
@article{SietalCELL-10,
    author = {Kausik Si and Yun-Beom Choi and Erica White-Grindley and Amitabha Majumdar and Eric R. Kandel},
     title = {Aplysia {CPEB} Can Form Prion-like Multimers in Sensory Neurons that Contribute to Long-Term Facilitation},
   journal = {Cell},
    volume = 140,
     issue = 3,
      year = 2010,
     pages = {421-435},
}
@inproceedings{SigaletalCVPR-04,
     author = {Leonid Sigal and Sidharth Bhatia and Stefan Roth and
               Michael J. Black and Michael Isard},
      title = {Tracking Loose-limbed People},
  booktitle = {Proceedings of IEEE Computer Vision and Pattern Recognition},
   location = {Washington, DC},
       year = 2004,
}
@article{SiklossyandRoachAIJ-75,
        title = {Model Verification and Improvement Using {DISPROVER}},
       author = {Laurent Sikl{\'{o}}ssy and John W. Roach},
      journal = {Artificial Intelligence},
       volume = {6},
       number = {1},
         year = {1975},
        pages = {41-52},
     abstract = {Confidence in the adequacy of a model is increased if tasks that are impossible in the world are shown to correspond to disprovable tasks in the model. DISPROVER has been used as a tool to test, in worlds of robots, the impossibility of tasks related to various conservation laws (objects, position, model consistency, etc.) and time constraints. The adequacy and sufficiency of operators can be established. Interacting with DISPROVER, the model designer can improve his axiomatization. The frontier between "acceptable" and "ridiculous" axiomatizations is shown, in many examples, to be a most tenuous one.}
}
@book{Silberschatzetal01,
     author = {Abraham Silberschatz and Peter Baer Galvin and Greg Gagne},
      title = {Operating System Concepts},
  publisher = {Sixth edition, John Wiley \& Sons},
    address = {New York, NY},
       year = 2001,
}
@book{Silberschatzetal01b,
     author = {Abraham Silberschatz and Henry F. Korth and S. Sudarshan},
      title = {Database System Concepts},
  publisher = {Fourth edition, McGraw-Hill},
    address = {New York, NY},
       year = 2001,
}
@inproceedings{SillaetaSMC-07,
       author = {Silla, C.N. and Kaestner, C.A.A. and Koerich, A.L.},
    booktitle = {IEEE International Conference on Systems, Man and Cybernetics},
        title = {Automatic music genre classification using ensemble of classifiers},
         year = 2007,
        pages = {1687-1692},
     abstract = {This paper presents a novel approach to the task of automatic music genre classification which is based on multiple feature vectors and ensemble of classifiers. Multiple feature vectors are extracted from a single music piece. First, three 30-second music segments, one from the beginning, one from the middle and one from end part of a music piece are selected and feature vectors are extracted from each segment. Individual classifiers are trained to account for each feature vector extracted from each music segment. At the classification, the outputs provided by each individual classifier are combined through simple combination rules such as majority vote, max, sum and product rules, with the aim of improving music genre classification accuracy. Experiments carried out on a large dataset containing more than 3,000 music samples from ten different Latin music genres have shown that for the task of automatic music genre classification, the features extracted from the middle part of the music provide better results than using the segments from the beginning or end part of the music. Furthermore, the proposed ensemble approach, which combines the multiple feature vectors, provides better accuracy than using single classifiers and any individual music segment.},
}
@article{SilvaCoRR-15,
       author = {Gabriel Silva},
        title = {Geometric constraints on the dynamics of networks},
      journal = {CoRR},
       volume = {arXiv:1510.08729},
         year = {2015},
     abstract = {In any physically constructible network there is a structure-function relationship between the geometry of the network and its dynamics. The network's geometry, i.e. its physical structure, constrains and bounds signaling and the flow of information through the network. In this paper we explore how the physical geometry of a network constrains and ultimately determines its dynamics. We construct a formal theoretical framework of the relationship between network structure and function and how information flows through a network. One interesting result of this work is that we show how a strictly local process at the scale of individual node pairs directly affects the behavior of the system at the whole network scale. Individual nodes responding to information from the upstream nodes it is connected to produce the observable emergent dynamics of the entire network at a global scale, independent of and without any knowledge of what all the other nodes in the network may be doing. This analysis does not rely on statistical metrics that characterize a network at the whole network scale, but rather, makes use of the nature by which the structure and geometry of physically constructible networks necessarily constrain information flow given the internal state dynamics of individual nodes. This is a fundamental yet intuitive result. This approach can also be used for the systematic design of networks that need to meet specific design criteria and behaviors. We then provide empirical evidence that at least some important examples of both naturally occurring and engineered networks are capable of approaching a state of optimal dynamical efficiency, one of the key results of the theory. While we progressively build up to what this formally means, informally it means that these networks have evolved or have been designed to optimize how they are able to handle the processing of information by matching the dynamical requirements of individual nodes to the flow of information (i.e. signals) between nodes in the network. In different ways, we use the theory to investigate properties of pyramidal neurons in biological neural networks in the visual cortex, the prevalence of the small world network topology, and the internet router network.}
}
@article{SilvaNATURE-06,
       author = {Silva, Gabriel A.},
        title = {Neuroscience nanotechnology: progress, opportunities and challenges},
      journal = {Nature Reviews Neuroscience},
         year = {2006},
       volume = {7},
        issue = {1},
        pages = {65-74},
     abstract = {Nanotechnologies exploit materials and devices with a functional organization that has been engineered at the nanometre scale. The application of nanotechnology in cell biology and physiology enables targeted interactions at a fundamental molecular level. In neuroscience, this entails specific interactions with neurons and glial cells. Examples of current research include technologies that are designed to better interact with neural cells, advanced molecular imaging technologies, materials and hybrid molecules used in neural regeneration, neuroprotection, and targeted delivery of drugs and small molecules across the blood-brain barrier.},
}
@article{SilvaetalSCIENCE-04,
       author = {Silva, G. A. and Czeisler, C. and Niece, K. L. and Beniash, E. and Harrington, D. A. and Kessler, J. A. and Stupp, S. I.},
        title = {{S}elective differentiation of neural progenitor cells by high-epitope density nanofibers},
      journal = {Science},
         year = {2004},
       volume = {303},
       number = {5662},
        pages = {1352-1355},
     abstract = {Neural progenitor cells were encapsulated in vitro within a three-dimensional network of nanofibers formed by self-assembly of peptide amphiphile molecules. The self-assembly is triggered by mixing cell suspensions in media with dilute aqueous solutions of the molecules, and cells survive the growth of the nanofibers around them. These nanofibers were designed to present to cells the neurite-promoting laminin epitope IKVAV at nearly van der Waals density. Relative to laminin or soluble peptide, the artificial nanofiber scaffold induced very rapid differentiation of cells into neurons, while discouraging the development of astrocytes. This rapid selective differentiation is linked to the amplification of bioactive epitope presentation to cells by the nanofibers.}
}
@misc{SilverDQN-15,
        title = {Advanced Topics in Reinforcement Learning},
       author = {David Silver},
         year = {2015},
 howpublished = {Course Notes},
          url = {http://www0.cs.ucl.ac.uk/staff/d.silver/web/Teaching.html},
     abstract = {Course on reinforcement learning taught at Cambridge University run by David Silver and including a useful tutorial on Deep Q Learning.}
}
@article{SilveretalCoRR-17,
       author = {David Silver and Hado van Hasselt and Matteo Hessel and Tom Schaul and Arthur Guez and Tim Harley and Gabriel Dulac{-}Arnold and David P. Reichert and Neil C. Rabinowitz and Andr{\'{e}} Barreto and Thomas Degris},
        title = {The Predictron: End-To-End Learning and Planning},
      journal = {CoRR},
       volume = {arXiv:1612.08810},
      comment = {ICML 2017 preprint},
         year = {2017},
     abstract = {One of the key challenges of artificial intelligence is to learn models that are effective in the context of planning. In this document we introduce the predictron architecture. The predictron consists of a fully abstract model, represented by a Markov reward process, that can be rolled forward multiple "imagined" planning steps. Each forward pass of the predictron accumulates internal rewards and values over multiple planning depths. The predictron is trained end-to-end so as to make these accumulated values accurately approximate the true value function. We applied the predictron to procedurally generated random mazes and a simulator for the game of pool. The predictron yielded significantly more accurate predictions than conventional deep neural network architectures.}
}
@article{SilveretalICML-17,
       author = {David Silver and Hado van Hasselt and Matteo Hessel and Tom Schaul and Arthur Guez and Tim Harley and Gabriel Dulac{-}Arnold and David P. Reichert and Neil C. Rabinowitz and Andr{\'{e}} Barreto and Thomas Degris},
        title = {The Predictron: End-To-End Learning and Planning},
      journal = {Proceedings of the 34th International Conference on Machine Learning},
         year = {2017},
}
@article{SilveretalNATURE-16,
       author = {Silver, David and Huang, Aja and Maddison, Chris J. and Guez, Arthur and Sifre, Laurent and van den Driessche, George and Schrittwieser, Julian and Antonoglou, Ioannis and Panneershelvam, Veda and Lanctot, Marc and Dieleman, Sander and Grewe, Dominik and Nham, John and Kalchbrenner, Nal and Sutskever, Ilya and Lillicrap, Timothy and Leach, Madeleine and Kavukcuoglu, Koray and Graepel, Thore and Hassabis, Demis},
        title = {Mastering the game of Go with deep neural networks and tree search},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = {529},
        issue = {7587},
         year = {2016},
        pages = {484-489},
     abstract = {The game of Go has long been viewed as the most challenging of classic games for artificial intelligence owing to its enormous search space and the difficulty of evaluating board positions and moves. Here we introduce a new approach to computer Go that uses "value networks" to evaluate board positions and "policy networks" to select moves. These deep neural networks are trained by a novel combination of supervised learning from human expert games, and reinforcement learning from games of self-play. Without any lookahead search, the neural networks play Go at the level of state-of-the-art Monte Carlo tree search programs that simulate thousands of random games of self-play. We also introduce a new search algorithm that combines Monte Carlo simulation with value and policy networks. Using this search algorithm, our program AlphaGo achieved a 99.8\% winning rate against other Go programs, and defeated the human European Go champion by 5 games to 0. This is the first time that a computer program has defeated a human professional player in the full-sized game of Go, a feat previously thought to be at least a decade away.},
}
@article{SilveretalNEUROPHYSIOLOGY-05,
    author = {Michael Silver and David Ress and David Heeger},
     title = {Topographic Maps of Visual Spatial Attention in Human Parietal Cortex},
   journal = {Journal of Neurophysiology}, 
    volume = 94,
     pages = {1358-1371},
      year = 2005,
}
@article{SilvermanCEO-09,
     title = {High-resolution ultrasound imaging of the eye --- a review},
    author = {Silverman, R.H.},
   journal = {Clinical Experimental Ophthalmology},
    volume = 37,
     issue = 1,
     pages = {54-67},
      year = 2009,
  abstract = {This report summarizes the physics, technology and clinical application of ultrasound biomicroscopy (UBM) of the eye, in which frequencies of 35 MHz and above provide over a threefold improvement in resolution compared with conventional ophthalmic ultrasound systems. UBM allows imaging of anatomy and pathology involving the anterior segment, including regions obscured by overlying optically opaque anatomic or pathologic structures. UBM provides diagnostically significant information in conditions such as glaucoma, cysts and neoplasms, trauma and foreign bodies. UBM also can provide crucial biometric information regarding anterior segment structures, including the cornea and its constituent layers and the anterior and posterior chambers. Although UBM has now been in use for over 15 years, new technologies, including transducer arrays, pulse encoding and combination of ultrasound with light, offer the potential for significant advances in high-resolution diagnostic imaging of the eye.}
}
@incollection{SimardLecunandDenkerNIPS-93,
    author = {Simard, P. and LeCun, Y. and Denker J.},
     title = {Efficient pattern recognition using a new transformation distance},
 Booktitle = {Advances in Neural Information Processing Systems},
    volume = 5,
 publisher = {Morgan Kaufmann},
    editor = {Hanson, S. and Cowan, J. and Giles, L.},
      year = 1993,
}
@techreport{SimmonsTR-83,
       author = {Simmons, Reid G.},
        title = {Representing and Reasoning About Change in Geologic Interpretation},
        issue = {AI-TR-749},
  institution = {{MIT} {AI} Laboratory},
         year = 1983,
}
@inproceedings{SimmonsandDavisIJCAI-87,
       author = {Simmons, Reid and Davis, Randall},
        title = {Generate, Test and Debug: Combining Associational Rules and Causal Models},
    booktitle = {Proceedings {IJCAI} 10},
     location = {Milan, Italy},
 organization = {IJCAII},
         year = 1987,
        pages = {1071-1078},
}
@book{Simon2015synthetic,
        title = {Synthetic Biology - A Primer (Revised Edition)},
       author = {Freemont Paul Simon and Kitney Richard},
         year = {2015},
    publisher = {World Scientific Publishing Company}
}
@article{Simon55,
       author = {Simon, Herbert A.},
        title = {A behavioral model of rational choice},
      journal = {Quarterly Journal of Economics},
         year = 1955,
       volume = 69,
        pages = {99-118},
}
@book{Simon81,
        title = {The Sciences of the Artificial},
       author = {Simon, Herbert A.},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1981,
}
@article{SimonAIJ-75,
       author = {Simon, Herbert A. and Kadane, Joseph B.},
        title = {Optimal Problem-Solving Search: All-or-None Solutions},
      journal = AIJ,
       volume = 6,
         year = 1975,
        pages = {235-247},
}
@inproceedings{Simoncelli-99,
       author = {E. P. Simoncelli},
        title = {Bayesian denoising of visual images in the wavelet domain},
    booktitle = {Bayesian Inference in Wavelet Based Models},
       editor = {P. M\"{u}ller and B. Vidakovic},
 series       = {Lecture Notes in Statistics},
       volume = 141,
 chapter      = 18,
        pages = {291-308},
    publisher = {Springer-Verlag},
         year = 1999,
}
@article{SimoncelliandHeegerVR-98,
     title = {A Model of Neuronal Responses in Visual Area {MT}},
    author = {Eero Simoncelli and David Heeger},
   journal = {Vision Research},
    volume = 38,
     issue = 5,
      year = 1998,
     pages = {743-761},
}
@book{Sinclair93,
       author = {Alistair Sinclair},
        title = {Algorithms for Random Generation and Counting: a {M}arkov Chain Approach},
    publisher = {Birkhauser},
      address = {Boston, Massachusetts},
         year = 1993,
}
@article{SingerNC-97,
       author = {Yoram Singer},
        title = {Adaptive Mixture of Probabilistic Transducers},
      journal = {Neural Computation},
       volume = 9,
        issue = 8,
        pages = {1711-1733},
         year = 1997,
}
@inproceedings{SingerNIPS-96,
       author = {Yoram Singer},
        title = {Adaptive Mixture of Probabilistic Transducers},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 8,
    publisher = {{MIT} Press},
       editor = {David S. Touretzky and Michael C. Mozer and Michael E. Hasselmo},
        pages = {381-387},
         year = 1996,
}
@book{Singh2012,
        title = {Fermat's Last Theorem},
       author = {Singh, Simon},
         year = {2012},
    publisher = {HarperCollins Publishers}
}
@article{SinghJMLR-92,
       author = {Singh, Satinder Pal},
        title = {Transfer of learning by composing solutions of elemental sequential tasks},
      journal = {Machine Learning},
         year = {1992},
       volume = {8},
       number = {3},
        pages = {323-339},
      comment = {Uses a gating bit vector to augment the state space vector to facilitate decomposition and transfer learning},
     abstract = {Although building sophisticated learning agents that operate in complex environments will require learning to perform multiple tasks, most applications of reinforcement learning have focused on single tasks. In this paper I consider a class of sequential decision tasks (SDTs), called composite sequential decision tasks, formed by temporally concatenating a number of elemental sequential decision tasks. Elemental SDTs cannot be decomposed into simpler SDTs. I consider a learning agent that has to learn to solve a set of elemental and composite SDTs. I assume that the structure of the composite tasks is unknown to the learning agent. The straightforward application of reinforcement learning to multiple tasks requires learning the tasks separately, which can waste computational resources, both memory and time. I present a new learning algorithm and a modular architecture that learns the decomposition of composite SDTs, and achieves transfer of learning by sharing the solutions of elemental SDTs across multiple composite SDTs. The solution of a composite SDT is constructed by computationally inexpensive modifications of the solutions of its constituent elemental SDTs. I provide a proof of one aspect of the learning algorithm.},
}
@inproceedings{SinghML-94,
       author = {Singh, Satinder and Jaakkola, Tommi and Jordan, Michael},
        title = {Learning Without State-Estimation in Partially Observable {M}arkovian Decision Processes},
    booktitle = {Proceedings of the Eleventh Machine Learning Conference},
         year = 1994,
}
@misc{SinghPOINTER-NETWORKS-17,
        title = {Understanding Pointer Networks - Medium, May 24, 2017},
       author = {Abhishek Singh},
 howpublished = {\url{https://towardsdatascience.com/understanding-pointer-networks-81fbbc1ddbc8}},
         year = {2017},
}
@incollection{SinghandCohnNIPS-98,
       author = {Singh, Satinder and Cohn, David},
        title = {How to Dynamically Merge {M}arkov Decision Processes},
    booktitle = {Advances in Neural Information Processing},
       volume = 11,
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1998,
}
@inproceedings{SinghandKohliSNAPL-17,
       author = {Singh, Rishabh and Kohli, Pushmeet},
        title = {{AP:Artificial Programming}},
    booktitle = {Summit on Advances in Programming Languages 2017},
         year = {2017},
     abstract = {The ability to automatically discover a program consistent with a given user intent (specification) is the holy grail of Computer Science. While significant progress has been made on the so-called problem of Program Synthesis, a number of challenges remain; particularly for the case of synthesizing richer and larger programs. This is in large part due to the difficulty of search over the space of programs. In this paper, we argue that the above-mentioned challenge can be tackled by learning synthesizers automatically from a large amount of training data. We present a first step in this direction by describing our novel synthesis approach based on two neural architectures for tackling the two key challenges of Learning to understand partial input-output specifications and Learning to search programs. The first neural architecture called the Spec Encoder computes a continuous representation of the specification, whereas the second neural architecture called the Program Generator incrementally constructs programs in a hypothesis space that is conditioned by the specification vector. The key idea of the approach is to train these architectures using a large set of (spec,P) pairs, where P denotes a program sampled from the DSL L and spec denotes the corresponding specification satisfied by P. We demonstrate the effectiveness of our approach on two preliminary instantiations. The first instantiation, called Neural FlashFill, corresponds to the domain of string manipulation programs similar to that of FlashFill. The second domain considers string transformation programs consisting of composition of API functions. We show that a neural system is able to perform quite well in learning a large majority of programs from few input-output examples. We believe this new approach will not only dramatically expand the applicability and effectiveness of Program Synthesis, but also would lead to the coming together of the Program Synthesis and Machine Learning research disciplines.},
}
@inproceedings{SinghetalCS-09,
       author = {Satinder Singh and Richard L. Lewis and Andrew G. Barto},
        title = {Where Do Rewards Come From?},
    booktitle = {Proceedings of the 31st Annual Conference of the Cognitive Science Society},
       editor = {Taatgen, N.A. and van Rijn, H.},
         year = {2009}
}
@article{SinghetalJoV-08,
       author = {Singh, Gurjeet and Memoli, Facundo and Ishkhanov, Tigran and Sapiro, Guillermo and Carlsson, Gunnar and Ringach, Dario L.},
        title = {Topological analysis of population activity in visual cortex},
      journal = {Journal of Vision},
       volume = {8},
       number = {8},
        pages = {11},
         year = {2008},
     abstract = {Information in the cortex is thought to be represented by the joint activity of neurons. Here we describe how fundamental questions about neural representation can be cast in terms of the topological structure of population activity. A new method, based on the concept of persistent homology, is introduced and applied to the study of population activity in primary visual cortex (V1). We found that the topological structure of activity patterns when the cortex is spontaneously active is similar to those evoked by natural image stimulation and consistent with the topology of a two sphere. We discuss how this structure could emerge from the functional organization of orientation and spatial frequency maps and their mutual relationship. Our findings extend prior results on the relationship between spontaneous and evoked activity in V1 and illustrates how computational topology can help tackle elementary questions about the representation of information in the nervous system.},
}
@inproceedings{SinghetalNIPS-99,
       author = {Satinder P. Singh and Michael J. Kearns and Diane J. Litman and Marilyn A. Walker},
        title = {Reinforcement Learning for Spoken Dialogue Systems},
    booktitle = {Advances in Neural Information Processing Systems 12},
         year = 1999,
        pages = {956-962},
}
@article{SinhaetalIEEE-06,
        title = {Face Recognition by Humans: Nineteen Results All Computer Vision Researchers Should Know About},
       author = {Pawan Sinha and Benjamin Balas and Yuri Ostrovsky and Richard Russell},
      journal = {Proceedings of the {IEEE}},
       volume = 94,
        issue = 11,
        pages = {1948-1962},
         year = 2006,
}
@article{SinhaetalPNAS-13,
       author = {Sinha, Supriyo and Liang, Liang and Ho, Eric T. W. and Urbanek, Karel E. and Luo, Liqun and Baer, Thomas M. and Schnitzer, Mark J.},
        title = {High-speed laser microsurgery of alert fruit flies for fluorescence imaging of neural activity},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {110},
       number = {46},
        pages = {18374-18379},
         year = {2013},
     abstract = {Intravital microscopy is a key means of monitoring cellular function in live organisms, but surgical preparation of a live animal for microscopy often is time-consuming, requires considerable skill, and limits experimental throughput. Here we introduce a spatially precise (< 1-µm edge precision), high-speed (<1 s), largely automated, and economical protocol for microsurgical preparation of live animals for optical imaging. Using a 193-nm pulsed excimer laser and the fruit fly as a model, we created observation windows (12- to 350-µm diameters) in the exoskeleton. Through these windows we used two-photon microscopy to image odor-evoked Ca2+ signaling in projection neuron dendrites of the antennal lobe and Kenyon cells of the mushroom body. The impact of a laser-cut window on fly health appears to be substantially less than that of conventional manual dissection, for our imaging durations of up to 18 h were \~5-20 times longer than prior in vivo microscopy studies of hand-dissected flies. This improvement will facilitate studies of numerous questions in neuroscience, such as those regarding neuronal plasticity or learning and memory. As a control, we used phototaxis as an exemplary complex behavior in flies and found that laser microsurgery is sufficiently gentle to leave it intact. To demonstrate that our techniques are applicable to other species, we created microsurgical openings in nematodes, ants, and the mouse cranium. In conjunction with emerging robotic methods for handling and mounting flies or other small organisms, our rapid, precisely controllable, and highly repeatable microsurgical techniques should enable automated, high-throughput preparation of live animals for optical experimentation.},
}
@techreport{SinhaetalTR-06,
     author = {Sudipta N. Sinha and Jan-michael Frahm and Marc Pollefeys and Yakup Genc},
      title = {{GPU}-based Video Feature Tracking and Matching},
institution = {In {EDGE 2006} Workshop on Edge Computing Using New Commodity Architectures},
       year = 2006
}
@article{SioetalMC-13,
       author = {Sio, Ut Na and Monaghan, Padraic and Ormerod, Tom},
        title = {Sleep on it, but only if it is difficult: Effects of sleep on problem solving},
      journal = {Memory {\&} Cognition},
       volume = {41},
       number = {2},
         year = {2013},
        pages = {159-166},
     abstract = {Previous research has shown that performance on problem solving improves over a period of sleep, as compared with wakefulness. However, these studies have not determined whether sleep is beneficial for problem solving or whether sleep merely mitigates against interference due to an interruption to solution attempts. Sleep-dependent improvements have been described in terms of spreading activation, which raises the prediction that an effect of sleep should be greater for problems requiring a broader solution search. We presented participants with a set of remote-associate tasks that varied in difficulty as a function of the strength of the stimuli-answer associations. After a period of sleep, wake, or no delay, participants reattempted previously unsolved problems. The sleep group solved a greater number of difficult problems than did the other groups, but no difference was found for easy problems. We conclude that sleep facilitates problem solving, most likely via spreading activation, but this has its primary effect for harder problems.},
}
@article {SirovichandKirbyJOSA-87,
       author = {L. Sirovich and M. Kirby},
        title = {Low Dimensional Procedure for the Characterization of Human Faces},
      journal = {Journal of the Optical Society of America},
        issue = 3,
        pages = {519-524},
       volume = 4,
         year = 1987,
}
@article{SiskindAIR-95,
       author = {Siskind, J.M.},
        title = {Grounding Language in Perception},
      journal = {Artificial Intelligence Review},
       volume = 8,
        pages = 1995,
         year = {371-391},
}
@article{SiskindCOGNITION-96,
       author = {Siskind, J.M.},
        title = {A Computational Study of Cross-Situational Techniques for Learning Word-to-Meaning Mappings},
      journal = {Cognition},
       volume = 61,
        issue = {1-2},
        pages = 1996,
         year = {39-91},
}
@article{SiuandMurphyEye_and_Brain-18,
       author = {Siu, Caitlin R. and Murphy, Kathryn M.},
        title = {The development of human visual cortex and clinical implications},
      journal = {Eye and brain},
         year = {2018},
    publisher = {Dove Medical Press},
       volume = {10},
        pages = {25-36},
     abstract = {The primary visual cortex (V1) is the first cortical area that processes visual information. Normal development of V1 depends on binocular vision during the critical period, and age-related losses of vision are linked with neurobiological changes in V1. Animal studies have provided important details about the neurobiological mechanisms in V1 that support normal vision or are changed by visual diseases. There is very little information, however, about those neurobiological mechanisms in human V1. That lack of information has hampered the translation of biologically inspired treatments from preclinical models to effective clinical treatments. We have studied human V1 to characterize the expression of neurobiological mechanisms that regulate visual perception and neuroplasticity. We have identified five stages of development for human V1 that start in infancy and continue across the life span. Here, we describe these stages, compare them with visual and anatomical milestones, and discuss implications for translating treatments for visual disorders that depend on neuroplasticity of V1 function.},
}
@inproceedings{SivicetalICCV-05,
     author = {Josef Sivic and Bryan Russell and Alexei A. Efros and Andrew Zisserman and Bill Freeman},
      title = {Discovering Objects and Their Location in Images},
  booktitle = {International Conference on Computer Vision (ICCV 2005)},
      month = {October},
       year = 2005,
}
@inproceedings{SivicetalIEEE-08,
    author = {Josef Sivic and Biliana Kaneva and Antonio Torralba and Shai Avidan and William T. Freeman},
     title = {Creating and Exploring a Large Photorealistic Virtual Space},
 booktitle = {First {IEEE} Workshop on Internet Vision},
      year = 2008
}
@article{SizemoreetalCoRR-16,
        title = {Closures and Cavities in the Human Connectome},
       author = {Ann Sizemore and Chad Giusti and Richard F. Betzel and Danielle S. Bassett},
      journal = {CoRR},
       volume = {arXiv:1608.03520},
         year = {2016},
     abstract = {Encoding brain regions and their connections as a network of nodes and edges captures many of the possible paths along which information can be transmitted as humans process and perform complex behaviors. Because cognitive processes involve large and distributed networks of brain areas, examinations of multi-node routes within larger connection patterns can offer fundamental insights into the complexities of brain function. Here, we investigate both densely connected groups of nodes that could perform local computations as well as larger patterns of interactions that would allow for parallel processing. Finding such structures necessitates we move from considering exclusively pairwise interactions to capturing higher order relations, considerations naturally expressed in the language of algebraic topology. In this parlance, clusters of all-to-all connected sets of brain regions that may readily share information to perform a coordinated function are called \emph{cliques}. We detect cliques in the average structural connectomes of 8 healthy adults scanned in triplicate and discover the presence of more high-dimensional cliques than expected in null networks constructed via wiring minimization. This provides architecture through which brain network can perform rapid, local processing. Complementary to theis study of locally dense structures, we employ a tool called \emph{persistent homology} to locate \emph{cycles}, topological cavities of different dimensions, around which information may flow in either diverging or converging patterns. These cycles exist consistently across subjects, differ from those observed in null model networks, and -- importantly -- link regions of early and late evolutionary origin in long loops, underscoring their unique role in controlling brain function.},
}
@book{Skinner71,
       author = {B. F. Skinner},
        title = {Beyond Freedom and Dignity},
    publisher = {Knopf},
      address = {New York, NY},
         year = 1971,
}
@book{SlateandAtkin77,
       author = {Slate, D. J. and Atkin, L. R.},
        title = {Chess 4.5: The Northwestern University Chess Program},
    publisher = {Springer-Verlag},
      address = {New York},
         year = 1977,
}
@article{SmaersetalCURRENT-BIOLOGY-17,
       author = {Jeroen B. Smaers and Aida G{\'{o}}mez-Robles and Ashley N. Parks and Chet C. Sherwood},
        title = {Exceptional Evolutionary Expansion of Prefrontal Cortex in Great Apes and Humans},
      journal = {Current Biology},
       volume = {27},
       number = {5},
         year = {2017},
        pages = {714-720},
     abstract = {One of the enduring questions that has driven neuroscientific enquiry in the last century has been the nature of differences in the prefrontal cortex of humans versus other animals [1]. The prefrontal cortex has drawn particular interest due to its role in a range of evolutionarily specialized cognitive capacities such as language [2], imagination [3], and complex decision making [4]. Both cytoarchitectonic [5] and comparative neuroimaging [6] studies have converged on the conclusion that the proportion of prefrontal cortex in the human brain is greatly increased relative to that of other primates. However, considering the tremendous overall expansion of the neocortex in human evolution, it has proven difficult to ascertain whether this extent of prefrontal enlargement follows general allometric growth patterns, or whether it is exceptional [1]. Species' adherence to a common allometric relationship suggests conservation through phenotypic integration, while species' deviations point toward the occurrence of shifts in genetic and/or developmental mechanisms. Here we investigate prefrontal cortex scaling across anthropoid primates and find that great ape and human prefrontal cortex expansion are non-allometrically derived features of cortical organization. This result aligns with evidence for a developmental heterochronic shift in human prefrontal growth [7, 8], suggesting an association between neurodevelopmental changes and cortical organization on a macroevolutionary scale. The evolutionary origin of non-allometric prefrontal enlargement is estimated to lie at the root of great apes (19-15 mya), indicating that selection for changes in executive cognitive functions characterized both great ape and human cortical organization.}
}
@article{SmallwoodandSondikOR-73,
       author = {Smallwood, Richard D. and Sondik, Edward J.},
        title = {The Optimal Control of Partially Observable {M}arkov Processes over a Finite Horizon},
      journal = {Operations Research},
       volume = 21,
         year = 1973,
        pages = {1071-1088},
}
@article{SmaouiandAl-EnezJCAM-04,
       author = {Nejib Smaoui and Suad Al-Enezi},
        title = {Modelling the dynamics of nonlinear partial differential equations using neural networks},
      journal = {Journal of Computational and Applied Mathematics},
       volume = {170},
       number = {1},
         year = {2004},
        pages = {27-32},
     abstract = {The dynamics of two nonlinear partial differential equations (PDEs) known as the Kuramoto-Sivashinsky (K-S) equation and the two-dimensional Navier-Stokes (N-S) equations are analyzed using Karhunen-Lo\'{e}ve (K-L) decomposition and artificial neural networks (ANN). For the K-S equation, numerical simulations using a pseudospectral Galerkin method is presented at a bifurcation parameter '=17.75, where a dynamical behavior represented by a heteroclinic connection is obtained. We apply K-L decomposition on the numerical simulation data with the task of reducing the data into a set of data coefficients. Then we use \{ANN\} to model, and predict the data coefficients at a future time. It is found that training the neural networks with only the first data coefficient is enough to capture the underlying dynamics, and to predict for the other remaining data coefficients. As for the two-dimensional N-S equation, a quasiperiodic behavior represented in phase space by a torus is analyzed at Re=14.0. Applying the symmetry observed in the two-dimensional N-S equations on the quasiperiodic behavior, eight different tori were obtained. We show that by exploiting the symmetries of the equation and using K-L decomposition in conjunction with neural networks, a smart neural model is obtained.}
}
@book{Smith1759,
        title = {The Theory of Moral Sentiments},
       author = {Smith, Adam},
         year = {1759},
    publisher = {McMaster University Archive for the History of Economic Thought},
}
@incollection{Smith2008episodic,
        title = {The hippocampus, context processing and episodic memory},
       author = {David M. Smith},
    booktitle = {Handbook of Episodic Memory},
       editor = {Ekrem Dere and Alexander Easton and Lynn Nadel and Joseph P. Huston},
       series = {Handbook of Behavioral Neuroscience},
    publisher = {Elsevier},
       volume = {18},
         year = {2008}
}
@book{Smith82,
       author = {John Maynard Smith},
        title = {Evolution and the Theory of Games},
    publisher = {Cambridge University Press},
      address = {Cambridge},
         year = 1982,
}
@book{Smith89,
       author = {John Maynard Smith},
        title = {Evolutionary Genetics},
    publisher = {Oxford University Press},
      address = {Oxford},
         year = 1989,
}
@inproceedings{SmithAAAI-88,
       author = {Ow, P. S. and Smith, S. F. and Thiriez, A.},
        title = {Reactive Plan Revision},
    booktitle = {Proceedings AAAI-88},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
         year = 1988,
}
@article{SmithNATURE-13,
     title = {Neuroscience: {M}ethod {Man} --- {K}arl {D}eisseroth is leaving his mark on brain science one technique at a time},
    author = {Kerri Smith},
   journal = {Nature},
      year = 2013,
}
@article{SmithSMC-81,
       author = {Smith, R. G. and Davis, R.},
        title = {Frameworks for Cooperation in Distributed Problem Solving},
      journal = {{IEEE} Transactions on Systems, Man, and Cybernetics},
       volume = 11,
        issue = 1,
         year = 1981,
}
@techreport{SmithTR-83,
       author = {Smith, Stephen F.},
        title = {Exploiting Temporal Knowledge to Organize Constraints},
        issue = {CMU-RI-TR-83-12},
  institution = {Carnegie-Mellon University Intelligent Systems Laboratory},
         year = 1983,
}
@techreport{SmithTR-88,
       author = {Smith, David E.},
        title = {A Decision-Theoretic Approach to the Control of Planning Search},
        issue = {LOGIC-87-11},
  institution = {Stanford Logic Group},
         year = 1988,
}
@techreport{SmithandArnoldIUI-09,
       author = {Dustin Smith and Kenneth Arnold},
        title = {Learning Hierarchical Plans by Reading Simple English Narratives},
        issue = {ACM International Conference on Intelligent User Interfaces (IUI-09), Commonsense Workshop},
  institution = {MIT},
         year = {2009},
     abstract = {We describe an approach for learning a rich plan representation from a parallel corpus of commonsense narratives. Each narrative is an ordered natural language description of the steps required to accomplish common domestic tasks, including "get the mail" and "make a bed", and there are tens to hundreds of differently written narratives for each task. With the goal of learning a single rich plan structure, we 1) convert each narrative from English statements into a sequence of logical predicates, 2) find a global alignment for the sequences, and 3) use the sequences to construct a single underlying plan representation that can be used in language understanding problems. Doing this requires being able to distinguish different ways to accomplish the same goal from missing information, and recognize and compactly represent recurring plan sub-sequences. We describe a simple algorithm that recursively finds graph cycles by applying rules to merge nodes to learn a sequential, parameterized composition (part-of) and abstraction (is-a) plan hierarchy. We hope that these plan representations will help us learn procedural knowledge from increasingly more sophisticated text, where the sub-goals for various actions are not stated.}
}
@article{SmithandCheesemanIJRR-86,
       author = {Smith, Randall and Peter Cheeseman},
        title = {On the Representation and Estimation of Spatial Uncertainty},
      journal = {The International Journal of Robotics Research},
       volume = 5,
         year = 1986,
        pages = {56-68},
}
@article{SmithandGeneserethAIJ-85,
       author = {Smith, David E. and Michael J. Genesereth},
        title = {Ordering Conjunctive Queries},
      journal = AIJ,
       volume = 26,
         year = 1985,
        pages = {171-216},
}
@incollection{SmithandLewickiNIPS-04,
       author = {Evan C. {Smith} and Michael S. {Lewicki}},
        title = {Learning Efficient Auditory Codes Using Spikes Predicts Cochlear Filters},
    booktitle = {Advances in Neural Information Processing Systems 17},
       editor = {Lawrence K. Saul and Yair Weiss and {L\'{e}on} Bottou},
    publisher = {MIT Press},
      address = {Cambridge, MA},
        pages = {1289-1296},
         year = 2005,
}
@article{SmithandWatermanJMB-81,
       author = {Smith, T. F. and Waterman, M. S.},
        title = {Identification of common molecular subsequences},
      journal = {Journal of Molecular Biology},
       number = {1},
         year = {1981},
       volume = {147},
        pages = {195-197},
}
@unpublished{SmithandWilliamsonSSS-95,
        title = {Representation and Evaluation of Plans with Loops},
       author = {Smith, David E. and Williamson, Mike Theories of Action},
         year = 1995,
}
@misc{Smithetal85,
       author = {Smith, Randall and Mathew Self and Peter Cheeseman},
        title = {Estimating Uncertain Spatial Relationships in Robotics},
    booktitle = {publication unknown},
         year = 1985,
}
@article{SmithetalCONTROL-CYBERNETICS-02,
       author = {Smith, K. and Palaniswami, M. and Krishnamoorthy, M.},
        title = {Neural Techniques for Combinatorial Optimization with Applications},
      journal = {Transactions on Neural Networks},
       volume = {9},
       number = {6},
         year = {1998},
        pages = {1301-1318},
    publisher = {IEEE Press},
     abstract = {After more than a decade of research, there now exist several neural-network techniques for solving NP-hard combinatorial optimization problems. Hopfield networks and self-organizing maps are the two main categories into which most of the approaches can be divided. Criticism of these approaches includes the tendency of the Hopfield network to produce infeasible solutions, and the lack of generalizability of the self-organizing approaches (being only applicable to Euclidean problems). The paper proposes two new techniques which have overcome these pitfalls: a Hopfield network which enables feasibility of the solutions to be ensured and improved solution quality through escape from local minima, and a self-organizing neural network which generalizes to solve a broad class of combinatorial optimization problems. Two sample practical optimization problems from Australian industry are then used to test the performances of the neural techniques against more traditional heuristic solutions}
}
@article{SmithetalCORTEX-01,
       author = {Smith, A.T. and Singh, K.D. and Williams, A.L. and Greenlee, M.W.},
        title = {Estimating Receptive Field Size from {fMRI} Data in Human Striate and Extrastriate Visual Cortex},
      journal = {Cerebral Cortex},
       volume = 11,
        issue = 12,
        pages = {1182-1190},
         year = 2001,
}
@article{SmithetalJEP-17,
        title = {Self-deception facilitates interpersonal persuasion},
      journal = {Journal of Economic Psychology},
         year = {2017},
       author = {Megan K. Smith and Robert Trivers and William von Hippel},
     abstract = {Self-deception is both commonplace and costly, which raises the question of what purpose it might serve. According to the dominant explanation in psychology and economics, self-deception is an intrapersonal process that fortifies and protects the self from threatening information. An alternative possibility is that self-deception evolved as an interpersonal strategy to persuade others. To investigate interpersonal aspects of self-deception, we gave people a persuasive task and measured their information processing biases and their persuasiveness. Results revealed that people who were financially motivated to persuade another person in a particular direction demonstrated a self-deceptive information processing bias consistent with their persuasive goals. This information processing bias led people to convince themselves of the veracity of their persuasive goal, and subsequently to be more persuasive to others. These findings suggest that self-deception has interpersonal benefits that offset its costs.}
}
@article{SmithetalNATURE-COMMUNICATIONS-17,
       author = {Smith, Daniel and Schlaepfer, Philip and Major, Katie and Dyble, Mark and Page, Abigail E. and Thompson, James and Chaudhary, Nikhil and Salali, Gul Deniz and Mace, Ruth and Astete, Leonora and Ngales, Marilyn and Vinicius, Lucio and Migliano, Andrea Bamberg},
        title = {Cooperation and the evolution of hunter-gatherer storytelling},
      journal = {Nature Communications},
         year = {2017},
       volume = {8},
        issue = {1},
        pages = {1853},
     abstract = "Storytelling is a human universal. From gathering around the camp-fire telling tales of ancestors to watching the latest television box-set, humans are inveterate producers and consumers of stories. Despite its ubiquity, little attention has been given to understanding the function and evolution of storytelling. Here we explore the impact of storytelling on hunter-gatherer cooperative behaviour and the individual-level fitness benefits to being a skilled storyteller. Stories told by the Agta, a Filipino hunter-gatherer population, convey messages relevant to coordinating behaviour in a foraging ecology, such as cooperation, sex equality and egalitarianism. These themes are present in narratives from other foraging societies. We also show that the presence of good storytellers is associated with increased cooperation. In return, skilled storytellers are preferred social partners and have greater reproductive success, providing a pathway by which group-beneficial behaviours, such as storytelling, can evolve via individual-level selection. We conclude that one of the adaptive functions of storytelling among hunter gatherers may be to organise cooperation.",
}
@article{SmolenskyAIJ-90,
       author = {Smolensky, P.},
        title = {Tensor Product Variable Binding and the Representation of Symbolic Structures in Connectionist Systems},
      journal = {Artificial Intelligence},
       volume = 46,
        issue = {1-2},
         year = 1990,
        pages = {159--216},
    publisher = {Elsevier Science Publishers Ltd.},
}
@book{SmolenskyandLegendre2011,
        title = {The Harmonic Mind: From Neural Computation to Optimality-theoretic Grammar. Linguistic and philosophical implications},
       author = {Smolensky, P. and Legendre, G.},
         year = {2011},
    publisher = {MIT Press}
}
@techreport{SmythHeckermanandJordanTR-96,
        title = {Probabilistic Independence Networks for Hidden {M}arkov Probability Models},
       author = {Smyth, Padhraic and Heckerman, David and Jordan, Michael},
  institution = {{MIT} Department of Brain and Cognitive Sciences},
        issue = 1565,
         year = 1996,
}
@misc{SnaiderPhD-12,
        title = {Integer Sparse Distributed Memory and Modular Composite Representation},
       author = {Xavier Snaider},
         year = 2012,
 howpublished = {Technical, Report University of Memphis, August 2012},
}
@article{SnaiderandFranklinMCR-14,
       author = {Snaider, Javier and Franklin, Stan},
        title = {Modular Composite Representation},
      journal = {Cognitive Computation},
    publisher = {Springer US},
         year = 2014,
        pages = {1-18},
     abstract = {High-dimensional vector spaces have noteworthy properties that make them attractive for representation models. A reduced description model is a mechanism for encoding complex structures as single high-dimensional vectors. Moreover, these vectors can be used to directly process complex operations such as analogies, inferences, and structural comparisons. Also, it is possible to reconstruct the whole structure from the reduced description vector. Here, we introduce the modular composite representation (MCR), a new reduced description model that employs long integer vectors. We also describe several experiments with them, and give a theoretical analysis of the distance distribution in this vector space, and of properties of this representation. Finally, we compare MCR with other two reduced description models: Spatter Code and holographic reduced representation.},
}
@book{SnedecorandCochran89,
       author = {George W. Snedecor and William G. Cochran},
        title = {Statistical Methods, Eighth Edition},
    publisher = {Iowa State University Press},
      address = {Ames, Iowa},
         year = 1989,
}
@article{SnideretalCOMPUTER-11,
       author = {Snider, G. and Amerson, R. and Carter, D. and Abdalla, H. and Qureshi, M.S. and Leveille, J. and Versace, M. and Ames, H. and Patrick, S. and Chandler, B. and Gorchetchnikov, A. and Mingolla, E.},
        title = {From Synapses to Circuitry: Using Memristive Memory to Explore the Electronic Brain},
      journal = {Computer},
       volume = 44,
        issue = 2,
         year = 2011,
        pages = {21-28},
     abstract = {In a synchronous digital platform for building large cognitive models, memristive nanodevices form dense, resistive memories that can be placed close to conventional processing circuitry. Through adaptive transformations, the devices can interact with the world in real time.},
}
@techreport{SnodgrassTR-85,
       author = {Snodgrass, Richard},
        title = {A Temporal Query Language},
        issue = {TR-85-013},
  institution = {University of North Carolina Department of Computer Science},
         year = 1985,
}
@inproceedings{SnoeketalCoRR-12,
       author = {Snoek, Jasper and Larochelle, Hugo and Adams, Ryan P.},
        title = {Practical Bayesian Optimization of Machine Learning Algorithms},
    booktitle = {Proceedings of the 25th International Conference on Neural Information Processing Systems - Volume 2},
    publisher = {Curran Associates Inc.},
      address = {USA},
         year = {2012},
        pages = {2951-2959},
     abstract = {The use of machine learning algorithms frequently involves careful tuning of learning parameters and model hyperparameters. Unfortunately, this tuning is often a "black art" requiring expert experience, rules of thumb, or sometimes brute-force search. There is therefore great appeal for automatic approaches that can optimize the performance of any given learning algorithm to the problem at hand. In this work, we consider this problem through the framework of Bayesian optimization, in which a learning algorithm's generalization performance is modeled as a sample from a Gaussian process (GP). We show that certain choices for the nature of the GP, such as the type of kernel and the treatment of its hyperparameters, can play a crucial role in obtaining a good optimizer that can achieve expertlevel performance. We describe new algorithms that take into account the variable cost (duration) of learning algorithm experiments and that can leverage the presence of multiple cores for parallel experimentation. We show that these proposed algorithms improve on previous automatic procedures and can reach or surpass human expert-level optimization for many algorithms including latent Dirichlet allocation, structured SVMs and convolutional neural networks.}
}
@inproceedings{SoboroffSIGIR02,
       author = {Ian Soboroff},
        title = {Does {WT10g} Look Like the {W}eb?},
    booktitle = {Proceedings of the 25th International Conference on Research and Development in Information Retrieval (SIGIR)},
        pages = {423-424},
         year = 2002,
}
@misc{SocherPhD-14,
        title = {Recursive Deep Learning for Natural Language Processing and Computer Vision},
       author = {Richard Socher},
         year = 2014,
 howpublished = {Ph.D. Thesis. Computer Science Department, Stanford University}
}
@inproceedings{SocherandManningDEEP-13,
       author = {Richard Socher and Christopher D. Manning},
        title = {Deep Learning for {NLP} (without Magic) Tutorial},
    booktitle = {Human Language Technologies: Conference of the North American Chapter of the Association of Computational Linguistics},
         year = 2013,
        pages = {1-3},
}
@inproceedings{SocheretalACL-13,
       author = {Richard Socher and John Bauer and Christopher D. Manning and Andrew Y. Ng},
        title = {Parsing with Compositional Vector Grammars},
    booktitle = {Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics},
    publisher = {Association for Computational Linguistics},
      address = {Stroudsburg, PA, USA},
         year = 2013,
     abstract = {Natural language parsing has typically been done with small sets of discrete cat- egories such as NP and VP, but this rep- resentation does not capture the full syntactic nor semantic richness of linguistic phrases, and attempts to improve on this by lexicalizing phrases or splitting cate- gories only partly address the problem at the cost of huge feature spaces and sparseness. Instead, we introduce a Compositional Vector Grammar (CVG), which combines PCFGs with a syntactically un- tied recursive neural network that learns syntactico-semantic, compositional vector representations. The CVG improves the PCFG of the Stanford Parser by 3.8\% to obtain an F1 score of 90.4\%. It is fast to train and implemented approximately as an efficient reranker it is about 20\% faster than the current Stanford factored parser. The CVG learns a soft notion of head words and improves performance on the types of ambiguities that require semantic information such as PP attachments.}
}
@article{SocheretalCoRR-13,
     title = {Zero-Shot Learning Through Cross-Modal Transfer},
    author = {Richard Socher and Milind Ganjoo and Hamsa Sridhar and Osbert Bastani and Christopher D. Manning and Andrew Y. Ng},
   journal = {CoRR},
    volume = {arXiv:1301.3666},
      year = 2013,
       URL = {http://arxiv.org/abs/1301.3666},
}
@incollection{SocheretalEMNLP-12,
       author = {Richard Socher and Brody Huval and Christopher D. Manning and Andrew Y. Ng},
        title = {Semantic Compositionality through Recursive Matrix-Vector Spaces},
    booktitle = {Proceedings of the 2012 Conference on Empirical Methods in Natural Language Processing},
    publisher = {Association for Computational Linguistics},
         year = 2012,
     abstract = {Single-word vector space models have been very successful at learning lexical information. However, they cannot capture the compositional meaning of longer phrases, preventing them from a deeper understanding of language. We introduce a recursive neural network (RNN) model that learns compositional vector representations for phrases and sentences of arbitrary syntactic type and length. Our model assigns a vector and a matrix to every node in a parse tree: the vector captures the inherent meaning of the constituent, while the matrix captures how it changes the meaning of neighboring words or phrases. This matrix-vector RNN can learn the meaning of operators in propositional logic and natural language. The model obtains state of the art performance on three different experiments: predicting fine-grained sentiment distributions of adverb-adjective pairs; classifying sentiment labels of movie reviews and classifying semantic relationships such as cause-effect or topic-message between nouns using the syntactic path between them.}
}
@incollection{SocheretalEMNLP-13,
       author = {Richard Socher and Alex Perelygin and Jean Y. Wu and Jason Chuang and 
                 Christopher D. Manning and Andrew Y. Ng and Christopher Potts},
        title = {Recursive Deep Models for Semantic Compositionality Over a Sentiment Treebank},
    booktitle = {Proceedings of the 2013 Conference on Empirical Methods in Natural Language Processing},
    publisher = {Association for Computational Linguistics},
      address = {Stroudsburg, PA, USA},
         year = 2013,
        pages = {1631-1642},
     abstract = {Semantic word spaces have been very useful but cannot express the meaning of longer phrases in a principled way. Further progress towards understanding compositionality in tasks such as sentiment detection requires richer supervised training and evaluation resources and more powerful models of composition. To remedy this, we introduce a Sentiment Treebank. It includes fine grained sentiment labels for 215,154 phrases in the parse trees of 11,855 sentences and presents new challenges for sentiment compositionality. To address them, we introduce the Recursive Neural Tensor Network. When trained on the new treebank, this model outperforms all previous methods on several metrics. It pushes the state of the art in single sentence positive/negative classification from 80\% up to 85.4\%. The accuracy of predicting fine-grained sentiment labels for all phrases reaches 80.7\%, an improvement of 9.7\% over bag of features baselines. Lastly, it is the only model that can accurately capture the effects of negation and its scope at various tree levels for both positive and negative phrases.},
}
@inproceedings{SocheretalICML-11,
       author = {Richard Socher and Cliff Chiung-Yu Lin and Andrew Y. Ng and Christopher D. Manning},
        title = {Parsing Natural Scenes and Natural Language with Recursive Neural Networks},
    booktitle = {Proceedings of the 28th International Conference on Machine Learning},
       editor = {Lise Getoor and Tobias Scheffer},
        pages = {129-136},
         year = 2011,
}
@inproceedings{SocheretalISBI-08,
        title = {A Learning Based Hierarchical Model for Vessel Segmentation},
       author = {Richard Socher and Adrian Barbu and Dorin Comaniciu},
    booktitle = {IEEE International Symposium on Biomedical Imaging: From Nano to Macro},
    publisher = {IEEE},
      address = {Paris, France},
         year = 2008,
     abstract = {In this paper we present a learning based method for vessel segmentation in angiographic videos. Vessel Segmentation is an important task in medical imaging and has been investigated extensively in the past. Traditional approaches often require pre-processing steps, standard conditions or manually set seed points. Our method is automatic, fast and robust towards noise often seen in low radiation X-ray images. Furthermore, it can be easily trained and used for any kind of tubular structure. We formulate the segmentation task as a hierarchical learning problem over 3 levels: border points, cross-segments and vessel pieces, corresponding to the vessel's position, width and length. Following the Marginal Space Learning paradigm the detection on each level is performed by a learned classifier. We use Probabilistic Boosting Trees with Haar and steerable features. First results of segmenting the vessel which surrounds a guide wire in 200 frames are presented and future additions are discussed.}
}
@inproceedings{SocheretalNIPS-10,
       author = {Richard Socher and Christopher D. and Manning, Andrew Y. Ng},
        title = {Learning Continuous Phrase Representations and Syntactic Parsing with Recursive Neural Networks},
    booktitle = {Advances in Neural Information Processing Systems 23},
         year = 2010,
     abstract = {Natural language parsing has typically been done with small sets of discrete categories such as NP and VP, but this representation does not capture the full syntactic nor semantic richness of linguistic phrases, and attempts to improve on this by lexicalizing phrases only partly address the problem at the cost of huge feature spaces and sparseness. To address this, we introduce a recursive neural network architecture for jointly parsing natural language and learning vector space representations for variable-sized inputs. At the core of our architecture are context-aware recursive neural networks (CRNN). These networks can induce distributed feature representations for unseen phrases and provide syntactic information to accurately predict phrase structure trees. Most excitingly, the representation of each phrase also captures semantic information: For instance, the phrases "decline to comment" and "would not disclose the terms" are close by in the induced embedding space. Our current system achieves an unlabeled bracketing F-measure of 92.1\% on the Wall Street Journal development dataset for sentences up to length 15.},
}
@inproceedings{SocheretalNIPS-11,
       author = {Richard Socher and Eric H. Huang and Jeffrey Pennington and Andrew Y. Ng and Christopher D. Manning},
        title = {Dynamic Pooling and Unfolding Recursive Autoencoders for Paraphrase Detection},
    booktitle = {Advances in Neural Information Processing Systems 24},
       editor = {John Shawe-Taylor and Richard S. Zemel and Peter L. Bartlett and Fernando C. N. Pereira and Kilian Q. Weinberger},
    booktitle = {Advances in Neural Information Processing Systems 24},
       editor = {J. Shawe-Taylor and R.S. Zemel and P. Bartlett and F. Pereira and K.Q. Weinberger},
    publisher = {MIT Press},
         year = 2011,
        pages = {801-809},
     abstract = {Paraphrase detection is the task of examining two sentences and determining whether they have the same meaning. In order to obtain high accuracy on this task, thorough syntactic and semantic analysis of the two statements is needed. We introduce a method for paraphrase detection based on recursive autoencoders (RAE). Our unsupervised RAEs are based on a novel unfolding objective and learn feature vectors for phrases in syntactic trees. These features are used to measure the word- and phrase-wise similarity between two sentences. Since sentences may be of arbitrary length, the resulting matrix of similarity measures is of variable size. We introduce a novel dynamic pooling layer which computes a fixed-sized representation from the variable-sized matrices. The pooled representation is then used as input to a classifier. Our method outperforms other state-of-the-art approaches on the challenging MSRP paraphrase corpus.},
}
@incollection{SocheretalNIPS-13,
        title = {Reasoning With Neural Tensor Networks for Knowledge Base Completion},
       author = {Socher, Richard and Chen, Danqi and Manning, Christopher D and Ng, Andrew},
    booktitle = {Advances in Neural Information Processing Systems 26},
       editor = {C. J. C. Burges and L. Bottou and M. Welling and Z. Ghahramani and K. Q. Weinberger},
        pages = {926-934},
         year = {2013},
    publisher = {Curran Associates, Inc.},
     abstract = {Knowledge bases are an important resource for question answering and other tasks but often suffer from incompleteness and lack of ability to reason over their discrete entities and relationships. In this paper we introduce an expressive neural tensor network suitable for reasoning over relationships between two entities. Previous work represented entities as either discrete atomic units or with a single entity vector representation. We show that performance can be improved when entities are represented as an average of their constituting word vectors. This allows sharing of statistical strength between, for instance, facts involving the "Sumatran tiger" and "Bengal tiger." Lastly, we demonstrate that all models improve when these word vectors are initialized with vectors learned from unsupervised large corpora. We assess the model by considering the problem of predicting additional true relations between entities given a subset of the knowledge base. Our model outperforms previous models and can classify unseen relationships in Word Net? and Free Base? with an accuracy of 86.2\% and 90.0\%, respectively.},
}
@article {SofroniewetalBIORXIV-16,
       author = {Sofroniew, Nicholas James and Flickinger, Daniel and King, Jonathan and Svoboda, Karel},
        title = {A large field of view two-photon mesoscope with subcellular resolution for in vivo imaging},
         year = {2016},
    publisher = {Cold Spring Harbor Labs Journals},
      journal = {bioRxiv},
     abstract = {Imaging is used to map activity across populations of neurons. Microscopes with cellular resolution have small (less than 1 millimeter) fields of view and cannot simultaneously image activity distributed across multiple brain areas. Typical large field of view microscopes do not resolve single cells, especially in the axial dimension. We developed a 2-photon random access mesoscope (2p-RAM) that allows high-resolution imaging anywhere within a volume spanning multiple brain areas (∅ 5 mm x 1 mm cylinder). 2p-RAM resolution is near diffraction limited (lateral, 0.66 \hmmicron{}, axial 4.09 \hmmicron{} at the center; excitation wavelength = 970 nm; numerical aperture = 0.6) over a large range of excitation wavelengths. A fast three-dimensional scanning system allows efficient sampling of neural activity in arbitrary regions of interest across the entire imaging volume. We illustrate the use of the 2p-RAM by imaging neural activity in multiple, non-contiguous brain areas in transgenic mice expressing protein calcium sensors.},
}
@article{Sohl-DicksteinetalCoRR-15,
       author = {Jascha Sohl-Dickstein and Eric A. Weiss and Niru Maheswaranathan and Surya Ganguli},
        title = {Deep Unsupervised Learning using Nonequilibrium Thermodynamics},
      journal = {CoRR},
       volume = {arXiv:1503.03585},
         year = 2015,
     abstract = {A central problem in machine learning involves modeling complex data-sets using highly flexible families of probability distributions in which learning, sampling, inference, and evaluation are still analytically or computationally tractable. Here, we develop an approach that simultaneously achieves both flexibility and tractability. The essential idea, inspired by non-equilibrium statistical physics, is to systematically and slowly destroy structure in a data distribution through an iterative forward diffusion process. We then learn a reverse diffusion process that restores structure in data, yielding a highly flexible and tractable generative model of the data. This approach allows us to rapidly learn, sample from, and evaluate probabilities in deep generative models with thousands of layers or time steps.}
}
@article{Soiza-ReillyetalFiNC-15,
       author = {Soiza-Reilly, Mariano and Saggau, Peter and Arenkiel, Benjamin R.},
        title = {Neural circuits revealed},
      journal = {Frontiers in Neural Circuits},
    publisher = {Frontiers Media S.A.},
       volume = {9},
         year = {2015},
        pages = {35},
     abstract = {The appropriate function of the nervous system relies on precise patterns of connectivity among hundreds to billions of neurons across different biological systems. Evolutionarily conserved patterns of neural circuit organization and connectivity between morphologically and functionally diverse sets of neurons emerge from a remarkably robust set of genetic blueprints, uniquely defining circuits responsible for planning and execution of behavioral repertoires [...]},
}
@article{SokolovetalTiCS-17,
       author = {Sokolov, Arseny A. and Miall, R. Chris and Ivry, Richard B.},
        title = {The Cerebellum: Adaptive Prediction for Movement and Cognition},
      journal = {Trends in cognitive sciences},
       volume = {21},
        issue = {5},
         year = {2017},
        pages = {313-332},
     abstract = {Over the past 30 years, cumulative evidence has indicated that cerebellar function extends beyond sensorimotor control. This view has emerged from studies of neuroanatomy, neuroimaging, neuropsychology, and brain stimulation, with the results implicating the cerebellum in domains as diverse as attention, language, executive function, and social cognition. Although the literature provides sophisticated models of how the cerebellum helps refine movements, it remains unclear how the core mechanisms of these models can be applied when considering a broader conceptualization of cerebellar function. In light of recent multidisciplinary findings, we examine how two key concepts that have been suggested as general computational principles of cerebellar function - prediction and error-based learning- might be relevant in the operation of cognitive cerebro-cerebellar loops.},
}
@article{SolariandStonerFIN-11,
    author = {Solari, Soren Van Hout  and  Stoner, Rich Martin},   
     title = {Cognitive consilience: Primate non-primary neuroanatomical circuits underlying cognition},     
   journal = {Frontiers in Neuroanatomy},     
    volume = 5,     
      year = 2011,     
     issue = 65, 
  abstract = {Interactions between the cerebral cortex, thalamus, and basal ganglia form the basis cognitive information processing in the mammalian brain. Understanding the principles of neuroanatomical organization in these structures is critical to understanding the functions they perform and ultimately how the human brain works. We have manually distilled and synthesized hundreds of primate neuroanatomy facts into a single interactive visualization. The resulting picture represents the fundamental neuroanatomical blueprint upon which cognitive functions must be implemented. Within this framework we hypothesize and detail 7 functional circuits corresponding to psychological perspectives on the brain: consolidated long-term declarative memory, short-term declarative memory, working memory/information processing, behavioral memory selection, behavioral memory output, cognitive control, and cortical information flow regulation. Each circuit is described in terms of distinguishable neuronal groups including the cerebral isocortex (9 pyramidal neuronal groups), parahippocampal gyrus and hippocampus, thalamus (4 neuronal groups), basal ganglia (7 neuronal groups), metencephalon, basal forebrain and other subcortical nuclei. We focus on neuroanatomy related to primate non-primary cortical systems to elucidate the basis underlying the distinct homotypical cognitive architecture. To display the breadth of this review, we introduce a novel method of integrating and presenting data in multiple independent visualizations: an interactive website (www.cognitiveconsilience.com) and standalone iPhone and iPad applications. With these tools we present a unique, annotated view of neuroanatomical consilience (integration of knowledge).}
}
@misc{Soloviev97,
       author = {S. Soloviev},
        title = {Shift-invariant recognition by the conjunction of basic invariant patterns},
          url = {http://citeseer.ist.psu.edu/soloviev97shiftinvariant.html},
         year = 1997,
}
@article{SoltaniandWangNATURE-10,
       author = {Soltani, Alireza and Wang, Xiao-Jing},
        title = {Synaptic computation underlying probabilistic inference},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {13},
        issue = {1},
         year = {2010},
        pages = {112-119},
     abstract = {We propose that synapses may be the workhorse of the neuronal computations that underlie probabilistic reasoning. We built a neural circuit model for probabilistic inference in which information provided by different sensory cues must be integrated and the predictive powers of individual cues about an outcome are deduced through experience. We found that bounded synapses naturally compute, through reward-dependent plasticity, the posterior probability that a choice alternative is correct given that a cue is presented. Furthermore, a decision circuit endowed with such synapses makes choices on the basis of the summed log posterior odds and performs near-optimal cue combination. The model was validated by reproducing salient observations of, and provides insights into, a monkey experiment using a categorization task. Our model thus suggests a biophysical instantiation of the Bayesian decision rule, while predicting important deviations from it similar to the 'base-rate neglect' observed in human studies when alternatives have unequal prior probabilities.}
}
@incollection{Solway2018decisions,
        title = {The Temporal Dynamics of Reward-Based Goal-Directed Decision-Making},
       author = {Alec Solway and Matthew M. Botvinick},
       editor = {Richard Morris and Aaron Bornstein and Amitai Shenhav},
    booktitle = {Goal-Directed Decision Making},
    publisher = {Academic Press},
         year = {2018},
        pages = {49-66},
     abstract = {This chapter reviews recent work on computational models describing the temporal dynamics of reward-based goal-directed decision-making, from a conceptual rather than technical perspective. Models of temporal dynamics make predictions about the time course and duration of deliberation in addition to the choices that people make. We begin by reviewing the simple one-step choice decision situation and the idea of evidence integration, a core feature of dynamic models, within the context of the drift-diffusion model. We then present more recent extensions of the evidence integration perspective to decisions that span more than one time step.}
}
@article{SomersetalNAS-99,
       author = {David C. Somers and Anders M. Dale and Adriane E. Seiffert and Roger B. H. Tootell},
        title = {Functional {MRI} reveals spatially specific attentional modulation in human primary visual cortex},
      journal = {Proceedings of the National Academy of Sciences},
       volume = 96,
        pages = {1663-1668},
         year = 1999,
}
@article{SompolinskyetalPRE-01,
       author = {Sompolinsky, H. and Yoon, H. and Kang, K. and Shamir, M.},
        title = {Population coding in neuronal systems with correlated noise},
      journal = {Physics Review E Statistical, Nonlinear, and Soft Matter Physics},
       volume = {64},
       number = {5},
         year = {2001},
        pages = {051904},
     abstract = {Neuronal representations of external events are often distributed across large populations of cells. We study the effect of correlated noise on the accuracy of these neuronal population codes. Our main question is whether the inherent error in the population code can be suppressed by increasing the size of the population N in the presence of correlated noise. We address this issue using a model of a population of neurons that are broadly tuned to an angular variable in two dimensions. The fluctuations in the neuronal activities are modeled as Gaussian noises with pairwise correlations that decay exponentially with the difference between the preferred angles of the correlated cells. We assume that the system is broadly tuned, which means that both the correlation length and the width of the tuning curves of the mean responses span a substantial fraction of the entire system length. The performance of the system is measured by the Fisher information (FI), which bounds its estimation error. By calculating the FI in the limit of a large N, we show that positive correlations decrease the estimation capability of the network, relative to the uncorrelated population. The information capacity saturates to a finite value as the number of cells in the population grows. In contrast, negative correlations substantially increase the information capacity of the neuronal population. These results are supplemented by the effect of correlations on the mutual information of the system. Our analysis provides an estimate of the effective number of statistically independent degrees of freedom, denoted N(eff), that a large correlated system can have. According to our theory N(eff) remains finite in the limit of a large N. Estimating the parameters of the correlations and tuning curves from experimental data in some cortical areas that code for angles, we predict that the number of effective degrees of freedom embedded in localized populations in these areas is less than or of the order of approximately 10(2).}
}
@article{SondikOR-82,
       author = {Sondik, Edward J.},
        title = {The Optimal Control of Partially Observable {M}arkov Processes over the Infinite Horizon: Discounted Cost},
      journal = {Operations Research},
       volume = 26,
        issue = 2,
         year = 1978,
        pages = {282-304},
}
@article{SongandChoiNANOMATERIALS-13,
       author = {Song, Edward and Choi, Jin-Woo},
        title = {Conducting Polyaniline Nanowire and Its Applications in Chemiresistive Sensing},
      journal = {Nanomaterials},
       volume = 3,
        issue = 3,
         year = 2013,
        pages = 498,
     abstract = {One dimensional polyaniline nanowire is an electrically conducting polymer that can be used as an active layer for sensors whose conductivity change can be used to detect chemical or biological species. In this review, the basic properties of polyaniline nanowires including chemical structures, redox chemistry, and method of synthesis are discussed. A comprehensive literature survey on chemiresistive/conductometric sensors based on polyaniline nanowires is presented and recent developments in polyaniline nanowire-based sensors are summarized. Finally, the current limitations and the future prospect of polyaniline nanowires are discussed.}
}
@inproceedings{SongetalECCV-12,
     title = {Sparselet Models for Efficient Multiclass Object Detection},
    author = {Song, H.O. and Zickler, S. and Althoff, T. and Girshick, R. and Fritz, M. and Geyer, C. and Felzenszwalb, P. and Darrell, T.},
 booktitle = {Proceedings of the 11th European Conference on Computer Vision},
      year = 2012
}
@article{SongetalNATURE-12,
      title = {Multiplexed volumetric bar-chart chip for point-of-care diagnostics},
     author = {Song, Yujun and Zhang, Yuanqing and Bernard, Paul E. and Reuben, James M. and Ueno, Naoto T. and Arlinghaus, Ralph B. and Zu, Youli and Qin, Lidong},
    journal = {Nature Communications},
  publisher = {Nature Publishing Group},
     volume = 3,
       year = 2012,
      pages = 1283,
   abstract = {Microfluidics have become an enabling technology for point-of-care and personalized diagnostics. Desirable capabilities of microfluidics-based diagnostic devices include simplicity, portability, low cost and the performance of multiplexed and quantitative measurements, ideally in a high-throughput format. Here we present the multiplexed volumetric bar-chart chip ({V-Chip}), which integrates all these capabilities in one device. A key feature of the {V-Chip} is that quantitative results are displayed as bar charts directly on the device--without the need for optical instruments or any data processing or plotting steps. This is achieved by directly linking oxygen production by catalase, which is proportional to the concentration of the analyte, with the displacement of ink along channels on the device. We demonstrate the rapid quantification of protein biomarkers in diverse clinical samples with the {V-Chip}. The development of the {V-Chip} thus opens up the possibility of greatly simplified point-of-care and personalized diagnostics.},
}
@article{SongetalNATURE-13,
       author = {Song, Young Min and Xie, Yizhu and Malyarchuk, Viktor and Xiao, Jianliang and Jung, Inhwa and Choi, Ki-Joong and Liu, Zhuangjian and Park, Hyunsung and Lu, Chaofeng and Kim, Rak-Hwan and Li, Rui and Crozier, Kenneth B. and Huang, Yonggang and Rogers, John A.},
        title = {Digital cameras with designs inspired by the arthropod eye},
      journal = {Nature},
    publisher = {Nature Publishing Group},
       volume = {497},
        issue = {7447},
         year = {2013},
        pages = {95-99},
}
@article{SongetalPAMI-03,
       author = {Song, Y. and Goncalves, L. and Perona, P.},
        title = {Unsupervised learning of human motion},
      journal = {Pattern Analysis and Machine Intelligence},
       volume = 25,
        issue = 7,
        pages = {814-827},
         year = 2003,
}
@article{SongetalPLoS-05,
     title = {Highly nonrandom features of synaptic connectivity in local cortical circuits.},
    author = {Song, S. and Sj\"{o}str\"{o}m, P.J. and Reigl, M. and Nelson, S. and Chklovskii, D.B.},
   journal = {{PLoS} Biology},
    volume = 3,
     issue = 3,
     pages = {e68},
      year = 2005,
  abstract = {How different is local cortical circuitry from a random network? To answer this question, we probed synaptic connections with several hundred simultaneous quadruple whole-cell recordings from layer 5 pyramidal neurons in the rat visual cortex. Analysis of this dataset revealed several nonrandom features in synaptic connectivity. We confirmed previous reports that bidirectional connections are more common than expected in a random network. We found that several highly clustered three-neuron connectivity patterns are overrepresented, suggesting that connections tend to cluster together. We also analyzed synaptic connection strength as defined by the peak excitatory postsynaptic potential amplitude. We found that the distribution of synaptic connection strength differs significantly from the Poisson distribution and can be fitted by a lognormal distribution. Such a distribution has a heavier tail and implies that synaptic weight is concentrated among few synaptic connections. In addition, the strengths of synaptic connections sharing pre- or postsynaptic neurons are correlated, implying that strong connections are even more clustered than the weak ones. Therefore, the local cortical network structure can be viewed as a skeleton of stronger connections in a sea of weaker ones. Such a skeleton is likely to play an important role in network dynamics and should be investigated further.}
}
@article{SongetalPLoS-16,
       author = {Song, H. Francis and Yang, Guangyu R. and Wang, Xiao-Jing},
       editor = {Sporns, Olaf},
        title = {Training Excitatory-Inhibitory Recurrent Neural Networks for Cognitive Tasks: A Simple and Flexible Framework},
      journal = {PLoS Comput Biol},
    publisher = {Public Library of Science},
       volume = {12},
        issue = {2},
         year = {2016},
        pages = {1004792},
    abstract = {The ability to simultaneously record from large numbers of neurons in behaving animals has ushered in a new era for the study of the neural circuit mechanisms underlying cognitive functions. One promising approach to uncovering the dynamical and computational principles governing population responses is to analyze model recurrent neural networks (RNNs) that have been optimized to perform the same tasks as behaving animals. Because the optimization of network parameters specifies the desired output but not the manner in which to achieve this output, trained networks serve as a source of mechanistic hypotheses and a testing ground for data analyses that link neural computation to behavior. Complete access to the activity and connectivity of the circuit, and the ability to manipulate them arbitrarily, make trained networks a convenient proxy for biological circuits and a valuable platform for theoretical investigation. However, existing RNNs lack basic biological features such as the distinction between excitatory and inhibitory units (Dale's principle), which are essential if RNNs are to provide insights into the operation of biological circuits. Moreover, trained networks can achieve the same behavioral performance but differ substantially in their structure and dynamics, highlighting the need for a simple and flexible framework for the exploratory training of RNNs. Here, we describe a framework for gradient descent-based training of excitatory-inhibitory RNNs that can incorporate a variety of biological knowledge. We provide an implementation based on the machine learning library Theano, whose automatic differentiation capabilities facilitate modifications and extensions. We validate this framework by applying it to well-known experimental paradigms such as perceptual decision-making, context-dependent integration, multisensory integration, parametric working memory, and motor sequence generation. Our results demonstrate the wide range of neural activity patterns and behavior that can be modeled, and suggest a unified setting in which diverse cognitive computations and mechanisms can be studied.},
}
@book{Sonkaetal1993,
    author = {Sonka, Milan and Vaclav Hlavac and Boyle, Richard},
     title = {Image Processing, Analysis and Machine Vision, Third Edition}, 
 publisher = {CL Engineering},
      year = 2007,
}
@book{SonnenburgandSonnenburg2015,
      author = {Justin Sonnenburg and Erica Sonnenburg}, 
       title = {The Good Gut Taking Control of your Weight, Your Mood and Your Long Term Health},
   publisher = {Penguin Press},
     address = {New York},
        year = 2015,
}
@article{SordonietalCoRR-15,
       author = {Alessandro Sordoni and Yoshua Bengio and Hossein Vahabi and Christina Lioma and Jakob Grue Simonsen and Jian{-}Yun Nie},
        title = {A Hierarchical Recurrent Encoder-Decoder For Generative Context-Aware Query Suggestion},
      journal = {CoRR},
       volume = {arXiv:1507.02221},
         year = {2015},
     abstract = {Users may strive to formulate an adequate textual query for their information need. Search engines assist the users by presenting query suggestions. To preserve the original search intent, suggestions should be context-aware and account for the previous queries issued by the user. Achieving context awareness is challenging due to data sparsity. We present a probabilistic suggestion model that is able to account for sequences of previous queries of arbitrary lengths. Our novel hierarchical recurrent encoder-decoder architecture allows the model to be sensitive to the order of queries in the context while avoiding data sparsity. Additionally, our model can suggest for rare, or long-tail, queries. The produced suggestions are synthetic and are sampled one word at a time, using computationally cheap decoding techniques. This is in contrast to current synthetic suggestion models relying upon machine learning pipelines and hand-engineered feature sets. Results show that it outperforms existing context-aware approaches in a next query prediction setting. In addition to query suggestion, our model is general enough to be used in a variety of other applications.}
}
@article{SorensenetalCC-13,
    author = {Sorensen, Staci A. and Bernard, Amy and Menon, Vilas and Royall, Joshua J. and Glattfelder, Katie J. and Hirokawa, Karla and Mortrud, Marty and Miller, Jeremy A. and Zeng, Hongkui and Hohmann, John G. and Jones, Allan R. and Lein, Ed S.}, 
     title = {Correlated Gene Expression and Target Specificity Demonstrate Excitatory Projection Neuron Diversity},
   journal = {Cerebral Cortex}, 
      year = 2013, 
  abstract = {The neocortex contains diverse populations of excitatory neurons segregated by layer and further definable by their specific cortical and subcortical projection targets. The current study describes a systematic approach to identify molecular correlates of specific projection neuron classes in mouse primary somatosensory cortex (S1), using a combination of in situ hybridization (ISH) data mining, marker gene colocalization, and combined retrograde labeling with ISH for layer-specific marker genes. First, we identified a large set of genes with specificity for each cortical layer, and that display heterogeneous patterns within those layers. Using these genes as markers, we find extensive evidence for the covariation of gene expression and projection target specificity in layer 2/3, 5, and 6, with individual genes labeling neurons projecting to specific subsets of target structures. The combination of gene expression and target specificity imply a great diversity of projection neuron classes that is similar to or greater than that of GABAergic interneurons. The covariance of these 2 phenotypic modalities suggests that these classes are both discrete and genetically specified.}, 
}
@book{Sorenson85,
       editor = {Sorenson, Harold W.},
        title = {Kalman Filtering: Theory and Application},
    booktitle = {Kalman Filtering: Theory and Application},
    publisher = {{IEEE} Press},
      address = {New York},
         year = 1985,
}
@article{SorrellsetalNATURE-18,
       author = {Sorrells, Shawn F. and Paredes, Mercedes F. and Cebrian-Silla, Arantxa and Sandoval, Kadellyn and Qi, Dashi and Kelley, Kevin W. and James, David and Mayer, Simone and Chang, Julia and Auguste, Kurtis I. and Chang, Edward F. and Gutierrez, Antonio J. and Kriegstein, Arnold R. and Mathern, Gary W. and Oldham, Michael C. and Huang, Eric J. and Garcia-Verdugo, Jose Manuel and Yang, Zhengang and Alvarez-Buylla, Arturo},
        title = {Human hippocampal neurogenesis drops sharply in children to undetectable levels in adults},
      journal = {Nature},
    publisher = {Macmillan Publishers Limited},
       volume = {555},
        pages = {377},
         year = {2018},
}
@article{Soto-TrevinoetalNATURE-06,
       author = {Cristina Soto-Trevi\~{n}o, Kurt A. Thoroughman and Eve Marder, L. F. Abbott},
        title = {Activity-dependent modification of inhibitory synapses in models of rhythmic neural networks},
      journal = {Nature},
       volume = 4,
        issue = 3,
         year = 2006,
        pages = {2102-2121},
     abstract = {The faithful production of rhythms by many neural circuits depends critically on the strengths of inhibitory synaptic connections. We propose a model in which the strengths of inhibitory synapses in a central pattern-generating circuit are subject to activity-dependent plasticity. The strength of each synapse is modified as a function of the global activity of the postsynaptic neuron and by correlated activity of the pre- and postsynaptic neurons. This allows the self-assembly, from random initial synaptic strengths, of two cells into reciprocal oscillation and three cells into a rhythmic triphasic motor pattern. This self-assembly illustrates that complex oscillatory circuits that depend on multiple inhibitory synaptic connections can be tuned via simple activity-dependent rules.}
}
@article{SoudryetalPLoS-15,
       author = {Soudry, Daniel and Keshri, Suraj and Stinson, Patrick and Oh, Min-hwan and Iyengar, Garud and Paninski, Liam},
      journal = {{PLoS} Computational Biology},
    publisher = {Public Library of Science},
        title = {Efficient 'Shotgun' Inference of Neural Connectivity from Highly Sub-sampled Activity Data},
       volume = {11},
       number = {10},
         year = {2015},
        pages = {1-30},
     abstract = {Optical imaging of the activity in a neuronal network is limited by the scanning speed of the imaging device. Therefore, typically, only a small fixed part of the network is observed during the entire experiment. However, in such an experiment, it can be hard to infer from the observed activity patterns whether (1) a neuron A directly affects neuron B, or (2) another, unobserved neuron C affects both A and B. To deal with this issue, we propose a "shotgun" observation scheme, in which, at each time point, we observe a small changing subset of the neurons from the network. Consequently, many fewer neurons remain completely unobserved during the entire experiment, enabling us to eventually distinguish between cases (1) and (2) given sufficiently long experiments. Since previous inference algorithms cannot efficiently handle so many missing observations, we develop a scalable algorithm for data acquired using the shotgun observation scheme, in which only a small fraction of the neurons are observed in each time bin. Using this kind of simulated data, we show the algorithm is able to quickly infer connectivity in spiking recurrent networks with thousands of neurons.},
}
@techreport{SoudryetalTR-13,
        title = {Memristor-based multilayer neural networks with online gradient descent training},
       author = {Daniel Soudry and Dotan Di Castro and Asaf Gal and Avinoam Kolodny and and Shahar Kvatinsky},
         year = 2013,
  institution = {Department of Statistics, Center for Theoretical Neuroscience, Columbia University},
     abstract = {Learning in Multilayer Neural Networks (MNNs) relies on continuous updating of large matrices of synaptic weights by local rules. Such locality can be exploited for massive parallelism when implementing MNNs in hardware. However, these update rules require a multiply and accumulate operation for each synaptic weight, which is challenging to implement compactly using CMOS. In this paper, a method for performing these update operations simultaneously ("incremental outer products") using memristor-based arrays is proposed. The method is based on the fact that, approximately, given a voltage pulse, the conductivity of a memristor will increment proportionally to the pulse duration multiplied by the pulse magnitude, if the increment is sufficiently small. The proposed method uses a synaptic circuit composed of a small number of components per synapse: one memristor and two CMOS transistors. This circuit is expected to consume between 2\% to 8\% of the area and static power of previous CMOS-only hardware alternatives. Such a circuit can compactly implement scalable training algorithms based on online gradient descent (e.g., Backpropagation). The utility and robustness of the proposed memristor-based circuit is demonstrated on standard supervised learning tasks.}
}
@incollection{Sowa2007languagegames,
       author = {Sowa, John},
        title = {Language Games, {F}oundation for Semantics and Ontology},
    booktitle = {Game Theory and Linguistic Meaning},
       editor = {Ahti-Veikko Pietarinen},
    publisher = {Elsevier},
         year = {2007},
        pages = {17-37},
}
@article{SowelletalNN-99,
    author = {Sowell, E. R. and Thompson, P. M. and Holmes, C. J. and Jernigan, T. L. and Toga, A. W.},
     title = {In vivo evidence for post-adolescent brain maturation in frontal and striatal regions},
   journal = {Nature Neuroscience},
     issue = 10,
     pages = {859-861},
    volume = 2,
      year = 1999,
}
@article{SpaldingetalCELL-13,
       author = {Spalding, Kirsty L. and Bergmann, Olaf and Alkass, Kanar and Bernard, Samuel and Salehpour, Mehran and Huttner, Hagen B. and Bostr\"{o}m, Emil and Westerlund, Isabelle and Vial, Celine and Buchholz, Bruce A. and Possnert, G\"{o}ran and Mash, Deborah C. and Druid, Henrik and Fris\'{e}n, Jonas},
        title = {Dynamics of hippocampal neurogenesis in adult humans},
      journal = {Cell},
       volume = {153},
        issue = {6},
         year = {2013},
        pages = {1219-1227},
     abstract = {Adult-born hippocampal neurons are important for cognitive plasticity in rodents. There is evidence for hippocampal neurogenesis in adult humans, although whether its extent is sufficient to have functional significance has been questioned. We have assessed the generation of hippocampal cells in humans by measuring the concentration of nuclear-bomb-test-derived C14 in genomic DNA, and we present an integrated model of the cell turnover dynamics. We found that a large subpopulation of hippocampal neurons constituting one-third of the neurons is subject to exchange. In adult humans, 700 new neurons are added in each hippocampus per day, corresponding to an annual turnover of 1.75\% of the neurons within the renewing fraction, with a modest decline during aging. We conclude that neurons are generated throughout adulthood and that the rates are comparable in middle-aged humans and mice, suggesting that adult hippocampal neurogenesis may contribute to human brain function.},
}
@article{SpalthoffFiNC-12,
       author = {Spalthoff, Christian and Gerdes, Ralf and Kurtz, Rafael},
        title = {Neuronal representation of visual motion and orientation in the fly medulla},
      journal = {Frontiers in Neural Circuits},
    publisher = {Frontiers Media S.A.},
       volume = {72},
        issue = {6},
         year = {2012},
     abstract = {In insects, the first extraction of motion and direction clues from local brightness modulations is thought to take place in the medulla. However, whether and how these computations are represented in the medulla stills remain widely unknown, because electrical recording of the neurons in the medulla is difficult. As an effort to overcome this difficulty, we employed local electroporation in vivo in the medulla of the blowfly (Calliphora vicina) to stain small ensembles of neurons with a calcium-sensitive dye. We studied the responses of these neuronal ensembles to spatial and temporal brightness modulations and found selectivity for grating orientation. In contrast, the responses to the two opposite directions of motion of a grating with the same orientation were similar in magnitude, indicating that strong directional selectivity is either not present in the types of neurons covered by our data set, or that direction-selective signals are too closely spaced to be distinguished by our calcium imaging. The calcium responses also showed a bell-shaped dependency on the temporal frequency of drifting gratings, with an optimum higher than that observed in one of the subsequent processing stages, i.e., the lobula plate. Medulla responses were elicited by on- as well as off-stimuli with some spatial heterogeneity in the sensitivity for "on" and "off", and in the polarity of the responses. Medulla neurons thus show similarities to some established principles of motion and edge detection in the vertebrate visual system.},
}
@misc{SpelkeBRAIN-INSPIRED-19,
        title = {What Makes Us Special?},
       author = {Elizabeth Spelke},
         year = {2019},
 howpublished = {Brain Inspired Podcast Episode \#48 September 25, 2019}
}
@misc{SpelkeNEW-YORK-TIMES-12,
        title = {{Insights From the Youngest Minds}},
       author = {Natalie Angier},
         year = {2012},
 howpublished = {Feature article on Elizabeth Spelke in the New York Times, April 30, 2012}
}
@article{SpelkeandKinzlerDEVELOPMENTAL-SCIENCE-07,
       author = {Spelke, E. S and Kinzler, K. D.},
        title = {Core knowledge},
      journal = {Developmental Science},
       volume = {10},
         year = {2007},
        pages = {89-96},
          url = {http://www.wjh.harvard.edu/~lds/pdfs/SpelkeKinzler07.pdf},
     abstract = {Human cognition is founded, in part, on four systems for representing objects, actions, number, and space. It may be based, as well, on a fifth system for representing social partners. Each system has deep roots in human phylogeny and ontogeny, and it guides and shapes the mental lives of adults. Converging research on human infants, non-human primates, children and adults in diverse cultures can aid both understanding of these systems and attempts to overcome their limits}
}
@book{SpenceandDriver2004,
        title = {{Crossmodal Space and Crossmodal Attention}},
       author = {Spence, Charles and Driver, Jon},
    publisher = {Oxford University Press},
         year = {2004},
     abstract = {Many organisms possess multiple sensory systems, such as vision, hearing, touch, smell, and taste. The possession of such multiple ways of sensing the world offers many benefits. These benefits arise not only because each modality can sense different aspects of the environment, but also because different senses can respond jointly to the same external object or event, thus enriching the overall experience - for example, looking at an individual while listening to them speak. However, combining information from different senses also poses many challenges for the nervous system. In recent years there has been dramatic progress in understanding how information from different sensory modalities gets integrated in order to construct useful representations of external space; and in how such multimodal representations constrain spatial attention. Such progress has involved numerous different disciplines, including neurophysiology, experimental psychology, neurological work with brain-damaged patients, neuroimaging studies, and computational modelling. This volume brings together the leading researchers from all these approaches, to present aan integrative overview of this central topic in cognitive neuroscience.}
}
@article{SpencerJSL-93,
       author = {J. Spencer},
        title = {0-1 Laws with Variable Probability},
      journal = {Journal of Symbolic Logic},
       volume = 58,
        pages = {1-14},
         year = 1993,
}
@article{SperlingetalCVGIP-85,
      title = {G. Sperling and M.S. Landy and Y. Cohen and M. Pavel}, 
     author = {Intelligible encoding of {ASL} image sequences at extremely low information rates}, 
    journal = {Computer Vision Graphics and Image Processing}, 
     volume = 31, 
      issue = 1, 
       year = 1985, 
      pages = {335-391}, 
}
@article{SpiraetalTIP-07,
    author = {Alon Spira and Ron Kimmel and Nir A. Sochen},
     title = {A Short-Time Beltrami Kernel for Smoothing Images and Manifolds},
   journal = {IEEE Transactions on Image Processing},
    volume = 16,
     issue = 6,
      year = 2007,
     pages = {1628-1636},
}
@book{Spirtesetal93,
       author = {Spirtes, Peter and Glymour, Clark and Scheines, Richard},
        title = {Causation, Prediction, and Search},
    publisher = {Springer-Verlag},
      address = {New York},
         year = 1993,
}
@book{SpohrerandRiesbeckTR-84,
        title = {Reasoning-Driven Memory Modification in the Economics Domain, Computer Science Technical Report Number 308},
       author = {Spohrer, James C. and Riesbeck, Christopher K.},
    publisher = {Yale University, Department of Computer Science},
         year = {1984},
}
@article{SpornsNATURE-NEUROSCIENCE-14,
       author = {Sporns, Olaf},
        title = {Contributions and challenges for network models in cognitive neuroscience},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {17},
        issue = {5},
        pages = {652-660},
         year = {2014},
     abstract = {The confluence of new approaches in recording patterns of brain connectivity and quantitative analytic tools from network science has opened new avenues toward understanding the organization and function of brain networks. Descriptive network models of brain structural and functional connectivity have made several important contributions; for example, in the mapping of putative network hubs and network communities. Building on the importance of anatomical and functional interactions, network models have provided insight into the basic structures and mechanisms that enable integrative neural processes. Network models have also been instrumental in understanding the role of structural brain networks in generating spatially and temporally organized brain activity. Despite these contributions, network models are subject to limitations in methodology and interpretation, and they face many challenges as brain connectivity data sets continue to increase in detail and complexity.},
}
@article{SpornsandKotterPLoS-04,
       author = {Sporns, Olaf and K\"{o}tter, Rolf},
      journal = {PLoS Biol},
    publisher = {Public Library of Science},
        title = {Motifs in Brain Networks},
       volume = {2},
       number = {11},
         year = {2004},
        pages = {1910-1918},
     abstract = {Complex brains have evolved a highly efficient network architecture whose structural connectivity is capable of generating a large repertoire of functional states. We detect characteristic network building blocks (structural and functional motifs) in neuroanatomical data sets and identify a small set of structural motifs that occur in significantly increased numbers. Our analysis suggests the hypothesis that brain networks maximize both the number and the diversity of functional motifs, while the repertoire of structural motifs remains small. Using functional motif number as a cost function in an optimization algorithm, we obtain network topologies that resemble real brain networks across a broad spectrum of structural measures, including small-world attributes. These results are consistent with the hypothesis that highly evolved neural architectures are organized to maximize functional repertoires and to support highly efficient integration of information.},
}
@article{SpornsandZwiNEUR0INFORMATICS-04,
       author = {Olaf Sporns and Jonathan D. Zwi},
        title = {The Small World of the Cerebral Cortex},
      journal = {Neuroinformatics},
    publisher = {Humana Press Incorporated},
        pages = {145-162},
       volume = 2,
        issue = 2,
         year = 2004,
          url = {SpornsandZwiNEUR0INFORMATICS-04.pdf},
}
@article{SpornsetalPLoS-05,
    author = {Sporns, O. and Tononi, G. and K\"{o}tter, R.},
     title = {The human connectome: A structural description of the human brain},
   journal = {{PLoS} Computataional Biology},
      year = 2005,
    volume = 1,
     issue = 4,
     pages = {e42},
  abstract = {The connection matrix of the human brain (the human "connectome") represents an indispensable foundation for basic and applied neurobiological research. However, the network of anatomical connections linking the neuronal elements of the human brain is still largely unknown. While some databases or collations of large-scale anatomical connection patterns exist for other mammalian species, there is currently no connection matrix of the human brain, nor is there a coordinated research effort to collect, archive, and disseminate this important information. We propose a research strategy to achieve this goal, and discuss its potential impact.}
}
@article{SpratlingCC-17,
       author = {Spratling, M. W.},
        title = {A Hierarchical Predictive Coding Model of Object Recognition in Natural Images},
      journal = {Cognitive Computation},
         year = {2017},
       volume = {9},
       number = {2},
        pages = {151-167},
     abstract = {Predictive coding has been proposed as a model of the hierarchical perceptual inference process performed in the cortex. However, results demonstrating that predictive coding is capable of performing the complex inference required to recognise objects in natural images have not previously been presented. This article proposes a hierarchical neural network based on predictive coding for performing visual object recognition. This network is applied to the tasks of categorising hand-written digits, identifying faces, and locating cars in images of street scenes. It is shown that image recognition can be performed with tolerance to position, illumination, size, partial occlusion, and within-category variation. The current results, therefore, provide the first practical demonstration that predictive coding (at least the particular implementation of predictive coding used here; the PC/BC-DIM algorithm) is capable of performing accurate visual object recognition.},
}
@article{SprechmannetalICLR-18,
        title = {Memory-based Parameter Adaptation},
       author = {Pablo Sprechmann and Siddhant M. Jayakumar and Jack W. Rae and Alexander Pritzel and Adria Puigdomenech Badia and Benigno Uria and Oriol Vinyals and Demis Hassabis and Razvan Pascanu and Charles Blundell},
      journal = {International Conference on Learning Representations},
         year = {2018},
     abstract = {Deep neural networks have excelled on a wide range of problems, from vision to language and game playing. Neural networks very gradually incorporate information into weights as they process data, requiring very low learning rates. If thetraining distribution shifts, the network is slow to adapt, and when it does adapt, it typically performs badly on the training distribution before the shift. Our method, Memory-based Parameter Adaptation, stores examples in memory and then uses a context-based lookup to directly modify the weights of a neural network. Much higher learning rates can be used for this local adaptation, reneging the need for many iterations over similar data before good predictions can be made. As our method is memory-based, it alleviates several shortcomings of neural networks, such as catastrophic forgetting, fast, stable acquisition of new knowledge, learning with an imbalanced class labels, and fast learning during evaluation. We demonstrate this on a range of supervised tasks: large-scale image classification andlanguage modelling.}
}
@article{SprengFiP-13,
       author = {Spreng, R. Nathan},
        title = {Examining the role of memory in social cognition},
      journal = {Frontiers in Psychology},
       volume = {4},
        pages = {437},
         year = {2013},
     abstract = {The function of memory is not only to recall the past, but also to form and update models of our experiences and use these models to navigate the world. Perhaps, the most complex environment for humans to navigate is the social one. Social dynamics are extraordinarily complex, unstructured, labile and difficult to predict. Successful navigation through our many social landscapes is essential to forming and maintaining the durable social bonds necessary for physical and mental health. Until recently, little research has examined the role that memory plays in social behavior and interpersonal sensitivity. There is growing evidence that recalling personally experienced events (autobiographical memory) and inferring the mental states of others (mentalizing or theory-of-mind) share an extensive functional neuroanatomy (Buckner and Carroll, 2007; Spreng et al., 2009; Spreng and Grady, 2010; Rabin et al., 2010) and may be critical for adaptive social cognition.},
}
@book{Squire2008history,
        title = {The History of Neuroscience in Autobiography},
       author = {Squire, Larry R.},
       volume = {6},
         year = {2008},
    publisher = {Oxford University Press},
     abstract = {The sixth volume of The History of Neuroscience in Autobiography is a collection of autobiographical essays by notable senior scientists who discuss the major events that shaped their discoveries and their influences, as well as the people who inspired them and helped shape their careers as neuroscientists. Each entry also includes a complete CV so that the interested reader may see their rise through the ranks as they achieved some of the highest honors in neuroscience.},
}
@article{SquireNEUROSCIENCE-11,
       author = {Squire, Larry R.},
        title = {Memory and brain systems: 1969-2009},
      journal = {Journal of Neuroscience},
       volume = {29},
       number = {41},
         year = {2009},
        pages = {12711-12716},
}
@article{SquireandZola-MorganTiNS-88,
       author = {Larry R. Squire and Stuart Zola-Morgan},
        title = {Memory: brain systems and behavior},
      journal = {Trends in Neurosciences},
       volume = {11},
       number = {4},
        pages = {170-175},
         year = {1988},
     abstract = {Certain questions about memory address a relatively global, structural level of analysis. Is there one kind of memory or many? What brain structures or systems are involved in memory and what jobs do they do? One useful approach to such questions has focused on studies of neurological patients with memory impairment and parallel studies with animal models.}
}
@inproceedings{SrikumarandManningNIPS-14,
       author = {Vivek Srikumar and Christopher Manning},
        title = {Learning Distributed Representations for Structured Output Prediction},
    booktitle = {Advances in Neural Information Processing Systems 27},
         year = 2014,
     abstract = {In recent years, distributed representations of inputs have led to performance gains in many applications by allowing statistical information to be shared across inputs. However, the predicted outputs (labels, and more generally structures) are still treated as discrete objects even though outputs are often not discrete units of meaning. In this paper, we present a new formulation for structured prediction where we represent individual labels in a structure as dense vectors and allow semantically similar labels to share parameters. We extend this representation to larger structures by defining compositionality using tensor products to give a natural generalization of standard structured prediction approaches. We define a learning objective for jointly learning the model parameters and the label vectors and propose an alternating minimization algorithm for learning. We show that our formulation outperforms structural SVM baselines in two tasks: multiclass document classification and part-of-speech tagging.},
}
@techreport{SrinivasTR-89,
       author = {Sampath Srinivas and Jack Breese},
        title = {IDEAL: Influence Diagram Evaluation and Analysis in Lisp},
  institution = {Rockwell International Science Center},
        issue = {Technical Memorandum No. 23},
         year = 1989,
}
@article{SrinivasetalCoRR-18,
       author = {Aravind Srinivas and Allan Jabri and Pieter Abbeel and Sergey Levine and Chelsea Finn},
        title = {Universal Planning Networks},
      journal = {CoRR},
       volume = {arXiv:1804.00645},
         year = {2018},
     abstract = {A key challenge in complex visuomotor control is learning abstract representations that are effective for specifying goals, planning, and generalization. To this end, we introduce universal planning networks (UPN). UPNs embed differentiable planning within a goal-directed policy. This planning computation unrolls a forward model in a latent space and infers an optimal action plan through gradient descent trajectory optimization. The plan-by-gradient-descent process and its underlying representations are learned end-to-end to directly optimize a supervised imitation learning objective. We find that the representations learned are not only effective for goal-directed visual imitation via gradient-based trajectory optimization, but can also provide a metric for specifying goals using images. The learned representations can be leveraged to specify distance-based rewards to reach new target states for model-free reinforcement learning, resulting in substantially more effective learning when solving new tasks described via image-based goals. We were able to achieve successful transfer of visuomotor planning strategies across robots with significantly different morphologies and actuation capabilities.}
}
@article{SrivastavaetalUAU-13,
     author = {Nitish Srivastava and Ruslan Salakhutdinov and Geoffrey E. Hinton},
      title = {Modeling Documents with Deep Boltzmann Machines},
    journal = {CoRR},
     volume = {arXiv:1309.6865},
       year = {2013},
}
@article{St-PierreetalNATURE-NEUROSCIENCE-14,
       author = {St-Pierre, Francois and Marshall, Jesse D. and Yang, Ying and Gong, Yiyang and Schnitzer, Mark J. and Lin, Michael Z.},
        title = {High-fidelity optical reporting of neuronal electrical activity with an ultrafast fluorescent voltage sensor},
      journal = {Nature Neuroscience},
         year = 2014,
    publisher = {Nature Publishing Group},
       volume = 17,
        issue = 6,
        pages = {884-889},
     abstract = {Accurate optical reporting of electrical activity in genetically defined neuronal populations is a long-standing goal in neuroscience. We developed Accelerated Sensor of Action Potentials 1 (ASAP1), a voltage sensor design in which a circularly permuted green fluorescent protein is inserted in an extracellular loop of a voltage-sensing domain, rendering fluorescence responsive to membrane potential. ASAP1 demonstrated on and off kinetics of [sim]2 ms, reliably detected single action potentials and subthreshold potential changes, and tracked trains of action potential waveforms up to 200 Hz in single trials. With a favorable combination of brightness, dynamic range and speed, ASAP1 enables continuous monitoring of membrane potential in neurons at kilohertz frame rates using standard epifluorescence microscopy.},
}
@article{StachenfeldetalNATURE-17,
       author = {Stachenfeld, Kimberly L. and Botvinick, Matthew M. and Gershman, Samuel J.},
        title = {The hippocampus as a predictive map},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
         year = {2017},
       volume = {20},
        pages = {1643},
     abstract = {A cognitive map has long been the dominant metaphor for hippocampal function, embracing the idea that place cells encode a geometric representation of space. However, evidence for predictive coding, reward sensitivity and policy dependence in place cells suggests that the representation is not purely spatial. We approach this puzzle from a reinforcement learning perspective: what kind of spatial representation is most useful for maximizing future reward? We show that the answer takes the form of a predictive representation. This representation captures many aspects of place cell responses that fall outside the traditional view of a cognitive map. Furthermore, we argue that entorhinal grid cells encode a low-dimensionality basis set for the predictive representation, useful for suppressing noise in predictions and extracting multiscale structure for hierarchical planning.}
}
@article{StachenfeldetalNATURE-NEUROSCIENCE-17,
       author = {Stachenfeld, Kimberly L. and Botvinick, Matthew M. and Gershman, Samuel J.},
        title = {The hippocampus as a predictive map},
      journal = {Nature Neuroscience},
       volume = {20},
       number = {1643},
         year = {2017},
        pages = {1643-1653},
    publisher = {Nature Publishing Group},
     abstract = {A cognitive map has long been the dominant metaphor for hippocampal function, embracing the idea that place cells encode a geometric representation of space. However, evidence for predictive coding, reward sensitivity, and policy dependence in place cells suggests that the representation is not purely spatial. We approach this puzzle from a reinforcement learning perspective: what kind of spatial representation is most useful for maximizing future reward? We show that the answer takes the form of a predictive representation. This representation captures many aspects of place cell responses that fall outside the traditional view of a cognitive map. Furthermore, we argue that entorhinal grid cells encode a low-dimensional basis set for the predictive representation, useful for suppressing noise in predictions and extracting multiscale structure for hierarchical planning.},
}
@incollection{StallmanandSussman79,
       author = {Stallman, Richard M. and Gerald J. Sussman},
        title = {Problem Solving About Electrical Circuits},
       editor = {Winston, Patrick H. and Brown, Richard H.},
       volume = {Volume 1},
    booktitle = {Artificial Intelligence: an {MIT} Perspective},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1979,
        pages = {33-92},
}
@article{StandingJEP-73,
    author = {Lionel Standing},
     title = {Learning 10,000 Pictures},
   journal = {Journal of Experimental Psychology},
    volume = 25, 
     issue = 2, 
      year = 1973, 
     pages = {207-222},
}
@article{StanleyetalJN-99,
     author = {Garrett B. Stanley and Fei-Fei Li and Yang Dan},
      title = {Reconstruction of Natural Scenes from Ensemble 
               Responses in the Lateral Geniculate Nucleus},
    journal = {The Journal of Neuroscience},
     volume = 19, 
      issue = 18,
      pages = {8036-8042},
       year = 1999, 
}
@book{Stanovich2009a,
      title = {Decision Making and Rationality in the Modern World},
     author = {Keith E. Stanovich},
  publisher = {Oxford University Press},
    address = {Oxford, UK},
       year = 2009
}
@book{Stanovich2009b,
      title = {What Intelligence Tests Miss: The Psychology of Rational Thought},
     author = {Keith E. Stanovich},
  publisher = {Yale University Press},
    address = {New Haven, CT},
       year = 2009
}
@article{StanovichSAM-09,
      title = {Rational and Irrational Thought: The Thinking that {IQ} Tests Miss},
     author = {Keith E. Stanovich},
    journal = {Scientific American},
     volume = {November/December},
       year = 2009,
      pages = {34-39},
}
@article{StanovichandWestJPSP-08, 
      title = {On the relative independence of thinking biases and cognitive ability}, 
     author = {Stanovich, Keith E. and West, Richard F.}, 
    journal = {Journal of Personality and Social Psychology}, 
     volume = 94, 
      issue = 4, 
  publisher = {American Psychological Association}, 
       year = 2008, 
      pages = {672-695}
}
@article{StaresinaetalNATURE-COMMUNICATIONS-19,
       author = {Staresina, Bernhard P. and Reber, Thomas P. and Niediek, Johannes and Bostr{\"{o}}m, Jan and Elger, Christian E. and Mormann, Florian},
        title = {Recollection in the human hippocampal-entorhinal cell circuitry},
      journal = {Nature Communications},
         year = {2019},
       volume = {10},
        issue = {1},
        pages = {1503},
     abstract = {Imagine how flicking through your photo album and seeing a picture of a beach sunset brings back fond memories of a tasty cocktail you had that night. Computational models suggest that upon receiving a partial memory cue ("beach"), neurons in the hippocampus coordinate reinstatement of associated memories ("cocktail") in cortical target sites. Here, using human single neuron recordings, we show that hippocampal firing rates are elevated from 500-1500 ms after cue onset during successful associative retrieval. Concurrently, the retrieved target object can be decoded from population spike patterns in adjacent entorhinal cortex (EC), with hippocampal firing preceding EC spikes and predicting the fidelity of EC object reinstatement. Prior to orchestrating reinstatement, a separate population of hippocampal neurons distinguishes different scene cues (buildings vs. landscapes). These results elucidate the hippocampal-entorhinal circuit dynamics for memory recall and reconcile disparate views on the role of the hippocampus in scene processing vs. associative memory.},
}
@inproceedings{StaufferandGrimsonCVPR-99,
      title = {Adaptive Background Mixture Models for Real-Time Tracking},
     author = {Stauffer, C. and Grimson, W.E.L.},
  booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
     volume = 2,
       year = 1999,
      pages = {2246-252},
}
@article {StavrinidouetalSCIADV-15,
       author = {Stavrinidou, Eleni and Gabrielsson, Roger and Gomez, Eliot and Crispin, Xavier and Nilsson, Ove and Simon, Daniel T. and Berggren, Magnus},
        title = {Electronic plants},
      journal = {Science Advances},
       volume = {1},
        issue = {10},
         year = {2015},
    publisher = {American Association for the Advancement of Science},
     abstract = {The roots, stems, leaves, and vascular circuitry of higher plants are responsible for conveying the chemical signals that regulate growth and functions. From a certain perspective, these features are analogous to the contacts, interconnections, devices, and wires of discrete and integrated electronic circuits. Although many attempts have been made to augment plant function with electroactive materials, plants' 'circuitry' has never been directly merged with electronics. We report analog and digital organic electronic circuits and devices manufactured in living plants. The four key components of a circuit have been achieved using the xylem, leaves, veins, and signals of the plant as the template and integral part of the circuit elements and functions. With integrated and distributed electronics in plants, one can envisage a range of applications including precision recording and regulation of physiology, energy harvesting from photosynthesis, and alternatives to genetic modification for plant optimization.},
}
@book{Steele89,
       author = {David J. Steele},
        title = {Golden Common Lisp: A Hands on Approach},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1989,
}
@book{Steele90,
       author = {{Steele Jr.}, Guy L.},
        title = {Common Lisp: The Language, Second Edition},
    publisher = {Digital Press},
      address = {Billerica, Massachusetts},
         year = 1990,
}
@article{StefikAIJ-81,
       author = {Stefik, Mark J.},
        title = {Planning with Constraints},
      journal = AIJ,
       volume = 16,
         year = 1981,
        pages = {111-140},
}
@inproceedings{SteinandHebertBEYOND_PATCHES-06,
       author = {Andrew Stein and Martial Hebert},
        title = {Using Spatio-Temporal Patches for Simultaneous Estimation of Edge Strength, Orientation and Motion},
    booktitle = {Beyond Patches Workshop at {IEEE} Conference on Computer Vision and Pattern Recognition},
        pages = 19,
         year = 2006,
}
@inproceedings{SteinandHebertBMVC-06,
       author = {Andrew Stein and Martial Hebert},
        title = {Local Detection of Occlusion Boundaries in Video},
    booktitle = {British Machine Vision Conference},
        volume = 1,
        pages = {407-416},
         year = 2006,
}
@inproceedings{SteinetalICCV-07,
    author = {Andrew Stein and Derek Hoiem and Martial Hebert},
     title = {Learning to Find Object Boundaries Using Motion Cues},
 booktitle = {{IEEE} International Conference on Computer Vision},
     month = {October},
      year = 2007,
}
@inproceedings{StenetorpNIPS-13,
       author = {Stenetorp, Pontus},
        title = {Transition-based Dependency Parsing Using Recursive Neural Networks},
    booktitle = {Deep Learning Workshop at the 2013 Conference on Neural Information Processing Systems},
      address = {Lake Tahoe, Nevada, USA},
         year = 2013,
}
@article{StentetalSCIENCE-78,
     title = {Neuronal generation of the leech swimming movement},
    author = {Stent, G.S. and Kristan, W.B. and Friesen, W.O. and Ort, C.A. and Poon, M. and Calabrese, R.L.}, 
   journal = {Science}, 
    volume = 200, 
     issue = 4348, 
      year = 1978, 
     pages = {1348-1357}, 
  abstract = {The swimming movement of the leech is produced by an ensemble of bilaterally symmetric, rhythmically active pairs of motor neurons present in each segmental ganglion of the ventral nerve cord. These motor neurons innervate the longitudinal muscles in dorsal or ventral sectors of the segmental body wall. Their duty cycles are phase-locked in  a manner such that the dorsal and ventral body wall sectors of any given segment undergo an antiphasic contractile rhythm and that the contractile rhythms of different segments form a rostrocaudal phase progression. This activity rhythm is imposed on the motor neurons by a central swim oscillator, of which four bilaterally symmetric pairs of interneurons present in each segmental ganglion appear to constitute the major component. These interneurons are linked intra- and intersegmentally via inhibitory connections to form a segmentally iterated and inter-segmentally concatenated cyclic neuronal network. The network appears to owe its oscillatory activity pattern to the mechanism of recurrent cyclic inhibition.}, 
}
@article{StephanetalFP-81,
    author = {Stephan, H. and Frahm, H. and Baron, G.},
     title = {{New and revised data on volumes of brain structures in insectivores and primates}},
   journal = {Folia Primatologia},
     pages = {1-29},
    volume = 35,
      year = 1981
}
@article{StephensetalPLoS-08,
       author = {Stephens, Greg J. and Johnson-Kerner, Bethany and Bialek, William and Ryu, William S.},
      journal = {PLoS Computational Biology},
    publisher = {Public Library of Science},
        title = {Dimensionality and Dynamics in the Behavior of {\it{Caenorhabditis elegans}}},
       volume = {4},
       number = {4},
         year = {2008},
        pages = {e1000028},
    abstract = {A great deal of work has been done in characterizing the genes, proteins, neurons, and circuits that are involved in the biology of behavior, but the techniques used to quantify behavior have lagged behind the advancements made in these areas. Here, we address this imbalance in a domain rich enough to allow complex, natural behavior yet simple enough so that movements can be explored exhaustively: the motions of Caenorhabditis elegans freely crawling on an agar plate. From measurements of the worm's curvature, we show that the space of natural worm postures is low dimensional and can be almost completely described by their projections along four principal "eigenworms." The dynamics along these eigenworms offer both a quantitative characterization of classical worm movement such as forward crawling, reversals, and Omega-turns, and evidence of more subtle behaviors such as pause states at particular postures. We can partially construct equations of motion for this shape space, and within these dynamics we find a set of attractors that can be used as a rigorous definition of behavioral state. Our observations of C. elegans reveal a precise and complete language of motion and new aspects of worm behavior. We believe this is a lesson with promise for other organisms.},
}
@article{StephensetalPLoS-10,
       author = {Stephens, Greg J. and Johnson-Kerner, Bethany and Bialek, William and Ryu, William S.},
        title = {From Modes to Movement in the Behavior of {\it{Caenorhabditis elegans}}},
      journal = {PLoS ONE},
    publisher = {Public Library of Science},
       volume = {5},
       number = {11},
         year = {2010},
        pages = {e13914},
    abstract = {Organisms move through the world by changing their shape, and here we explore the mapping from shape space to movements in the nematode Caenorhabditis elegans as it crawls on an agar plate. We characterize the statistics of the trajectories through the correlation functions of the orientation angular velocity, orientation angle and the mean-squared displacement, and we find that the loss of orientational memory has significant contributions from both abrupt, large amplitude turning events and the continuous dynamics between these events. Further, we discover long-time persistence of orientational memory in the intervals between abrupt turns. Building on recent work demonstrating that C. elegans movements are restricted to a low-dimensional shape space\cite{StephensetalPLoS-08}, we construct a map from the dynamics in this shape space to the trajectory of the worm along the agar. We use this connection to illustrate that changes in the continuous dynamics reveal subtle differences in movement strategy that occur among mutants defective in two classes of dopamine receptors.},
}
@article{StephensetalPNAS-10,
       author = {Stephens, Greg J. and Bueno de Mesquita, Matthew and Ryu, William S. and Bialek, William},
        title = {Emergence of long timescales and stereotyped behaviors in Caenorhabditis elegans},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {108},
       number = {18},
        pages = {7286-7289},
         year = {2011},
     abstract = {Animal behaviors often are decomposable into discrete, stereotyped elements, well separated in time. In one model, such behaviors are triggered by specific commands; in the extreme case, the discreteness of behavior is traced to the discreteness of action potentials in the individual command neurons. Here, we use the crawling behavior of the nematode Caenorhabditis elegans to demonstrate the opposite view, in which discreteness, stereotypy, and long timescales emerge from the collective dynamics of the behavior itself. In previous work, we found that as C. elegans crawls, its body moves through a "shape space" in which four dimensions capture approximately 95\% of the variance in body shape. Here we show that stochastic dynamics within this shape space predicts transitions between attractors corresponding to abrupt reversals in crawling direction. With no free parameters, our inferred stochastic dynamical system generates reversal timescales and stereotyped trajectories in close agreement with experimental observations. We use the stochastic dynamics to show that the noise amplitude decreases systematically with increasing time away from food, resulting in longer bouts of forward crawling and suggesting that worms can use noise to modify their locomotory behavior.},
}
@article{StephensetalPNAS-11,
     title = {Searching for simplicity in the analysis of neurons and behavior}, 
    author = {Stephens, Greg J. and Osborne, Leslie C. and Bialek, William}, 
   journal = {Proceedings of the National Academy of Sciences}, 
    volume = 108, 
     issue = 3, 
      year = 2011, 
     pages = {15565-15571}, 
  abstract = {What fascinates us about animal behavior is its richness and complexity, but understanding behavior and its neural basis requires a simpler description. Traditionally, simplification has been imposed by training animals to engage in a limited set of behaviors, by hand scoring behaviors into discrete classes, or by limiting the sensory experience of the organism. An alternative is to ask whether we can search through the dynamics of natural behaviors to find explicit evidence that these behaviors are simpler than they might have been. We review two mathematical approaches to simplification, dimensionality reduction and the maximum entropy method, and we draw on examples from different levels of biological organization, from the crawling behavior of Caenorhabditis elegans to the control of smooth pursuit eye movements in primates, and from the coding of natural scenes by networks of neurons in the retina to the rules of English spelling. In each case, we argue that the explicit search for simplicity uncovers new and unexpected features of the biological system and that the evidence for simplification gives us a language with which to phrase new questions for the next generation of experiments. The fact that similar mathematical structures succeed in taming the complexity of very different biological systems hints that there is something more general to be discovered.}, 
}
@article{Stephenson-JonesetalCURRENT-BIOLOGY-11,
       author = {Stephenson-Jones, M.  and Samuelsson, E.  and Ericsson, J.  and Robertson, B.  and Grillner, S.},
        title = {Evolutionary conservation of the basal ganglia as a common vertebrate mechanism for action selection},
      journal = {Current Biology},
         year = {2011},
       volume = {21},
       number = {13},
        pages = {1081-1091},
     abstract = {Although the basal ganglia are thought to play a key role in action selection in mammals, it is unknown whether this mammalian circuitry is present in lower vertebrates as a conserved selection mechanism. We aim here, using lamprey, to elucidate the basal ganglia circuitry in the phylogenetically oldest group of vertebrates (cyclostomes) and determine how this selection architecture evolved to accommodate the increased behavioral repertoires of advanced vertebrates.\\ We show, using immunohistochemistry, tract tracing, and whole-cell recordings, that all parts of the mammalian basal ganglia (striatum, globus pallidus interna [GPi] and externa [GPe], and subthalamic nucleus [STN]) are present in the lamprey forebrain. In addition, the circuit features, molecular markers, and physiological activity patterns are conserved. Thus, GABAergic striatal neurons expressing substance P project directly to the pallidal output layer, whereas enkephalin-expressing striatal neurons project indirectly via nuclei homologous to the GPe and STN. Moreover, pallidal output neurons tonically inhibit tectum, mesencephalic, and diencephalic motor regions.\\ These results show that the detailed basal ganglia circuitry is present in the phylogenetically oldest vertebrates and has been conserved, most likely as a mechanism for action selection used by all vertebrates, for over 560 million years. Our data also suggest that the mammalian basal ganglia evolved through a process of exaptation, where the ancestral core unit has been co-opted for multiple functions, allowing them to process cognitive, emotional, and motor information in parallel and control a broader range of behaviors.}
}
@book{Stephenson1996,
        title = {The Diamond Age},
       author = {Stephenson, Neal},
         year = {1996},
    publisher = {Bantam Books}
}
@book{SterlingandLaughlin2015,
        title = {Peter Sterling and Simon Laughlin},
       author = {Principles of Neural Design},
         year = {2015},
    publisher = {MIT Press},
     abstract = {Setting out to "reverse engineer" the brain--disassembling it to understand it--Sterling and Laughlin first consider why an animal should need a brain, tracing computational abilities from bacterium to protozoan to worm. They examine bigger brains and the advantages of "anticipatory regulation"; identify constraints on neural design and the need to "nanofy"; and demonstrate the routes to efficiency in an integrated molecular system, phototransduction. They show that the principles of neural design at finer scales and lower levels apply at larger scales and higher levels; describe neural wiring efficiency; and discuss learning as a principle of biological design that includes "save only what is needed." Sterling and Laughlin avoid speculation about how the brain might work and endeavor to make sense of what is already known. Their distinctive contribution is to gather a coherent set of basic rules and exemplify them across spatial and functional scales.}
}
@book{SterlingandShapiro86,
        title = {Qualitative Reasoning and Physical Systems},
       author = {Sterling, Leon and Shapiro, Ehud},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1986,
}
@book{Stevens1949,
        title = {An Ordinary Evening in New Haven},
       author = {Wallace Stevens},
         year = {1977},
    publisher = {Connecticut Academy of Arts \& Sciences}
}
@article{StevensetalJARMC-13,
       author = {Jeffrey R. Stevens and Thorsten Pachur and Lael J. Schooler},
        title = {Rational analysis of the adaptive and predictive nature of memory},
      journal = {Journal of Applied Research in Memory and Cognition},
       volume = {2},
       number = {4},
        pages = {251-253},
         year = {2013},
     abstract = {In his target article, Klein (2013) makes the important point that many approaches to studying memory neglect the function of memory, in particular its capacity to help predict the future. Here, we complement Klein's argument in two ways. First, we point to an existing and well-developed research program that formalizes a functional approach to memory, exploring its adaptive nature. Second, we illustrate how this approach can be applied to analyze regularities in social interactions, which memory might exploit to predict future interactions.},
}
@article{StevensonandKordingNEURON-11,
     title = {How advances in neural recording affect data analysis},
    author = {Stevenson, I.H. and Kording, K.P.},
   journal = {Nature Neuroscience},
    volume = 14,
     issue = 2,
     pages = {139-42},
      year = 2011,
  abstract = {Over the last five decades, progress in neural recording techniques has allowed the number of simultaneously recorded neurons to double approximately every seven years, mimicking Moore's law. Such exponential growth motivates us to ask how data analysis techniques are affected by progressively larger numbers of recorded neurons. Traditionally, neurons are analyzed independently on the basis of their tuning to stimuli or movement. Although tuning curve approaches are unaffected by growing numbers of simultaneously recorded neurons, newly developed techniques that analyze interactions between neurons become more accurate and more complex as the number of recorded neurons increases. Emerging data analysis techniques should consider both the computational costs and the potential for more accurate models associated with this exponential growth of the number of recorded neurons.}
}
@article{StevensonetalCOIN-08,
        title = {Inferring functional connections between neurons},
       author = {Ian H. Stevenson and James M. Rebesco and Lee E. Miller and Konrad P. Kording},
      journal = {Current Opinion in Neurobiology},
       volume = 18,
         year = 2008,
        pages = {1-7},
     abstract = {A central question in neuroscience is how interactions between neurons give rise to behavior. In many electrophysiological experiments, the activity of a set of neurons is recorded while sensory stimuli or movement tasks are varied. Tools that aim to reveal underlying interactions between neurons from such data can be extremely useful. Traditionally, neuroscientists have studied these interactions using purely descriptive statistics (cross-correlograms or joint peri-stimulus time histograms). However, the interpretation of such data is often difficult, particularly as the number of recorded neurons grows. Recent research suggests that model-based, maximum likelihood methods can improve these analyses. In addition to estimating neural interactions, application of these techniques has improved decoding of external variables, created novel interpretations of existing electrophysiological data, and may provide new insight into how the brain represents information.},
}
@book{Stewart2007,
     title = {The Courtier and the Heretic: Leibniz, Spinoza, and the Fate of God in the Modern World},
    author = {Matthew Stewart},
 publisher = {W.W. Norton \& Co.},
      year = 2007,
}
@book{Stewart2011discworld3,
        title = {{Science of Discworld III: Darwin's Watch}},
       author = {Stewart, Ian and Cohen, Jack and Pratchett, Terry},
         year = {2015},
    publisher = {Ebury Publishing}
}
@book{Stewart2011science2,
        title = {The Science Of Discworld II: The Globe},
       author = {Stewart, Ian and Cohen, Jack and Pratchett, Terry},
         year = {2015},
    publisher = {Ebury Publishing}
}
@article{StewartNATURE-88,
       author = {Stewart, Ian},
        title = {The ultimate in undecidability},
      journal = {Nature},
         year = {1988},
       volume = {332},
        issue = {6160},
        pages = {115-116},
}
@techreport {StewartNEF-12,
        title = {A Technical Overview of the Neural Engineering Framework},
       author = {Terrence C. Stewart},
         year = 2012,
          pdf = {http://compneuro.uwaterloo.ca/files/publications/stewart.2012d.pdf},
     abstract = {The Neural Engineering Framework (NEF) is a general methodology that allows you to build largescale, biologically plausible, neural models of cognition. In particular, it acts as a neural compiler: you specify the properties of the neurons, the values to be represented, and the functions to be computed, and it solves for the connection weights between components that will perform the desired functions. Importantly, this works not only for feed-forward computations, but recurrent connections as well, allowing for complex dynamical systems including integrators, oscillators, Kalman filters, and so on. It also incorporates realistic local error-driven learning rules, allowing for online adaptation and optimization of responses. The NEF has been used to model visual attention, inductive reasoning, reinforcement learning, and many other tasks. Recently, we used it to build Spaun, the world's largest functional brain model, using 2.5 million neurons to perform eight different cognitive tasks by interpreting visual input and producing hand-written output via a simulated 6-muscle arm. Our open-source software Nengo was used for all of these, and is available at http://nengo.ca, along with tutorials, demos, and downloadable models.},
  institution = {Centre for Theoretical Neuroscience},
}
@book{StewartandCohen1997figments,
        title = {Figments of Reality: The Evolution of the Curious Mind},
       author = {Stewart, Ian and Cohen, Jack},
         year = {1997},
    publisher = {Cambridge University Press}
}
@inproceedings{StewartetalCOGSCI-10,
       author = {Terrence C. Stewart and Xuan Choo and Chris Eliasmith},
        title = {Symbolic Reasoning in Spiking Neurons: A Model of the Cortex/Basal Ganglia/Thalamus Loop},
    booktitle = {Prodeedings of the 32nd Annual Meeting of the Cognitive Science Society},
    published = {Cognitive Science Society},
         year = 2010,
          url = {http://compneuro.uwaterloo.ca/files/publications/stewart.2010a.pdf}
}
@article{StewartetalFIN-12,
       author = {Stewart, Terrence C. and  Bekolay, Trevor  and  Eliasmith, Chris},
        title = {Learning to select actions with spiking neurons in the basal ganglia},
      journal = {Frontiers in Neuroscience},
       volume = 6,
        issue = 2,
         year = 2012,
     abstract = {We expand our existing spiking neuron model of decision making in the cortex and basal ganglia to include local learning on the synaptic connections between the cortex and striatum, modulated by a dopaminergic reward signal. We then compare this model to animal data in the bandit task, which is used to test rodent learning in conditions involving forced choice under rewards. Our results indicate a good match in terms of both behavioural learning results and spike patterns in the ventral striatum. The model successfully generalizes to learning the utilities of multiple actions, and can learn to choose different actions in different states. The purpose of our model is to provide both high-level behavioural predictions and low-level spike timing predictions while respecting known neurophysiology and neuroanatomy.}
}
@article{StewartetalVR-03,
       author = {Alan L. Stewart and Roger S. Pinkham and Thomas K. Bittner and Dean G. Purcell},
        title = {Hermite acuity gratings and models of space-variant acuity},
      journal = {Vision Research},
    publisher = {Elsevier},
       volume = 43,
        issue = 15,
        pages = {1597-1610},
         year = 2003,
           url = {StewartetalVR-03.pdf},
}
@article{StickgoldandWalkerNN-13,
     title = {Sleep-dependent memory triage: evolving generalization through selective processing},
    author = {Stickgold R. and Walker M. P.},
   journal = {Nature Neuroscience},
    volume = 2013, 
     issue = 16,
      year = 2,
     pages = {139-145},
  abstract = {The brain does not retain all the information it encodes in a day. Much is forgotten, and of those memories retained, their subsequent evolution can follow any of a number of pathways. Emerging data makes clear that sleep is a compelling candidate for performing many of these operations. But how does the sleeping brain know which information to preserve and which to forget? What should sleep do with that information it chooses to keep? For information that is retained, sleep can integrate it into existing memory networks, look for common patterns and distill overarching rules, or simply stabilize and strengthen the memory exactly as it was learned. We suggest such 'memory triage' lies at the heart of a sleep-dependent memory processing system that selects new information, in a discriminatory manner, and assimilates it into the brain's vast armamentarium of evolving knowledge, helping guide each organism through its own, unique life.}
}
@article{StickgoldetalPNAS-02,
       author = {Stickgold, Robert and Fosse, Roar and Walker, Matthew P.},
        title = {Linking brain and behavior in sleep-dependent learning and memory consolidation},
    publisher = {National Academy of Sciences},
       volume = {99},
       number = {26},
         year = {2002},
        pages = {16519-16521},
      journal = {Proceedings of the National Academy of Sciences},
     abstract = {}
}
@incollection{StilesandBartol2001,
       author = {Stiles, J.R. and Bartol, T.M.},
        title = {{Monte Carlo} methods for simulating realistic synaptic microphysiology using {MCell}},
    booktitle = {Computational Neuroscience: Realistic Modeling for Experimentalists},
       editor = {De Schutter, E.},
    publisher = {CRC Press},
      address = {Boca Raton},
         year = {2001},
        pages = {87-127},
}
@incollection{StilesetalMCELL-01,
      title = {Synaptic Variability: New Insights from Reconstructions and {Monte Carlo} Simulations with {MCell} Synapses},
     author = {Joel R. Stiles and Thomas M. Bartol and Miriam M. Salpeter and Edwin E. Salpeter, Terrence J. Sejnowski},
  booktitle = {Synapses},
     editor = {W. Maxwell Cowan and Thomas C. Sudhof and Charles F. Stevens},
  publisher = {Johns Hopkins University Press},
    address = {Baltimore, MD},
       year = 2001,
      pages = {681-731}
}
@inproceedings{Stillman90,
       author = {Stillman, Jonathan},
        title = {On heuristics for finding loop cutsets in multiply connected belief networks},
    booktitle = {Proceedings of the Sixth Conference on Uncertainty in Artificial Intelligence},
     location = {Cambridge, Massachusetts},
        pages = {265-272},
         year = 1990,
}
@article {StirmanetalBIORXIV-15,
       author = {Stirman, Jeffrey N. and Smith, Ikuko T. and Kudenov, Michael W. and Smith, Spencer L.},
        title = {Wide field-of-view, twin-region two-photon imaging across extended cortical networks},
      journal = {bioRxiv},
         year = {2015},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {We demonstrate a two-photon imaging system with corrected optics including a custom objective that provides cellular resolution across a 3.5 mm field of view (9.6 mm^2). Temporally multiplexed excitation pathways can be independently repositioned in XY and Z to simultaneously image regions within the expanded field of view. We used this new imaging system to measure activity correlations between neurons in different cortical areas in awake mice.},
}
@article{StoccoetalPR-10,
       author = {Stocco, Andrea and Lebiere, Christian and Anderson, John R.},
        title = {Conditional Routing of Information to the Cortex: A Model of the Basal Ganglia's Role in Cognitive Coordination},
      journal = {Psychology Review},
         year = {2010},
       volume = {117},
        issue = {2},
        pages = {541-574},
     abstract = {The basal ganglia play a central role in cognition and are involved in such general functions as action selection and reinforcement learning. Here, we present a model exploring the hypothesis that the basal ganglia implement a conditional information-routing system. The system directs the transmission of cortical signals between pairs of regions by manipulating separately the selection of sources and destinations of information transfers. We suggest that such a mechanism provides an account for several cognitive functions of the basal ganglia. The model also incorporates a possible mechanism by which subsequent transfers of information control the release of dopamine. This signal is used to produce novel stimulus-response associations by internalizing transferred cortical representations in the striatum. We discuss how the model is related to production systems and cognitive architectures. A series of simulations is presented to illustrate how the model can perform simple stimulus-response tasks, develop automatic behaviors, and provide an account of impairments in Parkinson's and Huntington's diseases.},
}
@article{StoccoetalPSYCHOLOGY-REVIEW-10,
       author = {Stocco, A. and Lebiere, C. and Anderson, J. R.},
        title = {{C}onditional routing of information to the cortex: a model of the basal ganglia's role in cognitive coordination},
      journal = {Psychology Review},
         year = {2010},
       volume = {117},
       number = {2},
        pages = {541-574},
     abstract = {The basal ganglia play a central role in cognition and are involved in such general functions as action selection and reinforcement learning. Here, we present a model exploring the hypothesis that the basal ganglia implement a conditional information-routing system. The system directs the transmission of cortical signals between pairs of regions by manipulating separately the selection of sources and destinations of information transfers. We suggest that such a mechanism provides an account for several cognitive functions of the basal ganglia. The model also incorporates a possible mechanism by which subsequent transfers of information control the release of dopamine. This signal is used to produce novel stimulus-response associations by internalizing transferred cortical representations in the striatum. We discuss how the model is related to production systems and cognitive architectures. A series of simulations is presented to illustrate how the model can perform simple stimulus-response tasks, develop automatic behaviors, and provide an account of impairments in Parkinson's and Huntington's diseases.}
}
@incollection{StolckeandOmohundroNIPS-93,
       author = {Stolcke, Andreas and Omohundro, Stephen},
        title = {Hidden {M}arkov Model Induction by {B}ayesian Model Merging},
       editor = {Hanson, S. J. and Cowan, J. D. and Giles, C. L.},
    booktitle = {Advances in Neural Information Processing},
       volume = 5,
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
        pages = {11-18},
         year = 1993,
}
@techreport{StolckeandOmohundroTR-94,
       author = {Stolcke, Andreas and Omohundro, Stephen},
        title = {Best-First Model Merging for Hidden {M}arkov Model Induction},
  institution = {International Computer Science Institute},
        issue = {TR-94-003},
      address = {Berkeley, California},
         year = 1994,
}
@article{StolckeetalACL-00,
     title = {Dialogue Act Modeling for Automatic Tagging and Recognition of Conversational Speech},
    author = {Stolcke, Andreas and Ries, Klaus and Coccaro, Noah and Shriberg, Elizabeth and Bates, Rebecca A. and Jurafsky, Daniel and Taylor, Paul and Martin, Rachel and Ess-Dykema, Carol Van and Meteer, Marie},
   journal = {Computational Linguistics},
    volume = 26,
     issue = 3,
     pages = {339-373},
}
@book{Stollnitzetal96,
       author = {Eric J. Stollnitz and Tony D. Derose and David H. Salesin},
        title = {Wavelets for computer graphics: theory and applications},
         year = 1996,
     publisher = {Morgan Kaufmann Publishers Inc.},
      address = {San Francisco, CA, USA},
}
@inproceedings{StougiannisetalSIGMOD-13,
     author = {Alexandros Stougiannis and Farhan Tauheed and Mirjana Pavlovic and Thomas Heinis and Anastasia Ailamak},
      title = {Data-driven Neuroscience: Enabling Breakthroughs Via Innovative Data Management},
  booktitle = {{ACM} {SIGMOD} International Conference on Management of Data},
  publisher = {ACM},
       year = 2013,
}
@article{StowersetalNATURE-MERHODS-17,
       author = {Stowers, J. R. and Hofbauer, M. and Bastien, R. and Griessner, J. and Higgins, P. and Farooqui, S. and Fischer, R. M. and Nowikovsky, K. and Haubensak, W. and Couzin, I. D.  and Tessmar-Raible, K. and Straw, A. D.},
        title = {Virtual reality for freely moving animals},
      journal = {Nature Methods},
         year = {2017},
     abstract = {Standard animal behavior paradigms incompletely mimic nature and thus limit our understanding of behavior and brain function. Virtual reality (VR) can help, but it poses challenges. Typical VR systems require movement restrictions but disrupt sensorimotor experience, causing neuronal and behavioral alterations. We report the development of FreemoVR, a VR system for freely moving animals. We validate immersive VR for mice, flies, and zebrafish. FreemoVR allows instant, disruption-free environmental reconfigurations and interactions between real organisms and computer-controlled agents. Using the FreemoVR platform, we established a height-aversion assay in mice and studied visuomotor effects in Drosophila and zebrafish. Furthermore, by photorealistically mimicking zebrafish we discovered that effective social influence depends on a prospective leader balancing its internally preferred directional choice with social interaction. FreemoVR technology facilitates detailed investigations into neural function and behavior through the precise manipulation of sensorimotor feedback loops in unrestrained animals.}
}
@techreport{StraussTR-86,
       author = {Strauss, Paul S.},
        title = {A Tutorial Guide to the {SCEFO} Language},
  institution = {Brown University Department of Computer Science},
         year = 1986,
}
@article{StreetetalAPP-18,
       author = {Street, C. N. H. and Bischof, W. F. and Kingstone, A.},
        title = {{{P}erspective taking and theory of mind in hide and seek}},
      journal = {Attention Perception Psychophysics},
       volume = {80},
       number = {1},
         year = {2018},
        pages = {21-26},
     abstract = {Does theory of mind play a significant role in where people choose to hide an item or where they search for an item that has been hidden? Adapting Anderson's "Hide-Find Paradigm" Anderson et al. (Action, Perception and Performance, 76, 907-913, 2014) participants viewed homogenous or popout visual arrays on a touchscreen table. Their task was to indicate where in the array they would hide an item, or to search for an item that had been hidden, by either a friend or a foe. Critically, participants believed that their sitting location at the touchtable was the same as-or opposite to-their partner's location. Replicating Anderson et al., participants tended to (1) select items nearer to themselves on homogenous displays, and this bias was stronger for a friend than foe; and (2) select popout items, and again, more for a friend than foe. These biases were observed only when participants believed that they shared the same physical perspective as their partner. Collectively, the data indicate that theory of mind plays a significant role in hiding and finding, and demonstrate that the hide-find paradigm is a powerful tool for investigating theory of mind in adults.}
}
@article{StricketalARN-09,
       author = {Peter L. Strick and Richard P. Dum and Julie A. Fiez},
        title = {Cerebellum and Nonmotor Function},
      journal = {Annual Review of Neuroscience},
       volume = {32},
       number = {1},
        pages = {413-434},
         year = {2009},
     abstract = {Does the cerebellum influence nonmotor behavior? Recent anatomical studies demonstrate that the output of the cerebellum targets multiple nonmotor areas in the prefrontal and posterior parietal cortex, as well as the cortical motor areas. The projections to different cortical areas originate from distinct output channels within the cerebellar nuclei. The cerebral cortical area that is the main target of each output channel is a major source of input to the channel. Thus, a closed-loop circuit represents the major architectural unit of cerebro-cerebellar interactions. The outputs of these loops provide the cerebellum with the anatomical substrate to influence the control of movement and cognition. Neuroimaging and neuropsychological data supply compelling support for this view. The range of tasks associated with cerebellar activation is remarkable and includes tasks designed to assess attention, executive control, language, working memory, learning, pain, emotion, and addiction. These data, along with the revelations about cerebro-cerebellar circuitry, provide a new framework for exploring the contribution of the cerebellum to diverse aspects of behavior. }
}
@book{StriedterandNorthcutt2020brainsthroughtime,
       author = {Georg F. Striedter and R. Glenn Northcutt},
        title = {Brains Through Time: A Natural History of Vertebrates},
    publisher = {Oxford University Press},
         year = {2020},
}
@article{StrocketalNC-20,
       author = {Strock, Anthony and Hinaut, Xavier and Rougier, Nicolas P.},
        title = {A Robust Model of Gated Working Memory},
      journal = {Neural Computation},
       volume = {32},
       number = {1},
        pages = {153-181},
         year = {2020},
     abstract = {Gated working memory is defined as the capacity of holding arbitrary information at any time in order to be used at a later time. Based on electrophysiological recordings, several computational models have tackled the problem using dedicated and explicit mechanisms. We propose instead to consider an implicit mechanism based on a random recurrent neural network. We introduce a robust yet simple reservoir model of gated working memory with instantaneous updates. The model is able to store an arbitrary real value at random time over an extended period of time. The dynamics of the model is a line attractor that learns to exploit reentry and a nonlinearity during the training phase using only a few representative values. A deeper study of the model shows that there is actually a large range of hyperparameters for which the results hold (e.g., number of neurons, sparsity, global weight scaling) such that any large enough population, mixing excitatory and inhibitory neurons, can quickly learn to realize such gated working memory. In a nutshell, with a minimal set of hypotheses, we show that we can have a robust model of working memory. This suggests this property could be an implicit property of any random population, that can be acquired through learning. Furthermore, considering working memory to be a physically open but functionally closed system, we give account on some counterintuitive electrophysiological recordings.}
}
@book{Strogatz2002,
    author = {Steven H. Strogatz},
     title = {Nonlinear Dynamics And Chaos: With Applications To Physics, Biology, Chemistry, And Engineering},
 publisher = {Wiley},
    address = {New York},
      year = 2002,
}
@book{Strogatz2014,
        title = {Nonlinear Dynamics and Chaos: With Applications to Physics, Biology, Chemistry, and Engineering},
       author = {Strogatz, Steven H.},
 series       = {Studies in Nonlinearity},
    publisher = {Westview Press},
         year = {2014},
}
@article{StrosniderandPaulAIM-95,
        title = {A Structured View of Real-Time Problem Solving},
       author = {Strosnider, Jay. K. and Paul, C. J.},
      journal = {AI Magazine},
       volume = {Summer},
        pages = {45-66},
         year = 1995,
}
@book{StrossACCELERANDO-05,
        title = {Accelerando},
       author = {Stross, Charles},
    publisher = {Ace Books},
         year = {2005},
}
@incollection{StrossLOBSTERS-97,
        title = {Lobsters},
       author = {Stross, Charles},
    booktitle = {Toast},
    publisher = {Wildside Press},
         year = {2006},
        pages = {2017-2044}
}
@book{StrossTOAST-00,
        title = {Toast},
       author = {Stross, Charles},
    publisher = {Wildside Press},
         year = {2006},
}
@article{StrotheretalCURRENT-BIOLOGY-14,
       author = {James A. Strother and Aljoscha Nern and Michael B. Reiser},
        title = {Direct Observation of {ON} and {OFF} Pathways in the Drosophila Visual System},
      journal = {Current Biology},
       volume = {24},
       number = {9},
         year = {2014},
        pages = {976-983},
     abstract = {Summary Visual motion perception is critical to many animal behaviors, and flies have emerged as a powerful model system for exploring this fundamental neural computation. Although numerous studies have suggested that fly motion vision is governed by a simple neural circuit, the implementation of this circuit has remained mysterious for decades. Connectomics and neurogenetics have produced a surge in recent progress, and several studies have shown selectivity for light increments (ON) or decrements (OFF) in key elements associated with this circuit. However, related studies have reached disparate conclusions about where this selectivity emerges and whether it plays a major role in motion vision. To address these questions, we examined activity in the neuropil thought to be responsible for visual motion detection, the medulla, of Drosophila melanogaster in response to a range of visual stimuli using two-photon calcium imaging. We confirmed that the input neurons of the medulla, the LMCs, are not responsible for light-on and light-off selectivity. We then examined the pan-neural response of medulla neurons and found prominent selectivity for light-on and light-off in layers of the medulla associated with two anatomically derived pathways (L1/L2 associated). We next examined the activity of prominent interneurons within each pathway (Mi1 and Tm1) and found that these neurons have corresponding selectivity for light-on or light-off. These results provide direct evidence that motion is computed in parallel light-on and light-off pathways, demonstrate that this selectivity emerges in neurons immediately downstream of the LMCs, and specify where crucial elements of motion computation occur.}
}
@article{StrotheretalNEURON-17,
       author = {James A. Strother and Shiuan-Tze Wu and Allan M. Wong and Aljoscha Nern and Edward M. Rogers and Jasmine Q. Le and Gerald M. Rubin and Michael B. Reiser},
        title = {The Emergence of Directional Selectivity in the Visual Motion Pathway of Drosophila},
      journal = {Neuron},
       volume = {94},
       number = {1},
         year = {2017},
        pages = {168-182},
     abstract = {Summary The perception of visual motion is critical for animal navigation, and flies are a prominent model system for exploring this neural computation. In Drosophila, the \{T4\} cells of the medulla are directionally selective and necessary for \{ON\} motion behavioral responses. To examine the emergence of directional selectivity, we developed genetic driver lines for the neuron types with the most synapses onto \{T4\} cells. Using calcium imaging, we found that these neuron types are not directionally selective and that selectivity arises in the \{T4\} dendrites. By silencing each input neuron type, we identified which neurons are necessary for \{T4\} directional selectivity and \{ON\} motion behavioral responses. We then determined the sign of the connections between these neurons and \{T4\} cells using neuronal photoactivation. Our results indicate a computational architecture for motion detection that is a hybrid of classic theoretical models.}
}
@misc{StrouseandSchwabVDIB-18,
        title = {Variational Deterministic Information Bottleneck},
       author = {D.J. Strouse and David Schwab},
         year = {2018},
 howpublished = {http://djstrouse.com/downloads/vdib.pdf},
     abstract = {Alemi et al. [2016] recently introduced a variational version of the information bottleneck (IB) [Tishby et al., 1999]. Here we introduce a variational version of the deterministic information bottleneck (DIB) [Strouse and Schwab, 2017].}
}
@article{StrukovandWilliamsPNAS-09,
       author = {Strukov, D. B. and Williams, R. S.},
        title = {Four-dimensional address topology for circuits with stacked multilayer crossbar arrays},
      journal = {Proceedings of the National Academy of Sciences},
         year = 2009,
       volume = 106,
        issue = 48,
        pages = {20155-20158},
     abstract = {We present a topological framework that provides a simple yet powerful electronic circuit architecture for constructing and using multilayer crossbar arrays, allowing a significantly increased integration density of memristive crosspoint devices beyond the scaling limits of lateral feature sizes. The truly remarkable feature of such circuits, which is an extension of the CMOL (Cmos + MOLecular-scale devices) concept for an area-like interface to a three-dimensional system, is that a large-feature-size complimentary metal-oxide-semiconductor (CMOS) substrate can provide high-density interconnects to multiple crossbar layers through a single set of vertical vias. The physical locations of the memristive devices are mapped to a four-dimensional logical address space such that unique access from the CMOS substrate is provided to every device in a stacked array of crossbars. This hybrid architecture is compatible with digital memories, field-programmable gate arrays, and biologically inspired adaptive networks and with state-of-the-art integrated circuit foundries.}
}
@techreport{StuartTR-85,
       author = {Stuart, Christopher},
        title = {An Implementation of a MultiAgent Plan Synchronizer Using a Temporal Logic Theorem Prover},
        issue = {Technical Note 350},
  institution = {SRI International},
         year = 1985,
}
@article{StuhlmuellerandGoodmanCS-13,
       author = {Stuhlm\"{u}ller, A. and Goodman, N. D.},
      journal = {Journal of Cognitive Systems Research},
        title = {Reasoning about Reasoning by Nested Conditioning: Modeling Theory of Mind with Probabilistic Programs},
       volume = {28},
         year = {2014},
        pages = {80-99},
     abstract = {A wide range of human reasoning patterns can be explained as conditioning in probabilistic models; however, conditioning has traditionally been viewed as an operation applied to such models, not represented in such models. We describe how probabilistic programs can explicitly represent conditioning as part of a model. This enables us to describe reasoning about others' reasoning using nested conditioning. Much of human reasoning is about the beliefs, desires, and intentions of other people; we use probabilistic programs to formalize these inferences in a way that captures the flexibility and inherent uncertainty of reasoning about other agents. We express examples from game theory, artificial intelligence, and linguistics as recursive probabilistic programs and illustrate how this representation language makes it easy to explore new directions in each of these fields. We discuss the algorithmic challenges posed by these kinds of models and describe how Dynamic Programming techniques can help address these challenges.}
}
@misc{SturmTCOLORBOX-19,
        title = {{TCOLORBOX Manual Version 4.22}},
       author = {Thomas F. Sturm},
         year = {2019},
 howpublished = {Manual for the LaTeX tcolorbox package version 4.22 November 11, 2019}
}
@article{SuMenuzandCarlsonCELL-09,
       author = {Su, C. Y. and Menuz, K. and Carlson, J. R.},
        title = {{{O}lfactory perception: receptors, cells, and circuits}},
      journal = {Cell},
       volume = {139},
       number = {1},
         year = {2009},
        pages = {45-59},
     abstract = {Remarkable advances in our understanding of olfactory perception have been made in recent years, including the discovery of new mechanisms of olfactory signaling and new principles of olfactory processing. Here, we discuss the insight that has been gained into the receptors, cells, and circuits that underlie the sense of smell.}
}
@article{SubiauletalSCIENCE-04,
       author = {Subiaul, Francys and Cantlon, Jessica F. and Holloway, Ralph L. and Terrace, Herbert S.},
        title = {Cognitive Imitation in Rhesus Macaques},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {305},
       number = {5682},
        pages = {407-410},
         year = {2004},
     abstract = {Experiments on imitation typically evaluate a student's ability to copy some feature of an expert's motor behavior. Here, we describe a type of observational learning in which a student copies a cognitive rule rather than a specific motor action. Two rhesus macaques were trained to respond, in a prescribed order, to different sets of photographs that were displayed on a touch-sensitive monitor. Because the position of the photographs varied randomly from trial to trial, sequences could not be learned by motor imitation. Both monkeys learned new sequences more rapidly after observing an expert execute those sequences than when they had to learn new sequences entirely by trial and error.},
}
@techreport{SubrahmoniaetalTR-92,
       author = {Subrahmonia, J. and Cooper, D. B. and Keren, D.},
        title = {Practical Reliable {B}ayesian Recognition of {2-D} and {3D} Objects Using Implicit Polynomials and Algebraic Invariants},
  institution = {Brown University},
         year = 1992,
}
@article{SubramaniametalVC-00,
     title = {Accurate identification but no priming and chance recognition memory for pictures in {RSVP} sequences}, 
    author = {Subramaniam, S. and Biederman, I. and Madigan, S. A.},
   journal = {Visual Cognition},
    volume = 7,
     issue = 4,
      year = 1975,
     pages = {511-535},
}
@article{SuddendorfetalPTRS_B-09,
       author = {Suddendorf, Thomas and Addis, Donna Rose and Corballis, Michael C.},
        title = {Mental time travel and the shaping of the human mind},
      journal = {Philosophical Transactions Royal Society London B Biological Science},
    publisher = {The Royal Society},
       volume = {364},
        issue = {1521},
         year = {2009},
        pages = {1317-1324},
     abstract = {Episodic memory, enabling conscious recollection of past episodes, can be distinguished from semantic memory, which stores enduring facts about the world. Episodic memory shares a core neural network with the simulation of future episodes, enabling mental time travel into both the past and the future. The notion that there might be something distinctly human about mental time travel has provoked ingenious attempts to demonstrate episodic memory or future simulation in non-human animals, but we argue that they have not yet established a capacity comparable to the human faculty. The evolution of the capacity to simulate possible future events, based on episodic memory, enhanced fitness by enabling action in preparation of different possible scenarios that increased present or future survival and reproduction chances. Human language may have evolved in the first instance for the sharing of past and planned future events, and, indeed, fictional ones, further enhancing fitness in social settings.},
}
@incollection{SudderthandJordanNIPS-08,
      title = {Shared Segmentation of Natural Scenes Using Dependent {P}itman-{Y}or Processes},
     author = {Erik Sudderth and Michael Jordan},
  booktitle = {Advances in Neural Information Processing Systems 21},
     editor = {D. Koller and D. Schuurmans and Y. Bengio and L. Bottou},
  publisher = {MIT Press},
      pages = {1585-1592},
       year = 2009
}
@inproceedings{SudderthetalCVPR-03,
       author = {Erik Sudderth and Alexander Ihler and William Freeman and Alan Willsky},
        title = {Nonparametric belief propagation},
    booktitle = {Proceedings of the 2003 {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
       volume = 1,
        pages = {605-612},
         year = 2003,
}
@article{SudderthetalIJCV-08,
     author = {Sudderth, Erik B. and Torralba, Antonio and Freeman, William T. and Willsky, Alan S.},
      title = {Describing Visual Scenes Using Transformed Objects and Parts},
    journal = {International Journal of Computer Vision},
  publisher = {Kluwer Academic Publishers},
     volume = 77,
      issue = {1-3},
       year = 2008,
      pages = {291-330},
}
@inproceedings{SudderthetalNIPS-05,
       author = {Erik B. Sudderth and Antonio B. Torralba and William T. Freeman and Alan S. Willsky},
        title = {Describing Visual Scenes using Transformed Dirichlet Processes.},
    booktitle = {Advances in Neural Information Processing Systems 18},
     location = {Vancouver, British Columbia},
         year = 2005,
}
@inproceedings{SuelandYuan01,
       author = {T. Suel and J. Yuan},
        title = {Compressing the graph structure of the {W}eb},
    booktitle = {Proceedings of the 2001 Data Compression Conference},
        pages = {213-222},
         year = 2001,
}
@inproceedings{Sueletal03,
       author = {T. Suel and C. Mathur and J. Wu and J. Zhang and A. Delis and M. Kharrazi and X. Long and K. Shanmugasunderam},
        title = {{ODISSEA}: A Peer-to-Peer Architecture for Scalable Web Search and Information Retrieval},
    booktitle = {6th International Workshop on the Web and Databases (WebDB)},
         year = 2003,
}
@article{SuermondCooperIJAR-90,
       author = {Suermondt, H. Jacques and Cooper, Gregory F.},
        title = {Probabilistic inference in multiply connected belief networks using loop cutsets},
      journal = {International Journal of Approximate Reasoning},
       volume = 4,
         year = 1990,
        pages = {283-306},
}
@inproceedings{SuermondtUAI-88,
       author = {Suermondt, H. J. and Cooper, G. F.},
        title = {Updating probabilities in multiply-connected belief networks},
    booktitle = UAI-88,
     location = {Minneapolis, Minnesota},
        pages = {335-343},
         year = 1988,
}
@article{SuetalBIODATA-13,
       author = {Su, Chengwei and Andrew, Angeline and Karagas, Margaret R. and Borsuk, Mark E.},
        title = {Using Bayesian networks to discover relations between genes, environment, and disease},
      journal = {BioData Mining},
    publisher = {BioMed Central},
       volume = {6},
         year = {2013},
        pages = {6-6},
     abstract = {We review the applicability of Bayesian networks (BNs) for discovering relations between genes, environment, and disease. By translating probabilistic dependencies among variables into graphical models and vice versa, BNs provide a comprehensible and modular framework for representing complex systems. We first describe the Bayesian network approach and its applicability to understanding the genetic and environmental basis of disease. We then describe a variety of algorithms for learning the structure of a network from observational data. Because of their relevance to real-world applications, the topics of missing data and causal interpretation are emphasized. The BN approach is then exemplified through application to data from a population-based study of bladder cancer in New Hampshire, USA. For didactical purposes, we intentionally keep this example simple. When applied to complete data records, we find only minor differences in the performance and results of different algorithms. Subsequent incorporation of partial records through application of the EM algorithm gives us greater power to detect relations. Allowing for network structures that depart from a strict causal interpretation also enhances our ability to discover complex associations including gene-gene (epistasis) and gene-environment interactions. While BNs are already powerful tools for the genetic dissection of disease and generation of prognostic models, there remain some conceptual and computational challenges. These include the proper handling of continuous variables and unmeasured factors, the explicit incorporation of prior knowledge, and the evaluation and communication of the robustness of substantive conclusions to alternative assumptions and data manifestations.},
}
@article{SuetalCoRR-16,
       author = {Pei{-}Hao Su and Milica Gasic and Nikola Mrksic and Lina Maria Rojas{-}Barahona and Stefan Ultes and David Vandyke and Tsung{-}Hsien Wen and Steve J. Young},
        title = {On-line Active Reward Learning for Policy Optimisation in Spoken Dialogue Systems},
      journal = {CoRR},
       volume = {arXiv:1605.07669},
         year = {2016},
     abstract = {The ability to compute an accurate reward function is essential for optimising a dialogue policy via reinforcement learning. In real-world applications, using explicit user feedback as the reward signal is often unreliable and costly to collect. This problem can be mitigated if the user's intent is known in advance or data is available to pre-train a task success predictor off-line. In practice neither of these apply for most real world applications. Here we propose an on-line learning framework whereby the dialogue policy is jointly trained alongside the reward model via active learning with a Gaussian process model. This Gaussian process operates on a continuous space dialogue representation generated in an unsupervised fashion using a recurrent neural network encoder-decoder. The experimental results demonstrate that the proposed framework is able to significantly reduce data annotation costs and mitigate noisy user feedback in dialogue policy learning.},
}
@article{SugaharaetalDGD-17,
       author = {Sugahara, Fumiaki and Murakami, Yasunori and Pascual-Anaya, Juan and Kuratani, Shigeru},
        title = {Reconstructing the ancestral vertebrate brain},
      journal = {Development, Growth \& Differentiation},
       volume = {59},
       number = {4},
         year = {2017},
        pages = {163-174},
     abstract = {Highly complicated morphologies and sophisticated functions of vertebrate brains have been established through evolution. However, the origin and early evolutionary history of the brain remain elusive, owing to lack of information regarding the brain architecture of extant and fossil species of jawless vertebrates (agnathans). Comparative analyses of the brain of less studied cyclostomes (only extant agnathan group, consisting of lampreys and hagfish) with the well-known sister group of jawed vertebrates (gnathostomes) are the only tools we have available to illustrate the ancestral architecture of the vertebrate brain. Previous developmental studies had shown that the lamprey lacked well-established brain compartments that are present in gnathostomes, such as the medial ganglionic eminence and the rhombic lip. The most accepted scenario suggested that cyclostomes had fewer compartments than that of the gnathostome brain and that gnathostomes thus evolved by a stepwise addition of innovations on its developmental sequence. However, recent studies have revealed that these compartments are present in hagfish embryos, indicating that these brain regions have been acquired before the split of cyclostomes and gnathostomes. By comparing two cyclostome lineages and gnathostomes, it has become possible to speculate about a more complex ancestral state of the brain, excluding derived traits in either of the lineages. In this review, we summarize recent studies on the brain development of the lamprey and hagfish. Then, we attempt to reconstruct the possible brain architecture of the last common ancestor of vertebrates. Finally, we discuss how the developmental plan of the vertebrate brain has been modified independently in different vertebrate lineages.},
}
@misc{SugdenMSc,
       author = {Daniel Sugden},
        title = {A novel dual modeling method for characterizing human nerve fiber activation},
         year = 2014,
 howpublished = {M.Sc. Biomedical Engineering, California Polytechnic State University},
     abstract = {Presented in this work is the investigation and successful illustration of a coupled model of the human nerve fiber. SPICE netlist code was utilized to describe the electrical properties of the human nervous membrane in tandem with COMSOL Multiphysics, a finite element analysis software tool. The initial research concentrated on the utilization of the Hodgkin-Huxley electrical circuit representation of the nerve fiber membrane. Further development of the project identified the need for a linear circuit model that more closely resembled the McNeal linearization model augmented by the work of Szlavik which better facilitated the coupling of both SPICE and COMSOL programs. Related literature was investigated and applied to validate the model. This combination of analysis tools allowed for the presentation of a consistent model and revealed that a coupled model produced not only a qualitatively comparable, but also a quantitatively comparable result to studies presented in the literature. All potential profiles produced during the simulation were compared against the literature in order to meet the purpose of presenting an advanced computational model of human neural recruitment and excitation. It was demonstrated through this process that the correct usage of neuron models within a two dimensional conductive space did allow for the approximate modeling of human neural electrical characteristics.}
}
@article{SugiharaetalSCIENCE-12,
       author = {Sugihara, G. and May, R. and Ye, H. and Hsieh, C. H. and Deyle, E. and Fogarty, M. and Munch, S.},
        title = {Detecting causality in complex ecosystems},
      journal = {Science},
         year = 2012,
       volume = 338,
        issue = 6106,
        pages = {496-500},
     abstract = {Identifying causal networks is important for effective policy and management recommendations on climate, epidemiology, financial regulation, and much else. We introduce a method, based on nonlinear state space reconstruction, that can distinguish causality from correlation. It extends to nonseparable weakly connected dynamic systems (cases not covered by the current Granger causality paradigm). The approach is illustrated both by simple models (where, in contrast to the real world, we know the underlying equations/relations and so can check the validity of our method) and by application to real ecological systems, including the controversial sardine-anchovy-temperature problem.}
}
@article{SukharevandSachsJCS-12,
       author = {Sukharev, S. and Sachs, F.},
        title = {Molecular force transduction by ion channels: diversity and unifying principles},
      journal = {Journal Cellular Science},
         year = 2012,
       volume = 125,
        issue = 13,
        pages = {3075-3083},
     abstract = {Cells perceive force through a variety of molecular sensors, of which the mechanosensitive ion channels are the most efficient and act the fastest. These channels apparently evolved to prevent osmotic lysis of the cell as a result of metabolite accumulation and/or external changes in osmolarity. From this simple beginning, nature developed specific mechanosensitive enzymes that allow us to hear, maintain balance, feel touch and regulate many systemic variables, such as blood pressure. For a channel to be mechanosensitive it needs to respond to mechanical stresses by changing its shape between the closed and open states. In that way, forces within the lipid bilayer or within a protein link can do work on the channel and stabilize its state. Ion channels have the highest turnover rates of all enzymes, and they can act as both sensors and effectors, providing the necessary fluxes to relieve osmotic pressure, shift the membrane potential or initiate chemical signaling. In this Commentary, we focus on the common mechanisms by which mechanical forces and the local environment can regulate membrane protein structure, and more specifically, mechanosensitive ion channels}
}
@article{SukhbaataretalCoRR-15,
       author = {Sainbayar Sukhbaatar and Arthur Szlam and Jason Weston and Rob Fergus},
        title = {Weakly Supervised Memory Networks},
      journal = {CoRR},
       volume = {arXiv:1503.08895},
         year = {2015},
     abstract = {We introduce a neural network with a recurrent attention model over a possibly large external memory. The architecture is a form of Memory Network (Weston et al., 2015) but unlike the model in that work, it is trained end-to-end, and hence requires significantly less supervision during training, making it more generally applicable in realistic settings. It can also be seen as an extension of RNNsearch to the case where multiple computational steps (hops) are performed per output symbol. The flexibility of the model allows us to apply it to tasks as diverse as (synthetic) question answering and to language modeling. For the former our approach is competitive with Memory Networks, but with less supervision. For the latter, on the Penn TreeBank and Text8 datasets our approach demonstrates comparable performance to RNNs and LSTMs. In both cases we show that the key concept of multiple computational hops yields improved results.}
}
@incollection{SukhbaataretalNIPS-15,
        title = {End-To-End Memory Networks},
       author = {Sukhbaatar, Sainbayar and Szlam, Arthur and Weston, Jason and Fergus, Rob},
    booktitle = {Advances in Neural Information Processing Systems 28},
       editor = {C. Cortes and N. D. Lawrence and D. D. Lee and M. Sugiyama and R. Garnett},
        pages = {2440-2448},
         year = {2015},
    publisher = {Curran Associates, Inc.},
     abstract = {We introduce a neural network with a recurrent attention model over a possibly large external memory. The architecture is a form of Memory Network (Weston et al., 2015) but unlike the model in that work, it is trained end-to-end, and hence requires significantly less supervision during training, making it more generally applicable in realistic settings. It can also be seen as an extension of RNNsearch to the case where multiple computational steps (hops) are performed per output symbol. The flexibility of the model allows us to apply it to tasks as diverse as (synthetic) question answering and to language modeling. For the former our approach is competitive with Memory Networks, but with less supervision. For the latter, on the Penn TreeBank and Text8 datasets our approach demonstrates comparable performance to RNNs and LSTMs. In both cases we show that the key concept of multiple computational hops yields improved results.}
}
@article{SultanetalNEURON-15,
       author = {Sebastien Sultan and Liyi Li and Jonathan Moss and Francesco Petrelli and Frederic Cassa and Elias Gebara and Jan Lopatar and Frank W. Pfrieger and Paola Bezzi and Josef Bischofberger and Nicolas Toni},
        title = {Synaptic Integration of Adult-Born Hippocampal Neurons Is Locally Controlled by Astrocytes},
      journal = {Neuron},
       volume = {In press.},
        issue = {00},
         year = {2015},
        pages = {000-000},
     abstract = {Adult neurogenesis is regulated by the neurogenic niche, through mechanisms that remain poorly defined. Here, we investigated whether niche-constituting astrocytes influence the maturation of adult-born hippocampal neurons using two independent transgenic approaches to block vesicular release from astrocytes. In these models, adult-born neurons but not mature neurons showed reduced glutamatergic synaptic input and dendritic spine density that was accompanied with lower functional integration and cell survival. By taking advantage of the mosaic expression of transgenes in astrocytes, we found that spine density was reduced exclusively in segments intersecting blocked astrocytes, revealing an extrinsic, local control of spine formation. Defects in \{NMDA\} receptor (NMDAR)-mediated synaptic transmission and dendrite maturation were partially restored by exogenous D-serine, whose extracellular level was decreased in transgenic models. Together, these results reveal a critical role for adult astrocytes in local dendritic spine maturation, which is necessary for the NMDAR-dependent functional integration of newborn neurons.}
}
@article{SumbuletalFiN-14,
       author = {S\"{u}mb\"{u}l, Uygar  and  Zlateski, Aleksandar  and  Vishwanathan, Ashwin  and  Masland, Richard H. and  Seung, H. Sebastian},
        title = {Automated computation of arbor densities: a step toward identifying neuronal cell types},
      journal = {Frontiers in Neuroanatomy},
       volume = 8,
         year = 2014,
        issue = 139,
      abstract = {The shape and position of a neuron convey information regarding its molecular and functional identity. The identification of cell types from structure, a classic method, relies on the time-consuming step of arbor tracing. However, as genetic tools and imaging methods make data-driven approaches to neuronal circuit analysis feasible, the need for automated processing increases. Here, we first establish that mouse retinal ganglion cell types can be as precise about distributing their arbor volumes across the inner plexiform layer as they are about distributing the skeletons of the arbors. Then, we describe an automated approach to computing the spatial distribution of the dendritic arbors, or arbor density, with respect to a global depth coordinate based on this observation. Our method involves three-dimensional reconstruction of neuronal arbors by a supervised machine learning algorithm, post-processing of the enhanced stacks to remove somata and isolate the neuron of interest, and registration of neurons to each other using automatically detected arbors of the starburst amacrine interneurons as fiducial markers. In principle, this method could be generalizable to other structures of the CNS, provided that they allow sparse labeling of the cells and contain a reliable axis of spatial reference.}
}
@article{SunandHevnerNATURE-14,
       author = {Sun, Tao and Hevner, Robert F.},
        title = {Growth and folding of the mammalian cerebral cortex: from molecules to malformations},
      journal = {Nature Reviews Neuroscience},
         year = {2014},
    publisher = {Nature Publishing Group},
       volume = {15},
        issue = {4},
        pages = {217-232},
     abstract = {The size and extent of folding of the mammalian cerebral cortex are important factors that influence a species' cognitive abilities and sensorimotor skills. Studies in various animal models and in humans have provided insight into the mechanisms that regulate cortical growth and folding. Both protein-coding genes and microRNAs control cortical size, and recent progress in characterizing basal progenitor cells and the genes that regulate their proliferation has contributed to our understanding of cortical folding. Neurological disorders linked to disruptions in cortical growth and folding have been associated with novel neurogenetic mechanisms and aberrant signalling pathways, and these findings have changed concepts of brain evolution and may lead to new medical treatments for certain disorders.},
}
@inproceedings{SundbergetalCVPR-11,
      title = {Occlusion Boundary Detection and Figure/Ground Assignment from Optical Flow},
     author = {Patrik Sundberg and Thomas Brox and Michael Maire and Pablo Arbelaez and Jitendra Malik},
  booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
       year = 2011,
}
@article{SunetalCoRR-19a,
       author = {Chen Sun and Austin Myers and Carl Vondrick and Kevin Murphy and Cordelia Schmid},
        title = {VideoBERT: {A} Joint Model for Video and Language Representation Learning},
      journal = {CoRR},
       volume = {arXiv:1904.01766},
         year = {2019},
     abstract = {Self-supervised learning has become increasingly important to leverage the abundance of unlabeled data available on platforms like YouTube. Whereas most existing approaches learn low-level representations, we propose a joint visual-linguistic model to learn high-level features without any explicit supervision. In particular, inspired by its recent success in language modeling, we build upon the BERT model to learn bidirectional joint distributions over sequences of visual and linguistic tokens, derived from vector quantization of video data and off-the-shelf speech recognition outputs, respectively. We use VideoBERT in numerous tasks, including action classification and video captioning. We show that it can be applied directly to open-vocabulary classification, and confirm that large amounts of training data and cross-modal information are critical to performance. Furthermore, we outperform the state-of-the-art on video captioning, and quantitative results verify that the model learns high-level semantic features.}
}
@article{SunetalCoRR-19b,
       author = {Chen Sun and Fabien Baradel and Kevin Murphy and Cordelia Schmid},
        title = {Contrastive Bidirectional Transformer for Temporal Representation Learning},
      journal = {CoRR},
       volume = {arXiv:1906.05743},
         year = {2019},
     abstract = {This paper proposes a self-supervised learning approach for video features that results in significantly improved performance on downstream tasks (such as video classification, captioning and segmentation) compared to existing methods. Our method extends the BERT model for text sequences to the case of sequences of real-valued feature vectors, by replacing the softmax loss with noise contrastive estimation (NCE). We also show how to learn representations from sequences of visual features and sequences of words derived from ASR (automatic speech recognition), and show that such cross-modal training (when possible) helps even more.},
}
@article{SunetalNATURE-12,
       author = {Sun, Jeong-Yun and Zhao, Xuanhe and Illeperuma, Widusha R. K. and Chaudhuri, Ovijit and Oh, Kyu Hwan and Mooney, David J. and Vlassak, Joost J. and Suo, Zhigang},
        title = {Highly stretchable and tough hydrogels},
      journal = {Nature},
    publisher = {Nature Publishing Group},
         year = 2012,
       volume = 489,
        issue = 7414,
        pages = {133-136},
     abstract = {Hydrogels are used as scaffolds for tissue engineering, vehicles for drug delivery, actuators for optics and fluidics, and model extracellular matrices for biological studies. The scope of hydrogel applications, however, is often severely limited by their mechanical behaviour5. Most hydrogels do not exhibit high stretchability; for example, an alginate hydrogel ruptures when stretched to about 1.2 times its original length. Some synthetic elastic hydrogels have achieved stretches in the range 10-20, but these values are markedly reduced in samples containing notches. Most hydrogels are brittle, with fracture energies of about 10J $m^{-2}$, as compared with \tilde{}1,000J $m^{-2}$ for cartilage and \tilde{}10,000J $m^{-2}$ for natural rubbers10. Intense efforts are devoted to synthesizing hydrogels with improved mechanical properties, certain synthetic gels have reached fracture energies of 100-1,000J $m^{-2}$. Here we report the synthesis of hydrogels from polymers forming ionically and covalently crosslinked networks. Although such gels contain \tilde{}90\% water, they can be stretched beyond 20 times their initial length, and have fracture energies of \tilde{}9,000J $m^{-2}$. Even for samples containing notches, a stretch of 17 is demonstrated. We attribute the gels' toughness to the synergy of two mechanisms: crack bridging by the network of covalent crosslinks, and hysteresis by unzipping the network of ionic crosslinks. Furthermore, the network of covalent crosslinks preserves the memory of the initial state, so that much of the large deformation is removed on unloading. The unzipped ionic crosslinks cause internal damage, which heals by re-zipping. These gels may serve as model systems to explore mechanisms of deformation and energy dissipation, and expand the scope of hydrogel applications.}
}
@article{SunetalNATURE-13,
     title = {Fast {GCaMP}s for improved tracking of neuronal activity},
    author = {Sun, Xiaonan R. and Badura, Aleksandra and Pacheco, Diego A. and Lynch, Laura A. and Schneider, Eve R. and Taylor, Matthew P. and Hogue, Ian B. and Enquist, Lynn W. and Murthy, Mala and Wang, Samuel S. -H.},
   journal = {Nature Communications},
 publisher = {Nature Publishing Group},
    volume = 4,
      year = 2013,
  abstract = {The use of genetically encodable calcium indicator proteins to monitor neuronal activity is hampered by slow response times and a narrow Ca2+-sensitive range. Here we identify three performance-limiting features of GCaMP3, a popular genetically encodable calcium indicator protein. First, we find that affinity is regulated by the calmodulin domain's Ca2+-chelating residues. Second, we find that off-responses to Ca2+ are rate-limited by dissociation of the RS20 domain from calmodulin's hydrophobic pocket. Third, we find that on-responses are limited by fast binding to the N-lobe at high Ca2+ and by slow binding to the C-lobe at lower Ca2+. We develop Fast-GCaMPs, which have up to 20-fold accelerated off-responses and show that they have a 200-fold range of KD, allowing coexpression of multiple variants to span an expanded range of Ca2+ concentrations. Finally, we show that Fast-GCaMPs track natural song in Drosophila auditory neurons and generate rapid responses in mammalian neurons, supporting the utility of our approach.},
}
@article{SunkinetalNUCLEIC-13,
    author = {Susan Sunkin and Lydia Ng and Christopher Lau and Tim Dolbeare and Terri L. Gilbert and 
              Carol L. Thompson and Michael Hawrylycz and Chinh Dang},
     title = {Allen Brain Atlas: an integrated spatio-temporal portal for exploring the central nervous system},
   journal = {Nucleic Acids Research},
    volume = 41,
     issue = {Database-Issue},
      year = 2013,
     pages = {996-1008},
}
@article{SuperandUylingsCEREBRAL-CORTEX-01,
       author = {Sup\`{e}r, H. and Uylings, H.B.M.},
        title = {{The Early Differentiation of the Neocortex: a Hypothesis on Neocortical Evolution}},
      journal = {Cerebral Cortex},
       volume = {11},
       number = {12},
         year = {2001},
        pages = {1101-1109},
     abstract = {During development, a cerebral cortex appears in the wall of the telencephalic vesicle in reptiles and mammals. It arises from a cell-dense cortical plate, which develops within a primordial preplate. The neurons of the preplate are essential for cortical development; they regulate the neuronal migration of the cortical plate neurons and form the first axonal connections. In the reptilian cortex and in the hippocampus of the mammalian cerebral cortex, most ingrowing afferent axons run above the cortical plate, in the zone where the receptive tufts of apical dendrites of the cortical pyramidal neurons branch extensively. In the mammalian neocortex, however, axons enter mainly from below the cortical plate where they do not encounter the apical tufts of these pyramidal neurons. In this paper, we discuss the idea that this difference in cortical development has relieved a functional constraint in the expansion of the cortex during evolution. We hypothesize that the entrance of axons below the cell-dense cortical plate, together with the inside-out migration of cortical neurons, ensures that the neocortex remains an 'open' system, able to differentiate into new (sub)layers and more cortical areas.},
}
@article{SurandSinhaIPJ-09,
       author = {Sur, Shravani and Sinha, V. K.},
        title = {Event-related potential: An overview},
      journal = {Industrial psychiatry journal},
         year = {2009},
    publisher = {Medknow Publications},
       volume = {18},
        issue = {1},
        pages = {70-73},
     abstract = {Electroencephalography (EEG) provides an excellent medium to understand neurobiological dysregulation, with the potential to evaluate neurotransmission. Time-locked EEG activity or event-related potential (ERP) helps capture neural activity related to both sensory and cognitive processes. In this article, we attempt to present an overview of the different waveforms of ERP and the major findings in various psychiatric conditions.},
}
@article{SuretalJoN-99,
       author = {Sur, M. and Angelucci, A. and Sharma, J.},
        title = {{{R}ewiring cortex: the role of patterned activity in development and plasticity of neocortical circuits}},
      journal = {Journal of Neurobiology},
         year = {1999},
       volume = {41},
       number = {1},
        pages = {33-43},
     abstract = {Visually driven activity is not required for the establishment of ocular dominance columns, orientation columns, and long-range horizontal connections in visual cortex, although spontaneous activity appears to be necessary. The role of activity may be instructive or simply permissive; evidence for an instructive role requires inquiry into the role of the pattern of activity in shaping cortical circuits. The few experiments that have probed the role of patterned activity include the effects of artificial strabismus, artificial stimulation of the optic nerve, and rewiring visual projections from the retina to the auditory thalamus and cortex. These experiments demonstrate that patterned activity is vital for the maintenance of thalamocortical, local intracortical, and long-range horizontal connections in cortex.}
}
@article{SusakietalCELL-14,
       author = {Susaki, E. A. and Tainaka, K. and Perrin, D. and Kishino, F. and Tawara, T. and Watanabe, T. M. and Yokoyama, C. and Onoe, H. and Eguchi, M. and Yamaguchi, S. and Abe, T. and Kiyonari, H. and Shimizu, Y. and Miyawaki, A. and Yokota, H. and Ueda, H. R.},
        title = {{{W}hole-brain imaging with single-cell resolution using chemical cocktails and computational analysis}},
      journal = {Cell},
         year = 2014,
       volume = 157,
        issue = 3,
        pages = {726-739},
     abstract = {Systems-level identification and analysis of cellular circuits in the brain will require the development of whole-brain imaging with single-cell resolution. To this end, we performed comprehensive chemical screening to develop a whole-brain clearing and imaging method, termed CUBIC (clear, unobstructed brain imaging cocktails and computational analysis). CUBIC is a simple and efficient method involving the immersion of brain samples in chemical mixtures containing aminoalcohols, which enables rapid whole-brain imaging with single-photon excitation microscopy. CUBIC is applicable to multicolor imaging of fluorescent proteins or immunostained samples in adult brains and is scalable from a primate brain to subcellular structures. We also developed a whole-brain cell-nuclear counterstaining protocol and a computational image analysis pipeline that, together with CUBIC reagents, enable the visualization and quantification of neural activities induced by environmental stimulation. CUBIC enables time-course expression profiling of whole adult brains with single-cell resolution.},
}
@article{SussilloCOIN-14,
       author = {David Sussillo},
        title = {Neural circuits as computational dynamical systems},
      journal = {Current Opinion in Neurobiology},
       volume = {25},
         year = {2014},
        pages = {156-163},
     abstract = {Many recent studies of neurons recorded from cortex reveal complex temporal dynamics. How such dynamics embody the computations that ultimately lead to behavior remains a mystery. Approaching this issue requires developing plausible hypotheses couched in terms of neural dynamics. A tool ideally suited to aid in this question is the recurrent neural network (RNN). RNNs straddle the fields of nonlinear dynamical systems and machine learning and have recently seen great advances in both theory and application. I summarize recent theoretical and technological advances and highlight an example of how RNNs helped to explain perplexing high-dimensional neurophysiological data in the prefrontal cortex.}
}
@article{SussilloandAbbottNEURON-09,
       author = {Sussillo, David and Abbott, L. F.},
        title = {Generating Coherent Patterns of Activity from Chaotic Neural Networks},
      journal = {Neuron},
    publisher = {Elsevier},
       volume = {63},
        issue = {4},
         year = {2009},
        pages = {544-557},
     abstract = {Neural circuits display complex activity patterns both spontaneously and when responding to a stimulus or generating a motor output. How are these two forms of activity related? We develop a procedure called FORCE learning for modifying synaptic strengths either external to or within a model neural network to change chaotic spontaneous activity into a wide variety of desired activity patterns. FORCE learning works even though the networks we train are spontaneously chaotic and we leave feedback loops intact and unclamped during learning. Using this approach, we construct networks that produce a wide variety of complex output patterns, input-output transformations that require memory, multiple outputs that can be switched by control inputs, and motor patterns matching human motion capture data. Our results reproduce data on premovement activity in motor and premotor cortex, and suggest that synaptic plasticity may be a more rapid and powerful modulator of network activity than generally appreciated.},
}
@article{SussilloandBarakNC-13,
       author = {Sussillo, David and Barak, Omri},
        title = {Opening the Black Box: Low-dimensional Dynamics in High-dimensional Recurrent Neural Networks},
      journal = {Neural Computation},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
       volume = {25},
       number = {3},
         year = {2013},
        pages = {626-649},
      comment = {This paper provides the critical link between viewing RNNs as neural networks and also as dynamical systems. Often RNNs are considered 'black-box' approaches, implying that their mechanism cannot be understood. However, the paper shows that in simple cases an RNN can be 'reverse engineered' to reeal its underlying dynamical mechanism.},
     abstract = {Recurrent neural networks (RNNs) are useful tools for learning nonlinear relationships between time-varying inputs and outputs with complex temporal dependencies. Recently developed algorithms have been successful at training RNNs to perform a wide variety of tasks, but the resulting networks have been treated as black boxes: their mechanism of operation remains unknown. Here we explore the hypothesis that fixed points, both stable and unstable, and the linearized dynamics around them, can reveal crucial aspects of how RNNs implement their computations. Further, we explore the utility of linearization in areas of phase space that are not true fixed points but merely points of very slow movement. We present a simple optimization technique that is applied to trained RNNs to find the fixed and slow points of their dynamics. Linearization around these slow regions can be used to explore, or reverse-engineer, the behavior of the RNN. We describe the technique, illustrate it using simple examples, and finally showcase it on three high-dimensional RNN examples: a 3-bit flip-flop device, an input-dependent sine wave generator, and a two-point moving average. In all cases, the mechanisms of trained networks could be inferred from the sets of fixed and slow points and the linearized dynamics around them.},
}
@article{SussilloetalCoRR-16,
       author = {David Sussillo and Rafal Jozefowicz and L.F. Abbott and Chethan Pandarinath},
        title = {{LFADS} - Latent Factor Analysis via Dynamical Systems},
      journal = {CoRR},
       volume = {arXiv:1608.06315},
         year = {2016},
     abstract = {Neuroscience is experiencing a data revolution in which many hundreds or thousands of neurons are recorded simultaneously. Currently, there is little consensus on how such data should be analyzed. Here we introduce LFADS (Latent Factor Analysis via Dynamical Systems), a method to infer latent dynamics from simultaneously recorded, single-trial, high-dimensional neural spiking data. LFADS is a sequential model based on a variational auto-encoder. By making a dynamical systems hypothesis regarding the generation of the observed data, LFADS reduces observed spiking to a set of low-dimensional temporal factors, per-trial initial conditions, and inferred inputs. We compare LFADS to existing methods on synthetic data and show that it significantly out-performs them in inferring neural firing rates and latent dynamics}
}
@article{SussilloetalNATURE-NEUROSCIENCE-15,
       author = {Sussillo, D. and Churchland, M. M. and Kaufman, M. T. and Shenoy, K. V.},
        title = {A neural network that finds a naturalistic solution for the production of muscle activity},
      journal = {Nature Neuroscience},
       volume = {18},
       number = {7},
         year = {2015},
        pages = {1025-1033},
     abstract = {It remains an open question how neural responses in motor cortex relate to movement. We explored the hypothesis that motor cortex reflects dynamics appropriate for generating temporally patterned outgoing commands. To formalize this hypothesis, we trained recurrent neural networks to reproduce the muscle activity of reaching monkeys. Models had to infer dynamics that could transform simple inputs into temporally and spatially complex patterns of muscle activity. Analysis of trained models revealed that the natural dynamical solution was a low-dimensional oscillator that generated the necessary multiphasic commands. This solution closely resembled, at both the single-neuron and population levels, what was observed in neural recordings from the same monkeys. Notably, data and simulations agreed only when models were optimized to find simple solutions. An appealing interpretation is that the empirically observed dynamics of motor cortex may reflect a simple solution to the problem of generating temporally patterned descending commands.}
}
@book{Sussman75,
       author = {Sussman, Gerald J.},
        title = {A Computer Model of Skill Acquisition},
    publisher = {American Elsevier},
 comment      = {HACKER},
      address = {New York},
         year = 1975,
}
@techreport{SussmanetalTR-71,
       author = {Sussman, Gerald J. and Terry Winograd and Eugene Charniak},
        title = {Micro-Planner Reference Manual},
        issue = 203,
  institution = {{MIT} {AI} Laboratory},
         year = 1971,
}
@misc{SutskeverPhD-13,
        title = {Training Recurrent Neural Networks},
       author = {Sutskever, Ilya},
         year = 2013,
 howpublished = {Ph.D. Thesis. Department of Computer Science, University of Toronto}
}
@article{SutskeverandHintonNC-08,
       author = {Sutskever, I. and Hinton, G. E.},
        title = {{D}eep, narrow sigmoid belief networks are universal approximators},
      journal = {Neural Computing},
         year = {2008},
       volume = {20},
       number = {11},
        pages = {2629-2636},
     abstract = {In this note, we show that exponentially deep belief networks can approximate any distribution over binary vectors to arbitrary accuracy, even when the width of each layer is limited to the dimensionality of the data. We further show that such networks can be greedily learned in an easy yet impractical way.}
}
@incollection{SutskeverandHintonNIPS-09,
        title = {Using matrices to model symbolic relationship},
       author = {Sutskever, Ilya and Hinton, Geoffrey E},
    booktitle = {Advances in Neural Information Processing Systems 21},
       editor = {D. Koller and D. Schuurmans and Y. Bengio and L. Bottou},
    publisher = {Curran Associates, Inc.},
         year = {2009},
        pages = {1593-1600},
}
@article{SutskeveretalCoRR-14,
       author = {Ilya Sutskever and Oriol Vinyals and Quoc V. Le},
        title = {Sequence to Sequence Learning with Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1409.3215},
         year = 2014,
          url = {http://arxiv.org/abs/1409.3215},
     abstract = {Deep Neural Networks (DNNs) are powerful models that have achieved excellent performance on difficult learning tasks. Although DNNs work well whenever large labeled training sets are available, they cannot be used to map sequences to sequences. In this paper, we present a general end-to-end approach to sequence learning that makes minimal assumptions on the sequence structure. Our method uses a multilayered Long Short-Term Memory (LSTM) to map the input sequence to a vector of a fixed dimensionality, and then another deep LSTM to decode the target sequence from the vector. Our main result is that on an English to French translation task from the WMT-14 dataset, the translations produced by the LSTM achieve a BLEU score of 34.8 on the entire test set, where the LSTM's BLEU score was penalized on out-of-vocabulary words. Additionally, the LSTM did not have difficulty on long sentences. For comparison, a phrase-based SMT system achieves a BLEU score of 33.3 on the same dataset. When we used the LSTM to rerank the hypotheses produced by the aforementioned SMT system, its BLEU score increases to 36.5, which is close to the previous state of the art. The LSTM also learned sensible phrase and sentence representations that are sensitive to word order and are relatively invariant to the active and the passive voice. Finally, we found that reversing the order of the words in all source sentences (but not target sentences) improved the LSTM's performance markedly, because doing so introduced many short term dependencies between the source and the target sentence which made the optimization problem easier.}
}
@inproceedings{SutskeveretalICML-11,
       author = {Ilya Sutskever and James Martens and Geoffrey Hinton},
        title = {Generating Text with Recurrent Neural Networks},
    booktitle = {Proceedings of the 28th International Conference on Machine Learning},
       editor = {Lise Getoor and Tobias Scheffer},
     location = {Bellevue, Washington, USA},
    publisher = {ACM},
      address = {New York, NY, USA},
         year = 2011,
        pages = {1017-1024},
}
@incollection{SutskeveretalNIPS-09,
        title = {Modelling Relational Data using Bayesian Clustered Tensor Factorization},
       author = {Ilya Sutskever and Ruslan Salakhutdinov and Joshua Tenenbaum},
    booktitle = {Advances in Neural Information Processing Systems 22},
       editor = {Y. Bengio and D. Schuurmans and J. Lafferty and C. K. I. Williams and A. Culotta},
    publisher = {{MIT} Press},
         year = 2009,
        pages = {1821-1828},
}
@incollection{SutskeveretalNIPS-14,
        title = {Sequence to Sequence Learning with Neural Networks},
       author = {Sutskever, Ilya and Vinyals, Oriol and Le, Quoc V},
    booktitle = {Advances in Neural Information Processing Systems 27},
       editor = {Z. Ghahramani and M. Welling and C. Cortes and N. D. Lawrence and K. Q. Weinberger},
    publisher = {Curran Associates, Inc.},
         year = {2014},
        pages = {3104-3112},
}
@article{SuttonML-88,
       author = {Sutton, Richard S.},
        title = {Learning to Predict by the Methods of Temporal Differences},
      journal = {Machine Learning},
       volume = 3,
         year = 1988,
        pages = {9-44},
}
@inproceedings{SuttonML-90,
       author = {Richard S. Sutton},
        title = {Integrated Architectures for Learning, Planning, and Reacting Based on Approximating Dynamic Programming},
    booktitle = {In Proceedings of the Seventh International Conference on Machine Learning},
    publisher = {Morgan Kaufmann},
     location = {Austin, Texas},
         year = {1990},
        pages = {216-224},
     abstract = {This paper extends previous work with Dyna, a class of architectures for intelligent systems based on approximating dynamic programming methods. Dyna architectures integrate trial-and-error (reinforcement) learning and execution-time planning into a single process operating alternately on the world and on a learned model of the world. In this paper, I present and show results for two Dyna architectures. The Dyna-PI architecture is based on dynamic programming's policy iteration method and can be related to existing AI ideas such as evaluation functions and universal plans (reactive systems). Using a navigation task, results are shown for a simple Dyna-PI system that simultaneously learns by trial and error, learns a world model, and plans optimal routes using the evolving world model. The Dyna-Q architecture is based on Watkins's Q-learning, a new kind of reinforcement learning. Dyna-Q uses a less familiar set of data structures than does Dyna-PI, but is arguably simpler to implement and use. We show that Dyna-Q architectures are easy to adapt for use in changing environments.}
}
@inproceedings{SuttonML-95,
       author = {Sutton, Richard},
        title = {{TD} Models: Modeling the World at a Mixture of Time Scales},
    booktitle = {Proceedings Twelfth International Conference on Machine Learning},
     location = {Lake Tahoe, California},
        pages = {531-539},
         year = 1995,
}
@article{SuttonSIGART-91,
       author = {Richard S. Sutton},
        title = {Dyna, an integrated architecture for learning, planning and Reacting},
      journal = {{SIGART} Bulletin},
    publisher = {ACM},
       volume = {4},
       number = {2},
         year = {1991},
        pages = {160-163},
     abstract = {Dyna is an AI architecture that integrates learning, planning, and reactive execution. Learning methods are used in Dyna both for compiling planning results and for updating a model of the effects of the agent's actions on the world. Planning is incremental and can use the probabilistic and ofttimes incorrect world models generated by learning processes. Execution is fully reactive in the sense that no planning intervenes between perception and action. Dyna relies on machine learning methods for learning from examples---these are among the basic building blocks making up the architecture---yet is not tied to any particular method. This paper briefly introduces Dyna and discusses its strengths and weaknesses with respect to other architectures.}
}
@techreport{SuttonTR-87,
       author = {Sutton, Richard S.},
        title = {Learning to Predict by the Methods of Temporal Differences},
        issue = {TR87-5091},
  institution = {GTE Laboratories},
      address = {Waltham, Massachusetts},
         year = 1987,
}
@book{SuttonandBarto2018sedition,
       author = {Sutton, Richard and Barto, Andrew},
        title = {Reinforcement Learning: An Introduction},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 2018,
}
@book{SuttonandBarto98,
       author = {Sutton, Richard and Barto, Andrew},
        title = {Reinforcement Learning: An Introduction},
    publisher = {{MIT} Press},
      address = {Cambridge, Massachusetts},
         year = 1998,
}
@incollection{SuttonandMcCallum06,
       author = {Charles Sutton and Andrew McCallum},
        title = {An Introduction to Conditional Random Fields for Relational Learning},
    booktitle = {Introduction to Statistical Relational Learning},
       editor = {Lise Getoor and Ben Taskar},
    publisher = {MIT Press},
         year = 2006,
}
@incollection{SuttonandTannerNIPS-05,
       author = {Sutton, Richard S. and Tanner, Brian T.},
        title = {Temporal Difference Networks},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 17,
       editor = {Lawrence K. Saul and Yair Weiss and {L\'{e}on} Bottou},
    publisher = {{MIT} Press},
      address = {Cambridge, MA},
         year = 2005,
}
@article{SuttonetalAIJ-99,
       author = {Sutton, Richard S. and Precup, Doina and Singh, Satinder},
        title = {Between {MDP}s and semi-{MDP}s: A Framework for Temporal Abstraction in Reinforcement Learning},
      journal = {Artificial. Intelligence},
       volume = {112},
       number = {1-2},
         year = {1999},
        pages = {181-211},
    publisher = {Elsevier Science Publishers Ltd.},
     abstract = {Learning, planning, and representing knowledge at multiple levels of temporal abstraction are key, longstanding challenges for AI. In this paper we consider how these challenges can be addressed within the mathematical framework of reinforcement learning and Markov decision processes (MDPs). We extend the usual notion of action in this framework to include options \emdash{} closed-loop policies for taking action over a period of time. Examples of options include picking up an object, going to lunch, and traveling to a distant city, as well as primitive actions such as muscle twitches and joint torques. Overall, we show that options enable temporally abstract knowledge and action to be included in the reinforcement learning framework in a natural and general way. In particular, we show that options may be used interchangeably with primitive actions in planning methods such as dynamic programming and in learning methods such as Q-learning. Formally, a set of options defined over an MDP constitutes a semi-Markov decision process (SMDP), and the theory of SMDPs provides the foundation for the theory of options. However, the most interesting issues concern the interplay between the underlying MDP and the SMDP and are thus beyond SMDP theory. We present results for three such cases: (1) we show that the results of planning with options can be used during execution to interrupt options and thereby perform even better than planned, (2) we introduce new intra-option methods that are able to learn about an option from fragments of its execution, and (3) we propose a notion of subgoal that can be used to improve the options themselves. All of these results have precursors in the existing literature; the contribution of this paper is to establish them in a simpler and more general setting with fewer changes to the existing reinforcement learning framework. In particular, we show that these results can be obtained without committing to (or ruling out) any particular approach to state abstraction, hierarchy, function approximation, or the macroutility problem.}
}
@inproceedings{SuttonetalHORDE-11,
       author = {Sutton, Richard S. and Modayil, Joseph and Delp, Michael and Degris, Thomas and Pilarski, Patrick M. and White, Adam and Precup, Doina},
        title = {Horde: A Scalable Real-time Architecture for Learning Knowledge from Unsupervised Sensorimotor Interaction},
    booktitle = {The 10th International Conference on Autonomous Agents and Multiagent Systems - Volume 2},
    publisher = {International Foundation for Autonomous Agents and Multiagent Systems},
         year = {2011},
        pages = {761-768},
     abstract = {Maintaining accurate world knowledge in a complex and changing environment is a perennial problem for robots and other artificial intelligence systems. Our architecture for addressing this problem, called Horde, consists of a large number of independent reinforcement learning sub-agents, or demons. Each demon is responsible for answering a single predictive or goal-oriented question about the world, thereby contributing in a factored, modular way to the system's overall knowledge. The questions are in the form of a value function, but each demon has its own policy, reward function, termination function, and terminal-reward function unrelated to those of the base problem. Learning proceeds in parallel by all demons simultaneously so as to extract the maximal training information from whatever actions are taken by the system as a whole. Gradient-based temporal-difference learning methods are used to learn efficiently and reliably with function approximation in this off-policy setting. Horde runs in constant time and memory per time step, and is thus suitable for learning online in real-time applications such as robotics. We present results using Horde on a multi-sensored mobile robot to successfully learn goal-oriented behaviors and long-term predictions from off-policy experience. Horde is a significant incremental step towards a real-time architecture for efficient learning of general knowledge from unsupervised sensorimotor interaction.},
}
@book{Svoboda1948,
        title = {Computing mechanisms and linkages},
       author = {Svoboda, A. and James, H.M.},
       series = {Radiation Laboratory series},
         year = 1948,
    publisher = {McGraw-Hill}
}
@book{SwainandDavis78,
       author = {Swain, P. H. and Davis, S. M.},
        title = {Remote Sensing: The Quantitative Approach},
    publisher = {McGraw-Hill},
      address = {New York},
         year = 1978,
} 
@article{SwansonTiN-95,
       author = {Larry W. Swanson},
        title = {Mapping the human brain: past, present, and future},
      journal = {Trends in Neurosciences},
       volume = {18},
       number = {11},
        pages = {471-474},
         year = {1995},
     abstract = {The ability to map functional activity in the living human brain has rekindled interest in the organization of human neural circuitry specifically and in animal neural circuitry generally. Faced with incredible complexity, researchers are turning to the power of computer graphics for two- and three-dimensional interactive mapping, to the mathematical modeling of dynamics in proposed circuits, and to databases with powerful discovery engines.}
}
@inproceedings{SwartoutAAAI-83,
       author = {Swartout, William},
        title = {The GIST Behavior Explainer},
    booktitle = {Proceedings AAAI-83},
     location = {Washington, DC, USA},
 organization = {AAAI},
         year = 1983,
        pages = {402-407},
}
@book{Swenson2006,
        title = {Review of Clinical and Functional Neuroscience},
       author = {Rand S. Swenson},
    publisher = {Dartmouth Medical School},
         year = {2006},
}
@article{SwineyandSousaFiHN-14,
       author = {Swiney, Lauren and Sousa, Paulo},
        title = {A new comparator account of auditory verbal hallucinations: how motor prediction can plausibly contribute to the sense of agency for inner speech},
      journal = {Frontiers in Human Neuroscience},
       volume = {8},
        pages = {675},
         year = {2014},
     abstract = {The comparator account holds that processes of motor prediction contribute to the sense of agency by attenuating incoming sensory information and that disruptions to this process contribute to misattributions of agency in schizophrenia.  Over the last 25 years this simple and powerful model has gained widespread support not only as it relates to bodily actions but also as an account of misattributions of agency for inner speech, potentially explaining the etiology of auditory verbal hallucination (AVH).  In this paper we provide a detailed analysis of the traditional comparator account for inner speech, pointing out serious problems with the specification of inner speech on which it is based and highlighting inconsistencies in the interpretation of the electrophysiological evidence commonly cited in its favor.  In light of these analyses we propose a new comparator account of misattributed inner speech.  The new account follows leading models of motor imagery in proposing that inner speech is not attenuated by motor prediction, but rather derived directly from it.  We describe how failures of motor prediction would therefore directly affect the phenomenology of inner speech as well as triggering a mismatch in the comparison between motor prediction and motor intention, contributing to abnormal feelings of agency.  We argue that the new account fits with the emerging phenomenological evidence that AVH are both distinct from ordinary inner speech and heterogeneous.  Finally, we explore the possibility that the new comparator account may extend to explain disruptions across a range of imagistic modalities, and outline avenues for future research.}
}
@article {SwingleyPTRS_B-09,
       author = {Swingley, Daniel},
        title = {Contributions of infant word learning to language development},
      journal = {Philosophical Transactions of the Royal Society of London B: Biological Sciences},
       volume = 364,
        issue = 1536,
        pages = {3617-3632},
         year = 2009,
    publisher = {The Royal Society},
     abstract = {Infants learn the forms of words by listening to the speech they hear. Though little is known about the degree to which these forms are meaningful for young infants, the words still play a role in early language development. Words guide the infant to his or her first syntactic intuitions, aid in the development of the lexicon, and, it is proposed, may help infants learn phonetic categories.},
}
@article{SykovaandNicholsonPR-08,
       author = {Sykova, E. and Nicholson, C.},
        title = {Diffusion in brain extracellular space},
      journal = {Physiology Review},
       volume = {88},
       number = {4},
         year = {2008},
        pages = {1277-1340},
     abstract = {Diffusion in the extracellular space (ECS) of the brain is constrained by the volume fraction and the tortuosity and a modified diffusion equation represents the transport behavior of many molecules in the brain. Deviations from the equation reveal loss of molecules across the blood-brain barrier, through cellular uptake, binding, or other mechanisms. Early diffusion measurements used radiolabeled sucrose and other tracers. Presently, the real-time iontophoresis (RTI) method is employed for small ions and the integrative optical imaging (IOI) method for fluorescent macromolecules, including dextrans or proteins. Theoretical models and simulations of the ECS have explored the influence of ECS geometry, effects of dead-space microdomains, extracellular matrix, and interaction of macromolecules with ECS channels. Extensive experimental studies with the RTI method employing the cation tetramethylammonium (TMA) in normal brain tissue show that the volume fraction of the ECS typically is approximately 20\% and the tortuosity is approximately 1.6 (i.e., free diffusion coefficient of TMA is reduced by 2.6), although there are regional variations. These parameters change during development and aging. Diffusion properties have been characterized in several interventions, including brain stimulation, osmotic challenge, and knockout of extracellular matrix components. Measurements have also been made during ischemia, in models of Alzheimer's and Parkinson's diseases, and in human gliomas. Overall, these studies improve our conception of ECS structure and the roles of glia and extracellular matrix in modulating the ECS microenvironment. Knowledge of ECS diffusion properties is valuable in contexts ranging from understanding extrasynaptic volume transmission to the development of paradigms for drug delivery to the brain.}
}
@book{Syski79,
       author = {Syski, Ryszard},
        title = {Random Processes},
    publisher = {Marcel Dekker},
      address = {New York},
         year = 1979,
}
@article{SzaboetalNEURON-14,
       author = {Szabo, V. and Ventalon, C. and de Sars, V. and Bradley, J. and Emiliani, V.},
        title = {Spatially Selective Holographic Photoactivation and Functional Fluorescence Imaging in Freely Behaving Mice with a Fiberscope},
      journal = {Neuron},
        issue = {14},
         year = {2014},
        pages = {S0896-6273},
     abstract = {Correlating patterned neuronal activity to defined animal behaviors is a core goal in neuroscience. Optogenetics is one large step toward achieving this goal, yet optical methods to control neural activity in behaving rodents have so far been limited to perturbing all light-sensitive neurons in a large volume. Here we demonstrate an all-optical method for precise spatial control and recording of neuronal activity in anesthetized and awake freely behaving mice. Photoactivation patterns targeted to multiple neuronal somata, produced with computer-generated holography, were transmitted to the mouse brain using a micro-objective-coupled fiber bundle. Fluorescence imaging through the same device, via epifluorescence, structured illumination, or scanless multipoint confocal microscopy, allowed imaging of neurons and recording of neuronal activity. The fiberscope was tested in mice coexpressing ChR2-tdTomato and GCaMP5-G in cerebellar interneurons, delivering near-cellular resolution photoactivation in freely behaving mice.}
}
@article{SzlaviketalTNSRE-06,
       author = {R. B. Szlavik and A. K. Bhuiyan and A. Carver and F. Jenkins},
      journal = {IEEE Transactions on Neural Systems and Rehabilitation Engineering},
        title = {Neural-electronic inhibition simulated with a neuron model implemented in SPICE},
       volume = {14},
        issue = {1},
         year = {2006},
        pages = {109-115},
     abstract = {There have been numerous studies presented in the literature related to the simulation of the interaction between biological neurons and electronic devices. A complicating factor associated with these simulations is the algebraic complexity involved in implementation. This complication has impeded simulation of more involved neural-electronic circuitry and consequently has limited potential advancements in the integration of biological neurons with synthetic electronics. In this paper, we describe a modification to a previously proposed SPICE based Hodgkin-Huxley neuron model that demonstrates more physiologically relevant electrical behavior. We utilize this SPICE based neuron model in conjunction with an external circuit that allows for artificial selective inhibition of neural spiking. The neural firing control scheme proposed herein would allow for action potential frequency modulation of neural activity that, if developed further, could potentially be applied to suppress undesirable neural activity that manifests symptomatically as the tremors or seizures associated with specific pathologies of the nervous system.}
}
@incollection{Szolovits82,
        title = {Artificial Intelligence in Medicine},
       author = {Szolovits, Peter},
    booktitle = {Artificial Intelligence in Medicine},
       editor = {Szolovits, Peter},
    publisher = {Westview},
      address = {Boulder, Colorado},
        pages = {1-19},
         year = 1982,
}
@article{TPU-NIPS-CoRR-17,
  author    = {Norman P. Jouppi and
               Cliff Young and
               Nishant Patil and
               David Patterson and
               Gaurav Agrawal and
               Raminder Bajwa and
               Sarah Bates and
               Suresh Bhatia and
               Nan Boden and
               Al Borchers and
               Rick Boyle and
               Pierre{-}luc Cantin and
               Clifford Chao and
               Chris Clark and
               Jeremy Coriell and
               Mike Daley and
               Matt Dau and
               Jeffrey Dean and
               Ben Gelb and
               Tara Vazir Ghaemmaghami and
               Rajendra Gottipati and
               William Gulland and
               Robert Hagmann and
               Richard C. Ho and
               Doug Hogberg and
               John Hu and
               Robert Hundt and
               Dan Hurt and
               Julian Ibarz and
               Aaron Jaffey and
               Alek Jaworski and
               Alexander Kaplan and
               Harshit Khaitan and
               Andy Koch and
               Naveen Kumar and
               Steve Lacy and
               James Laudon and
               James Law and
               Diemthu Le and
               Chris Leary and
               Zhuyuan Liu and
               Kyle Lucke and
               Alan Lundin and
               Gordon MacKean and
               Adriana Maggiore and
               Maire Mahony and
               Kieran Miller and
               Rahul Nagarajan and
               Ravi Narayanaswami and
               Ray Ni and
               Kathy Nix and
               Thomas Norrie and
               Mark Omernick and
               Narayana Penukonda and
               Andy Phelps and
               Jonathan Ross and
               Amir Salek and
               Emad Samadiani and
               Chris Severn and
               Gregory Sizikov and
               Matthew Snelham and
               Jed Souter and
               Dan Steinberg and
               Andy Swing and
               Mercedes Tan and
               Gregory Thorson and
               Bo Tian and
               Horia Toma and
               Erick Tuttle and
               Vijay Vasudevan and
               Richard Walter and
               Walter Wang and
               Eric Wilcox and
               Doe Hyun Yoon},
  title     = {In-Datacenter Performance Analysis of a Tensor Processing Unit},
  journal   = {CoRR},
  volume    = {arXiv:1704.04760},
  year      = {2017},
  url       = {http://arxiv.org/abs/1704.04760},
}
@misc{TWO_PHOTON_TUTORIAL_A-14,
       author = {Glenn J. Goldey and Mark L. Andermann},
        title = {Simultaneous Two-Photon Calcium Imaging of Entire Cortical Columns},
         year = {2014},
 howpublished = {Society for Neuroscience}
}
@misc{TWO_PHOTON_TUTORIAL_B-14,
       author = {Jeffrey N. Stirman and Spencer L. Smith},
        title = {Mesoscale Two-Photon Microscopy: Engineering a Wide Field of View with Cellular Resolution},
         year = {2014},
 howpublished = {Society for Neuroscience}
}
@article{TaatgenandFrankHUMAN-FACTORS-03,
       author = {Taatgen, Niels and Lee, Frank},
        title = {Production Compilation: A Simple Mechanism to Model Complex Skill Acquisition},
      journal = {Human Factors},
       volume = {45},
         year = {2003},
        pages = {61-76},
     abstract = {In psychology many theories of skill acquisition have had great success in addressing the find details of learning relatively simple tasks, but can they scale up to complex tasks that are more typical of human learning in the world? In this paper we describe production composition , a theory of skill acquisition that combines aspects of the theories forwarded by Anderson (1982) and Newell and Roosenbloom (1981), that we believe can model the fine details of learning in complex and dynamic tasks. We use production composition to model in detail learning in a simulated air-traffic controller task.}
}
@inproceedings{TadepalietalICML-04,
       author = {Tadepalli, P. and Givan, R. and Driessens, K.},
        title = {Relational Reinforcement Learning An Overview},
    booktitle = {Proceedings of the ICML-2004 workshop on relational reinforcement learning},
      address = {Banff, Canada},
         year = {2004},
}
@article{TaietalCoRR-15,
        title = {Improved Semantic Representations From Tree-Structured Long Short-Term Memory Networks},
       author = {Kai Sheng Tai, Richard Socher, Christopher D. Manning},
      journal = {CoRR},
       volume = {arXiv:1503.00075},
         year = 2015,
     abstract = {A Long Short-Term Memory (LSTM) network is a type of recurrent neural network architecture which has recently obtained strong results on a variety of sequence modeling tasks. The only underlying LSTM structure that has been explored so far is a linear chain. However, natural language exhibits syntactic properties that would naturally combine words to phrases. We introduce the Tree-LSTM, a generalization of LSTMs to tree-structured network topologies. Tree-LSTMs outperform all existing systems and strong LSTM baselines on two tasks: predicting the semantic relatedness of two sentences (SemEval 2014, Task 1) and sentiment classification (Stanford Sentiment Treebank).}
}
@article{TainakaetalCELL-14,
       author = {Tainaka, K. and Kubota, S. I. and Suyama, T. Q. and Susaki, E. A. and Perrin, D. and Ukai-Tadenuma, M. and Ukai, H. and Ueda, H. R.},
        title = {{W}hole-body imaging with single-cell resolution by tissue decolorization},
      journal = {Cell},
         year = 2014,
       volume = 159,
        issue = 4,
        pages = {911-924},
     abstract = {The development of whole-body imaging at single-cell resolution enables system-level approaches to studying cellular circuits in organisms. Previous clearing methods focused on homogenizing mismatched refractive indices of individual tissues, enabling reductions in opacity but falling short of achieving transparency. Here, we show that an aminoalcohol decolorizes blood by efficiently eluting the heme chromophore from hemoglobin. Direct transcardial perfusion of an aminoalcohol-containing cocktail that we previously termed CUBIC coupled with a 10 day to 2 week clearing protocol decolorized and rendered nearly transparent almost all organs of adult mice as well as the entire body of infant and adult mice. This CUBIC-perfusion protocol enables rapid whole-body and whole-organ imaging at single-cell resolution by using light-sheet fluorescent microscopy. The CUBIC protocol is also applicable to {3D} pathology, anatomy, and immunohistochemistry of various organs. These results suggest that whole-body imaging of colorless tissues at high resolution will contribute to organism-level systems biology.},
}
@article{TaitetalCoRR-16.,
        title = {Neuromorphic Silicon Photonics},
       author = {Alexander N. Tait and Ellen Zhou and Thomas Ferreira de Lima and Allie X. Wu and Mitchell A. Nahmias and Bhavin J. Shastri and Paul R. Prucnal},
      journal = {CoRR},
       volume = {arXiv:1611.02272},
         year = {2016},
     abstract = {We report first observations of an integrated analog photonic network, in which connections are configured by microring weight banks, as well as the first use of electro-optic modulators as photonic neurons. A mathematical isomorphism between the silicon photonic circuit and a continuous neural model is demonstrated through dynamical bifurcation analysis. Exploiting this isomorphism, existing neural engineering tools can be adapted to silicon photonic information processing systems. A 49-node silicon photonic neural network programmed using a "neural compiler" is simulated and predicted to outperform a conventional approach 1,960-fold in a toy differential system emulation task. Photonic neural networks leveraging silicon photonic platforms could access new regimes of ultrafast information processing for radio, control, and scientific computing.}
}
@article{TakedaetalFiP-18,
       author = {Takeda, Kazuyoshi and Sumiyoshi, Tomiki and Matsumoto, Madoka and Murayama, Kou and Ikezawa, Satoru and Matsumoto, Kenji and Nakagome, Kazuyuki},
        title = {Neural Correlates for Intrinsic Motivational Deficits of Schizophrenia; Implications for Therapeutics of Cognitive Impairment},
      journal = {Frontiers in Psychiatry},
       volume = {9},
        pages = {178},
         year = {2018},
     abstract = {The ultimate goal of the treatment of schizophrenia is recovery, a notion related to improvement of cognitive and social functioning. Cognitive remediation therapies (CRT), one of the most effective cognition enhancing methods, have been shown to moderately improve social functioning. For this purpose, intrinsic motivation, related to internal values such as interest and enjoyment, has been shown to play a key role. Although the impairment of intrinsic motivation is one of the characteristics of schizophrenia, its neural mechanisms remain unclear. This is related to the lack of feasible measures of intrinsic motivation, and its response to treatment. According to the self-determination theory (SDT), not only intrinsic motivation, but extrinsic motivation has been reported to enhance learning and memory in healthy subjects to some extent. This finding suggests the contribution of different types of motivation to potentiate the ability of the CRT to treat cognitive impairment of schizophrenia. In this paper, we provide a review of psychological characteristics, assessment methods, and neural correlates of intrinsic motivation in healthy subjects and patients with schizophrenia. Particularly, we focus on neuroimaging studies of intrinsic motivation, including our own. These considerations are relevant to enhancement of functional outcomes of schizophrenia.}
}
@inproceedings{TakedaetalICIP-06,
     title = {Robust Kernel Regression for Restoration and Reconstruction of Images from Sparse Noisy Data},
    author = {Hiroyuki Takeda and Sina Farsiu and Peyman Milanfar},
 booktitle = {Proceedings of the International Conference on Image Processing},
      year = 2006,
     pages = {1257-1260},
}
@article{TakedaetalTIP-07,
     title = {Kernel Regression for Image Processing and Reconstruction},
    author = {Hiroyuki Takeda and Sina Farsiu and Peyman Milanfar},
   journal = {{IEEE} Transactions on Image Processing},
    volume = 16,
     issue = 2,
      year = 2007,
     pages = {349-366},
}
@article{TakemuraetalNATURE-13,
       author = {Takemura, Shin-ya and Bharioke, Arjun and Lu, Zhiyuan and Nern, Aljoscha and Vitaladevuni, Shiv and Rivlin, Patricia K. and Katz, William T. and Olbris, Donald J. and Plaza, Stephen M. and Winston, Philip and Zhao, Ting and Horne, Jane Anne and Fetter, Richard D. and Takemura, Satoko and Blazek, Katerina and Chang, Lei-Ann and Ogundeyi, Omotara and Saunders, Mathew A. and Shapiro, Victor and Sigmund, Christopher and Rubin, Gerald M. and Scheffer, Louis K. and Meinertzhagen, Ian A. and Chklovskii, Dmitri B.},
        title = {A visual motion detection circuit suggested by Drosophila connectomics},
      journal = {Nature},
         year = {2013},
    publisher = {Nature Publishing Group},
       volume = {500},
        issue = {7461},
        pages = {175-181},
     abstract = {Animal behaviour arises from computations in neuronal circuits, but our understanding of these computations has been frustrated by the lack of detailed synaptic connection maps, or connectomes. For example, despite intensive investigations over half a century, the neuronal implementation of local motion detection in the insect visual system remains elusive. Here we develop a semi-automated pipeline using electron microscopy to reconstruct a connectome, containing 379 neurons and 8,637 chemical synaptic contacts, within the Drosophila optic medulla. By matching reconstructed neurons to examples from light microscopy, we assigned neurons to cell types and assembled a connectome of the repeating module of the medulla. Within this module, we identified cell types constituting a motion detection circuit, and showed that the connections onto individual motion-sensitive neurons in this circuit were consistent with their direction selectivity. Our results identify cellular targets for future functional investigations, and demonstrate that connectomes can provide key insights into neuronal computations.}
}
@article{TakemuraetalPNAS-15,
       author = {Takemura, S. Y. and Xu, C. S. and Lu, Z. and Rivlin, P. K. and Parag, T. and Olbris, D. J. and Plaza, S. and Zhao, T. and Katz, W. T. and Umayam, L. and Weaver, C. and Hess, H. F. and Horne, J. A. and Nunez-Iglesias, J. and Aniceto, R. and Chang, L. A. and Lauchie, S. and Nasca, A. and Ogundeyi, O. and Sigmund, C. and Takemura, S. and Tran, J. and Langille, C. and Le Lacheur, K. and McLin, S. and Shinomiya, A. and Chklovskii, D. B. and Meinertzhagen, I. A. and Scheffer, L. K.},
        title = {Synaptic circuits and their variations within different columns in the visual system of {D}rosophila},
      journal = {Proceedings of the National Academy of Science},
         year = {2015},
       volume = {112},
       number = {44},
        pages = {13711-13716},
     abstract = {We reconstructed the synaptic circuits of seven columns in the second neuropil or medulla behind the fly's compound eye. These neurons embody some of the most stereotyped circuits in one of the most miniaturized of animal brains. The reconstructions allow us, for the first time to our knowledge, to study variations between circuits in the medulla's neighboring columns. This variation in the number of synapses and the types of their synaptic partners has previously been little addressed because methods that visualize multiple circuits have not resolved detailed connections, and existing connectomic studies, which can see such connections, have not so far examined multiple reconstructions of the same circuit. Here, we address the omission by comparing the circuits common to all seven columns to assess variation in their connection strengths and the resultant rates of several different and distinct types of connection error. Error rates reveal that, overall, < 1\% of contacts are not part of a consensus circuit, and we classify those contacts that supplement (E+) or are missing from it (E-). Autapses, in which the same cell is both presynaptic and postsynaptic at the same synapse, are occasionally seen; two cells in particular, Dm9 and Mi1, form > 20-fold more autapses than do other neurons. These results delimit the accuracy of developmental events that establish and normally maintain synaptic circuits with such precision, and thereby address the operation of such circuits. They also establish a precedent for error rates that will be required in the new science of connectomics.}
}
@incollection{Takens81,
        title = {Detecting Strange Attractors in Turbulence},
       author = {Takens, F.},
    booktitle = {Dynamical Systems and Turbulence: Lecture Notes in Mathematics, Volume 898},
       editor = {Rand, D. A. and Young, L.-S.},
    publisher = {Springer-Verlag},
      address = {Berlin},
        pages = {336-381},
         year = 1981,
}
@inproceedings{TakeuchiaandDeValoisSPIE-05,
       title = {Sharpening image motion based on the spatio-temporal characteristics of human vision}, 
      author = {Tatsuto Takeuchia and Karen K. De Valois}, 
   booktitle = {Proceedings of SPIE Conference on Human Vision and Electronic Imaging}, 
      editor = {Bernice E. Rogowitz},
        year = 2005,
       pages = {83-94}
}
@article{TakeuchietalSCIENCE-11,
       author = {Takeuchi, D. and Hirabayashi, T. and Tamura, K. and Miyashita, Y.},
        title = {{R}eversal of interlaminar signal between sensory and memory processing in monkey temporal cortex},
      journal = {Science},
         year = 2011,
       volume = 331,
        issue = 6023,
        pages = {1443-1447},
     abstract = {The primate temporal cortex implements visual long-term memory. However, how its interlaminar circuitry executes cognitive computations is poorly understood. Using linear-array multicontact electrodes, we simultaneously recorded unit activities across cortical layers in the perirhinal cortex of macaques performing a pair-association memory task. Cortical layers were estimated on the basis of current source density profiles with histological verifications, and the interlaminar signal flow was determined with cross-correlation analysis between spike trains. During the cue period, canonical "feed-forward" signals flowed from granular to supragranular layers and from supragranular to infragranular layers. During the delay period, however, the signal flow reversed to the "feed-back" direction: from infragranular to supragranular layers. This reversal of signal flow highlights how the temporal cortex differentially recruits its laminar circuits for sensory and mnemonic processing.}
}
@article{TallisWSJ-11,
     title = {Rethinking Thinking: How a lumpy bunch of tissue lets us plan, perceive, calculate, reflect, imagine and exercise free-will},
    author = {Raymond Tallis},
   journal = {Wall Street Journal},
      date = {November 12},
      year = 2011,
     pages = {C5-C8},
}
@inproceedings{TalwalkaretalCVPR-08,
     author = {Ameet Talwalkar and Sanjiv Kumar and Henry A. Rowley},
      title = {Large-scale manifold learning},
  booktitle = {2008 IEEE Computer Society Conference on Computer Vision and Pattern Recognition},
  publisher = {IEEE Computer Society},
       year = 2008,
}
@article{TamaretalCoRR-16,
       author = {Aviv Tamar and Sergey Levine and Pieter Abbeel},
        title = {Value Iteration Networks},
      journal = {CoRR},
         year = {2016},
       volume = {arXiv:1602.02867},
     abstract = {We introduce the value iteration network (VIN): a fully differentiable neural network with a `planning module' embedded within. VINs can learn to plan, and are suitable for predicting outcomes that involve planning-based reasoning, such as policies for reinforcement learning. Key to our approach is a novel differentiable approximation of the value-iteration algorithm, which can be represented as a convolutional neural network, and trained end-to-end using standard backpropagation. We evaluate VIN based policies on discrete and continuous path-planning domains, and on a natural-language based search task. We show that by learning an explicit planning computation, VIN policies generalize better to new, unseen domains.}
}
@incollection{TamaretalNIPS-16,
        title = {Value Iteration Networks},
       author = {Tamar, Aviv and Wu, Yi and Thomas, Garrett and Levine, Sergey and Abbeel, Pieter},
    booktitle = {Advances in Neural Information Processing Systems 29},
       editor = {D. D. Lee and M. Sugiyama and U. V. Luxburg and I. Guyon and R. Garnett},
        pages = {2154-2162},
         year = {2016},
     location = {Barcelona, Spain},
    publisher = {Curran Associates, Inc.},
}
@article{TamsiretalNATURE-11,
        title = {Robust multicellular computing using genetically encoded NOR gates and chemical 'wires'},
       author = {Tamsir, A. and Tabor, J.~J. and Voigt, C.A.},
      journal = {Nature},
       volume = 469,
        issue = 7329,
        pages = {212-5},
         year = 2011,
     abstract = {Computation underlies the organization of cells into higher-order structures, for example during development or the spatial association of bacteria in a biofilm. Each cell performs a simple computational operation, but when combined with cell-cell communication, intricate patterns emerge. Here we study this process by combining a simple genetic circuit with quorum sensing to produce more complex computations in space. We construct a simple NOR logic gate in Escherichia coli by arranging two tandem promoters that function as inputs to drive the transcription of a repressor. The repressor inactivates a promoter that serves as the output. Individual colonies of E. coli carry the same NOR gate, but the inputs and outputs are wired to different orthogonal quorum-sensing 'sender' and 'receiver' devices. The quorum molecules form the wires between gates. By arranging the colonies in different spatial configurations, all possible two-input gates are produced, including the difficult XOR and EQUALS functions. The response is strong and robust, with 5- to >300-fold changes between the 'on' and 'off' states. This work helps elucidate the design rules by which simple logic can be harnessed to produce diverse and complex calculations by rewiring communication between cells.}
}
@article{TanakaandOhzawaJN-09,
       author = {Hiroki Tanaka and Izumi Ohzawa},
        title = {Surround Suppression of {V1} Neurons Mediates Orientation-Based Representation of High-Order Visual Features},
      journal = {Journal of Neurophysiology},
       volume = 101,
        pages = {1444-1462},
         year = 2009,
}
@inproceedings{TanandEswaranICMS-08,
       author = {Chun Chet Tan and C. Eswaran},
        title = {Performance Comparison of Three Types of Autoencoder Neural Networks},
    booktitle = {Proceedings of the 2nd Asia International Conference on Modeling and Simulation},
        pages = {213-218},
         year = 2008,
}
@book{Tanenbaum96,
       author = {Andrew S. Tanenbaum},
        title = {Computer Networks},
    publisher = {Third edition, Prentice Hall},
      address = {Upper Saddle River, NJ},
         year = 1996,
}
@article{TangenetalPERCEPTION-11,
        title = {Flashed face distortion effect: Grotesque faces from relative spaces},
       author = {Jason M. Tangen and Sean C. Murphy and Matthew B. Thompson},
      journal = {Perception},
       volume = 40,
        issue = 1,
         year = 2011,
        pages = {628-630},
}
@article{TangermannCOMPUTER-12,
       author = {Michael Tangermann, and Fabien Lotte, and Jan B. F. van Erp},
        title = {Brain-Computer Interfaces: Beyond Medical Applications},
      journal = {Computer},
    publisher = {IEEE Computer Society},
       volume = {45},
         year = {2012},
        pages = {26-34},
     abstract = {Brain-computer interaction has already moved from assistive care to applications such as gaming. Improvements in usability, hardware, signal processing, and system integration should yield applications in other nonmedical areas.}
}
@article{TangetalCoRR-17,
       author = {Luming Tang and Boyang Deng and Haiyu Zhao and Shuai Yi},
        title = {Hierarchical Deep Recurrent Architecture for Video Understanding},
      journal = {CoRR},
       volume = {arXiv:1707.03296},
         year = {2017},
     abstract = {This paper introduces the system we developed for the Youtube-8M Video Understanding Challenge, in which a large-scale benchmark dataset was used for multi-label video classification. The proposed framework contains hierarchical deep architecture, including the frame-level sequence modeling part and the video-level classification part. In the frame-level sequence modelling part, we explore a set of methods including Pooling-LSTM (PLSTM), Hierarchical-LSTM (HLSTM), Random-LSTM (RLSTM) in order to address the problem of large amount of frames in a video. We also introduce two attention pooling methods, single attention pooling (ATT) and multiply attention pooling (Multi-ATT) so that we can pay more attention to the informative frames in a video and ignore the useless frames. In the video-level classification part, two methods are proposed to increase the classification performance, i.e. Hierarchical-Mixture-of-Experts (HMoE) and Classifier Chains (CC). Our final submission is an ensemble consisting of 18 sub-models. In terms of the official evaluation metric Global Average Precision (GAP) at 20, our best submission achieves 0.84346 on the public 50\% of test dataset and 0.84333 on the private 50\% of test data.}
}
@article{TangetalNEURON-14,
       author = {Tang, G.  and Gudsnuk, K.  and Kuo, S. H.  and Cotrina, M. L.  and Rosoklija, G.  and Sosunov, A.  and Sonders, M. S.  and Kanter, E.  and Castagna, C.  and Yamamoto, A.  and Yue, Z.  and Arancio, O.  and Peterson, B. S.  and Champagne, F.  and Dwork, A. J.  and Goldman, J.  and Sulzer, D. },
        title = {{{L}oss of m{T}{O}{R}-dependent macroautophagy causes autistic-like synaptic pruning deficits}},
      journal = {Neuron},
       volume = {83},
       number = {5},
         year = {2014},
        pages = {1131-1143},
     abstract = {Developmental alterations of excitatory synapses are implicated in autism spectrum disorders (ASDs). Here, we report increased dendritic spine density with reduced developmental spine pruning in layer V pyramidal neurons in postmortem ASD temporal lobe. These spine deficits correlate with hyperactivated mTOR and impaired autophagy. In Tsc2 ± ASD mice where mTOR is constitutively overactive, we observed postnatal spine pruning defects, blockade of autophagy, and ASD-like social behaviors. The mTOR inhibitor rapamycin corrected ASD-like behaviors and spine pruning defects in Tsc2 ± mice, but not in Atg7(CKO) neuronal autophagy-deficient mice or Tsc2 ± :Atg7(CKO) double mutants. Neuronal autophagy furthermore enabled spine elimination with no effects on spine formation. Our findings suggest that mTOR-regulated autophagy is required for developmental spine pruning, and activation of neuronal autophagy corrects synaptic pathology and social behavior deficits in ASD models with hyperactivated mTOR.}
}
@article{TaniJCS-98,
       author = {Tani, Jun},
        title = {An interpretation of the 'self' from the dynamical systems perspective: {A} constructivist approach},
    booktitle = {Journal of Consciousness Studies},
       volume = {5},
         year = {1998},
        pages = {516-542},
     abstract = {This study attempts to describe the notion of the 'self' using dynamical systems language based on the results of our robot learning experiments. A neural network model consisting of multiple modules is proposed, in which the interactive dynamics between the bottom-up perception and the top-down prediction are investigated. Our experiments with a real mobile robot showed that the incremental learning of the robot switches spontaneously between steady and unsteady phases. In the steady phase, the top-down prediction for the bottom-up perception works well when coherence is achieved between the internal and the environmental dynamics. In the unsteady phase, conflicts arise between the bottom-up perception and the top-down prediction; the coherence is lost, and a chaotic attractor is observed in the internal neural dynamics. By investigating possible analogies between this result and the phenomenological literature on the 'self', we draw the conclusions that (1) the structure of the 'self' corresponds to the 'open dynamic structure' which is characterized by co-existence of stability in terms of goal-directedness and instability caused by embodiment; (2) the open dynamic structure causes the system's spontaneous transition to the unsteady phase where the 'self' becomes aware.}
}
@article{TarjanJADM-85,
       author = {Tarjan, R. E.},
        title = {Amortized Computational Complexity},
      journal = {SIAM Journal of Algebraic Discrete Methods},
       volume = 6,
         year = 1985,
        pages = {306-318},
}
@incollection{TarrSINGLE-MECHANISM-03,
       author = {Tarr, M. J.},
        title = {Visual Object Recognition: {C}an a Single Mechanism Suffice?},
       editor = {M. A. Peterson and G. Rhodes},
    booktitle = {Perception of Faces, Objects, and Scenes: Analytic and Holistic Processes},
    publisher = {Oxford University Press},
      address = {Oxford, UK},
        pages = {177-211},
         year = 2003,
}
@article{TarrandBulthoffCOGNITION-98,
        title = {Image-based object recognition in man, monkey and machine},
       author = {Michael Tarr and Heinrich B\"{u}lthoff},
      journal = {Cognition},
       volume = 67,
        pages = {1-20},
         year = 1998,
}
@inproceedings{TashandRussellAAAI-94,
       author = {Tash, Jonathan and Russell, Stuart},
        title = {Control Strategies for a Stochastic Planner},
    booktitle = {Proceedings AAAI-94},
     location = {Seattle, Washington},
 organization = {AAAI},
         year = 1994,
}
@article{TassaetalCoRR-18,
       author = {Yuval Tassa and Yotam Doron and Alistair Muldal and Tom Erez and Yazhe Li and Diego de Las Casas and David Budden and Abbas Abdolmaleki and Josh Merel and Andrew Lefrancq and Timothy P. Lillicrap and Martin A. Riedmiller},
        title = {DeepMind Control Suite},
      journal = {CoRR},
       volume = {arXiv:1801.00690},
         year = {2018},
     abstract = {The DeepMind Control Suite is a set of continuous control tasks with a standardised structure and interpretable rewards, intended to serve as performance benchmarks for reinforcement learning agents. The tasks are written in Python and powered by the MuJoCo physics engine, making them easy to use and modify. We include benchmarks for several learning algorithms. The Control Suite is publicly available at this https URL. A video summary of all tasks is available at this http URL.}
}
@inproceedings{TateIJCAI-77,
       author = {Tate, Austin},
        title = {Generating Project Networks},
    booktitle = {Proceedings {IJCAI} 5},
     location = {Cambridge, Massachusetts},
      comment = {NONLIN},
 organization = {IJCAII},
         year = 1977,
        pages = {888-893},
}
@inproceedings{TateandWhiter84,
       author = {Tate, A. and A. M. Whiter},
        title = {Planning with Multiple Resource Constraints and an Application to a Naval Planning Problem},
    booktitle = {Proceedings of the First Conference on the Applications of Artificial Intelligence},
         year = 1984,
}
@article{TatmanSMC-90,
       author = {Tatman, Joseph A. and Shachter, Ross D.},
        title = {Dynamic Programming and Influence Diagrams},
      journal = {{IEEE} Transactions on Systems, Man, and Cybernetics},
       volume = 20,
        issue = 2,
        pages = {365-379},
         year = 1990,
}
@article{TauandPetersonNEUROPSYCHOPHARMACOLOGY-10,
       author = {Tau, Gregory Z. and Peterson, Bradley S.},
        title = {Normal development of brain circuits},
      journal = {Neuropsychopharmacology},
         year = {2010},
    publisher = {Nature Publishing Group},
       volume = {35},
        issue = {1},
        pages = {147-168},
     abstract = {Spanning functions from the simplest reflex arc to complex cognitive processes, neural circuits have diverse functional roles. In the cerebral cortex, functional domains such as visual processing, attention, memory, and cognitive control rely on the development of distinct yet interconnected sets of anatomically distributed cortical and subcortical regions. The developmental organization of these circuits is a remarkably complex process that is influenced by genetic predispositions, environmental events, and neuroplastic responses to experiential demand that modulates connectivity and communication among neurons, within individual brain regions and circuits, and across neural pathways. Recent advances in neuroimaging and computational neurobiology, together with traditional investigational approaches such as histological studies and cellular and molecular biology, have been invaluable in improving our understanding of these developmental processes in humans in both health and illness. To contextualize the developmental origins of a wide array of neuropsychiatric illnesses, this review describes the development and maturation of neural circuits from the first synapse through critical periods of vulnerability and opportunity to the emergent capacity for cognitive and behavioral regulation, and finally the dynamic interplay across levels of circuit organization and developmental epochs.},
}
@inproceedings{TauheedetalICDE-12,
       author = {Tauheed, F. and Biveinis, L. and Heinis, T. and Sch\"{u}rmann, F. and Markram, H. and Ailamaki, A.},
        title = {Accelerating Range Queries for Brain Simulations},
    booktitle = {IEEE International Conference on Data Engineering},
    publisher = {{IEEE} Computer Society},
         year = 2012,
        pages = {941-952},
     abstract = {Neuroscientists increasingly use computational tools in building and simulating models of the brain. The amounts of data involved in these simulations are immense and efficiently managing this data is key. One particular problem in analyzing this data is the scalable execution of range queries on spatial models of the brain. Known indexing approaches do not perform well even on today's small models which represent a small fraction of the brain, containing only few millions of densely packed spatial elements. The problem of current approaches is that with the increasing level of detail in the models, also the overlap in the tree structure increases, ultimately slowing down query execution. The neuroscientists' need to work with bigger and more detailed (denser) models thus motivates us to develop a new indexing approach. To this end we develop FLAT, a scalable indexing approach for dense data sets. We base the development of FLAT on the key observation that current approaches suffer from overlap in case of dense data sets. We hence design FLAT as an approach with two phases, each independent of density. In the first phase it uses a traditional spatial index to retrieve an initial object efficiently. In the second phase it traverses the initial object's neighborhood to retrieve the remaining query result. Our experimental results show that FLAT not only outperforms R-Tree variants from a factor of two up to eight but that it also achieves independence from data set size and density.},
}
@article {TawficketalAM-12,
       author = {Tawfick, Sameh and De Volder, Michael and Copic, Davor and Park, Sei Jin and Oliver, C. Ryan and Polsen, Erik S. and Roberts, Megan J. and Hart, A. John},
        title = {Engineering of Micro- and Nanostructured Surfaces with Anisotropic Geometries and Properties},
      journal = {Advanced Materials},
       volume = 24,
        issue = 13,
         year = 2012,
        pages = {1628-1674},
     abstract = {Widespread approaches to fabricate surfaces with robust micro- and nanostructured topographies have been stimulated by opportunities to enhance interface performance by combining physical and chemical effects. In particular, arrays of asymmetric surface features, such as arrays of grooves, inclined pillars, and helical protrusions, have been shown to impart unique anisotropy in properties including wetting, adhesion, thermal and/or electrical conductivity, optical activity, and capability to direct cell growth. These properties are of wide interest for applications including energy conversion, microelectronics, chemical and biological sensing, and bioengineering. However, fabrication of asymmetric surface features often pushes the limits of traditional etching and deposition techniques, making it challenging to produce the desired surfaces in a scalable and cost-effective manner. We review and classify approaches to fabricate arrays of asymmetric 2D and {3D} surface features, in polymers, metals, and ceramics. Analytical and empirical relationships among geometries, materials, and surface properties are discussed, especially in the context of the applications mentioned above. Further, opportunities for new fabrication methods that combine lithography with principles of self-assembly are identified, aiming to establish design principles for fabrication of arbitrary {3D} surface textures over large areas.}
}
@book{Taylor2016language,
        title = {The Language Animal: {T}he Full Shape of the Human Linguistic Capacity},
       author = {Taylor, Charles},
    publisher = {Harvard University Press},
      address = {Cambridge, MA},
         year = {2016},
}
@article{TaylorJQ-53,
       author = {Taylor, Wilson L.},
        title = {The Cloze procedure: a new tool for measuring readability},
      journal = {Journalism Quarterly},
       volume = 30,
         year = {415-433},
        pages = 1953,
     abstract = {Cloze Procedure involves no formula or "element counting," but consists of sampling all potential readability influences. Although similar to sentence-completion tests, the cloze method demands deletion of random words from a passage. After administration to a group the correctly identified omissions are tallied. Experimental results show: (1) the cloze method consistently ranked three selected passages in the same way as the Flesch and Dale-Chall formulas; (2) the method was reliable; (3) the cloze method seemed to handle specialized passages more adequately than other methods; (4) the same rankings of readability were obtained when words were deleted at random or every nth word; (5) the cloze procedure could be used for comparing reading abilities of different individuals.}
}
@article{TaylorTR-79,
       author = {Taylor, Russell H.},
        title = {Planning and execution of straight-line manipulator trajectories},
      journal = {IBM Journal of Research and Development},
       volume = 24,
         year = 1979,
        pages = {424-436},
}
@inproceedings{TayloretalECCV-10,
       author = {G. Taylor and R. Fergus and Y. LeCun and C. Bregler},
        title = {Convolutional Learning of Spatio-temporal Features},
    booktitle = {11th European Conference on Computer Vision},
         year = 2010,
}
@article{TayloretalIJRR-82,
       author = {Taylor, R. H. and Summers, P. D. and Meyer, J. M.},
        title = {AML: A Manufacturing Language},
      journal = {International Journal Robotics Research},
       volume = 1,
        pages = {19-41},
         year = 1982,
}
@book{Teegarden1965,
       author = {David M. Teegarden},
        title = {Polymer Chemistry: Introduction to an Indispensable Science},
    publisher = {National Science Teachers Association Press},
         year = 1965,
      address = {Arlington, Virginia},
}
@incollection{TefferandSemendeferiPBR-12,
       author = {Kate Teffer and Katerina Semendeferi},
        title = {Chapter 9 - Human prefrontal cortex: Evolution, development and pathology},
       editor = {Michel A. Hofman and Dean Falk},
       series = {Progress in Brain Research},
    publisher = {Elsevier},
       volume = {195},
        pages = {191-218},
         year = {2012},
    booktitle = {Evolution of the Primate Brain},
     abstract = {The prefrontal cortex is critical to many cognitive abilities that are considered particularly human, and forms a large part of a neural system crucial for normal socio-emotional and executive functioning in humans and other primates. In this chapter, we survey the literature regarding prefrontal development and pathology in humans as well as comparative studies of the region in humans and closely related primate species. The prefrontal cortex matures later in development than more caudal regions, and some of its neuronal subpopulations exhibit more complex dendritic arborizations. Comparative work suggests that the human prefrontal cortex differs from that of closely related primate species less in relative size than it does in organization. Specific reorganizational events in neural circuitry may have taken place either as a consequence of adjusting to increases in size or as adaptive responses to specific selection pressures. Living in complex environments has been recognized as a considerable factor in the evolution of primate cognition. Normal frontal lobe development and function are also compromised in several neurological and psychiatric disorders. A phylogenetically recent reorganization of frontal cortical circuitry may have been critical to the emergence of human-specific executive and social-emotional functions, and developmental pathology in these same systems underlies many psychiatric and neurological disorders, including autism and schizophrenia.}
}
@book{Tegmark2017life,
        title = {Life 3.0: Being Human in the Age of Artificial Intelligence},
       author = {Tegmark, Max},
         year = {2017},
    publisher = {Penguin Books Limited}
}
@article{TehetalJASA-06,
    author = {Y. W. Teh and M. I. Jordan and  M. J. Beal and D. M. Blei},
     title = {Hierarchical {D}irichlet processes},
   journal = {Journal of the American Statistical Association},
      year = 2006,
}
@article{TekietalFiIN-12,
       author = {Teki, Sundeep and Grube, Manon and Griffiths, Timothy D.},
        title = {A unified model of time perception accounts for duration-based and beat-based timing mechanisms},
      journal = {Frontiers in integrative neuroscience},
         year = {2012},
    publisher = {Frontiers Research Foundation},
       volume = {5},
        pages = {90-90},
     abstract = {Accurate timing is an integral aspect of sensory and motor processes such as the perception of speech and music and the execution of skilled movement. Neuropsychological studies of time perception in patient groups and functional neuroimaging studies of timing in normal participants suggest common neural substrates for perceptual and motor timing. A timing system is implicated in core regions of the motor network such as the cerebellum, inferior olive, basal ganglia, pre-supplementary, and supplementary motor area, pre-motor cortex as well as higher-level areas such as the prefrontal cortex. In this article, we assess how distinct parts of the timing system subserve different aspects of perceptual timing. We previously established brain bases for absolute, duration-based timing and relative, beat-based timing in the olivocerebellar and striato-thalamo-cortical circuits respectively (Teki et al., 2011). However, neurophysiological and neuroanatomical studies provide a basis to suggest that timing functions of these circuits may not be independent. Here, we propose a unified model of time perception based on coordinated activity in the core striatal and olivocerebellar networks that are interconnected with each other and the cerebral cortex through multiple synaptic pathways. Timing in this unified model is proposed to involve serial beat-based striatal activation followed by absolute olivocerebellar timing mechanisms.},
}
@article{Tel-VeredAEMB-12,
       author = {Tel-Vered, R. and Yehezkeli, O. and Willner, I.},
        title = {Biomolecule/nanomaterial hybrid systems for nanobiotechnology},
      journal = {Advances in Experimental Medicine and Biology},
       volume = 733,
         year = 2012,
        pages = {1-16},
     abstract = {The integration of biomolecules with metallic or semiconductor nanoparticles or carbon nanotubes yields new hybrid nanostructures of unique features that combine the properties of the biomolecules and of the nano-elements. These unique features of the hybrid biomolecule/nanoparticle systems provide the basis for the rapid development of the area of nanobiotechnology. Recent advances in the implementation of hybrid materials consisting of biomolecules and metallic nanoparticles or semiconductor quantum dots will be discussed. The following topics will be exemplified: (i) The electrical wiring of redox enzymes with electrodes by means of metallic nanoparticles or carbon nanotubes, and the application of the modified electrodes as amperometric biosensors or for the construction of biofuel cells. (ii) The biocatalytic growth of metallic nanoparticles as a means to construct optical or electrical sensors. (iii) The functionalization of semiconductor quantum dots with biomolecules and the application of the hybrid nanostructures for developing different optical sensors, including intracellular sensor systems. (iv) The use of biomolecule-metallic nanoparticle nanostructures as templates for growing metallic nanowires, and the construction of fuel-driven nano-transporters.}
}
@article{TenenbaumandBarrowAIJ-77,
     author = {Jay Tenenbaum and Harry Barrow},
      title = {Experiments in Interpretation-Guided Segmentation},
    journal = AIJ,
     volume = 8,
      pages = {241-277},
       year = 1977,
}
@article {TenenbaumetalSCIENCE-11,
       author = {Tenenbaum, Joshua B. and Kemp, Charles and Griffiths, Thomas L. and Goodman, Noah D.},
        title = {How to Grow a Mind: Statistics, Structure, and Abstraction},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {331},
       number = {6022},
        pages = {1279-1285},
         year = {2011},
     abstract = {In coming to understand the world{\textemdash}in learning concepts, acquiring language, and grasping causal relations{\textemdash}our minds make inferences that appear to go far beyond the data available. How do we do it? This review describes recent approaches to reverse-engineering human learning and cognitive development and, in parallel, engineering more humanlike machine learning systems. Computational models that perform probabilistic inference over hierarchies of flexibly structured representations can address some of the deepest questions about the nature and origins of human thought: How does abstract knowledge guide learning and reasoning from sparse data? What forms does our knowledge take, across different domains and tasks? And how is that abstract knowledge itself acquired?},
}
@incollection{Tenenberg91,
       author = {Josh D. Tenenberg},
        title = {Abstraction in Planning},
       editor = {J. F. Allen and H. A. Kautz and R. N. Pelavin and J. D. Tenenberg},
    booktitle = {Reasoning about Plans},
        pages = {213-280},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1991,
}
@inproceedings{TenenbergKR-89,
       author = {Tenenberg, Josh D.},
        title = {Inheritance in automated planning},
     crossref = {KR-89},
        pages = {475-485},
}
@book{TerHaarRomeny03,
       author = {B. M. Ter Haar Romeny},
        title = {Front-End Vision and Multi-Scale Image Analysis: Multi-scale Computer Vision Theory and Applications, written in Mathematica},
    publisher = {Springer},
          url = {http://www.bmi2.bmt.tue.nl/image-analysis/People/BRomeny/},
      address = {New York},
         year = 2003,
}
@book{Terrence_W_Deacon_The_Symbolic_Species,
        title = {The Symbolic Species: The Co-evolution of Language and the Brain},
       author = {Deacon, Terrence W.},
         year = {1998},
    publisher = {W. W. Norton},
      comment = {https://uberty.org/wp-content/uploads/2016/02/Terrence_W._Deacon_The_Symbolic_Species.pdf}
}
@inproceedings{TerzopoulosICCV-95,
       author = {D. Terzopoulos and T. Rabie},
        title = {Animat vision: Active vision with artificial animals},
    booktitle = {Proceedings of the Fifth {IEEE} International Conference on Computer Vision},
       volume = 1,
         year = 1995,
        pages = {801-808},
}
@book{TerzopoulosSIGGRAPH-97,
       editor = {Demetri Terzopoulos},
        title = {Artificial Life for Graphics, Animation, Multimedia, and Virtual Reality},
    publisher = {{ACM SIGGRAPH}},
       volume = 23,
         year = 1997,
        pages = {2-19},
     abstract = {The article focuses on artificial life for computer graphics. Computer graphics modeling for image synthesis, animation and virtual reality has advanced dramatically over the past decade, revolutionizing the motion picture, interactive game and multimedia industries. The field has advanced from first generation purely geometric models to more elaborate physics-based models. The modeling and simulation of living systems for computer graphics resonates with an emerging field of scientific inquiry called artificial life or Alife, a discipline that transcends the traditional boundaries of computer science and biological science. The synergy between computer graphics and artificial life defines the leading edge of advanced graphics modeling. Artificial life concepts play a central role in the construction of advanced graphics models for image synthesis, animation, multimedia and virtual reality. New graphics models have taken bold steps toward the realistic emulation of a variety of living things including plants and animals, from lower organisms all the way up the evolutionary ladder to humans.}
}
@article{TesauroML-92,
       author = {Tesauro, Gerald},
        title = {Practial Issues in Temporal Difference Learning},
      journal = {Machine Learning},
       volume = 8,
        issue = {3-4},
 month        = May,
        pages = {257-278},
         year = 1992,
}
@incollection{TesauroandSejnowski87,
    author = {Tesauro, G and Sejnowski, T. J.},
     title = {A Neural Network that Learns to Play Backgammon},
    editor = {Anderson, D. Z.},
 booktitle = {Neural Information Processing Systems},
 publisher = {American Institute of Physics},
   address = {New York},
      year = 1987,
}
@article{TeshomeetalNANOSCALE-14,
       author = {Teshome, Bezuayehu and Facsko, Stefan and Keller, Adrian},
        title = {Topography-controlled alignment of DNA origami nanotubes on nanopatterned surfaces},
      journal = {Nanoscale},
    publisher = {The Royal Society of Chemistry},
         year = 2014,
       volume = 6,
        issue = 3,
        pages = {1790-1796},
     abstract = {The controlled positioning of DNA nanostructures on technologically relevant surfaces represents a major goal along the route toward the full-scale integration of DNA-based materials into nanoelectronic and sensor devices. Previous attempts to arrange DNA nanostructures into defined arrays mostly relied on top-down lithographic patterning techniques combined with chemical surface functionalization. Here we combine two bottom-up techniques for nanostructure fabrication, i.e., self-organized nanopattern formation and DNA origami self-assembly, in order to demonstrate the electrostatic self-alignment of DNA nanotubes on topographically patterned silicon surfaces. Self-organized nanoscale ripple patterns with periodicities ranging from 20 nm to 50 nm are fabricated by low-energy ion irradiation and serve as substrates for DNA origami adsorption. Electrostatic interactions with the charged surface oxide during adsorption direct the DNA origami nanotubes to the ripple valleys and align them parallel to the ripples. By optimizing the pattern dimensions and the Debye length of the adsorption buffer, we obtain an alignment yield of [similar]70\%. Since this novel and versatile approach does not rely on any chemical functionalization of the surface or the DNA nanotubes, it can be applied to virtually any substrate material and any top-down or bottom-up nanopatterning technique. This technique thus may enable the wafer-scale fabrication of ordered arrays of functional DNA-based nanowires.}
}
@article{TetreaultetalJADD-12,
       author = {Tetreault, N. A.  and Hakeem, A. Y.  and Jiang, S.  and Williams, B. A.  and Allman, E.  and Wold, B. J.  and Allman, J. M. },
        title = {Microglia in the cerebral cortex in autism},
      journal = {Journal of Autism Developmental Disorders},
       volume = {42},
       number = {12},
         year = {2012},
        pages = {2569-2584},
     abstract = {We immunocytochemically identified microglia in fronto-insular (FI) and visual cortex (VC) in autopsy brains of well-phenotyped subjects with autism and matched controls, and stereologically quantified the microglial densities. Densities were determined blind to phenotype using an optical fractionator probe. In FI, individuals with autism had significantly more microglia compared to controls (p = 0.02). One such subject had a microglial density in FI within the control range and was also an outlier behaviorally with respect to other subjects with autism. In VC, microglial densities were also significantly greater in individuals with autism versus controls (p = 0.0002). Since we observed increased densities of microglia in two functionally and anatomically disparate cortical areas, we suggest that these immune cells are probably denser throughout cerebral cortex in brains of people with autism.}
}
@article{TeylerandDiScennaBEHAVIORAL-NEUROSCIENCE-86,
       author = {Teyler, T. J. and DiScenna, P.},
        title = {{T}he hippocampal memory indexing theory},
      journal = {Behavioral Neuroscience},
         year = {1986},
       volume = {100},
       number = {2},
        pages = {147-154},
     abstract = {The hippocampal formation (comprising the hippocampus proper, the dentate gyrus, and the subiculum) has been repeatedly implicated in information storage models of the mammalian brain. The precise nature of the hippocampal role in the storage of information has, however, remained elusive. Here it is proposed that the role of the hippocampus is to form and retain an index of neocortical areas activated by experiential events. The hippocampal index, thus, represents those unique cortical regions activated by specific events. The neuronal mechanism underlying the memory index is hypothesized to be long-term potentiation. It is asserted that the reactivation of the stored hippocampal memory index will serve to also reactivate the associated unique array of neocortical areas and thus will result in a memorial experience. This hippocampal reactivation of a neocortical array may also be involved in establishing a cortically based memory trace.}
}
@article{TeylerandRudyHIPPOCAMPUS-07,
       author = {Teyler, T. J. and Rudy, J. W.},
        title = {The hippocampal indexing theory and episodic memory: updating the index},
      journal = {Hippocampus},
         year = {2007},
       volume = {17},
       number = {12},
        pages = {1158-1169},
     abstract = {A little over 20 years ago, (Teyler and DiScenna,1986; Behav Neurosci 100:147-152) proposed the hippocampal memory index theory. It offered an account of episodic memory based on the intrinsic organization of the hippocampus, its synaptic physiology and its anatomical relationship to other regions of the brain. The essence of their idea was that the hippocampus was functionally designed and anatomically situated to capture information about neocortical activity generated by the individual features of behavioral episode. Moreover, because the hippocampus projects back to these neocortical regions the information it stored could serve as an index to the pattern of neocortical activity produced by the episode. Consequently, a partial cue that activated the index could activate the neocortical patterns and thus retrieve the memory of the episode. In this article we revisit and update indexing theory. Our conclusion is that it has aged very well. Its core ideas can be seen in many contemporary theories and there is a wealth of data that support this conceptual framework.}
}
@article{ThaissandDeBleserCORTEX-92,
       author = {Thaiss, L. and De Bleser, R.},
        title = {Visual agnosia: a case of reduced attentional "spotlight"?},
      journal = {Cortex},
       volume = 28,
        issue = 4,
         year = 1992,
        pages = {601-621},
     abstract = {A case study is presented of a patient with presenile dementia, for whom the dominant clinical feature from onset was a visual agnosia. The characteristics of the patient's visual agnosia were investigated in light of her apparent use of a "feature-by-feature" strategy to identify objects. Results from various tasks showed that the patient was unable to use global shape information or other grossly defined property cues characteristic of a "wide angle" attentional processing stage (Treisman, 1988) in object recognition. The patient appeared instead to rely on 'parts' or identifying features of the objects for object recognition. The patient showed significant improvement when the size of the drawing was reduced in size, thus suggesting that the disorder may be functionally localized to a reduction of the patient's attentional "spotlight".}
}
@book{ThalerandSunstein2008,
     author = {Richard H. Thaler and Cass R. Sunstein},
      title = {Nudge: Improving Decisions About Health, Wealth, and Happiness},
  publisher = {Yale University Press},
    address = {New Haven, CT},
       year = 2008,
    comment = {If you know what people like you tend to like, you might well be 
               comfortable in selecting products you don't know, because people 
               like you tend to like them. For many of us, collaborative filtering 
               is making difficult choices easier. Cautionary note: suprise and 
               serendipity can be fun for people, and good for them too, and it 
               may not be entirely wonderful if our primary source of information 
               is about what people like us like. --- pages 96--97},
}
@article {ThavabalasingametalPNAS-19,
       author = {Thavabalasingam, Sathesan and O`Neil, Edward B. and Tay, Jonathan and Nestor, Adrian and Lee, Andy C. H.},
        title = {Evidence for the incorporation of temporal duration information in human hippocampal long-term memory sequence representations},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
         year = {2019},
     abstract = {We demonstrate that multivariate patterns of activity in the human hippocampus during the recognition and cued mental replay of long-term sequence memories contain temporal structure information in the order of seconds. By using an experimental paradigm that required participants to remember the durations of empty intervals between visually presented scene images, our study provides evidence that the human hippocampus can represent elapsed time within a sequence of events in conjunction with other forms of information, such as event content. Our findings complement rodent studies that have shown that hippocampal neurons fire at specific times during the empty delay between two events and suggest a common hippocampal neural mechanism for representing temporal information in the service of episodic memory. There has been much interest in how the hippocampus codes time in support of episodic memory. Notably, while rodent hippocampal neurons, including populations in subfield CA1, have been shown to represent the passage of time in the order of seconds between events, there is limited support for a similar mechanism in humans. Specifically, there is no clear evidence that human hippocampal activity during long-term memory processing is sensitive to temporal duration information that spans seconds. To address this gap, we asked participants to first learn short event sequences that varied in image content and interval durations. During fMRI, participants then completed a recognition memory task, as well as a recall phase in which they were required to mentally replay each sequence in as much detail as possible. We found that individual sequences could be classified using activity patterns in the anterior hippocampus during recognition memory. Critically, successful classification was dependent on the conjunction of event content and temporal structure information (with unsuccessful classification of image content or interval duration alone), and further analyses suggested that the most informative voxels resided in the anterior CA1. Additionally, a classifier trained on anterior CA1 recognition data could successfully identify individual sequences from the mental replay data, suggesting that similar activity patterns supported participants recognition and recall memory. Our findings complement recent rodent hippocampal research, and provide evidence that long-term sequence memory representations in the human hippocampus can reflect duration information in the order of seconds.},
}
@book{The_Future_of_Supercomputing_PNAS-05,
       author = {National Research Council},
       editor = {Susan L. Graham and Marc Snir and Cynthia A. Patterson},
        title = {Getting Up to Speed: The Future of Supercomputing},
         year = 2005,
    publisher = {The National Academies Press},
      address = {Washington, DC},
     abstract = {Supercomputers play a significant and growing role in a variety of areas important to the nation.  They are used to address challenging science and technology problems.   In recent years, however, progress in supercomputing in the United States has slowed.  The development of the Earth Simulator supercomputer by Japan that the United States could lose its competitive advantage and, more importantly, the national competence needed to achieve national goals.  In the wake of this development, the Department of Energy asked the NRC to assess the state of U.S. supercomputing capabilities and relevant R&D.  Subsequently, the Senate directed DOE in S. Rpt. 107-220 to ask the NRC to evaluate the Advanced Simulation and Computing program of the National Nuclear Security Administration at DOE in light of the development of the Earth Simulator.  This report provides an assessment of the current status of supercomputing in the United States including a review of current demand and technology, infrastructure and institutions, and international activities.  The report also presents a number of recommendations to enable the United States to meet current and future needs for capability supercomputers.},
}
@article{The_Life_and_Work_of_Donald_O_Hebb,
        title = {The Life and Work of Donald O. Hebb},
       author = {John E. Brown},
         year = {2006},
    publisher = {Acta Neurologica},
}
@incollection{TheisandBethgeNIPS-15,
        title = {Generative Image Modeling Using Spatial {LSTM}s},
       author = {Theis, Lucas and Bethge, Matthias},
    booktitle = {Advances in Neural Information Processing Systems 28},
       editor = {C. Cortes and N. D. Lawrence and D. D. Lee and M. Sugiyama and R. Garnett},
    publisher = {Curran Associates, Inc.},
         year = {2015},
        pages = {1918-1926},
     abstract = {Modeling the distribution of natural images is challenging, partly because of strong statistical dependencies which can extend over hundreds of pixels. Re- current neural networks have been successful in capturing long-range dependen- cies in a number of problems but only recently have found their way into gener- ative image models. We here introduce a recurrent image model based on multi- dimensional long short-term memory units which are particularly suited for image modeling due to their spatial structure. Our model scales to images of arbitrary size and its likelihood is computationally tractable. We find that it outperforms the state of the art in quantitative comparisons on several image datasets and produces promising results when used for texture synthesis and inpainting.}
}
@article {TheisetalBIORXIV-15,
       author = {Theis, Lucas and Berens, Philipp and Froudarakis, Emmanouil and Reimer, Jacob and Roman-Roson, Miroslav and Baden, Thomas and Euler, Thomas and Tolias, Andreas S and Bethge, Matthias},
        title = {Supervised learning sets benchmark for robust spike detection from calcium imaging signals},
    publisher = {Cold Spring Harbor Labs Journals},
         year = 2015,
      journal = {bioRxiv},
     abstract = {A fundamental challenge in calcium imaging has been to infer the timing of action potentials from the measured noisy calcium fluorescence traces. We systematically evaluate a range of spike inference algorithms on a large benchmark dataset recorded from varying neural tissue (V1 and retina) using different calcium indicators (OGB-1 and GCamp6). We show that a new algorithm based on supervised learning in flexible probabilistic models outperforms all previously published techniques, setting a new standard for spike inference from calcium signals. Importantly, it performs better than other algorithms even on datasets not seen during training. Future data acquired in new experimental conditions can easily be used to further improve its spike prediction accuracy and generalization performance. Finally, we show that comparing algorithms on artificial data is not informative about performance on real population imaging data, suggesting that a benchmark dataset may greatly facilitate future algorithmic developments.},
}
@article{TheisetalJMLR-11,
     title = {In all likelihood, deep belief is not enough},
    author = {Lucas Theis and Sebastian Gerwinn and Fabian Sinz and Matthias Bethge},
   journal = {Journal of Machine Learning Research},
    volume = 12,
   comment = {http://jmlr.csail.mit.edu/papers/volume12/theis11a/theis11a.pdf},
      year = 2011,
     pages = {3071-3096},
}
@article{TheisetalPLoS-12,
    author = {Lucas Theis and R. Hosseini and Matthias Bethge},
     title = {Mixtures of conditional Gaussian scale mixtures applied to multiscale image representations},
   journal = {{PLoS} ONE},
 publisher = {Public Library of Science},
      year = 2012,
}
@inproceedings{TheocharousetalICRA-02,
    author = {G. Theocharous and K. Rohanimanesh and S. Mahadevan},
     title = {Learning hierarchical partially observable {M}arkov decision processes for robot navigation},
 booktitle = {IEEE Conference on Robotics and Automation (ICRA)},
  location = {Seoul, Korea},
      year = 2001,
}
@article{ThibeaultandSrinivasaFiCN-13,
       author = {Thibeault, C. M. and Srinivasa, N.},
        title = {Using a hybrid neuron in physiologically inspired models of the basal ganglia},
      journal = {Frontier Computational Neuroscience},
         year = {2013},
       volume = {7},
        pages = {88},
     abstract = {Our current understanding of the basal ganglia (BG) has facilitated the creation of computational models that have contributed novel theories, explored new functional anatomy and demonstrated results complementing physiological experiments. However, the utility of these models extends beyond these applications. Particularly in neuromorphic engineering, where the basal ganglia's role in computation is important for applications such as power efficient autonomous agents and model-based control strategies. The neurons used in existing computational models of the BG, however, are not amenable for many low-power hardware implementations. Motivated by a need for more hardware accessible networks, we replicate four published models of the BG, spanning single neuron and small networks, replacing the more computationally expensive neuron models with an Izhikevich hybrid neuron. This begins with a network modeling action-selection, where the basal activity levels and the ability to appropriately select the most salient input is reproduced. A Parkinson's disease model is then explored under normal conditions, Parkinsonian conditions and during subthalamic nucleus deep brain stimulation (DBS). The resulting network is capable of replicating the loss of thalamic relay capabilities in the Parkinsonian state and its return under DBS. This is also demonstrated using a network capable of action-selection. Finally, a study of correlation transfer under different patterns of Parkinsonian activity is presented. These networks successfully captured the significant results of the originals studies. This not only creates a foundation for neuromorphic hardware implementations but may also support the development of large-scale biophysical models. The former potentially providing a way of improving the efficacy of DBS and the latter allowing for the efficient simulation of larger more comprehensive networks.},
}
@incollection{ThiebauxetalEWSP-94,
       author = {Thiebaux, Sylvie and Hertzberg, Joachim and Shoaf, William and Schneider, Moti},
        title = {A Stochastic Model of Actions and Plans for for Anytime Planning under Uncertainty},
       editor = {Sandewall, E. and Backstrom, C.},
    booktitle = {Current Trends in {AI} Planning},
    publisher = {IOS Press},
      address = {Amsterdam},
         year = 1994,
}
@article{ThomaetalPP-15,
       author = {Thoma, N. and Pilecki, B. and McKay, D.},
        title = {Contemporary Cognitive Behavior Therapy: {A} Review of Theory, History, and Evidence},
      journal = {Psychodynamic Psychiatry},
         year = {2015},
       volume = {43},
       number = {3},
        pages = {423-461},
     abstract = {Cognitive behavior therapy (CBT) has come to be a widely practiced psychotherapy throughout the world. The present article reviews theory, history, and evidence for CBT. It is meant as an effort to summarize the forms and scope of CBT to date for the uninitiated. Elements of CBT such as cognitive therapy, behavior therapy, and so-called "third wave" CBT, such as dialectical behavior therapy (DBT) and acceptance and commitment therapy (ACT) are covered. The evidence for the efficacy of CBT for various disorders is reviewed, including depression, anxiety disorders, personality disorders, eating disorders, substance abuse, schizophrenia, chronic pain, insomnia, and child/adolescent disorders. The relative efficacy of medication and CBT, or their combination, is also briefly considered. Future directions for research and treatment development are proposed.}
}
@article{ThomasPHYSICS-D-13,
       author = {Andy Thomas},
        title = {Memristor-based neural networks},
      journal = {Journal of Physics D: Applied Physics},
       volume = 46,
        issue = 9,
         year = 2013,
        pages = 093001,
     abstract = {The synapse is a crucial element in biological neural networks, but a simple electronic equivalent has been absent. This complicates the development of hardware that imitates biological architectures in the nervous system. Now, the recent progress in the experimental realization of memristive devices has renewed interest in artificial neural networks. The resistance of a memristive system depends on its past states and exactly this functionality can be used to mimic the synaptic connections in a (human) brain. After a short introduction to memristors, we present and explain the relevant mechanisms in a biological neural network, such as long-term potentiation and spike time-dependent plasticity, and determine the minimal requirements for an artificial neural network. We review the implementations of these processes using basic electric circuits and more complex mechanisms that either imitate biological systems or could act as a model system for them.}
}
@inproceedings{ThomasandKarmiloff-SmithHCD-02,
        title = {Modelling Typical and Atypical Cognitive Development: Computational constraints on mechanisms of change},
       author = {Michael S. C. Thomas and Annette Karmiloff-Smith},
    booktitle = {Handbook of Childhood Development},
       editor = {Goswami, U.},
    publisher = {Blackwells Publishers},
         year = {2002},
        pages = {575-599},
     abstract = {In this chapter, we examine the use of computational models for studying development from one main perspective. This is the approach that employs connectionist models, also known as artificial neural networks. Although we relate these models to other types of computational modelling, much of the chapter is taken up with considering the range of cognitive developmental phenomena to which connectionist models have so far been applied, both in typical and atypical populations. We start with a very brief introduction to the basic concepts of connectionist modelling and then consider a single model in some detail, that of children's performance in reasoning about balance scale problems. Subsequently we look at models proposed to account for the development of other aspects of reasoning in children, development in infancy, and the acquisition of language. We then pause to examine some of the theoretical issues raised by these models. In the second half, we consider a recent extension of connectionist networks to capture behavioural deficits in developmental disorders.}
}
@article{ThomasandKarmiloff-SmithPR-03,
        title = {Modeling language acquisition in atypical phenotypes},
       author = {Michael S. C. Thomas and Annette Karmiloff-Smith},
      journal = {Psychological Review},
       volume = {110},
       number = {4},
         year = {2003},
        pages = {647-82},
     abstract = {An increasing number of connectionist models have been proposed to explain behavioral deficits in developmental disorders. These simulations motivate serious consideration of the theoretical implications of the claim that a developmental disorder fits within the parameter space of a particular computational model of normal development. The authors examine these issues in depth with respect to a series of new simulations investigating past-tense formation in Williams syndrome. This syndrome and the past-tense domain are highly relevant because both have been used to make strong theoretical claims about the processes underlying normal language acquisition. The authors conclude that computational models have great potential to advance psychologists' understanding of developmental deficits because they focus on the developmental process itself as a pivotal causal factor in producing atypical phenotypic outcomes.}
}
@techreport{ThomasetalTR-89,
      author = {Thomas, Becky and Shoham, Yoav and Schwartz, Anton},
       title = {Modalities in Agent-Oriented Programming},
       issue = {Working Paper},
 institution = {Stanford Knowledge Systems Laboratory},
        year = 1989,
}
@inproceedings{ThomasonetalIJCAI-15,
       author = {Thomason, Jesse and Zhang, Shiqi and Mooney, Raymond and Stone, Peter},
        title = {Learning to Interpret Natural Language Commands Through Human-robot Dialog},
    booktitle = {Proceedings of the 24th International Conference on Artificial Intelligence},
         year = {2015},
        pages = {1923-1929},
    publisher = {AAAI Press},
     abstract = {Intelligent robots frequently need to understand requests from naive users through natural language. Previous approaches either cannot account for language variation, e.g., keyword search, or require gathering large annotated corpora, which can be expensive and cannot adapt to new variation. We introduce a dialog agent for mobile robots that understands human instructions through semantic parsing, actively resolves ambiguities using a dialog manager, and incrementally learns from human-robot conversations by inducing training data from user paraphrases. Our dialog agent is implemented and tested both on a web interface with hundreds of users via Mechanical Turk and on a mobile robot over several days, tasked with understanding navigation and delivery requests through natural language in an office environment. In both contexts, We observe significant improvements in user satisfaction after learning from conversations.}
}
@book{Thompson2013smarter,
        title = {Smarter Than You Think: How Technology is Changing Our Minds for the Better},
       author = {Thompson, Clive},
         year = {2013},
    publisher = {HarperCollins Publishers}
}
@book{Thomson2013,
        title = {Statistical Methods for Spoken Dialogue Management},
       author = {Thomson, Blaise},
         year = {2013},
    publisher = {Springer London}
}
@book{Thomson2018unthinkable,
        title = {Unthinkable: {A}n Extraordinary Journey Through the World's Strangest Brains},
       author = {Thomson, Helen},
    publisher = {Hodder \& Stoughton},
         year = {2018},
}
@article{ThomsonandBannisterCEREBRAL-CORTEX-03,
       author = {Thomson, Alex M. and Bannister, A. Peter},
        title = {Interlaminar Connections in the Neocortex},
      journal = {Cerebral Cortex},
       volume = 13,
        issue = 1,
        pages = {5-14},
         year = 2003,
     abstract = {This review summarizes the local circuit, interlaminar connections in adult mammalian neocortex. These were first demonstrated with anatomical techniques, which indicate some of the exquisite spatial precision present in the circuitry. Details, such as the class(es) of neurons targeted by some of these projections, have begun to be added in studies that combine paired/triple intracellular recordings with dye-filling of connected neurons. Clear patterns are emerging from these studies, with 'forward' projections from layer 4 to 3 and from 3 to 5 targeting both selected pyramidal cells and interneurons, while 'back' projections from layer 5 to 3 and from 3 to 4 target only interneurons. To place these data in a wider context, the major afferent inputs to and efferent outputs from each of the layers are discussed first.},
}
@article{ThorpeCN-04,
       author = {Simon Thorpe},
        title = {Emergence of filters from natural scenes in a sparse spike coding scheme},
      journal = {Neurocomputing},
       volume = {58-60},
        pages = {821-826},
         year = 2004,
      comment = {Computational Neuroscience: Trends in Research 2004},
}
@techreport{ThorpeTR-81,
       author = {Thorpe, C.},
        title = {Sonar Image Processing - An Application of Template Matching Through Relaxation},
        issue = {CMU-RI-TR-81-6},
  institution = {Carnegie-Mellon University Robotics Institute},
         year = 1981,
}
@article{ThorpeetalNATURE-96,
        title = {Speed of processing in the human visual system},
       author = {Thorpe, S. and Fize, D. and Marlot, C. and others},
      journal = {Nature},
       volume = 381,
        issue = 6582,
         year = 1996,
        pages = {520-522},
}
@book{Thrun96,
       author = {Thrun, Sebastian},
        title = {Explanation-Based Neural Network Learning: A Lifelong Learning Approach},
    publisher = {Kluwer},
      address = {Boston, Massachusetts},
         year = 1996,
}
@article{ThrunAIJ-00,
       author = {Thrun, S. and Fox, D. and Burgard, W. and Dellaert, F.},
        title = {Robust Monte Carlo Localization for Mobile Robots},
      journal = {Artificial Intelligence},
       volume = 128,
       number = {1-2},
        pages = {99-141},
         year = 2000,
}
@techreport{ThrunTR-95,
       author = {Thrun, Sebastian},
        title = {Lifelong Learning: A Case Study},
  institution = {Carnegie Mellon University},
        issue = {CS-95-208},
         year = 1995,
}
@book{ThrunandPratt97,
       author = {Thrun, S. and Pratt, L. Y.},
        title = {Learning To Learn},
    publisher = {Kluwer},
      address = {Boston, Massachusetts},
         year = 1997,
}
@inproceedings{ThrunetalWAFR-02,
       author = {S. Thrun and D. Koller and Z. Ghahramani and H. Durrant-Whyte and A. Ng},
        title = {Simultaneous mapping and localization with sparse extended information filters},
    booktitle = {Proceedings of {WAFR} 2002},
          url = {http://citeseer.ist.psu.edu/thrun02simultaneous.html},
         year = 2002,
}
@book{Thurston,
       author = {Robert H. Thurston},
        title = {A History of the Growth of the Steam-Engine},
    publisher = {D. Appleton},
      address = {New York, NY},
         year = 1878,
}
@inproceedings{TianetalAAAI-14,
       author = {Learning Deep Representations for Graph Clustering},
        title = {Fei Tian and Bin Gao and Qing Cui and Enhong Chen and Tie-Yan Liu},
    booktitle = {Proceedings of the Twenty-Eighth {AAAI} Conference on Artificial Intelligence},
    publisher = {AAAI Press},
 organization = {AAAI},
         year = {2014},
        pages = {1293-1299},
}
@article{TianetalNATURE-METHODS-09,
       author = {Tian, Lin and Hires, S. Andrew and Mao, Tianyi and Huber, Daniel and Chiappe, M. Eugenia and Chalasani, Sreekanth H. and Petreanu, Leopoldo and Akerboom, Jasper and McKinney, Sean A. and Schreiter, Eric R. and Bargmann, Cornelia I. and Jayaraman, Vivek and Svoboda, Karel and Looger, Loren L.},
        title = {Imaging neural activity in worms, flies and mice with improved GCaMP calcium indicators},
      journal = {Nature Methods},
    publisher = {Nature Publishing Group},
       volume = {6},
        issue = {12},
         year = {2009},
        pages = {875-881},
}
@article{TibshiraniRSSB-96,
     abstract = {We propose a new method for estimation in linear models. The `lasso' minimizes the residual sum of squares subject to the sum of the absolute value of the coefficients being less than a constant. Because of the nature of this constraint it tends to produce some coefficients that are exactly 0 and hence gives interpretable models. Our simulation studies suggest that the lasso enjoys some of the favourable properties of both subset selection and ridge regression. It produces interpretable models like subset selection and exhibits the stability of ridge regression. There is also an interesting relationship with recent work in adaptive function estimation by Donoho and Johnstone. The lasso idea is quite general and can be applied in a variety of statistical models: extensions to generalized regression models and tree-based models are briefly described.},
       author = {Robert Tibshirani},
        title = {Regression Shrinkage and Selection via the Lasso},
      journal = {Journal of the Royal Statistical Society. Series B (Methodological)},
    publisher = {[Royal Statistical Society, Wiley]},
       volume = {58},
       number = {1},
         year = {1996},
        pages = {267-288},
}
@inproceedings{TielemanICML-08,
    author = {Tieleman, Tijmen},
     title = {Training Restricted Boltzmann Machines using Approximations to the Likelihood Gradient},
 booktitle = {Proceedings of the 25th International Conference on Machine learning},
  location = {Helsinki, Finland},
    editor = {Andrew McCallum and Sam Roweis},
 publisher = {Omnipress},
      year = 2008,
}
@article{TiemeieretalNI-10,
    author = {Henning Tiemeier and Rhoshel K. Lenroot and Deanna K. Greenstein and Lan Tran and Ronald Pierson and Jay N. Giedd},
     title = {Cerebellum development during childhood and adolescence: A longitudinal morphometric {MRI} study},
   journal = {Neuroimage},
    volume = 49,
     issue = 1,
     pages = {63-70},
      year = 2010,
}
@article{TierneyandNelsonZERO-THREE-09,
       author = {Tierney, Adrienne L. and Nelson, Charles A. 3rd},
        title = {Brain Development and the Role of Experience in the Early Years},
      journal = {Zero Three},
         year = {2009},
       volume = {30},
        issue = {2},
        pages = {9-13},
     abstract = {Research over the past several decades has provided insight into the processes that govern early brain development and how those processes contribute to behavior. In the following article, we provide an overview of early brain development beginning with a summary of the prenatal period. We then turn to postnatal development and examine how brain functions are built and how experience mediates this process. Specifically, we discuss findings from research on speech and on face processing. The results of this research highlight how the first few years of life are a particularly important period of development of the brain.},
}
@article{Tikhonov63,
    author = {Tikhonov, A. N.},
     title = {Solution of Incorrectly Formulated Problems and the Regularization Method},
   journal = {Soviet Math. Dokl.},
    volume = 4,
     pages = {1035-1038},
      year = 1963,
}
@article{Tikidji-HamburyanetalFiN17,
       author = {Tikidji-Hamburyan, Ruben A. and Narayana, Vikram and Bozkus, Zeki and El-Ghazawi, Tarek A.},
        title = {Software for Brain Network Simulations: A Comparative Study},
      journal = {Frontiers in Neuroinformatics},
       volume = {11},
        pages = {46},
         year = {2017},
     abstract = {Numerical simulations of brain networks are a critical part of our efforts in understanding brain functions under pathological and normal conditions. For several decades, the community has developed many software packages and simulators to accelerate research in computational neuroscience. In this article, we select the three most popular simulators, as determined by the number of models in the ModelDB database, such as NEURON, GENESIS, and BRIAN, and perform an independent evaluation of these simulators. In addition, we study NEST, one of the lead simulators of the Human Brain Project. First, we study them based on one of the most important characteristics, the range of supported models. Our investigation reveals that brain network simulators may be biased toward supporting a specific set of models. However, all simulators tend to expand the supported range of models by providing a universal environment for the computational study of individual neurons and brain networks. Next, our investigations on the characteristics of computational architecture and efficiency indicate that all simulators compile the most computationally intensive procedures into binary code, with the aim of maximizing their computational performance. However, not all simulators provide the simplest method for module development and/or guarantee efficient binary code. Third, a study of their amenability for high-performance computing reveals that NEST can almost transparently map an existing model on a cluster or multicore computer, while NEURON requires code modification if the model developed for a single computer has to be mapped on a computational cluster. Interestingly, parallelization is the weakest characteristic of BRIAN, which provides no support for cluster computations and limited support for multicore computers. Fourth, we identify the level of user support and frequency of usage for all simulators. Finally, we carry out an evaluation using two case studies: a large network with simplified neural and synaptic models and a small network with detailed models. These two case studies allow us to avoid any bias toward a particular software package. The results indicate that BRIAN provides the most concise language for both cases considered. Furthermore, as expected, NEST mostly favors large network models, while NEURON is better suited for detailed models. Overall, the case studies reinforce our general observation that simulators have a bias in the computational performance toward specific types of the brain network models.}
}
@incollection{TimpetalNANOPORES-11,
      title = {Third Generation {DNA} Sequencing with a Nanopore},
     author = {Timp, Gregory and Mirsaidov, Utkur and Timp, Winston and Shim, Jiwook and Wang, 
               Deqiang and Dimitrov, Valentin and Scrimgeour, Jan and Lin, Chunchen and Comer, 
               Jeffrey and Ho, AnthonyH. and Zou, Xueqing and Aksimentiev, Aleksei and Schulten, Klaus},
  booktitle = {Nanopores},
     editor = {Iqbal, Samir M. and Bashir, Rashid},
  publisher = {Springer US},
       year = 2011,
      pages = {287-311},
}
@article{TippingandBishopNC-99,
    author = {Michael Tipping and Christopher Bishop},
     title = {Mixtures of probabilistic principal component analyzers},
   journal = {Neural Computation},
    volume = 11,
     issue = 2,
     pages = {443-482},
      year = 1999,
}
@inproceedings{TishbyandPolaniITDA-11,
        title = {Information Theory of Decisions and Actions},
       author = {Naftali Tishby and Daniel Polani},
    booktitle = {Perception-Action Cycle: Models, Architectures, and Hardware},
       editor = {Cutsuridis, V. and Hussain, A. and Taylor, J.G.},
    publisher = {Springer New York},
         year = {2011},
     abstract = {The perception-action cycle is often defined as "the circular flow of information between an organism and its environment in the course of a sensory guided sequence of actions towards a goal" (Fuster 2001, 2006). The question we address in this paper is in what sense this "flow of information" can be described by Shannon's measures of information introduced in his mathematical theory of communication. We provide an affirmative answer to this question using an intriguing analogy between Shannon's classical model of communication and the Perception-Action-Cycle. In particular, decision and action sequences turn out to be directly analogous to codes in communication, and their complexity - the minimal number of (binary) decisions required for reaching a goal - directly bounded by information measures, as in communication. This analogy allows us to extend the standard Reinforcement Learning framework. The latter considers the future expected reward in the course of a behaviour sequence towards a goal (value-to-go). Here, we additionally incorporate a measure of information associated with this sequence: the cumulated information processing cost or bandwidth required to specify the future decision and action sequence (information-to-go). Using a graphical model, we derive a recursive Bellman optimality equation for information measures, in analogy to Reinforcement Learning; from this, we obtain new algorithms for calculating the optimal trade-off between the value-to-go and the required information-to-go, unifying the ideas behind the Bellman and the Blahut-Arimoto iterations. This trade-off between value-to-go and information-togo provides a complete analogy with the compression-distortion trade-off in source coding. The present new formulation connects seemingly unrelated optimization problems. The algorithm is demonstrated on grid world examples.}
}
@article{TishbyandZaslavskyCoRR-15,
       author = {Naftali Tishby and Noga Zaslavsky},
        title = {Deep Learning and the Information Bottleneck Principle},
      journal = {CoRR},
       volume = {arXiv:1503.02406},
         year = {2015},
     abstract = {Deep Neural Networks (DNNs) are analyzed via the theoretical framework of the information bottleneck (IB) principle. We first show that any DNN can be quantified by the mutual information between the layers and the input and output variables. Using this representation we can calculate the optimal information theoretic limits of the DNN and obtain finite sample generalization bounds. The advantage of getting closer to the theoretical limit is quantifiable both by the generalization bound and by the network's simplicity. We argue that both the optimal architecture, number of layers and features/connections at each layer, are related to the bifurcation points of the information bottleneck tradeoff, namely, relevant compression of the input layer with respect to the output layer. The hierarchical representations at the layered network naturally correspond to the structural phase transitions along the information curve. We believe that this new insight can lead to new optimality bounds and deep learning algorithms.}
}
@inproceedings{TishbyetalCCC-99,
    author = {Naftali Tishby and Fernando Pereira and William Bialek},
     title = {The information bottleneck method},
 booktitle = {Proceedings of the 37th Annual Allerton Conference on Communication, Control and Computing},
     pages = {368-377},
      year = 1999,
}
@article{TkaciketalPLoS-14,
       author = {Tka\v{c}ki, Ga\v{s}per and Marre, Olivier and Amodei, Dario and Schneidman, Elad and Bialek, William and Berry, II, Michael J.},
      journal = {PLoS Computational Biology},
    publisher = {Public Library of Science},
        title = {Searching for Collective Behavior in a Large Network of Sensory Neurons},
         year = {2014},
       volume = {10},
        pages = {1-23},
     abstract = {Sensory neurons encode information about the world into sequences of spiking and silence. Multi-electrode array recordings have enabled us to move from single units to measuring the responses of many neurons simultaneously, and thus to ask questions about how populations of neurons as a whole represent their input signals. Here we build on previous work that has shown that in the salamander retina, pairs of retinal ganglion cells are only weakly correlated, yet the population spiking activity exhibits large departures from a model where the neurons would be independent. We analyze data from more than a hundred salamander retinal ganglion cells and characterize their collective response using maximum entropy models of statistical physics. With these models in hand, we can put bounds on the amount of information encoded by the neural population, constructively demonstrate that the code has error correcting redundancy, and advance two hypotheses about the neural code: that collective states of the network could carry stimulus information, and that the distribution of neural activity patterns has very nontrivial statistical properties, possibly related to critical systems in statistical physics.},
}
@article{TkaciketalPLoS-2011,
      title = {Natural Images from the Birthplace of the Human Eye},
     author = {Tkacik, Gasper and Garrigan, Patrick and Ratliff, Charles and Milcinski, Grega and Klein, Jennifer M. and Seyfarth, Lucia H. and Sterling, Peter and Brainard, David H. and Balasubramanian, Vijay},
    journal = {{PLoS} ONE},
  publisher = {Public Library of Science},
       year = 2011,
     volume = 6,
      pages = {e20409},
}
@article {TobinetalBIORXIV-17,
       author = {Tobin, William F. and Wilson, Rachel I. and Lee, Wei-Chung Allen},
        title = {Wiring variations that enable and constrain neural computation in a sensory microcircuit},
      journal = {bioRxiv},
    publisher = {Cold Spring Harbor Labs Journals},
         year = {2017},
     abstract = {Neural network function can be shaped by varying the strength of synaptic connections. One way to achieve this is to vary connection structure. To investigate how structural variation among synaptic connections might affect neural computation, we examined primary afferent connections in the Drosophila olfactory system. We used large-scale serial section electron microscopy to reconstruct all the olfactory receptor neuron (ORN) axons that target a left-right pair of glomeruli, as well as all the projection neurons (PNs) postsynaptic to these ORNs. We found three variations in ORN{\textrightarrow}PN connectivity. First, we found a systematic co-variation in synapse number and PN dendrite size, suggesting total synaptic conductance is tuned to postsynaptic excitability. Second, we discovered that PNs receive more synapses from ipsilateral than contralateral ORNs, providing a structural basis for odor lateralization behavior. Finally, we found evidence of imprecision in ORN{\textrightarrow}PN connections and show how this can diminish network performance.},
}
@article{TobinetalCoRR-17,
       author = {Joshua Tobin and Rachel Fong and Alex Ray and Jonas Schneider and Wojciech Zaremba and Pieter Abbeel},
        title = {Domain Randomization for Transferring Deep Neural Networks from Simulation to the Real World},
      journal = {CoRR},
       volume = {arXiv:1703.06907},
         year = {2017},
     abstract = {ridging the 'reality gap' that separates simulated robotics from experiments on hardware could accelerate robotic research through improved data availability. This paper explores domain randomization, a simple technique for training models on simulated images that transfer to real images by randomizing rendering in the simulator. With enough variability in the simulator, the real world may appear to the model as just another variation. We focus on the task of object localization, which is a stepping stone to general robotic manipulation skills. We find that it is possible to train a real-world object detector that is accurate to 1.5cm and robust to distractors and partial occlusions using only data from a simulator with non-realistic random textures. To demonstrate the capabilities of our detectors, we show they can be used to perform grasping in a cluttered environment. To our knowledge, this is the first successful transfer of a deep neural network trained only on simulated RGB images (without pre-training on real images) to the real world for the purpose of robotic control.},
}
@incollection{TodorovCN-09, 
     title = {Parallels between sensory and motor information processing},
    author = {Todorov, Emanuel}, 
 booktitle = {The Cognitive Neurosciences, 4th Edition}, 
 publisher = {MIT Press}, 
   address = {Cambridge, MA},
    editor = {Gazzaniga, Michael}, 
      year = 2009, 
     pages = {613-624},
}
@inproceedings{TodorovandGhahramaniEMBC-03,
       author = {E. Todorov and Z. Ghahramani},
    booktitle = {Proceedings of the 25th Annual International Conference of the IEEE Engineering in Medicine and Biology Society},
        title = {Unsupervised learning of sensory-motor primitives},
         year = {2003},
       volume = {2},
        pages = {1750-1753},
     abstract = {The search for motor primitives has captured the attention of researchers in both biological and computational motor control. Yet a theory of how to construct such primitives from first principles is lacking. Here we propose to do that by building a compact forward model of the sensory-motor periphery via unsupervised learning. We also propose a method for probabilistic inversion of the forward model, which yields low-level feedback loops that can simplify control. The idea is applied to simulated biomechanical systems of varying levels of detail.},
}
@article{TodorovandJordonNATURE-NEUROSCIENCE-02,
       author = {Todorov, Emanuel and Jordan, Michael I.},
        title = {Optimal feedback control as a theory of motor coordination},
      journal = {Nature Neuroscience},
         year = {2002},
       volume = {5},
        issue = {11},
        pages = {1226-1235},
     abstract = {A central problem in motor control is understanding how the many biomechanical degrees of freedom are coordinated to achieve a common goal. An especially puzzling aspect of coordination is that behavioral goals are achieved reliably and repeatedly with movements rarely reproducible in their detail. Existing theoretical frameworks emphasize either goal achievement or the richness of motor variability, but fail to reconcile the two. Here we propose an alternative theory based on stochastic optimal feedback control. We show that the optimal strategy in the face of uncertainty is to allow variability in redundant (task-irrelevant) dimensions. This strategy does not enforce a desired trajectory, but uses feedback more intelligently, correcting only those deviations that interfere with task goals. From this framework, task-constrained variability, goal-directed corrections, motor synergies, controlled parameters, simplifying rules and discrete coordination modes emerge naturally. We present experimental results from a range of motor tasks to support this theory.},
}
@article{TodorovetalPLoS-ONE-19,
       author = {Todorov, Dmitrii I. and Capps, Robert A. and Barnett, William H. and Latash, Elizaveta M. and Kim, Taegyo and Hamade, Khaldoun C. and Markin, Sergey N. and Rybak, Ilya A. and Molkov, Yaroslav I.},
      journal = {PLOS ONE},
    publisher = {Public Library of Science},
        title = {The interplay between cerebellum and basal ganglia in motor adaptation: A modeling study},
       volume = {14},
       number = {4},
         year = {2019},
        pages = {1-36},
     abstract = {Motor adaptation to perturbations is provided by learning mechanisms operating in the cerebellum and basal ganglia. The cerebellum normally performs motor adaptation through supervised learning using information about movement error provided by visual feedback. However, if visual feedback is critically distorted, the system may disengage cerebellar error-based learning and switch to reinforcement learning mechanisms mediated by basal ganglia. Yet, the exact conditions and mechanisms of cerebellum and basal ganglia involvement in motor adaptation remain unknown. We use mathematical modeling to simulate control of planar reaching movements that relies on both error-based and non-error-based learning mechanisms. We show that for learning to be efficient only one of these mechanisms should be active at a time. We suggest that switching between the mechanisms is provided by a special circuit that effectively suppresses the learning process in one structure and enables it in the other. To do so, this circuit modulates learning rate in the cerebellum and dopamine release in basal ganglia depending on error-based learning efficiency. We use the model to explain and interpret experimental data on error- and non-error-based motor adaptation under different conditions.},
}
@article{TolmanPSYCHOLOGICAL-REVIEW-48,
       author = {TOLMAN, Edward C.},
        title = {Cognitive maps in rats and men},
      journal = {Psychological Review},
       volume = {55},
       number = {4},
         year = {1948},
        pages = {189-208},
}
@techreport{TomasiandKanade91,
    author = {Carlo Tomasi and Takeo Kanade},
     title = {Detection and Tracking of Point Features},
institution = {Carnegie Mellon University},
   comment = {International Journal of Computer Vision},
      Year = 1991
}
@article{TommerdahletalCC-93,
       author = {Tommerdahl, M. and Favorov, O. and Whitsel, B. L. and Nakhle, B. and Gonchar, Y. A.},
        title = {Minicolumnar Activation Patterns in Cat and Monkey {SI} Cortex},
      journal = {Cerebral Cortex},
       volume = 3,
        issue = 5,
         year = 1993,
        pages = {399-411},
     abstract = {The distribution of stimulus-evoked 14C-2-deoxyglucose (2DG) labeling in primary somatosensory cortex (SI) of monkey (Macaca fasciculans) and cat was investigated. Reconstructions of the global pattern of labeling reveal that discrete skin stimuli evoke activity within an extensive region of SI, and that the activation pattern typically consists of multiple, elongated regions of above-background labeling ("modules," typically 0.5-1.0 mm wide, and 1-4 mm long). Evidence obtained using recently developed methods (Tommerdahl, 1989) for quantitative analysis of 2DG activity patterns is shown to be consistent with the idea (Whitsel et al., 1991) that SI modules typically are bounded by zones dominated by stimulus-evoked inhibition. The labeling pattern within individual 2DG modules in SI of both cats and monkeys is analyzed quantitatively (in the frequency domain). Within-module spatial activation patterns are demonstrated to be periodic, consisting of radially on ent.d profiles of above-background labeling separated from each other by less strongly labeled radial profiles. The spectral characteristics of within-module 2DG la beling change systematically with location along the module's long axis: spatial frequencies htween 18 and 35 cycles/mm are prominent in the labeling that occupies both the middle and upper layers at central locations in the module, but are a less obvious component of the labeling in both the middle and upper layers at locations remote to the module center. Since the radially oriented periodic variation both (1) in 2DG labeling in regions of SI outside modules and (2) in optical density in images of Nissl-stained sections of SI consists pre dominantly of spatial frequencies in the range of 18-35 cycles/mm, it is concluded that the radial profiles of labeling within individual 2DG modules correspond to groupings of minicolumns distinguishable from their neighbors on the basis of labeling intensity. The findings raise the possibility that highly structured, within-module spatial patterns of SI minicolumnar activation encode information about the physical properties of tactile stimuli.},
}
@book{TompkinsandWilson69,
       author = {Tompkins, C. B. and Wilson, W. L.},
        title = {Elementary Numerical Analysis},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, New Jersey},
         year = 1969,
}
@book{Tong1990,
       author = {Tong, Howell},
        title = {Nonlinear Time Series: A Dynamical System Approach},
    publisher = {Oxford University Press},
      address = {New York},
         year = 1990
}
@article{ToniandAlejandroCSHL-2015,
       author = {Toni, Nicolas and Schinder, Alejandro F.},
        title = {Maturation and Functional Integration of New Granule Cells into the Adult Hippocampus},
      journal = {Cold Spring Harbor perspectives in biology},
         year = {2015},
    publisher = {Cold Spring Harbor Laboratory Press},
       volume = {8},
        issue = {1},
        pages = {a018903-a018903},
     abstract = {The adult hippocampus generates functional dentate granule cells (GCs) that release glutamate onto target cells in the hilus and cornus ammonis (CA)3 region, and receive glutamatergic and gama-aminobutyric acid (GABA)ergic inputs that tightly control their spiking activity. The slow and sequential development of their excitatory and inhibitory inputs makes them particularly relevant for information processing. Although they are still immature, new neurons are recruited by afferent activity and display increased excitability, enhanced activity-dependent plasticity of their input and output connections, and a high rate of synaptogenesis. Once fully mature, new GCs show all the hallmarks of neurons generated during development. In this review, we focus on how developing neurons remodel the adult dentate gyrus and discuss key aspects that illustrate the potential of neurogenesis as a mechanism for circuit plasticity and function.},
}
@article{TonnesenetalCELL-18,
       author = {T{\o}nnesen, Jan and Inavalli, V. V. G. Krishna and N\"{a}erl, U. Valentin},
        title = {Super-Resolution Imaging of the Extracellular Space in Living Brain Tissue},
      journal = {Cell},
         year = {2018},
    publisher = {Elsevier},
       volume = {172},
        issue = {5},
        pages = {1108-1121},
     abstract = {The extracellular space (ECS) of the brain has an extremely complex spatial organization, which has defied conventional light microscopy. Consequently, despite a marked interest in the physiological roles of brain ECS, its structure and dynamics remain largely inaccessible for experimenters. We combined 3D-STED microscopy and fluorescent labeling of the extracellular fluid to develop super-resolution shadow imaging (SUSHI) of brain ECS in living organotypic brain slices. SUSHI enables quantitative analysis of ECS structure and reveals dynamics on multiple scales in response to a variety of physiological stimuli. Because SUSHI produces sharp negative images of all cellular structures, it also enables unbiased imaging of unlabeled brain cells with respect to their anatomical context. Moreover, the extracellular labeling strategy greatly alleviates problems of photobleaching and phototoxicity associated with traditional imaging approaches. As a straightforward variant of STED microscopy, SUSHI provides unprecedented access to the structure and dynamics of live brain ECS and neuropil.},
}
@article{TononiARCHIVES-ITALIENNES-DE_BIOLOGIE-12a,
       author = {Tononi, G.},
        title = {{I}ntegrated information theory of consciousness: {A}n updated account},
      journal = {Archives Italiennes de Biologie},
         year = {2012},
       volume = {150},
       number = {2-3},
        pages = {56-90},
     abstract = {This article presents an updated account of integrated information theory of consciousness (IIT) and some of its implications. IIT stems from thought experiments that lead to phenomenological axioms and ontological postulates. The information axiom asserts that every experience is one out of many, i.e. specific - it is what it is by differing in its particular way from a large repertoire of alternatives. The integration axiom asserts that each experience is one, i.e. unified - it cannot be reduced to independent components. The exclusion axiom asserts that every experience is definite - it is limited to particular things and not others and flows at a particular speed and resolution. IIT formalizes these intuitions with three postulates. The information postulate states that only "differences that make a difference" from the intrinsic perspective of a system matter: a mechanism generates cause-effect information if its present state has specific past causes and specific future effects within a system. The integration postulate states that only information that is irreducible matters: mechanisms generate integrated information only to the extent that the information they generate cannot be partitioned into that generated within independent components. The exclusion postulate states that only maxima of integrated information matter: a mechanism specifies only one maximally irreducible set of past causes and future effects - a concept. A complex is a set of elements specifying a maximally irreducible constellation of concepts, where the maximum is evaluated at the optimal spatio-temporal scale. Its concepts specify a maximally integrated conceptual information structure or quale, which is identical with an experience. Finally, changes in information integration upon exposure to the environment reflect a system's ability to match the causal structure of the world. After introducing an updated definition of information integration and related quantities, the article presents some theoretical considerations about the relationship between information and causation and about the relational structure of concepts within a quale. It also explores the relationship between the temporal grain size of information integration and the dynamic of metastable states in the corticothalamic complex. Finally, it summarizes how IIT accounts for empirical findings about the neural substrate of consciousness, and how various aspects of phenomenology may in principle be addressed in terms of the geometry of information integration.}
}
@article{TononiARCHIVES-ITALIENNES-DE_BIOLOGIE-12b,
       author = {Tononi, G.},
        title = {{I}ntegrated information theory of consciousness: {A}n updated account},
      journal = {Archives Italiennes de Biologie},
         year = {2012},
       volume = {150},
       number = {4},
        pages = {293-329},
     abstract = {This article presents an updated account of integrated information theory of consciousness (liT) and some of its implications. /IT stems from thought experiments that lead to phenomenological axioms (existence, compositionality, information, integration, exclusion) and corresponding ontological postulates. The information axiom asserts that every experience is spec~fic - it is what it is by differing in its particular way from a large repertoire of alternatives. The integration axiom asserts that each experience is unified- it cannot be reduced to independent components. The exclusion axiom asserts that every experience is definite - it is limited to particular things and not others and flows at a particular speed and resolution. /IT formalizes these intuitions with postulates. The information postulate states that only "differences that make a difference" from the intrinsic perpective of a system matter: a mechanism generates cause-effect information if its present state has selective past causes and selective future effects within a system. The integration postulate states that only information that is irreducible matters: mechanisms generate integrated information only to the extent that the information they generate cannot be partitioned into that generated within independent components. The exclusion postulate states that only maxima of integrated information matter: a mechanism specifies only one maximally irreducible set of past causes and future effects - a concept. A complex is a set of elements specifying a maximally irreducible constellation of concepts, where the maximum is evaluated over elements and at the optimal spatiatemporal scale. Its concepts specify a maximally integrated conceptual information structure or quale, which is identical with an experience. Finally, changes in information integration upon exposure to the environment reflect a system's ability to match the causal structure of the world. After introducing an updated definition of information integration and related quantities, the article presents some theoretical considerations about the relationship between information and causation and about the relational structure of concepts within a qua/e. It also explores the relationship between the temporal grain size of information integration and the dynamic of metastable states in the corticothalamic complex. Finally, it summarizes how liT accounts for empirical findings about the neural substrate of consciousness, and how various aspects of phenomenology may in principle be addressed in terms of the geometry of information integration.}
}
@article{TononiBIOLOGICAL-BULLETIN-08,
       author = {Tononi, G.},
        title = {{C}onsciousness as integrated information: {A} provisional manifesto},
      journal = {Biological Bulletin},
         year = {2008},
       volume = {215},
       number = {3},
        pages = {216-242},
     abstract = {The integrated information theory (IIT) starts from phenomenology and makes use of thought experiments to claim that consciousness is integrated information. Specifically: (i) the quantity of consciousness corresponds to the amount of integrated information generated by a complex of elements; (ii) the quality of experience is specified by the set of informational relationships generated within that complex. Integrated information (Phi) is defined as the amount of information generated by a complex of elements, above and beyond the information generated by its parts. Qualia space (Q) is a space where each axis represents a possible state of the complex, each point is a probability distribution of its states, and arrows between points represent the informational relationships among its elements generated by causal mechanisms (connections). Together, the set of informational relationships within a complex constitute a shape in Q that completely and univocally specifies a particular experience. Several observations concerning the neural substrate of consciousness fall naturally into place within the IIT framework. Among them are the association of consciousness with certain neural systems rather than with others; the fact that neural processes underlying consciousness can influence or be influenced by neural processes that remain unconscious; the reduction of consciousness during dreamless sleep and generalized seizures; and the distinct role of different cortical architectures in affecting the quality of experience. Equating consciousness with integrated information carries several implications for our view of nature.}
}
@incollection{ToobyandCosmides1992,
        title = {The Psychological Foundations of Culture},
       author = {J. Tooby and L. Cosmides},
    booktitle = {The Adapted Mind: Evolutionary psychology and the generation of culture},
       editor = {J. Barkow and L. Cosmides and J. Tooby},
    publisher = {Oxford University Press},
      address = {New York},
         year = 1992,
}
@inproceedings{TookeretalEMBS-14,
        title = {Towards a large-scale recording system: demonstration of polymer-based penetrating array for chronic neural recording},
       author = {Tooker A. and Liu D. and Anderson E.B. and Felix S. and Shah K.G. and Lee K.Y. and Chung J.E. and Pannu S. and Frank L. and Tolosa V.},
    booktitle = {Conference Proceedings IEEE Engineering in Medical Biology Society},
         year = {2014},
        pages = {6830-6833},
     abstract = {The brain is a massively interconnected network of specialized circuits. Even primary sensory areas, once thought to support relatively simple, feed-forward processing, are now known to be parts of complex feedback circuits. All brain functions depend on millisecond timescale interactions across these brain networks. Current approaches cannot measure or manipulate such large-scale interactions. Here we demonstrate that polymer-based, penetrating, micro-electrode arrays can provide high quality neural recordings from awake, behaving animals over periods of months. Our results indicate that polymer electrodes are a viable substrate for the development of systems that can record  from thousands of channels across months to years. This is our first step towards developing a 1000-plus electrode system capable of providing high-quality, long-term neural recordings.},
}
@article{TopalidouetalENEURO-18,
       author = {Topalidou, Meropi and Kase, Daisuke and Boraud, Thomas and Rougier, Nicolas P.},
        title = {A Computational Model of Dual Competition between the Basal Ganglia and the Cortex},
      journal = {eNeuro},
       volume = {5},
       number = {6},
         year = {2018},
    publisher = {Society for Neuroscience},
     abstract = {We propose a model that includes interactions between the cortex, the basal ganglia (BG), and the thalamus based on a dual competition. We hypothesize that the striatum, the subthalamic nucleus (STN), the internal globus pallidus (GPi), the thalamus, and the cortex are involved in closed feedback loops through the hyperdirect and direct pathways. These loops support a competition process that results in the ability of BG to make a cognitive decision followed by a motor one. Considering lateral cortical interactions, another competition takes place inside the cortex allowing the latter to make a cognitive and a motor decision. We show how this dual competition endows the model with two regimes. One is driven by reinforcement learning and the other by Hebbian learning. The final decision is made according to a combination of these two mechanisms with a gradual transfer from the former to the latter. We confirmed these theoretical results on primates (Macaca mulatta) using a novel paradigm predicted by the model.},
}
@article{TorralbaIJCV-03,
       author = {Antonio Torralba},
        title = {Contextual priming for object detection},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 53,
        issue = 2,
        pages = {169-191},
         year = 2003,
}
@inproceedings{TorralbaetalCVPR-04,
       author = {Antonio Torralba and Keving Murphy and William Freeman},
        title = {Sharing features: efficient boosting procedures for multiclass object detection},
    booktitle = {Proceedings of IEEE Computer Vision and Pattern Recognition},
     location = {Washington, DC},
         year = 2004,
}
@inproceedings{TorralbaetalCVPR-08,
       author = {Antonio Torralba and Rob Fergus and Yair Weiss},
        title = {Small Codes and Large Image Databases for Recognition},
    booktitle = {Proceedings of {IEEE} Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
        pages = {1-8},
         year = 2008,
}
@article{TorralbaetalJoV-07,
       author = {Torralba, Antonio and Fergus, Rob and Freeman, William},
        title = {Object and scene recognition in tiny images},
      journal = {Journal of Vision},
       volume = 7,
        issue = 9,
        pages = {193-193},
         year = 2007,
}
@incollection{TorralbaetalNIPS-05,
       author = {Antonio Torralba and Kevin P. Murphy and William Freeman},
        title = {Contextual Models for Object Detection Using Boosted Random Fields},
    booktitle = {Advances in Neural Information Processing Systems 17},
       editor = {Lawrence K. Saul and Yair Weiss and {L\'{e}on} Bottou},
    publisher = {MIT Press},
      address = {Cambridge, MA},
        pages = {1401-1408},
         year = 2005
}
@article{TorralbaetalPR-06,
       author = {Torralba, A. and Oliva, A. and Castelhano, M.S. and Henderson, J.M.},
        title = {Contextual guidance of attention in natural scenes},
      journal = {Psychological Review},
       volume = 113,
        pages = {766-786},
         year = 2006,
}
@book{Touretsky84,
        title = {{LISP}, A Gentle Introduction to Symbolic Computation},
       author = {Touretsky, David},
    publisher = {Harper and Row},
      address = {New York},
         year = 1984,
}
@article{TouretzkyandHintonDCPS-88,
       author = {David S. Touretzky and Geoffrey E. Hinton},
        title = {A distributed connectionist production system},
      journal = {Cognitive Science},
         year = 1988,
       volume = 12,
        pages = {423-466},
     abstract = {DCPS is a connectionist production system interpreter that uses distributed repre-sentations. As a connectionist model it consists of many simple, richly intercon-nected neuron-like computing units that cooperate to solve problems in parallel. One motivation far constructing DCPS was to demonstrate that connectionist models ore copable of representing and using explicit rules. A second motivation was to show how "coarse coding" or "distributed representations" can be used to construct a working memory that requires far fewer units than the number of different facts that can potentially be stored. The simulation we present is intended as a detailed demonstration of the feasibility of certain ideas and should not be viewed as a full implementation of production systems. Our current model only has o few of the many interesting emergent properties that we eventually hope to demonstrate: It is damage-resistant, it performs matching and variable bind-ing by massively parallel constraint satisfaction, and the capacity of its working memory is dependent on the similarity of the items being stored.}
}
@inproceedings{TouretzkyandHintonIJCAI-85,
       author = {Touretzky, David S. and Hinton, Geoffrey E.},
        title = {Symbols among the neurons: details of a connectionist inference architecture},
    booktitle = {Proceedings of the 9th International Joint Conference on Artificial Intelligence},
     location = {Los Angeles, California},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, CA, USA},
         year = 1985,
        pages = {238-243},
     abstract = {Pattern matching and variable binding are easily implemented in conventional computer architectures, but not necessarily in all architectures. In a distributed neural network architecture each symbol is represented by activity in many units and each unit contributes to the representation of many symbols. Manipulating symbols using this type of distributed representation is not as easy as with a local representation where each unit denotes one symbol, but there is evidence that the distributed approach is the one chosen by nature. We describe a working implementation of a production system interpreter in a neural network using distributed representations for both symbols and rules. The research provides a detailed account of two important symbolic reasoning operations, pattern matching and variable binding, as emergent properties of collections of neuron-like elements. The success of our production system implementation goes some way towards answering a common criticism of connectionist theories: that they aren't powerful enough to do symbolic reasoning.},
}
@article{TrachtenbergNATURE-02,
       author = {Trachtenberg, Joshua T. and Chen, Brian E. and Knott, Graham W. and Feng, Guoping and Sanes, Joshua R. and Welker, Egbert and Svoboda, Karel},
        title = {Long-term in vivo imaging of experience-dependent synaptic plasticity in adult cortex},
      journal = {Nature},
         year = {2002},
       volume = {420},
        issue = {6917},
        pages = {788-794},
     abstract = {Do new synapses form in the adult cortex to support experience-dependent plasticity? To address this question, we repeatedly imaged individual pyramidal neurons in the mouse barrel cortex over periods of weeks. We found that, although dendritic structure is stable, some spines appear and disappear. Spine lifetimes vary greatly: stable spines, about 50\% of the population, persist for at least a month, whereas the remainder are present for a few days or less. Serial-section electron microscopy of imaged dendritic segments revealed retrospectively that spine sprouting and retraction are associated with synapse formation and elimination. Experience-dependent plasticity of cortical receptive fields was accompanied by increased synapse turnover. Our measurements suggest that sensory experience drives the formation and elimination of synapses and that these changes might underlie adaptive remodelling of neural circuits.}
}
@inproceedings{TranetalACL-17,
       author = {Tran, Quan Hung and Zukerman, Ingrid and Haffari, Gholamreza},
        title = {A Hierarchical Neural Model for Learning Sequences of Dialogue Acts},
    booktitle = {Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers},
         year = {2017},
    publisher = {Association for Computational Linguistics},
        pages = {428-437},
     location = {Valencia, Spain},
     abstract = {We propose a novel hierarchical Recurrent Neural Network (RNN) for learning se- quences of Dialogue Acts (DAs). The in- put in this task is a sequence of utterances (i.e., conversational contributions) com- prising a sequence of tokens, and the out- put is a sequence of DA labels (one label per utterance). Our model leverages the hi- erarchical nature of dialogue data by using two nested RNNs that capture long-range dependencies at the dialogue level and the utterance level. This model is combined with an attention mechanism that focuses on salient tokens in utterances. Our exper- imental results show that our model out- performs strong baselines on two popular datasets, Switchboard and MapTask; and our detailed empirical analysis highlights the impact of each aspect of our model.}
}
@article{TraversandMilgram69,
        title = {{An Experimental Study of the Small World Problem}},
       author = {Travers, Jeffrey and Milgram, Stanley},
      journal = {Sociometry},
       volume = 32,
        issue = 4,
        pages = {425-443},
    publisher = {American Sociological Association},
         year = 1969,
     abstract = {Arbitrarily selected individuals (N = 296) in Nebraska and Boston are asked to generate acquaintance chains to a target person in Massachusetts, employing "the small world method"  (Milgram, 1967). Sixty-four chains reach the target person. Within this group the mean number of intermediaries between starters and targets is 5.2. Boston starting chains reach the target person with fewer intermediaries than those starting in Nebraska; subpopulations in the Nebraska group do not differ among themselves. The funneling of chains through sociometric "stars" is noted, with 48 per cent of the chains passing through three persons before reaching the target. Applications of the method to studies of large scale social structure are discussed.}, 
}
@article{TravisetalHBM-15,
       author = {Travis and Katherine E. and Leitner, Yael and Feldman, Heidi M. and Ben-Shachar, Michal},
        title = {Cerebellar white matter pathways are associated with reading skills in children and adolescents},
      journal = {Human Brain Mapping},
       volume = {36},
       number = {4},
         year = {2015},
        pages = {1536-1553},
     abstract = {Abstract Reading is a critical life skill in the modern world. The neural basis of reading incorporates a distributed network of cortical areas and their white matter connections. The cerebellum has also been implicated in reading and reading disabilities. However, little is known about the contribution of cerebellar white matter pathways to major component skills of reading. We used diffusion magnetic resonance imaging (dMRI) with tractography to identify the cerebellar peduncles in a group of 9- to 17-year-old children and adolescents born full term (FT, n = 19) or preterm (PT, n = 26). In this cohort, no significant differences were found between fractional anisotropy (FA) measures of the peduncles in the PT and FT groups. FA of the cerebellar peduncles correlated significantly with measures of decoding and reading comprehension in the combined sample of FT and PT subjects. Correlations were negative in the superior and inferior cerebellar peduncles and positive in the middle cerebellar peduncle. Additional analyses revealed that FT and PT groups demonstrated similar patterns of reading associations within the left superior cerebellar peduncle, middle cerebellar peduncle, and left inferior cerebellar peduncle. Partial correlation analyses showed that distinct sub-skills of reading were associated with FA in segments of different cerebellar peduncles. Overall, the present findings are the first to document associations of microstructure of the cerebellar peduncles and the component skills of reading.},
}
@article{TreismanBMB-64,
       author = {Treisman, Anne. M.},
        title = {Selective attention in man},
      journal = {British Medical Bulletin},
       volume = {20},
       number = {1},
         year = {1964},
        pages = {12-16},
}
@article{TreismanNEURON-99,
       author = {Treisman, Anne. M.},
        title = {Solutions to the Binding Problem: Review Progress through Controversy Summary and Convergence},
      journal = {Neuron},
    publisher = {Cell Press},
       volume = {24},
         year = {1999},
        pages = {105-110},
     abstract = {The binding problem, or constellation of problems, concerns our capacity to integrate information across time, space, attributes, and ideas. The goal of research in this area is to understand how we can respond to relations within relevant subsets of the world but not to relations between arbitrarily selected parts or properties. Language comprehension and thinking critically depend on correct binding of syntactic and semantic structures. Binding is required when we select an action to perform in a particular context. We must, for example, reach in the right direction, lift the glass with the correct muscle tension, and drink the water it contains rather than eat or inhale it. The mediating "event file" binds stimulus to response (Hommel, 1998; Shadlen and Movshon, 1999 [this issue of Neuron]). The most extensive discussion so far has focused on the problem of binding in visual perception. How does the brain segregate the correct sensory data to represent the objects that are actually present and not some illusory recombinations of their features? In considering perceptual binding, it is important to note that "seeing" an object is not the same as identifying it (Kahneman et al., 1992; Treisman, 1992a; Treisman and Kanwisher, 1998). To generate a perceptual experience and to be able to act on it, we need to specify the current details of how an object looks, where it is, how it is oriented, and many other often arbitrary details of its current instantiation. Thus we must construct a temporary token ("object file") that binds together these current features with the more permanent identifying characteristics of its type.}
}
@article{TreismanetalQJEP-94,
       author = {Treisman, Michel and Cook, Norman and Naish, Peter L. N. and MacCrone, Janice K.},
        title = {The internal clock: Electroencephalographic evidence for oscillatory processes underlying time perception},
      journal = {The Quarterly Journal of Experimental Psychology Section A},
       volume = 47,
        issue = 2,
        pages = {241-289},
         year = 1994,
     abstract = {It has been proposed that temporal perception and performance depend on a biological source of temporal information. A model for a temporal oscillator put forward by Treisman, Faulkner, Naish, and Brogan (1990) predicted that if intense sensory pulses (such as auditory clicks) were presented to subjects at suitable rates they would perturb the frequency at which the temporal oscillator runs and so cause over- or underestimation of time. The resulting pattern of interference between sensory pulse rates and time judgments would depend on the frequency of the temporal oscillator and so might allow that frequency to be estimated. Such interference patterns were found using auditory clicks and visual flicker (Treisman \& Brogan, 1992; Treisman et al., 1990). The present study examines time estimation together with the simultaneously recorded electroencephalogram to examine whether evidence of such an interference pattern can be found in the EEG. Alternative models for the organization of a temporal system consisting of an oscillator or multiple oscillators are considered and predictions derived from them relating to the EEG. An experiment was run in which time intervals were presented for estimation, auditory clicks being given during those intervals, and the EEG was recorded concurrently. Analyses of the EEG revealed interactions between auditory click rates and certain EEG components which parallel the interference patterns previously found. The overall pattern of EEG results is interpreted as favouring a model for the organization of the temporal system in which sets of click-sensitive oscillators spaced at intervals of about 12.8 Hz contribute to the EEG spectrum. These are taken to represent a series of harmonically spaced distributions of oscillators involved in time-keeping.}
}
@article{TremblayetalFiN-15,
       author = {Tremblay, Marie-Eve and Lecours, Cynthia and Samson, Louis and S\'{a}nchez-Zafra, Víctor and Sierra, Amanda},
        title = {From the Cajal alumni Achúcarro and Río-Hortega to the rediscovery of never-resting microglia},
      journal = {Frontiers in Neuroanatomy},
       volume = {9},
         year = {2015},
       number = {45},
     abstract = {Under the guidance of Ram\'{o}n y Cajal, a plethora of students flourished and began to apply his silver impregnation methods to study brain cells other than neurons: the neuroglia. In the first decades of the twentieth century, Nicol\'{a}s Achúcarro was one of the first researchers to visualize the brain cells with phagocytic capacity that we know today as microglia. Later, his pupil Pío del Río-Hortega developed modifications of Achúcarro's methods and was able to specifically observe the fine morphological intricacies of microglia. These findings contradicted Cajal's own views on cells that he thought belonged to the same class as oligodendroglia (the so called "third element" of the nervous system), leading to a long-standing discussion. It was only in 1924 that Río-Hortega's observations prevailed worldwide, thus recognizing microglia as a unique cell type. This late landing in the Neuroscience arena still has repercussions in the twenty first century, as microglia remain one of the least understood cell populations of the healthy brain. For decades, microglia in normal, physiological conditions in the adult brain were considered to be merely "resting," and their contribution as "activated" cells to the neuroinflammatory response in pathological conditions mostly detrimental. It was not until microglia were imaged in real time in the intact brain using two-photon in vivo imaging that the extreme motility of their fine processes was revealed. These findings led to a conceptual revolution in the field: "resting" microglia are constantly surveying the brain parenchyma in normal physiological conditions. Today, following Cajal's school of thought, structural and functional investigations of microglial morphology, dynamics, and relationships with neurons and other glial cells are experiencing a renaissance and we stand at the brink of discovering new roles for these unique immune cells in the healthy brain, an essential step to understand their causal relationship to diseases.}
}
@article{TremblayetalJoVE-10,
       author = {Tremblay, Marie-Eve and Riad, Mustapha and Majewska, Ania},
        title = {Preparation of mouse brain tissue for immunoelectron microscopy},
      journal = {Journal of Visualualized Experiments},
         year = {2010},
        issue = {41},
        pages = {2021},
     abstract = {Transmission electron microscopy (TEM) is extremely useful for visualizing microglial, oligodendrocytic, astrocytic, and neuronal subcellular compartments (dendrite, dendritic spine, axon, axon terminal, perikaryon), as well as their intracellular organelles and cytoskeleton, in the central nervous system at high spatial resolution. Combined with TEM, pre-embedding immunocytochemistry allows the discrimination of cellular elements with few distinctive features and identification criteria (e.g., microglial perikarya and processes, when using an antibody against the microglia-specific marker Iba1 (ionized calcium binding adaptor molecule 1; as presented here)), identifying the neurotransmitter contents of cellular elements (e.g., serotonergic) and their ultrastructural localization of soluble or membrane-bound proteins (e.g., 5 HT1A and EphA4 receptors). Here, we describe a protocol for transcardiac perfusion of mice with acrolein fixative, removal and sectioning of the brain, as well as immunoperoxidase-diaminobenzidine (DAB) staining, resin embedding, and ultrathin sectioning of the brain sections. Upon completion of these procedures, the immunostained material is ready for examination with TEM. When rigorously performed, this technique provides an excellent compromise between optimal ultrastructural preservation and immunocytochemical detection.},
}
@article{TremblayetalNEURON-15,
       author = {Tremblay, S\'{e}bastien and Pieper, Florian Sachs, Adam and Martinez-Trujillo, Julio},
        title = {Attentional Filtering of Visual Information by Neuronal Ensembles in the Primate Lateral Prefrontal Cortex},
      journal = {Neuron},
         year = 2015,
    publisher = {Elsevier},
       volume = 85,
        issue = 1,
        pages = {202-215},
     abstract = {The activity of neurons in the primate lateral prefrontal cortex (LPFC) is strongly modulated by visual attention. Such a modulation has mostly been documented by averaging the activity of independently recorded neurons over repeated experimental trials. However, in realistic settings, ensembles of simultaneously active LPFC neurons must generate attentional signals on a single-trial basis, despite the individual and correlated variability of neuronal responses. Whether, under these circumstances, the LPFC can reliably generate attentional signals is unclear. Here, we show that the simultaneous activity of neuronal ensembles in the primate LPFC can be reliably decoded to predict the allocation of attention on a single-trial basis. Decoding was sensitive to the noise correlation structure of the ensembles. Additionally, it was resilient to distractors, predictive of behavior, and stable over weeks. Thus, LPFC neuronal ensemble activity can reliably encode attention within behavioral timeframes, despite the noisy and correlated nature of neuronal activity.},
}
@article{TrevesandRollHIPPOCAMPUS-92,
        title = {Computational Constraints Suggest the Need for Two Distinct Input Systems to the Hippocampal CA3 Network},
       author = {Alessandro Treves and Edmund T. Rolls},
      journal = {Hippocampus},
       volume = {2},
       number = {2},
         year = {1992},
        pages = {189-200},
     abstract = {The CA3 network in the hippocampus may operate as an autoassociator, in which declarative memories, known to be dependent on hippocampal processing, could be stored, and subsequently retrieved, using modifiable synaptic efficacies in the CA3 recurrent collateral system. On the basis of this hypothesis, the authors explore the computational relevance of the extrinsic afferents to the CA3 network. A quantitative statistical analysis of the information that may be relayed by such afferent connections reveals the need for two distinct systems of input synapses. The synapses of the first system need to be strong(but not associatively modifiable) in order to force, during learning, the CA3 cells into a pattern of activity relatively independent of any inputs being received from the recurrent collaterals, and which thus reflects sizable amounts of new information. It is proposed that the mossy fiber system performs this function. A second system, with a large number of associatively modifiable synapses on each receiving cell, is needed in order to relay a signal specific enough to initiate the retrieval process. This may be identified, we propose, with the perforant path input to CA3.},
}
@incollection{TrimbleandCavannaHBN-08,
       author = {Michael R. Trimble and Andrea E. Cavanna},
       editor = {Dere, E. and Easton, A. and Nadel, L. and Huston, J.P.},
        title = {Chapter 3.7 The role of the precuneus in episodic memory},
    booktitle = {Handbook of Behavioral Neuroscience},
    publisher = {Elsevier Science},
         year = {2008},
        pages = {378-392}
}
@incollection{TripathyetalDCIT-09,
        title = {Web Users' Personality Traits Analysis},
       author = {Tripathy, AmiyaK. and Gomes, Colton and Jain, Ashok},
    booktitle = {Distributed Computing and Internet Technology},
       editor = {Parashar, Manish and Aggarwal, Sanjeev K.},
       volume = 5375,
       series = {Lecture Notes in Computer Science},
    publisher = {Springer Berlin Heidelberg},
         year = 2009,
        pages = {57-70},
     abstract = {Determining web users' personality traits plays a vital role in understanding user navigation patterns \& behaviours on the web. This information can then be used by various online firms to provide various personalized services to online customers. There are already numerous approaches towards analyzing user behavioural patterns. The client side tracking, server-side tracking, user remote tracker etc. are some of these. However all of them have certain inherent drawbacks which make them inappropriate for multi-user environments. We propose a new integrated approach and thereby develop a unique conceptual model which would prove to be extremely advantageous even in multi-user environments. We use user clicks to analyze the user behaviour along with the content on the various websites that the user visits. We also use a personality questionnaire which would help to further enhance our results and give a better understanding of the user's personality traits.}
}
@article{TriplettetalCELL-09,
       author = {Jason W. Triplett and Melinda T. Owens and Jena Yamada and Greg Lemke and Jianhua Cang and Michael P. Stryker and David A. Feldheim},
        title = {Retinal input instructs alignment of visual topographic maps},
      journal = {Cell},
       volume = 139,
        issue = 1,
         year = 2009,
        pages = {175-185},
}
@book{Trivers2011,
       author = {Robert Trivers},
        title = {The Folly of Fools: The Logic of Deceit and Self-Deception in Human Life},
         year = 2011,
    publisher = {Basic Books},
      address = {New York, NY},
}
@techreport{TrivinoandMoralesTR-01,
       author = {J.L. {Trivi\~{n}o} and R. Morales},
        title = {{MPSG}s ({M}ultiattribute {P}rediction {S}uffix {G}raphs)},
        issue = {Research Tecchnical Report ITI-2001-6},
  institution = {Universidad de {M\'{a}laga} ({Spa\~{n}a})},
         year = 2001,
}
@article{TrubutschekBIORXIV-16,
       author = {Tr\"{u}butschek, Darinka and Marti, S\'{e}bastien and Ojeda, Andr\'{e}s and King, Jean-R\'{e}mi and Mi, Yuanyuan and Tsodyks, Misha and Dehaene, Stanislas},
        title = {A theory of working memory without consciousness or sustained activity},
      journal = {Biorxiv},
         year = {2016},
     abstract = {Working memory and conscious perception are thought to share similar brain mechanisms, yet recent reports of non-conscious working memory challenge this view. Combining visual masking with magnetoencephalography, we demonstrate the reality of nonconscious working memory and dissect its neural mechanisms. In a spatial delayed-response task, participants reported the location of a subjectively unseen target above chance-level after a long delay. Conscious perception and conscious working memory were characterized by similar signatures: a sustained desynchronization in the alpha/beta band over frontal cortex, and a decodable representation of target location in posterior sensors. During non-conscious working memory, such activity vanished. Our findings contradict models that identify working memory with sustained neural firing, but are compatible with recent proposals of 'activity-silent' working memory. We present a theoretical framework and simulations showing how slowly decaying synaptic changes allow cell assemblies to go dormant during the delay, yet be retrieved above chance-level after several seconds.}
}
@article{TruccoloetalNATURE-10,
       author = {Truccolo, Wilson and Hochberg, Leigh R. and and Donoghue, John P.},
        title = {Collective dynamics in human and monkey sensorimotor cortex: predicting single neuron spikes},
      journal = {Nature Neuroscience},
         year = 2010,
    publisher = {Nature Publishing Group},
       volume = 13,
        issue = 1,
        pages = {105-111},
}
@article{TruccoloetalNEUROSCIENCE-08,
        title = {Primary motor cortex tuning to intended movement kinematics in humans with tetraplegia},
       author = {Truccolo, W. and Friehs, G.M. and Donoghue, J.P. and Hochberg, L.R.},
      journal = {Journal Neurosci},
       volume = {28},
        issue = {5},
         year = {2008},
        pages = {1163-78},
     abstract = {The relationship between spiking activities in motor cortex and movement kinematics has been well studied in neurologically intact nonhuman primates. We examined the relationship between spiking activities in primary motor cortex (M1) and intended movement kinematics (position and velocity) using 96-microelectrode arrays chronically implanted in two humans with tetraplegia. Study participants were asked to perform two different tasks: imagined pursuit tracking of a cursor moving on a computer screen and a "neural cursor center-out" task in which cursor position was controlled by the participant's neural activity. In the pursuit tracking task, the majority of neurons were significantly tuned: 90\% were tuned to velocity and 86\% were tuned to position in one participant; 95\% and 84\%, respectively, in the other. Additionally, velocity and position of the tracked cursor could be decoded from the ensemble of neurons. In the neural cursor center-out task, tuning to direction of the intended target was well captured by a log-linear cosine function. Neural spiking soon after target appearance could be used to classify the intended target with an accuracy of 95\% in one participant, and 80\% in the other. It was also possible to extract information about the direction of the difference vector between the target position and the instantaneous neural cursor position. Our results indicate that correlations between spiking activity and intended movement velocity and position are present in human M1 after the loss of descending motor pathways, and that M1 spiking activities share many kinematic tuning features whether movement is imagined by humans with tetraplegia, or is performed as shown previously in able-bodied nonhuman primates.}
}
@article{TruongetalCoRR-17,
        title = {{MACA:} {A} Modular Architecture for Conversational Agents},
       author = {Hoai Phuoc Truong and Prasanna Parthasarathi and Joelle Pineau},
      journal = {CoRR},
       volume = {arXiv:1705.00673},
         year = {2017},
     abstract = {We propose a software architecture designed to ease the implementation of dialogue systems. The Modular Architecture for Conversational Agents (MACA) uses a plug-n-play style that allows quick prototyping, thereby facilitating the development of new techniques and the reproduction of previous work. The architecture separates the domain of the conversation from the agent's dialogue strategy, and as such can be easily extended to multiple domains. MACA provides tools to host dialogue agents on Amazon Mechanical Turk (mTurk) for data collection and allows processing of other sources of training data. The current version of the framework already incorporates several domains and existing dialogue strategies from the recent literature.}
}
@article{TsaiSaxeCoxCoRR-16,
       author = {Chuan-Yung Tsai and Andrew Saxe and David Cox},
        title = {Tensor Switching Networks},
      journal = {CoRR},
       volume = {arXiv:1610.10087},
         year = {2016},
     abstract = {We present a novel neural network algorithm, the Tensor Switching (TS) network, which generalizes the Rectified Linear Unit (ReLU) nonlinearity to tensor-valued hidden units. The TS network copies its entire input vector to different locations in an expanded representation, with the location determined by its hidden unit activity. In this way, even a simple linear readout from the TS representation can implement a highly expressive deep-network-like function. The TS network hence avoids the vanishing gradient problem by construction, at the cost of larger representation size. We develop several methods to train the TS network, including equivalent kernels for infinitely wide and deep TS networks, a one-pass linear learning algorithm, and two backpropagation-inspired representation learning algorithms. Our experimental results demonstrate that the TS network is indeed more expressive and consistently learns faster than standard ReLU networks.}
}
@book{Tsang93,
       author = {Edward Tsang},
        title = {Foundations of Constraint Satisfaction},
    publisher = {Academic Press},
      address = {San Diego, California},
         year = 1993,
}
@article{TschoppetalCoRR-18,
       author = {A Connectome Based Hexagonal Lattice Convolutional Network Model of the Drosophila Visual System},
        title = {Fabian David Tschopp and Michael B. Reiser and Srinivas C. Turaga},
      journal = {CoRR},
       volume = {arXiv:1806.04793},
         year = {2018},
          url = {https://arxiv.org/abs/1806.04793},
     abstract = {What can we learn from a connectome? We constructed a simplified model of the first two stages of the fly visual system, the lamina and medulla. The resulting hexagonal lattice convolutional network was trained using backpropagation through time to perform object tracking in natural scene videos. Networks initialized with weights from connectome reconstructions automatically discovered well-known orientation and direction selectivity properties in T4 neurons and their inputs, while networks initialized at random did not. Our work is the first demonstration, that knowledge of the connectome can enable in silico predictions of the functional properties of individual neurons in a circuit, leading to an understanding of circuit function from structure alone.}
}
@article{TsengOR-90,
        title = {Solving {$H$}-Horizon, Stationary {M}arkov decision problems in time proportional to {$\log(H)$}},
       author = {Paul Tseng},
      journal = {Operations Research Letters},
       volume = 9,
        issue = 5,
    publisher = {North-Holland},
        pages = {287-297},
         year = 1990,
}
@article{TsengandNotargiacomoJNN-05,
       author = {Tseng, A. A. and Notargiacomo, A.},
        title = {Nanoscale fabrication by nonconventional approaches},
      journal = {Journal of Nanoscience and Nanotechnology},
         year = 2005,
       volume = 5,
        issue = 5,
        pages = {683-702},
     abstract = {An overview of three major nonconventional approaches in nanofabrication, scanning probe microscopy lithography, self-assembly, and imprint lithography, is presented. Typically, these nonconventional approaches are emerging technologies based on simple principles with potential cost-effective manufacturability, as compared to those conventional processes that are widely used and highly developed for making microelectronic circuits. Following the introduction of nonconventional technologies and their significances in nanofabrication, the details of each approach are presented. The lithographic applications of scanning probe microscopy, which involves three major variations: scanning tunneling microscopy, atomic force microscopy, and scanning near-field optical microscopy, are first evaluated with emphasis on their abilities in making structures with subnanoscale resolutions. The principles and processes for each technique are presented while their differences are also discussed. For the second approach, self-assembly, which uses a bottom-up fabrication strategy, is reported starting with an introduction of its basic principle. Self-assembly, with and without externally controlled forces for patterning nanoscale structures, is then examined. The associated principles and procedures of key assembling processes are presented. The third one, imprint lithography, is addressed with an emphasis on its recent progress and challenges. The nanolithographic abilities of different techniques developed using the general imprinting principle are examined. Finally, concluding remarks are provided to summarize the major technologies studied and to recommend the scopes for technology improvement and future research.}
}
@incollection{TsienHBN-08,
       author = {Joe Z. Tsien},
       editor = {Dere, E. and Easton, A. and Nadel, L. and Huston, J.P.},
        title = {Chapter 4.1 Neural coding of episodic memory},
    booktitle = {Handbook of Behavioral Neuroscience},
    publisher = {Elsevier Science},
         year = {2008},
        pages = {412-431}
}
@article{TsienetalCELL-96,
       author = {Tsien JZ and Chen DF and Gerber D and Tom C and Mercer EH and Anderson DJ, Mayford M and Kandel ER and Tonegawa S},
        title = {Subregion- and cell type-restricted gene knockout in mouse brain},
      journal = {Cell},
       volume = 87,
        issue = 7,
         year = 1996,
        pages = {1317-26},
     abstract = {Using the phage P1 derived Cre/loxP recombination system, we have developed a method to create mice in which the deletion (knockout) of virtually any gene of interest is restricted to a subregion or a specific cell type in the brain such as the pyramidal cells of the hippocampal CA1 region. The Cre/loxP recombination-based gene deletion appears to require a certain level of Cre protein expression. The brain subregional restricted gene knockout should allow a more precise analysis of the impact of a gene mutation on animal behaviors.}
}
@article{TsitsiklisandVanRoyML-96,
       author = {Tsitsiklis, John N. and Van Roy, Benjamin},
        title = {Feature-Based Methods for Large Scale Dynamic Programming},
      journal = {Machine Learning},
       volume = 22,
        pages = {59-94},
         year = 1996,
}
@article{TsodyksandMarkramPNAS-97,
       author = {Tsodyks, Misha V. and Markram, Henry},
        title = {The neural code between neocortical pyramidal neurons depends on neurotransmitter release probability},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {94},
        issue = {2},
         year = {1997},
        pages = {719-723},
}
@article{TsodyksetalJN-00,
       author = {Tsodyks, M. and Uziel, A. and Markram, H.},
        title = {Synchrony generation in recurrent networks with frequency-dependent synapses},
      journal = {Journal of Neuroscience},
         year = {2000},
       volume = {20},
       number = {1},
        pages = {RC50 1-5},
     abstract = {Throughout the neocortex, groups of neurons have been found to fire synchronously on the time scale of several milliseconds. This near coincident firing of neurons could coordinate the multifaceted information of different features of a stimulus. The mechanisms of generating such synchrony are not clear. We simulated the activity of a population of excitatory and inhibitory neurons randomly interconnected into a recurrent network via synapses that display temporal dynamics in their transmission; surprisingly, we found a behavior of the network where action potential activity spontaneously self-organized to produce highly synchronous bursts involving virtually the entire network. These population bursts were also triggered by stimuli to the network in an all-or-none manner. We found that the particular intensities of the external stimulus to specific neurons were crucial to evoke population bursts. This topographic sensitivity therefore depends on the spectrum of basal discharge rates across the population and not on the anatomical individuality of the neurons, because this was random. These results suggest that networks in which neurons are even randomly interconnected via frequency-dependent synapses could exhibit a novel form of reflex response that is sensitive to the nature of the stimulus as well as the background spontaneous activity.}
}
@book{Tsotsos2011,
        title = {A Computational Perspective on Visual Attention},
       author = {Tsotsos, John K.},
         year = {2011},
    publisher = {MIT Press}
}
@article{TsotsosetalCVIU-05,
       author = {J. Tsotsos and Y. Liu and J. Martinez-Trujillo and M. Pomplun and E. Simine and K. Zhou},
        title = {Attending to visual motion},
      journal = {Computer Vision and Image Understanding},
        pages = {3-40},
       volume = 100,
        issue = {1-2},
         year = 2005,
}
@article {TsubokawaandRossNEUROPHYSIOLOGY-96,
       author = {Tsubokawa, H. and Ross, W. N.},
        title = {{IPSP}s modulate spike backpropagation and associated {[Ca2+]i} changes in the dendrites of hippocampal {CA1} pyramidal neurons},
      journal = {Journal of Neurophysiology},
    publisher = {American Physiological Society},
       volume = 76,
        issue = 5,
        pages = {2896-2906},
         year = 1996,
     abstract = {1. We studied the effects of synaptic inhibition on backpropagating Na+ spikes in the apical dendrites of CA1 pyramidal neurons in transverse slices from the rat hippocampus. Action potentials were evoked synaptically by stimulation in the stratum radiatum or antidromically by stimulation in the alveus. 2. Inhibitory postsynaptic potentials, evoked by stimulation in the stratum lacunosum moleculare, reduced the amplitude of single spikes in the distal dendrites but did not change the amplitudes in the somatic or proximal regions. Inhibition also reduced the spike-associated [Ca2+]i changes in the distal dendrites but had little effect on the changes in the proximal part of the cell. Both of these results are consistent with inhibition converting actively backpropagating spikes into passively spreading potentials at some point in the arbor. 3. In most cells, the spike amplitude reduction in the distal dendrites was blocked by bicuculline methiodide (10 microM) and inhibition was most effective when evoked in a time window < 10 ms preceding the action potential. This suggests that the amplitude reduction was due to a conductance shunt activated by gamma-aminobuturic acid-A (GABAA) receptors. Synaptically evoked GABAB responses were detected but usually did not block spike propagation. 4. Direct hyperpolarization in the distal dendrites was also effective in blocking antidromically evoked spike backpropagation but probably does not contribute when the action potentials are evoked synaptically. 5. This effect of inhibition is different from its usual function in synaptic integration because spike generation and propagation down the axon are not significantly affected. This kind of inhibition might be important in regulating transient [Ca2+]i changes in the dendrites including individual dendritic branches.},
}
@article{TsunodaetalNATURE-01,
       author = {Tsunoda, K. and Yamane, Y. and Nishizaki, M. and Tanifuji, M.},
        title = {Complex objects are represented in macaque inferotemporal cortex by the combination of feature columns},
      journal = {Nature Neuroscience},
       volume = 4,
        pages = {832-838},
         year = 2001,
}
@incollection{Tucker2008microgenesis,
       author = {Don M. Tucker and Gwen Frishkoff and Phan Luu},
        title = {{Chapter 5 - Microgenesis of Language: Vertical Integration of Linguistic Mechanisms Across the Neuraxis}},
       editor = {Brigitte Stemmer and Harry A. Whitaker},
    booktitle = {Handbook of the Neuroscience of Language},
    publisher = {Elsevier},
      address = {San Diego},
         year = {2008},
        pages = {45-55},
}
@article{TuckerandHintonIEEE-11,
       author = {Tucker, R.S. and Hinton, K.},
      journal = {Photonics Journal, IEEE},
        title = {Energy Consumption and Energy Density in Optical and Electronic Signal Processing},
       volume = 3,
        issue = 5,
         year = 2011,
        pages = {821-833},
     abstract = {We compare the energy consumption of digital optical and digital electronic signal processing circuits, including the contributions to energy consumption of the optical to electrical (O/E) converters and electrical to optical (E/O) converters, and the demultiplexers (DEMUXs) and multiplexers (MUXs) required for electronic circuits to process high-speed optical signals. This paper focuses on three key practical considerations, namely, energy consumption, energy density, and the complexity of processing. We show that optical signal processing is potentially competitive with electronics in very high-speed circuits that provide only limited processing, i.e., when only a small number of processing operations are performed on each bit of data. However, in applications that require anything more than limited processing, electronics provides better energy efficiency and occupies a smaller footprint. In these applications, electronics is likely to remain the technology of choice. More attention needs to be paid to energy consumption issues in the research and development of new digital optical technologies.},
}
@misc{TuckeretalGEODESICS-09,
        title = {Neural Mechanisms of Inhibitory Specification in Cognitive and Linguistic Complexity},
       author = {Don M. Tucker and Phan Luu and Catherine Poulsen},
         year = {2009},
 howpublished = {http://www.ruf.rice.edu/~eivs/sympo/papers/Tucker.pdf},
 abstract = {We theorize that the ability to organize a complex linguistic structure such as a hierarchic clause may depend upon the capacity for what may be called inhibitory specification, in which certain meanings are isolated within working memory, such that they can be sustained and grouped within hierarchic structures. Clues to the neural mechanisms of inhibitory specification can be gained from studying the sequencing and routinization of action within the motor system, and from examining the unique properties of object memory wtihin the ventral corticolimbic pathway.}
}
@article{TuckeretalIEEE_SMC-01,
       author = {A. Tucker and S. Swift and X. Liu},
      journal = {IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics)},
        title = {Variable grouping in multivariate time series via correlation},
         year = {2001},
       volume = {31},
       number = {2},
        pages = {235-245},
     abstract = {The decomposition of high-dimensional multivariate time series (MTS) into a number of low-dimensional MTS is a useful but challenging task because the number of possible dependencies between variables is likely to be huge. This paper is about a systematic study of the "variable groupings" problem in MTS. In particular, we investigate different methods of utilizing the information regarding correlations among MTS variables. This type of method does not appear to have been studied before. In all, 15 methods are suggested and applied to six datasets where there are identifiable mixed groupings of MTS variables. This paper describes the general methodology, reports extensive experimental results, and concludes with useful insights on the strength and weakness of this type of grouping method},
}

@book{Tulving1972,
        title = {Organization of memory},
       author = {Tulving, E. and Donaldson, W. and Bower, G.H.},
         year = {1972},
    publisher = {Academic Press}
}
@book{Tulving1985,
        title = {Elements of Episodic Memory},
       author = {Tulving, Endel},
       series = {Oxford Psychology Series},
         year = {1985},
    publisher = {Oxford University Press}
}
@article{TumaetalNATURE-NANOTECHNOLOGY-16,
       author = {Tuma, Tomas and Pantazi, Angeliki and Le Gallo, Manuel and Sebastian, Abu and Eleftheriou, Evangelos},
        title = {Stochastic phase-change neurons},
      journal = {Nature Nanotechnology},
         year = {2016},
    publisher = {Nature Publishing Group},
       volume = {11},
        issue = {8},
        pages = {693-699},
     abstract = {Artificial neuromorphic systems based on populations of spiking neurons are an indispensable tool in understanding the human brain and in constructing neuromimetic computational systems. To reach areal and power efficiencies comparable to those seen in biological systems, electroionics-based and phase-change-based memristive devices have been explored as nanoscale counterparts of synapses. However, progress on scalable realizations of neurons has so far been limited. Here, we show that chalcogenide-based phase-change materials can be used to create an artificial neuron in which the membrane potential is represented by the phase configuration of the nanoscale phase-change device. By exploiting the physics of reversible amorphous-to-crystal phase transitions, we show that the temporal integration of postsynaptic potentials can be achieved on a nanosecond timescale. Moreover, we show that this is inherently stochastic because of the melt-quench-induced reconfiguration of the atomic structure occurring when the neuron is reset. We demonstrate the use of these phase-change neurons, and their populations, in the detection of temporal correlations in parallel data streams and in sub-Nyquist representation of high-bandwidth signals.},
}
@inproceedings{TuragaetalNIPS-13,
        title = {Inferring neural population dynamics from multiple partial recordings of the same neural circuit},
       author = {S.C. Turaga and L. Buesing and A.M. Packer and H. Dalgleish and N. Petit and M. Hausser and J.H. Macke},
    booktitle = {Advances in Neural Information Processing Systems 26},
      address = {Tahoe, CA},
         year = 2013,
        pages = {3093-3101},
}
@article{Turing50,
       author = {Turing, Alan},
        title = {Computing Machinery and Intelligence},
      journal = {Mind},
       volume = 59,
         year = 1950,
        pages = {434-460},
}
@inproceedings{TurkandPentlandCVPR-91,
       author = {Turk, M. A. and Pentland, A. P.},
        title = {Face Recognition Using Eigenfaces},
    booktitle = {Proceedings of the 1991 {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
        pages = {586-591},
         year = 1991,
}
@article{TurkandPentlandJCN-91,
       author = {Turk, Matthew and Pentland, Alex},
        title = {Eigenfaces for Recognition},
      journal = {Journal Cognitive Neuroscience},
    publisher = {MIT Press},
      address = {Cambridge, MA, USA},
       volume = {3},
       number = {1},
         year = {1991},
        pages = {71-86},
     abstract = {We have developed a near-real-time computer system that can locate and track a subject's head, and then recognize the person by comparing characteristics of the face to those of known individuals. The computational approach taken in this system is motivated by both physiology and information theory, as well as by the practical requirements of near-real-time performance and accuracy. Our approach treats the face recognition problem as an intrinsically two-dimensional (2-D) recognition problem rather than requiring recovery of three-dimensional geometry, taking advantage of the fact that faces are normally upright and thus may be described by a small set of 2-D characteristic views. The system functions by projecting face images onto a feature space that spans the significant variations among known face images. The significant features are known as "eigenfaces," because they are the eigenvectors (principal components) of the set of faces; they do not necessarily correspond to features such as eyes, ears, and noses. The projection operation characterizes an individual face by a weighted sum of the eigenface features, and so to recognize a particular face it is necessary only to compare these weights to those of known individuals. Some particular advantages of our approach are that it provides for the ability to learn and later recognize new faces in an unsupervised manner, and that it is easy to implement using a neural network architecture.}
}
@article{Turner-EvansandJayaramanCURRENT-BIOLOGY-16,
       author = {Daniel B. Turner-Evans and Vivek Jayaraman},
        title = {The insect central complex},
      journal = {Current Biology},
       volume = {26},
       number = {11},
        pages = {R453-R457},
         year = {2016},
     abstract = {Summary Hordes of tourists flock to Washington, D.C. every spring to see the cherry trees blossom. Once in the city, they must find their way to the Tidal Basin where the Japanese trees grow. Fortunately, a number of visual landmarks can help them to navigate. In 1910, the United States Congress passed The Height of Buildings Act, limiting the elevation of commercial and residential structures in D.C. to 130 feet. Thus, the 555-foot-tall Washington Monument often looms large against the horizon, serving as an anchor point to help set the tourists' sense of direction. Once their heading is set, they can lose sight of the monument behind buildings or groups of tall Scandinavian visitors and still use their internal compass to navigate to the Basin. This compass keeps track of their paces and turns and updates their sense of where they are and where they need to go. Yet while their heading informs their actions, it does not dictate them. Tourists who have been to D.C. in the past can, for example, use remembered views to alter their routes to avoid crowds. On an even finer scale, their leg movements also depend on their current state -- they might increase the frequency and length of their strides if hunger pangs compete with their desire to see cherry blossoms, for example. The way in which these disparate cues and motivations influence exploration is a neuroscience mystery across creatures large and small.}
}
@article{TusnadyandCsiszarSDSI-84,
       author = {Sanjeev Arora and Rong Ge and Tengyu Ma and Ankur Moitra},
        title = {Information geometry and alternating minimization procedures},
      journal = {Statistics \& Decisions: Supplement Issues},
       volume = {1},
         year = {1984},
        pages = {205-237},
     abstract = {Keywords: Incomplete data, EM algorithm, Kullback-Leibler informational divergence}
}
@article{TuthilletalNEURON-14,
        title = {Wide-Field Feedback Neurons Dynamically Tune Early Visual Processing},
      journal = {Neuron},
       volume = {82},
       number = {4},
         year = {2014},
        pages = {887-895},
       author = {John C. Tuthill and Aljoscha Nern and Gerald M. Rubin and Michael B. Reiser},
     abstract = {An important strategy for efficient neural coding is to match the range of cellular responses to the distribution of relevant input signals. However, the structure and relevance of sensory signals depend on behavioral state. Here, we show that behavior modifies neural activity at the earliest stages of fly vision. We describe a class of wide-field neurons that provide feedback to the most peripheral layer of the Drosophila visual system, the lamina. Using in vivo patch-clamp electrophysiology, we found that lamina wide-field neurons respond to low-frequency luminance fluctuations. Recordings in flying flies revealed that the gain and frequency tuning of wide-field neurons change during flight, and that these effects are mimicked by the neuromodulator octopamine. Genetically silencing wide-field neurons increased behavioral responses to slow-motion stimuli. Together, these findings identify a cell type that is gated by behavior to enhance neural coding by subtracting low-frequency signals from the inputs to motion detection circuits.}
}
@book{Tversky2019mindinmotion,
        title = {Mind in Motion: {H}ow Action Shapes Thought},
       author = {Tversky, Barbara},
         year = {2019},
    publisher = {Basic Books}
}
@article{TverskyandKahnemanPB-71,
        title = {Belief in the law of small numbers},
       author = {Tversky, Amos and Kahneman, Daniel},
      journal = {Psychological Bulletin},
       volume = 76,
        issue = 2,
         year = 1971,
        pages = {105-110}
}
@article{TverskyandKahnemanSCIENCE-74,
      title = {Judgment under Uncertainty: Heuristics and Biases},
     author = {Amos Tversky and Daniel Kahneman},
    journal = {Science},
     volume = 185,
      issue = 24157,
       year = 1974,
      pages = {1124-1131},
}
@article{TyeetallNATURE-12,
        title = {Dopamine neurons modulate neural encoding and expression of depression-related behaviour},
       author = {Tye, Kay M. and Mirzabekov, Julie J. and Warden, Melissa R. and Ferenczi, Emily A. and Tsai, Hsing-Chen and Finkelstein, Joel and Kim, Sung-Yon and Adhikari, Avishek and Thompson, Kimberly R. and Andalman, Aaron S. and Gunaydin, Lisa A. and Witten, Ilana B. and Deisseroth, Karl},
      journal = {Nature},
    publisher = {Nature Publishing Group,},
         year = {2012},
       volume = {493},
        pages = {537-541},
     abstract = {Major depression is characterized by diverse debilitating symptoms that include hopelessness and anhedonia. Dopamine neurons involved in reward and motivation are among many neural populations that have been hypothesized to be relevant10, and certain antidepressant treatments, including medications and brain stimulation therapies, can influence the complex dopamine system. Until now it has not been possible to test this hypothesis directly, even in animal models, as existing therapeutic interventions are unable to specifically target dopamine neurons. Here we investigated directly the causal contributions of defined dopamine neurons to multidimensional depression-like phenotypes induced by chronic mild stress, by integrating behavioural, pharmacological, optogenetic and electrophysiological methods in freely moving rodents. We found that bidirectional control (inhibition or excitation) of specified midbrain dopamine neurons immediately and bidirectionally modulates (induces or relieves) multiple independent depression symptoms caused by chronic stress. By probing the circuit implementation of these effects, we observed that optogenetic recruitment of these dopamine neurons potently alters the neural encoding of depression-related behaviours in the downstream nucleus accumbens of freely moving rodents, suggesting that processes affecting depression symptoms may involve alterations in the neural encoding of action in limbic circuitry.}
}
@article{TzengML-92,
       author = {Tzeng, Wen-Guey},
        title = {Learning Probabilistic Automata and {M}arkov Chains via Queries},
      journal = {Machine Learning},
       volume = 8,
         year = 1992,
        pages = {151-166},
}
@book{UAI-86,
       editor = {Kanal, Laveen F. and Lemmer, John F.},
        title = {Uncertainty in Artificial Intelligence},
    booktitle = {Uncertainty in Artificial Intelligence},
    publisher = {North-Holland},
         year = 1986,
}
@book{UAI-88,
       editor = {Kanal, Laveen F. and Lemmer, John F.},
        title = {Uncertainty in Artificial Intelligence 2},
    booktitle = {Uncertainty in Artificial Intelligence 2},
    publisher = {North-Holland},
         year = 1988,
}
@book{UAI-89,
    editor = {Max Henrion and Ross D. Shachter and Laveen N. Kanal and John F. Lemmer},
     title = {Uncertainty in Artificial Intelligence},
 booktitle = {Uncertainty in Artificial Intelligence: Volume 5},
 publisher = {North-Holland},
      year = 1990,
}
@book{ULPB_TOC_AND_CHAPTERS-01-02-07-14-19-22-23-24-26,
        title = {Ultra Low Power Bioelectronics: Fundamentals, Biomedical Applications, and Bio-inspired Systems},
       author = {Rahul Sarpeshkar},
         year = {2010},
    publisher = {Cambridge University Press},
}
@book{Uglow02,
      author = {The Lunar Men: Five Friends Whose Curiosity Changed the World},
       title = {Jenny Uglow},
   publisher = {Farrar, Straus \& Giroux},
     address = {New York, NY},
        year = 2002,
}
@book{Ullman79,
   author = {Shimon Ullman},
    title = {The Interpretation of Visual Motion},
publisher = {MIT Press},
  address = {Cambridge, MA, USA},
     year = 1979,
}
@book{Ullman82,
    author = {Ullman, Jeffrey D.},
     title = {Principles of Database Systems},
 publisher = {Computer Science Press},
  location = {Rockville, MD},
      year = 1982,
}
@incollection{Ullman87,
   author = {Shimon Ullman},
    title = {Analysis of Visual Motion by Biological and Computer Systems},
booktitle = {Readings in Computer Vision: Issues, Problems, Principles, and Paradigms},
   editor = {M. A. Fischler and O. Firschein},
publisher = {Kaufmann},
  address = {Los Altos, CA.},
    pages = {132-144},
     year = 1987,
}
@book{Ullman88,
     author = {Jeffrey D. Ullman},
      title = {Principles of Database and Knowledge-Base Systems},
  publisher = {W. H. Freeman},
    comment = {Originally Computer Science Press},
    address = {New York, NY},
       year = 1988,
}
@article{UllmanCOGNITION-84,
        title = {Visual Routines},
       author = {Ullman, Shimon},
      journal = {Cognition},
       volume = {18},
         year = {1984},
        pages = {97-159},
     abstract = {This paper examines the processing of visual information beyond the creation of the early representations. A fundamental requirement at this level is the capacity to establish visually abstract shape properties and spatial relations. This capacity plays a major role in object recognition, visually guided manipulation, and more abstract visual thinking.

For the human visual system, the perception of spatial properties and relations that are complex from a computational standpoint nevertheless often appears deceivingly immediate and effortless. The proficiency of the human system in analyzing spatial information far surpasses the capacities of current artificial systems. The study of the computations that underlie this competence may therefore lead to the development of new more efficient methods for the spatial analysis of visual information.

The perception of abstract shape properties and spatial relations raises fundamental difficulties with major implications for the overall processing of visual information. It will be argued that the computation of spatial relations divides the analysis of visual information into two main stages. The first is the bottom-up creation of certain representations of the visible environment. The second stage involves the application of processes called 'visual routines' to the representations constructed in the first stage. These routines can establish properties and relations that cannot be represented explicitly in the initial representations.

Visual routines are composed of sequences of elemental operations. Routines for different properties and relations share elemental operations. Using a fixed set of basic operations, the visual system can assemble different routines to extract an unbounded variety of shape properties and spatial relations. }
}
@techreport{UllmanTR-83,
     author = {Ullman, Shimon},
      title = {Visual Routines},
       type = {{AI} Memo},
      issue = 723,
institution = {{MIT} {AI} Laboratory},
       year = 1983,
}
@article{UllmanandSolovievNN-99,
     title = {Computation of pattern invariance in brain-like structures},
    author = {S. Ullman and S. Soloviev},
   journal = {Neural Networks},
    volume = 12,
     pages = {1021-1036},
       url = {http://citeseer.ist.psu.edu/ullman99computation.html},
      year = 1999
}
@article{UllmanetalNN-02,
     title = {Visual Features of Intermediate Complexity and their use in Classification},
    author = {Shimon Ullman and Michel Vidal-Naquet and Erez Sali},
   journal = {Nature Neuroscience},
    volume = 5,
     issue = 7,
     pages = {682-687},
       url = {http://www.nature.com/neuro/journal/v5/n7/full/nn870.html},
      year = 2002,
}
@article{UllmanetalPNAS-12,
       author = {Ullman, Shimon and Harari, Daniel and Dorfman, Nimrod},
        title = {From simple innate biases to complex visual concepts},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {109},
       number = {44},
         year = {2012},
        pages = {18215-18220},
     abstract = {Early in development, infants learn to solve visual problems that are highly challenging for current computational methods. We present a model that deals with two fundamental problems in which the gap between computational difficulty and infant learning is particularly striking: learning to recognize hands and learning to recognize gaze direction. The model is shown a stream of natural videos and learns without any supervision to detect human hands by appearance and by context, as well as direction of gaze, in complex natural scenes. The algorithm is guided by an empirically motivated innate mechanism--the detection of "mover" events in dynamic images, which are the events of a moving image region causing a stationary region to move or change after contact. Mover events provide an internal teaching signal, which is shown to be more effective than alternative cues and sufficient for the efficient acquisition of hand and gaze representations. The implications go beyond the specific tasks, by showing how domain-specific "proto concepts" can guide the system to acquire meaningful concepts, which are significant to the observer but statistically inconspicuous in the sensory input.},
}
@inproceedings{Upfal97,
    author = {Eli Upfal},
     title = {Stochastic Analysis of Dynamic Processes},
 booktitle = {Proceedings of the 11th International Symposium on Fundamentals of Computation Theory},
     pages = {85-92},
      year = 1997,
}
@inproceedings{Upfal98,
    author = {Eli Upfal},
     title = {Design and Analysis of Dynamic Processes: A Stochastic Approach},
 booktitle = {Algorithm - 6th Annual European Symposium},
     pages = {26-34},
      year = 1998,
}
@article{UrbanetalCoRR-16,
       author = {Gregor Urban and Krzysztof J. Geras and Samira Ebrahimi Kahou and \"{O}zlem Aslan and Shengjie Wang and Rich Caruana and Abdelrahman Mohamed and Matthai Philipose and Matthew Richardson},
        title = {Do Deep Convolutional Nets Really Need to be Deep (Or Even Convolutional)?},
      journal = {CoRR},
       volume = {arXiv:1603.05691},
         year = {2016},
}
@article{UtgoffNC-02,
  author = {Paul Utgoff},
   title = {Many-Layered Learning},
 journal = {Neural Computation},
  volume = 14,
   pages = {2497-2539},
    year = 2002,
}
@book{Uttal2003,
    author = {Uttal, William R.}, 
     title = {The New Phrenology: The Limits of Localizing Cognitive Processes in the Brain},
   address = {Cambridge, MA},
 publisher = {MIT Press},
      year = 2003,
}
@article{VaddirajuetalNIH-10,
        title = {Emerging Synergy between Nanotechnology and Implantable Biosensors: A Review},
       author = {Santhisagar Vaddiraju and Ioannis Tomazos and Diane J. Burgess and Faquir C. Jain and Fotios Papadimitrakopoulos},
      journal = {Biosensors and Bioelectronics},
       volume = 25,
        issue = 7,
         year = 2010,
        pages = {1553-1565},
}
@article{VaismanetalEMBO-05,
    author = {Vaisman, Alexandra and Ling, Hong and Woodgate, Roger and Yang, Wei},
     title = {Fidelity of {Dpo4}: effect of metal ions, nucleotide selection and pyrophosphorolysis},
   journal = {European Molecular Biology Organization Journal},
    volume = 24,
     issue = 17,
      year = 2005,
     pages = {2957-2967},
}
@article{ValiantCACM-84,
       author = {Valiant, L. G.},
        title = {A Theory of the Learnable},
      journal = {Communications of the {ACM}},
       volume = 27,
        pages = {1134-1142},
         year = 1984,
}
@article{ValiantJACM-00,
       author = {Valiant, Leslie G.},
        title = {A neuroidal architecture for cognitive computation},
      journal = {Journal of the {ACM}},
       volume = 47,
        issue = 5,
        pages = {854-882},
         year = 2000,
}
@article{ValiantNC-05,
        title = {Memorization and association on a realistic neural model},
       author = {Valiant, Leslie G.},
      journal = {Neural Compututation},
       volume = 17,
        issue = 3,
         year = 2005,
        pages = {527-555},
     abstract = {A central open question of computational neuroscience is to identify the data structures and algorithms that are used in mammalian cortex to support successive acts of the basic cognitive tasks of memorization and association. This letter addresses the simultaneous challenges of realizing these two distinct tasks with the same data structure, and doing so while respecting the following four basic quantitative parameters of cortex: the neuron number, the synapse number, the synapse strengths, and the switching times. Previous work has not succeeded in reconciling these opposing constraints, the low values of synapse strengths that are typically observed experimentally having contributed a particular obstacle. In this article, we describe a computational scheme that supports both memory formation and association and is feasible on networks of model neurons that respect the widely observed values of the four quantitative parameters. Our scheme allows for both disjoint and shared representations. The algorithms are simple, and in one version both memorization and association require just one step of vicinal or neighborly influence. The issues of interference among the different circuits that are established, of robustness to noise  and of the stability of the hierarchical memorization process are addressed. A calculus therefore is implied for analyzing the capabilities of particular neural systems and subsystems, in terms of their basic numerical parameters.},
}
@incollection{ValiantetalTAMC-09,
        title = {Neural Computations That Support Long Mixed Sequences of Knowledge Acquisition Tasks},
       author = {Valiant, Leslie G. and Chen, Jianer and Cooper, S.},
    booktitle = {Lecture Notes in Computer Science: Theory and Applications of Models of Computation},
    publisher = {Springer},
      address = {Berlin},
         year = 2009,
}
@article{VallortigaraCN-12,
       author = {Giorgio Vallortigara},
        title = {Core knowledge of object, number, and geometry: A comparative and neural approach},
      journal = {Cognitive Neuropsychology},
       volume = {29},
       number = {1-2},
        pages = {213-236},
         year = {2012},
    publisher = {Routledge},
     abstract = {Studies on the ontogenetic origins of human knowledge provide evidence for a small set of separable systems of core knowledge dealing with the representation of inanimate and animate objects, number, and geometry. Because core knowledge systems are evolutionarily ancient, they can be investigated from a comparative perspective, making use of various animal models. In this review, I discuss evidence showing precocious abilities in nonhuman species to represent (a) objects that move partly or fully out of view and their basic mechanical properties such as solidity, (b) the cardinal and ordinal/sequential aspects of numerical cognition and rudimentary arithmetic with small numerosities, and (c) the geometrical relationships among extended surfaces in the surrounding layout. Controlled rearing studies suggest that the abilities associated with core knowledge systems of objects, number, and geometry are observed in animals in the absence (or with very reduced) experience, supporting a nativistic foundation of such cognitive mechanisms. Animal models also promise a fresh approach to the issue of the neurobiological and genetic mechanisms underlying the expression of core knowledge systems.}
}
@book{VanBenthem83,
       author = {Van Benthem, J.},
        title = {The Logic of Time},
    publisher = {Kluwer},
      address = {Boston, Massachusetts},
         year = 1983,
}
@article{VanGaalandLammeTHE-NEUROSCIENTIST-12,
       author = {Simon van Gaal and Victor A. F. Lamme},
        title = {Unconscious High-Level Information Processing: {I}mplication for Neurobiological Theories of Consciousness},
      journal = {The Neuroscientist},
       volume = {18},
       number = {3},
        pages = {287-301},
         year = {2012},
     abstract = {Theories about the neural correlates and functional relevance of consciousness have traditionally assigned a crucial role to the prefrontal cortex in generating consciousness as well as in orchestrating high-level conscious control over behavior. However, recent neuroscientific findings show that prefrontal cortex can be activated unconsciously. The depth, direction, and scope of these activations depend on several top-down factors such as the task being probed (task-set, strategy) and on (temporal/spatial) attention. Regardless, such activations-when mediated by feedforward activation only-do not lead to a conscious sensation. Although unconscious, these prefrontal activations are functional, in the sense that they are associated with behavioral effects of cognitive control, such as response inhibition, task switching, conflict monitoring, and error detection. These findings challenge the pivotal role of the prefrontal cortex in consciousness. Instead, it appears that specific brain areas (or cognitive modules) may support specific cognitive functions but that consciousness is independent of this. Conscious sensations arise only when the brain areas involved engage in recurrent interactions enabling the long-lasting exchange of information between brain regions. Moreover, recent evidence suggests that also the state of consciousness, for example, in vegetative state patients or during sleep and anesthesia, is closely related to the scope and extent of residual recurrent interactions among brain regions.}
}
@article{VanGaaletalRiHN-12,
       author = {van Gaal, Simon and de Lange, Floris P. and Cohen, Michael X.},
        title = {The role of consciousness in cognitive control and decision making},
      journal = {Frontiers in Human Neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = {6},
         year = {2012},
        pages = {121-121},
     abstract = {Here we review studies on the complexity and strength of unconscious information processing. We focus on empirical evidence that relates awareness of information to cognitive control processes (e.g., response inhibition, conflict resolution, and task-switching), the life-time of information maintenance (e.g., working memory) and the possibility to integrate multiple pieces of information across space and time. Overall, the results that we review paint a picture of local and specific effects of unconscious information on various (high-level) brain regions, including areas in the prefrontal cortex. Although this neural activation does not elicit any conscious experience, it is functional and capable of influencing many perceptual, cognitive (control) and decision-related processes, sometimes even for relatively long periods of time. However, recent evidence also points out interesting dissociations between conscious and unconscious information processing when it comes to the duration, flexibility and the strategic use of that information for complex operations and decision-making. Based on the available evidence, we conclude that the role of task-relevance of subliminal information and meta-cognitive factors in unconscious cognition need more attention in future work.},
}
@article{VanHoecketalRiHN-15,
       author = {Van Hoeck, Nicole and Watson, Patrick D. and Barbey, Aron K.},
        title = {Cognitive neuroscience of human counterfactual reasoning},
      journal = {Frontiers in Human Neuroscience},
         year = {2015},
    publisher = {Frontiers Media S.A.},
       volume = {9},
        pages = {420-420},
     abstract = {Counterfactual reasoning is a hallmark of human thought, enabling the capacity to shift from perceiving the immediate environment to an alternative, imagined perspective. Mental representations of counterfactual possibilities (e.g., imagined past events or future outcomes not yet at hand) provide the basis for learning from past experience, enable planning and prediction, support creativity and insight, and give rise to emotions and social attributions (e.g., regret and blame). Yet remarkably little is known about the psychological and neural foundations of counterfactual reasoning. In this review, we survey recent findings from psychology and neuroscience indicating that counterfactual thought depends on an integrative network of systems for affective processing, mental simulation, and cognitive control. We review evidence to elucidate how these mechanisms are systematically altered through psychiatric illness and neurological disease. We propose that counterfactual thinking depends on the coordination of multiple information processing systems that together enable adaptive behavior and goal-directed decision making and make recommendations for the study of counterfactual inference in health, aging, and disease.},
}
@inproceedings{VanKleefetalDATE-13,
       author = {Van Kleef, J. and Massey, T. and Ledochowitsch, P. and Muller, R. and Tiefenauer, R. and Blanche, T. and Sato, Hirotaka and Maharbiz, M. M.},
        title = {Cyborg Insects, Neural Interfaces and Other Things: Building Interfaces Between the Synthetic and the Multicellular},
    booktitle = {Proceedings of the Conference on Design, Automation and Test in Europe},
         year = 2013,
        pages = {1546-1546},
     abstract = {As the computation and communication circuits we build radically miniaturize (i.e. become so low power that 1 pJ is sufficient to bang out a bit of information over a wireless transceiver; become so small that 500 \hmmu{}m$^2$ of thinned CMOS can hold a reasonable sensor front-end and digital engine), the barrier to introducing all sorts of interfaces and control loops into organisms will lower radically. Put another way, the rapid pace of computation and communication miniaturization is swiftly blurring the line between the technological base that created us and the technological based we've created. This talk will provide an overview of recent work in the Maharbiz lab that touches on this concern. Some of this will cover our ongoing exploration of the remote control of insects in free flight via implantable radio-equipped miniature neural stimulating systems and more recent work in next generation mammalian neural interfaces for brain machine interface (BMI) applications.}
}
@inproceedings{VanLehnAAAI-83,
       author = {Kurt VanLehn},
        title = {Human Procedural Skill Acquisition: Theory, Model and Psychological Validation},
    booktitle = {Proceedings of AAAI-83},
     location = {Washington, DC, USA},
 organization = {AAAI},
         year = 1983,
     abstract = {It is thought by some that conversations are governed by tacit conventions, called felicity conditions or conversational postulates (Austin. 1962: Grice. 1975: Gordon \& Lakoff, 1975). Learning a procedural skill is also a communication act. The teacher communicates a procedure to the student over the course of several lessons. The central idea of the theory to be presented is that there are specific felicity conditions that govern learning. In particular, five newly discovered felicity conditions govern the kind of skill acquisition studied here. The theory has been embedded in a learning model - a large AI-based computer program. The model's performance has been compared to data from several thousand students learning ordinary mathematical procedures: subtracting multidigit numbers, adding fractions, and solving simple algebraic equations. A key criterion for the theory is that the set of procedures that the model learns should exactly match the set of procedures that students actually acquire, including their 'buggy' procedures. However, much more is needed for psychological validation of this theory, or any complex Al-based theory, than merely testing its predictions. The method used with this theory is presented.}
}
@article{VanOverwalleetalNEUROIMAGE-14,
        title = {Social cognition and the cerebellum: a meta-analysis of over 350 fMRI studies},
       author = {Van Overwalle, Frank and Baetens, Kris and Mari{\"e}n, Peter and Vandekerckhove, Marie},
      journal = {Neuroimage},
       volume = {86},
        pages = {554-572},
         year = {2014},
    publisher = {Elsevier}
}
@article{VanhaterenandrudermanRSL-B-98,
       author = {J. H. van Hateren and D. L. Ruderman},
        title = {Independent component analysis of natural image sequences yields spatiotemporal filters similar to simple cells in primary visual cortex},
      journal = {Proceedings Royal Society London {B}},
         year = 1998,
       volume = 265,
        pages = {2315-2320}
}
@article{VankovandBowersPTRS_B-19,
       author = {Ivan Vankov and Jeffrey S. Bowers},
        title = {Training neural networks to encode symbols enables combinatorial generalization},
      journal = {Philosophical Transactions of the Royal Society of London. B, Biological Sciences},
       volume = {375},
         year = {2019},
     abstract = {Combinatorial generalization - the ability to understand and produce novel combinations of already familiar elements - is considered to be a core capacity of the human mind and a major challenge to neural network models. A significant body of research suggests that conventional neural networks cannot solve this problem unless they are endowed with mechanisms specifically engineered for the purpose of representing symbols. In this paper, we introduce a novel way of representing symbolic structures in connectionist terms - the vectors approach to representing symbols (VARS), which allows training standard neural architectures to encode symbolic knowledge explicitly at their output layers. In two simulations, we show that neural networks not only can learn to produce VARS representations, but in doing so they achieve combinatorial generalization in their symbolic and non-symbolic output. This adds to other recent work that has shown improved combinatorial generalization under some training conditions, and raises the question of whether specific mechanisms or training routines are needed to support symbolic processing. This article is part of the theme issue 'Towards mechanistic models of meaning composition'.}
}
@article{VapnikandChervonenkisSIAM-71,
       author = {Vapnik, V. N. and Chervonenkis, A. Ya.},
        title = {On the Uniform Convergence of Relative Frequencies of Events to Their Probabilities},
      journal = {Theory of Probability and its Applications},
       volume = 16,
        issue = 2,
    publisher = {SIAM},
         year = 1971,
        pages = {264-280},
}
@book{VarelaThompsonRoschTHE_EMBODIED_MIND-91,
        title = {The Embodied Mind: Cognitive Science and Human Experience},
       author = {Varela, F.J. and Rosch, Eleanor. and Thompson, E.},
         year = {1991},
    publisher = {MIT Press}
}
@article{VarelaetalNATURE-01,
       author = {Varela, F. and Lachaux, J. P. and Rodriguez, E. and Martinerie, J.},
        title = {The {B}rainweb: phase synchronization and large-scale integration},
      journal = {Nature Reviews Neuroscience},
         year = 2001,
       volume = 2,
        issue = 4,
        pages = {229-239},
     abstract = {The emergence of a unified cognitive moment relies on the coordination of scattered mosaics of functionally specialized brain regions. Here we review the mechanisms of large-scale integration that counterbalance the distributed anatomical and functional organization of brain activity to enable the emergence of coherent behaviour and cognition. Although the mechanisms involved in large-scale integration are still largely unknown, we argue that the most plausible candidate is the formation of dynamic links mediated by synchrony over multiple frequency bands.}
}
@article{VasilakietalCoRR-13,
       author = {Eleni Vasilaki and Michele Giugliano},
        title = {Emergence of Connectivity Motifs in Networks of Model Neurons with Short- and Long-term Plastic Synapses},
      journal = {CoRR},
       volume = {arXiv:1301.7187},
         year = {2013},
     abstract = {Recent evidence in rodent cerebral cortex and olfactory bulb suggests that short-term dynamics of excitatory synaptic transmission is correlated to the occurrence of stereotypical connectivity motifs. In particular, it was observed that neurons with short-term facilitating synapses form predominantly reciprocal pairwise connections, while neurons with short-term depressing synapses form unidirectional pairwise connections. The cause of these structural differences in excitatory synaptic microcircuits is unknown.

                 We propose that these connectivity motifs emerge from the interactions between short-term synaptic dynamics (SD) and long-term spike-timing dependent plasticity (STDP). While the im- pact of STDP on SD was shown in simultaneous neuronal pair recordings in vitro, the mutual interactions between STDP and SD in large networks is still the subject of intense research. Our approach combines a SD phenomenological model with a STDP model that captures faithfully long-term plasticity dependence on both spike times and frequency. As a proof of concept, we explore in silico the impact of SD-STDP in recurrent networks of spiking neurons with random ini- tial connection efficacies and where synapses are either all short-term facilitating or all depressing. For identical background inputs, and as a direct consequence of internally generated activity, we find that networks with depressing synapses evolve unidirectional connectivity motifs, while net- works with facilitating synapses evolve reciprocal connectivity motifs. The same results hold for heterogeneous networks including both facilitating and depressing synapses. Our study highlights the conditions under which SD-STDP explains the correlation between facilitation and reciprocal connectivity motifs, as well as between depression and unidirectional motifs. These conditions may lead to the design of experiments for the validation of the proposed mechanism.},
}
@article{VassalloetalPNAS-15,
       author = {Vassallo, Christopher and Pathak, Darshankumar T. and Cao, Pengbo and Zuckerman, David M. and Hoiczyk, Egbert and Wall, Daniel},
        title = {Cell rejuvenation and social behaviors promoted by {LPS} exchange in myxobacteria},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {112},
       number = {22},
        pages = {E2939-E2946},
         year = {2015},
     abstract = {Bacterial cells in their native environments must cope with factors that compromise the integrity of the cell. The mechanisms of coping with damage in a social or multicellular context are poorly understood. Here we investigated how a model social bacterium, Myxococcus xanthus, approaches this problem. We focused on the social behavior of outer membrane exchange (OME), in which cells transiently fuse and exchange their outer membrane (OM) contents. This behavior requires TraA, a homophilic cell surface receptor that identifies kin based on similarities in a polymorphic region, and the TraB cohort protein. As observed by electron microscopy, TraAB overexpression catalyzed a prefusion OM junction between cells. We then showed that damage sustained by the OM of one population was repaired by OME with a healthy population. Specifically, LPS mutants that were defective in motility and sporulation were rescued by OME with healthy donors. In addition, a mutant with a conditional lethal mutation in lpxC, an essential gene required for lipid A biosynthesis, was rescued by Tra-dependent interactions with a healthy population. Furthermore, lpxC cells with damaged OMs, which were more susceptible to antibiotics, had resistance conferred to them by OME with healthy donors. We also show that OME has beneficial fitness consequences to all cells. Here, in merged populations of damaged and healthy cells, OME catalyzed a dilution of OM damage, increasing developmental sporulation outcomes of the combined population by allowing it to reach a threshold density. We propose that OME is a mechanism that myxobacteria use to overcome cell damage and to transition to a multicellular organism.},
}
@article{VaswanietalCoRR-17,
       author = {Ashish Vaswani and Noam Shazeer and Niki Parmar and Jakob Uszkoreit and Llion Jones and Aidan N. Gomez and Lukasz Kaiser and Illia Polosukhin},
        title = {Attention Is All You Need},
      journal = {CoRR},
       volume = {arXiv:1706.03762},
         year = {2017},
     abstract = {The dominant sequence transduction models are based on complex recurrent or convolutional neural networks in an encoder-decoder configuration. The best performing models also connect the encoder and decoder through an attention mechanism. We propose a new simple network architecture, the Transformer, based solely on attention mechanisms, dispensing with recurrence and convolutions entirely. Experiments on two machine translation tasks show these models to be superior in quality while being more parallelizable and requiring significantly less time to train. Our model achieves 28.4 BLEU on the WMT 2014 English-to-German translation task, improving over the existing best results, including ensembles by over 2 BLEU. On the WMT 2014 English-to-French translation task, our model establishes a new single-model state-of-the-art BLEU score of 41.8 after training for 3.5 days on eight GPUs, a small fraction of the training costs of the best models from the literature. We show that the Transformer generalizes well to other tasks by applying it successfully to English constituency parsing both with large and limited training data.}
}
@article{VaughnetalJCN-77,
       author = {Vaughn, James E. and Sims, Terry and Nakashima, Mariko},
        title = {A comparison of the early development of axodendritic and axosomatic synapses upon embryonic mouse spinal motor neurons},
      journal = {The Journal of Comparative Neurology},
    publisher = {The Wistar Institute of Anatomy and Biology},
       volume = {175},
        issue = {1},
         year = {1977},
        pages = {79-100},
     abstract = {The early embryonic development of axodendritic and axosomatic synapses upon motor neurons has been investigated in mouse spinal cord. This investigation had two main objectives: (a) to determine the earliest embryonic day on which recognizable synaptic contacts occur upon motor neurons, and (b) to compare the development of synaptic contacts upon the dendrites and somata of motor neurons during the early synaptogenic period. The broad goal of this investigation was to determine whether the location of early-forming synapses upon motor neurons is consistent with the possibility that axosomatic synapses might be involved in the primary induction of motor neuronal dendrogenesis. Embryonic day 11 (E11) was the earliest developmental time at which synaptic contacts were observed in developing mouse spinal cord. All of these synaptic contacts appeared to be located upon motor neuronal dendrites within the lateral and ventral marginal zones. The number of synaptic contacts observed on E11 was too small to be detected by the quantitative sampling procedure used in this investigation, but the procedure was sufficiently sensitive to detect synaptic contacts on embryonic day 12. Both axodendritic and axosomatic synapses were found upon E12 motor neurons, but there were about four times as many synaptic contacts per unit length of dendritic membrane as there were per equivalent length of somal membrane. Furthermore, dendritic membranes continued to exhibit a higher density of synaptic contacts on all of the remaining embryonic days (i.e., 13-16) examined in this investigation. These quantitative data and the E11 observations indicate that axosomatic synaptic contacts are not a necessary prelude to the formation of motor neuronal dendrites. Therefore, it is suggested that axosomatic synapses do not play an obligatory role in the primary induction of motor neuronal dendrogenesis. The experimental findings of other investigators, however, have provided reasons to suspect that early-forming axosomatic synapses may somehow facilitate dendritic development once it has been induced. This possibility is discussed in terms of our observation that early-forming axosomatic synapses rather commonly occur at sites which may represent somal growth regions. This relationship leads us to suggest that early axosomatic synapses may facilitate dendritic development by signalling the motor somata that the formation of a synaptogenic axonal field is underway. Furthermore, we speculate that the positioning of early axosomatic contacts might be providing directive cues as to the location of the developing synaptogenic field. Thus a directive facilitation of dendritic growth is suggested as a function of early axosomatic synapses rather than one involved with the primary induction of dendrogenesis.},
}
@misc{VedaldiandFulkersonVLFEAT-08,
     author = {Andrea Vedaldi and Brian Fulkerson},
      title = {{VLFeat}: An Open and Portable Library of Computer Vision Algorithms},
      year  = 2008,
howpublished = {\url{http://www.vlfeat.org/}}
}
@inproceedings{VedaldietalICCV-09,
     title = {Multiple kernels for object detection},
    author = {Andrea Vedaldi and Varun Gulshan and Manik Varma and Andrew Zisserman},
 booktitle = {Proceedings of {IEEE} International Conference on Computer Vision},
 publisher = {IEEE},
      year = 2009,
     pages = {606-613},
}
@article{VedantametalCoRR-17,
       author = {Ramakrishna Vedantam and Samy Bengio and Kevin Murphy and Devi Parikh and Gal Chechik},
        title = {Context-aware Captions from Context-agnostic Supervision},
      journal = {CoRR},
       volume = {abs/1701.02870},
         year = {2017},
     abstract = {We introduce an inference technique to produce discriminative context-aware image captions (captions that describe differences between images or visual concepts) using only generic context-agnostic training data (captions that describe a concept or an image in isolation). For example, given images and captions of "siamese cat" and "tiger cat", we generate language that describes the "siamese cat" in a way that distinguishes it from "tiger cat". Our key novelty is that we show how to do joint inference over a language model that is context-agnostic and a listener which distinguishes closely-related concepts. We first apply our technique to a justification task, namely to describe why an image contains a particular fine-grained category as opposed to another closely-related category of the CUB-200-2011 dataset. We then study discriminative image captioning to generate language that uniquely refers to one of two semantically-similar images in the COCO dataset. Evaluations with discriminative ground truth for justification and human studies for discriminative image captioning reveal that our approach outperforms baseline generative and speaker-listener approaches for discrimination.},
}
@inproceedings{VelikovichetalACL-10,
        title = {The viability of web-derived polarity lexicons},
       author = {Velikovich, Leonid and Blair-Goldensohn, Sasha and Hannan, Kerry and McDonald, Ryan},
    booktitle = {Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics},
    publisher = {Association for Computational Linguistics},
     location = {Los Angeles, California},
         year = 2010,
        pages = {777-785},
}
@article{VentaetalNANO-13,
        title = {Differentiation of Short, Single-Stranded DNA Homopolymers in Solid-State Nanopores},
       author = {Venta, Kimberly and Shemer, Gabriel and Puster, Matthew and Rodriguez-Manzo, Julio A. and Balan, Adrian and Rosenstein, Jacob K. and Shepard, Ken and Drndic, Marija},
      journal = {ACS Nano},
       volume = {7},
        issue = {5},
        pages = {4629-4636},
         year = {2013},
     abstract = {In the last two decades, new techniques that monitor ionic current modulations as single molecules pass through a nanoscale pore have enabled numerous single-molecule studies. While biological nanopores have recently shown the ability to resolve single nucleotides within individual DNA molecules, similar developments with solid-state nanopores have lagged, due to challenges both in fabricating stable nanopores of similar dimensions as biological nanopores and in achieving sufficiently low-noise and high-bandwidth recordings. Here we show that small silicon nitride nanopores (0.8- to 2-nm diameter in 5- to 8-nm-thick membranes) can resolve differences between ionic current signals produced by short (30 base) ssDNA homopolymers (poly(dA), poly(dC), poly(dT)), when combined with measurement electronics that allow a signal-to-noise ratio of better than 10 to be achieved at 1-MHz bandwidth. While identifying intramolecular DNA sequences with silicon nitride nanopores will require further improvements in nanopore sensitivity and noise levels, homopolymer differentiation represents an important milestone in the development of solid-state nanopores.}
}
@book{Venter2013,
        title = {Life at the Speed of Light: From the Double Helix to the Dawn of Digital Life},
       author = {Venter, J.C.},
    publisher = {Little, Brown Book Group Limited},
         year = {2013},
}
@article{VentolaPT-12a,
      title = {The Nanomedicine Revolution, {Part 1}: Emerging Concepts},
     author = {C. Lee Ventola},
    journal = {Pharmacy and Therapeutics},
     volume = 37,
      issue = 9,
       year = 2012,
      pages = {512-525},
}
@article{VentolaPT-12b,
      title = {The Nanomedicine Revolution, {Part 2}: Current and Future Clinical Applications},
     author = {C. Lee Ventola},
    journal = {Pharmacy and Therapeutics},
     volume = 37,
      issue = 10,
       year = 2012,
      pages = {582-591},
}
@misc{Verbeeketal01,
   author = {J. Verbeek and N. Vlassis and B. Kr{\"o}se},
    title = {Greedy gaussian mixture learning for texture segmentation},
     text = {In Proceedings the of Workshop on Kernel and subspace methods for computer vision at International Conference on Artificial Neural Networks},
      url = {http://citeseer.ist.psu.edu/verbeek01greedy.html},
     year = 2001,
}
@article{VerbeeketalNC-03,
        title = {Efficient Greedy Learning of Gaussian Mixtures},
       author = {J.J. Verbeek and N. Vlassis and B. Kr{\"o}se},
      journal = {Neural Computation},
    publisher = {MIT Press},
       volume = 15,
        issue = 2,
         year = 2003,
        pages = {469-485},
}
@inproceedings{VereIJCAI-85a,
       author = {Vere, Steven},
        title = {Splicing Plans to Achieve Misordered Goals},
    booktitle = {Proceedings {IJCAI} 9},
     location = {Los Angeles, California},
 organization = {IJCAII},
         year = 1985,
        pages = {1016-1021},     
}
@inproceedings{VereIJCAI-85b,
       author = {Vere, Steven},
        title = {Temporal Scope of Assertions and Window Cutoff},
    booktitle = {Proceedings {IJCAI} 9},
     location = {Los Angeles, California},
 organization = {IJCAII},
         year = 1985,
        pages = {1055-1059},
}
@article{VerePAMI-83,
       author = {Vere, Steven},
        title = {Planning in Time: Windows and Durations for Activities and Goals},
      journal = PAMI,
       volume = 5,
      comment = {DEVISER},
         year = 1983,
        pages = {246-267},
}
@book{Veres2008,
        title = {{Natural Language Programming of Agents and Robotic Devices: Publishing for agents and humans in sEnglish}},
       author = {S\'{a}ndor M. Veres},
    publisher = {SysBrain Ltd},
      address = {London},
         year = {2008},
     abstract = {This paper is an application of ontology theory, conceptual graphs and programming language theories to develop the theoretical foundations of natural language programming (NLP) that has in recent years been used to produce natural language documents for intelligent agents and human readers. The analysis given reveals three benefits of NLP. First, it is "conceptualized" programming that enables developers to write less bug prone programs due to clarity of code presentation and enforced structuring of data. Secondly, NLP can aid programming of the all important abstractions for robots: event, action and world model abstractions can be created by sentences. Thirdly, NLP can be used to publish natural language documents by researchers, i.e. English language documents on control theory and procedures, on the Internet or in printed documents. This theoretical paper also defines a large class of intelligent agents that can read such documents. This enables human users and agents to have shared understanding of how application systems work.}
}
@article{VerhoefetalNEURON-12,
        title = {Inferotemporal Cortex Subserves Three-Dimensional Structure Categorization},
       author = {Bram-Ernst Verhf and Rufin Vogels and Peter Janssen},
      journal = {Neuron},
       volume = 73,
         year = 2012,
        pages = {171-182}
}
@article{VermaetalCoRR-18,
       author = {Programmatically Interpretable Reinforcement Learning},
        title = {Abhinav Verma and Vijayaraghavan Murali and Rishabh Singh and Pushmeet Kohli and Swarat Chaudhuri},
      journal = {CoRR},
       volume = {arXiv:1804.02477},
         year = {2018},
     abstract = {We study the problem of generating interpretable and verifiable policies through reinforcement learning. Unlike the popular Deep Reinforcement Learning (DRL) paradigm, in which the policy is represented by a neural network, the aim in Programmatically Interpretable Reinforcement Learning is to find a policy that can be represented in a high-level programming language. Such programmatic policies have the benefits of being more easily interpreted than neural networks, and being amenable to verification by symbolic methods. We propose a new method, called Neurally Directed Program Search (NDPS), for solving the challenging nonsmooth optimization problem of finding a programmatic policy with maxima reward. NDPS works by first learning a neural policy network using DRL, and then performing a local search over programmatic policies that seeks to minimize a distance from this neural "oracle". We evaluate NDPS on the task of learning to drive a simulated car in the TORCS car-racing environment. We demonstrate that NDPS is able to discover human-readable policies that pass some significant performance bars. We also find that a well-designed policy language can serve as a regularizer, and result in the discovery of policies that lead to smoother trajectories and are more easily transferred to environments not encountered during training.}
}
@inproceedings{VermaetalICML-18,
       author = {Programmatically Interpretable Reinforcement Learning},
        title = {Abhinav Verma and Vijayaraghavan Murali and Rishabh Singh and Pushmeet Kohli and Swarat Chaudhuri},
    booktitle = {Proceedings of the 35th International Conference on Machine Learning},
         year = {2018},
}
@article{Verschureetal03,
       author = {Verschure, P.F. and T. Voegtlin and R.J. Douglas},
        title = {Environmentally mediated synergy between perception and behaviour in mobile robots},
      journal = {Nature},
       volume = 425,
        issue = 6958,
        pages = {620-624},
         year = 2003,
}
@misc{VictorandKnight02,
       author = {Jonathan D. Victor and Bruce W. Knight},
        title = {Simultaneously Band and Space Limited Functions in Two Dimensions, and Receptive Fields of Visual Neurons},
 howpublished = {http://www-users.med.cornell.edu/~jdvicto/labonweb.html},
         year = 2002,
      comment = {Recommended by Alan Stewart},
          url = {VictorandKnight02.pdf},
}
@inproceedings{Vidal-NaquetandUllman,
       author = {Michel Vidal-Naquet and Shimon Ullman},
        title = {Object Recognition with Informative Features and Linear Classification},
    booktitle = {ICCV},
        pages = {281-288},
         year = 2003,
}
@inproceedings{ViegasetalCHI-06,
        title = {Visualizing email content: portraying relationships from conversational histories},
       author = {Fernanda B. Vi{\'e}gas and Scott A. Golder and Judith S. Donath},
       editor = {Rebecca E. Grinter and Tom Rodden and Paul M. Aoki and Edward Cutrell and Robin Jeffries and Gary M. Olson},
    booktitle = {Proceedings of the 2006 Conference on Human Factors in Computing Systems},
     location = {Montr{\'e}al, Qu{\'e}bec, Canada},
         year = 2006,
        pages = {979-988},
}
@inproceedings{VijayanarasimhanandKapoorCVPR-10,
       author = {Sudheendra Vijayanarasimhan and Ashish Kapoor},
        title = {Visual recognition and detection under bounded computational resources},
    booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE},
         year = {2010},
        pages = {1006-1013},
}
@inproceedings{VilainAAAI-82,
       author = {Vilain, Marc},
        title = {A System for Reasoning About Time},
    booktitle = {Proceedings AAAI-82},
     location = {Pittsburgh, Pennsylvania},
 organization = {AAAI},
         year = 1982,
}
@inproceedings{VilainAAAI-86,
       author = {Vilain, Marc and Kautz, Henry},
        title = {Constraint Propagation Algorithms for Temporal Reasoning},
    booktitle = {Proceedings AAAI-86},
     location = {Philadelphia, Pennsylvania},
 organization = {AAAI},
         year = 1986,
        pages = {377-382},
}
@book{Villani2015,
        title = {Birth of a Theorem: A Mathematical Adventure},
       author = {Villani, C\`{e}dric and Mouhot, Cl\`{e}ment},
         year = {2015},
    publisher = {Farrar, Straus and Giroux}
}
@article{VillarrubiaetalNEUROCOMPUTING-18,
        title = {Artificial neural networks used in optimization problems},
      journal = {Neurocomputing},
       volume = {272},
        pages = {10-16},
         year = {2018},
       author = {Gabriel Villarrubia and Juan F. De Paz and Pablo Chamoso and Fernando De la Prieta},
     abstract = {Optimization problems often require the use of optimization methods that permit the minimization or maximization of certain objective functions. Occasionally, the problems that must be optimized are not linear or polynomial; they cannot be precisely resolved, and they must be approximated. In these cases, it is necessary to apply heuristics, which are able to resolve these kinds of problems. Some algorithms linearize the restrictions and objective functions at a specific point of the space by applying derivatives and partial derivatives for some cases, while in other cases evolutionary algorithms are used to approximate the solution. This work proposes the use of artificial neural networks to approximate the objective function in optimization problems to make it possible to apply other techniques to resolve the problem. The objective function is approximated by a non-linear regression that can be used to resolve an optimization problem. The derivate of the new objective function should be polynomial so that the solution of the optimization problem can be calculated.}
}
@inproceedings{VincentetalICML-08,
       author = {Pascal Vincent and Hugo Larochelle and Yoshua Bengio and Pierre-Antoine Manzagol},
        title = {Extracting and Composing Robust Features with Denoising Autoencoders},
    booktitle = {Proceedings of the 25th International Conference on Machine learning},
     location = {Helsinki, Finland},
       editor = {Andrew McCallum and Sam Roweis},
    publisher = {Omnipress},
        pages = {1096-1103},
         year = 2008,
}
@inproceedings{VinyalsandLeICML-15,
        title = {A Neural Conversational Model},
       author = {Oriol Vinyals and Quoc V. Le},
         year = 2015,
          url = {http://arxiv.org/pdf/1506.05869v3.pdf},
    booktitle = {ICML Deep Learning Workshop},
     abstract = {Conversational modeling is an important task in natural language understanding and machine intelligence. Although previous approaches exist, they are often restricted to specific domains (e.g., booking an airline ticket) and require handcrafted rules. In this paper, we present a simple approach for this task which uses the recently proposed sequence to sequence framework. Our model converses by predicting the next sentence given the previous sentence or sentences in a conversation. The strength of our model is that it can be trained end-to-end and thus requires much fewer hand-crafted rules. We find that this straightforward model can generate simple conversations given a large conversational training dataset. Our preliminary results suggest that, despite optimizing the wrong objective function, the model is able to converse well. It is able extract knowledge from both a domain specific dataset, and from a large, noisy, and general domain dataset of movie subtitles. On a domainspecific IT helpdesk dataset, the model can find a solution to a technical problem via conversations. On a noisy open-domain movie transcript dataset, the model can perform simple forms of common sense reasoning. As expected, we also find that the lack of consistency is a common failure mode of our model.}
}
@article{VinyalsetalCoRR-15,
       author = {Oriol Vinyals and Lukasz Kaiser and Terry Koo and Slav Petrov and Ilya Sutskever and Geoffrey E. Hinton},
        title = {Grammar as a Foreign Language},
      journal = {CoRR},
       volume = {arXiv:1412.7449},
          url = {http://arxiv.org/abs/1412.7449},
         year = 2014,
     abstract = {Syntactic parsing is a fundamental problem in computational linguistics and natural language processing. Traditional approaches to parsing are highly complex and problem specific. Recently, Sutskever et al. (2014) presented a task-agnostic method for learning to map input sequences to output sequences that achieved strong results on a large scale machine translation problem. In this work, we show that precisely the same sequence-to-sequence method achieves results that are close to state-of-the-art on syntactic constituency parsing, whilst making almost no assumptions about the structure of the problem. To achieve these results we need to mitigate the lack of domain knowledge in the model by providing it with a large amount of automatically parsed data.}
}
@article{VinyalsetalCoRR-15b,
       author = {Oriol Vinyals and Alexander Toshev and Samy Bengio and Dumitru Erhan},
        title = {Show and Tell: {A} Neural Image Caption Generator},
      journal = {CoRR},
       volume = {arXiv:1411.4555v2},
         year = 2015,
     abstract = {Automatically describing the content of an image is a fundamental problem in artificial intelligence that connects computer vision and natural language processing. In this paper, we present a generative model based on a deep recurrent architecture that combines recent advances in computer vision and machine translation and that can be used to generate natural sentences describing an image. The model is trained to maximize the likelihood of the target description sentence given the training image. Experiments on several datasets show the accuracy of the model and the fluency of the language it learns solely from image descriptions. Our model is often quite accurate, which we verify both qualitatively and quantitatively. For instance, while the current state-of-the-art BLEU-1 score (the higher the better) on the Pascal dataset is 25, our approach yields 59, to be compared to human performance around 69. We also show BLEU-1 score improvements on Flickr30k, from 56 to 66, and on SBU, from 19 to 28. Lastly, on the newly released COCO dataset, we achieve a BLEU-4 of 27.7, which is the current state-of-the-art.}
}
@inproceedings{VinyalsetalICLR-16,
       author = {Oriol Vinyals and Samy Bengio and Manjunath Kudlur},
        title = {Order matters: {S}equence to sequence for sets},
    booktitle = {International Conference on Learning Representations},
         year = {2016},
     abstract = {Sequences have become first class citizens in supervised learning thanks to the resurgence of recurrent neural networks. Many complex tasks that require mapping from or to a sequence of observations can now be formulated with the sequence-to-sequence (seq2seq) framework which employs the chain rule to efficiently represent the joint probability of sequences. In many cases, however, variable sized inputs and/or outputs might not be naturally expressed as sequences. For instance, it is not clear how to input a set of numbers into a model where the task is to sort them; similarly, we do not know how to organize outputs when they correspond to random variables and the task is to model their unknown joint probability. In this paper, we first show using various examples that the order in which we organize input and/or output data matters significantly when learning an underlying model. We then discuss an extension of the seq2seq framework that goes beyond sequences and handles input sets in a principled way. In addition, we propose a loss which, by searching over possible orders during training, deals with the lack of structure of output sets. We show empirical evidence of our claims regarding ordering, and on the modifications to the seq2seq framework on benchmark language modeling and parsing tasks, as well as two artificial tasks -- sorting numbers and estimating the joint probability of unknown graphical models.}
}
@article{VinyalsetalNATURE-19,
       author = {Vinyals, Oriol and Babuschkin, Igor and Czarnecki, Wojciech M. and Mathieu, Michal and Dudzik, Andrew and Chung, Junyoung and Choi, David H. and Powell, Richard and Ewalds, Timo and Georgiev, Petko and Oh, Junhyuk and Horgan, Dan and Kroiss, Manuel and Danihelka, Ivo and Huang, Aja and Sifre, Laurent and Cai, Trevor and Agapiou, John P. and Jaderberg, Max and Vezhnevets, Alexander S. and Leblond, R{\'{e}}mi and Pohlen, Tobias and Dalibard, Valentin and Budden, David and Sulsky, Yury and Molloy, James and Paine, Tom L. and Gulcehre, Caglar and Wang, Ziyu and Pfaff, Tobias and Wu, Yuhuai and Ring, Roman and Yogatama, Dani and Wunsch, Dario and McKinney, Katrina and Smith, Oliver and Schaul, Tom and Lillicrap, Timothy and Kavukcuoglu, Koray and Hassabis, Demis and Apps, Chris and Silver, David},
        title = {Grandmaster level in StarCraft II using multi-agent reinforcement learning},
      journal = {Nature},
         year = {2019},
       volume = {575},
        issue = {7782},
        pages = {350-354},
     abstract = {Many real-world applications require artificial agents to compete and coordinate with other agents in complex environments. As a stepping stone to this goal, the domain of StarCraft has emerged as an important challenge for artificial intelligence research, owing to its iconic and enduring status among the most difficult professional esports and its relevance to the real world in terms of its raw complexity and multi-agent challenges. Over the course of a decade and numerous competitions, the strongest agents have simplified important aspects of the game, utilized superhuman capabilities, or employed hand-crafted sub-systems4. Despite these advantages, no previous agent has come close to matching the overall skill of top StarCraft players. We chose to address the challenge of StarCraft using general-purpose learning methods that are in principle applicable to other complex domains: a multi-agent reinforcement learning algorithm that uses data from both human and agent games within a diverse league of continually adapting strategies and counter-strategies, each represented by deep neural networks5,6. We evaluated our agent, AlphaStar, in the full game of StarCraft II, through a series of online games against human players. AlphaStar was rated at Grandmaster level for all three StarCraft races and above 99.8\% of officially ranked human players.},
}
@incollection{VinyalsetalNIPS-15,
        title = {Pointer Networks},
       author = {Vinyals, Oriol and Fortunato, Meire and Jaitly, Navdeep},
    booktitle = {Advances in Neural Information Processing Systems 28},
       editor = {C. Cortes and N. D. Lawrence and D. D. Lee and M. Sugiyama and R. Garnett},
    publisher = {Curran Associates, Inc.},
         year = {2015},
        pages = {2692-2700},
     abstract = {We introduce a new neural architecture to learn the conditional probability of an output sequence with elements that are discrete tokens corresponding to positions in an input sequence. Such problems cannot be trivially addressed by existent approaches such as sequence-to-sequence and Neural Turing Machines, because the number of target classes in each step of the output depends on the length of the input, which is variable. Problems such as sorting variable sized sequences, and various combinatorial optimization problems belong to this class. Our model solves the problem of variable size output dictionaries using a recently proposedmechanism of neural attention. It differs from the previous attentionattempts in that, instead of using attention to blend hidden units of anencoder to a context vector at each decoder step, it uses attention asa pointer to select a member of the input sequence as the output. We call this architecture a Pointer Net (Ptr-Net). We show Ptr-Nets can be used to learn approximate solutions to three challenging geometric problems -- finding planar convex hulls, computing Delaunay triangulations, and the planar Travelling Salesman Problem-- using training examples alone. Ptr-Nets not only improve over sequence-to-sequence with input attention, but also allow us to generalize to variable size output dictionaries. We show that the learnt models generalize beyond the maximum lengths they were trained on. We hope our results on these tasks will encourage a broader exploration of neural learning for discrete problems.}
}
@inproceedings{VinyalsetalNIPS-16,
       author = {Oriol Vinyals and Charles Blundell and Timothy P. Lillicrap and Koray Kavukcuoglu and Daan Wierstra},
        title = {Matching networks for one shot learning},
    booktitle = {Advances in Neural Information Processing Systems},
    publisher = {Curran Associates, Inc.},
         year = {2017},
        pages = {3630-3638},
     abstract = {Learning from a few examples remains a key challenge in machine learning. Despite recent advances in important domains such as vision and language, the standard supervised deep learning paradigm does not offer a satisfactory solution for learning new concepts rapidly from little data. In this work, we employ ideas from metric learning based on deep neural features and from recent advances that augment neural networks with external memories. Our framework learns a network that maps a small labelled support set and an unlabelled example to its label, obviating the need for fine-tuning to adapt to new class types. We then define one-shot learning problems on vision (using Omniglot, ImageNet) and language tasks. Our algorithm improves one-shot accuracy on ImageNet from 87.6\% to 93.2\% and from 88.0\% to 93.8\% on Omniglot compared to competing approaches. We also demonstrate the usefulness of the same model on language modeling by introducing a one-shot task on the Penn Treebank.}
}
@article{ViolaandJonesIJCV-01,
       author = {Paul Viola and Michael Jones},
        title = {Robust Real-time Object Detection},
      journal = {International Journal of Computer Vision},
    publisher = {Kluwer Academic Publishers},
       volume = 57,
        issue = 2,
        pages = {137-574},
         year = 2002,
}
@article{VitterTMS-85,
       author = {Vitter, Jeffrey S.},
        title = {Random Sampling with a Reservoir},
      journal = {ACM Transactions on Mathematical Software},
    publisher = {ACM},
       volume = {11},
       number = {1},
         year = {1985},
        pages = {37-57},
     abstract = {We introduce fast algorithms for selecting a random sample of n records without replacement from a pool of N records, where the value of N is unknown beforehand. The main result of the paper is the design and analysis of Algorithm Z; it does the sampling in one pass using constant space and in O(n(1 + log(N/n))) expected time, which is optimum, up to a constant factor. Several optimizations are studied that collectively improve the speed of the naive version of the algorithm by an order of magnitude. We give an efficient Pascal-like implementation that incorporates these modifications and that is suitable for general use. Theoretical and empirical results indicate that Algorithm Z outperforms current methods by a significant margin.}
}
@article{VlijmenetalCOI-11,
       author = {van Vlijmen, O.J. and Rangel, F.A. and Berge, S.J. and Bronkhorst, E.M. and Becking, A.G. and Kuijpers-Jagtman A.M.},
        title = {Measurements on {3D} models of human skulls derived from two different cone beam {CT} scanners},
      journal = {Clinical Oral Investigation},
       volume = 15,
        issue = 5,
         year = 2011,
        pages = {721-727},
}
@article{Vo-DinhandZhangNANOBIO-11,
      title = {Single-cell monitoring using fiberoptic nanosensors},
     author = {Vo-Dinh, Tuan and Zhang, Yan},
    journal = {Wiley Interdisciplinary Reviews: Nanomedicine and Nanobiotechnology},
     volume = 3,
      issue = 1,
  publisher = {John Wiley \& Sons, Inc.},
       year = 2011,
      pages = {79-85},
}
@article{VogelsteinetalBJ-09,
    author = {Vogelstein, Joshua T. and Watson, Brendon O. and Packer, Adam M. and Yuste, Rafael and Jedynak, Bruno and Paninski, Liam},
     title = {Spike Inference from Calcium Imaging Using Sequential Monte Carlo Methods},
   journal = {Biophysical Journal},
 publisher = {Cell Press},
    volume = 97,
     issue = 2,
      year = 2009,
     pages = {636-655},
  abstract = {As recent advances in calcium sensing technologies facilitate simultaneously imaging action potentials in neuronal populations, complementary analytical tools must also be developed to maximize the utility of this experimental paradigm. Although the observations here are fluorescence movies, the signals of interest spike trains and/or time varying intracellular calcium concentrations are hidden. Inferring these hidden signals is often problematic due to noise, nonlinearities, slow imaging rate, and unknown biophysical parameters. We overcome these difficulties by developing sequential Monte Carlo methods (particle filters) based on biophysical models of spiking, calcium dynamics, and fluorescence. We show that even in simple cases, the particle filters outperform the optimal linear (i.e., Wiener) filter, both by obtaining better estimates and by providing error bars. We then relax a number of our model assumptions to incorporate nonlinear saturation of the fluorescence signal, as well external stimulus and spike history dependence (e.g., refractoriness) of the spike trains. Using both simulations and in vitro fluorescence observations, we demonstrate temporal superresolution by inferring when within a frame each spike occurs. Furthermore, the model parameters may be estimated using expectation maximization with only a very limited amount of data (e.g., < $5^{10}$ s or $5^{40}$ spikes), without the requirement of any simultaneous electrophysiology or imaging experiments.}
}
@article{VogelsteinetalSCIENCE-14,
       author = {Vogelstein, Joshua T. and Park, Youngser and Ohyama, Tomoko and Kerr, Rex A. and Truman, James W. and Priebe, Carey E. and Zlatic, Marta},
        title = {Discovery of Brainwide Neural-Behavioral Maps via Multiscale Unsupervised Structure Learning},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {344},
       number = {6182},
         year = {2014},
        pages = {386-392},
     abstract = {A single nervous system can generate many distinct motor patterns. Identifying which neurons and circuits control which behaviors has been a laborious piecemeal process, usually for one observer-defined behavior at a time. We present a fundamentally different approach to neuron-behavior mapping. We optogenetically activated 1054 identified neuron lines in Drosophila larvae and tracked the behavioral responses from 37,780 animals. Application of multiscale unsupervised structure learning methods to the behavioral data enabled us to identify 29 discrete, statistically distinguishable, observer-unbiased behavioral phenotypes. Mapping the neural lines to the behavior(s) they evoke provides a behavioral reference atlas for neuron subsets covering a large fraction of larval neurons. This atlas is a starting point for connectivity- and activity-mapping studies to further investigate the mechanisms by which neurons mediate diverse behaviors.},
}
@article{VogesetalPiN-10,
        title = {A modeler's view on the spatial structure of intrinsic horizontal connectivity in the neocortex},
       author = {Nicole Voges and Almut Sch{\"u}z and Ad Aertsen and Stefan Rotter},
      journal = {Progress in Neurobiology},
         year = {2010},
       volume = {92},
        pages = {277-292},
     abstract = {Most current computational models of neocortical networks assume a homogeneous and isotropic arrangement of local synaptic couplings between neurons. Sparse, recurrent connectivity is typically implemented with simple statistical wiring rules. For spatially extended networks, however, such random graph models are inadequate because they ignore the traits of neuron geometry, most notably various distance dependent features of horizontal connectivity. It is to be expected that such non-random structural attributes have a great impact, both on the spatio-temporal activity dynamics and on the biological function of neocortical networks. Here we review the neuroanatomical literature describing long-range horizontal connectivity in the neocortex over distances of up to eight millimeters, in various cortical areas and mammalian species. We extract the main common features from these data to allow for improved models of large-scale cortical networks. Such models include, next to short-range neighborhood coupling, also long-range patchy connections. We show that despite the large variability in published neuroanatomical data it is reasonable to design a generic model which generalizes over different cortical areas and mammalian species. Later on, we critically discuss this generalization, and we describe some examples of how to specify the model in order to adapt it to specific properties of particular cortical areas or species.}
}
@article{VolkmannetalNATURE_REVIEWS_NEUROLOGY-10,
       author = {Volkmann, J. and Daniels, C. and Witt, K.},
        title = {{N}europsychiatric effects of subthalamic neurostimulation in {P}arkinson disease},
      journal = {Nature Reviews Neurology},
         year = {2010},
       volume = {6},
       number = {9},
        pages = {487-498},
     abstract = {Neurostimulation of the subthalamic nucleus (STN) is an established treatment for motor symptoms in advanced Parkinson disease (PD), although concerns exist regarding the safety of this therapy in terms of cognitive and psychiatric adverse effects. The basal ganglia are considered to be part of distributed cortico-subcortical networks that are involved in the selection, facilitation and inhibition of movements, emotions, behaviors and thoughts. The STN has a central role in these networks, probably providing a global 'no-go' signal. The behavioral and cognitive effects observed following STN high-frequency stimulation (HFS) probably reflect the intrinsic role of this nucleus in nonmotor functional domains. Nevertheless, postoperative behavioral changes are seldom caused by such stimulation alone. PD is a progressive neurodegenerative disorder with motor, cognitive, behavioral and autonomic symptoms. The pattern of neurodegeneration and expression of these symptoms are highly variable across individuals. The preoperative neuropsychiatric state can be further complicated by sensitization phenomena resulting from long-term dopaminergic treatment, which include impulse control disorders, punding, and addictive behaviors (dopamine dysregulation syndrome). Finally, personality traits, the social environment, culture and learned behaviors might be important determinants explaining why behavioral symptoms differ between patients after surgery. Here, we summarize the neuropsychiatric changes observed after STN HFS and try to disentangle their various etiologies.}
}
@inproceedings{VolkovaetalACL-13,
    author = {Svitlana Volkova and Pallavi Choudhury and Chris Quirk and Bill Dolan and Luke Zettlemoyer},
     title = {Lightly Supervised Learning of Procedural Dialog Systems},
 booktitle = {Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics},
 publisher = {Association for Computational Linguistics},
   address = {Stroudsburg, PA, USA},
      year = 2013,
}
@article{VollimetalNEUROIMAGE-06,
       author = {Birgit A. V\"{o}llm and Alexander N.W. Taylor and Paul Richardson and Rhiannon Corcoran and John Stirling and Shane McKie and John F.W. Deakin and Rebecca Elliott},
        title = {Neuronal correlates of theory of mind and empathy: A functional magnetic resonance imaging study in a nonverbal task},
      journal = {NeuroImage},
       volume = 29,
        issue = 1,
        pages = {90-98},
         year = 2006,
     abstract = {Theory of Mind (ToM), the ability to attribute mental states to others, and empathy, the ability to infer emotional experiences, are important processes in social cognition. Brain imaging studies in healthy subjects have described a brain system involving medial prefrontal cortex, superior temporal sulcus and temporal pole in ToM processing. Studies investigating networks associated with empathic responding also suggest involvement of temporal and frontal lobe regions. In this fMRI study, we used a cartoon task derived from Sarfati et al. (1997) [Sarfati, Y., Hardy-Bayle, M.C., Besche, C., Widlocher, D. 1997. Attribution of intentions to others in people with schizophrenia: a non-verbal exploration with comic strips. Schizophrenia Research 25, 199-209.]with both ToM and empathy stimuli in order to allow comparison of brain activations in these two processes. Results of 13 right-handed, healthy, male volunteers were included. Functional images were acquired using a 1.5 T Phillips Gyroscan. Our results confirmed that ToM and empathy stimuli are associated with overlapping but distinct neuronal networks. Common areas of activation included the medial prefrontal cortex, temporoparietal junction and temporal poles. Compared to the empathy condition, ToM stimuli revealed increased activations in lateral orbitofrontal cortex, middle frontal gyrus, cuneus and superior temporal gyrus. Empathy, on the other hand, was associated with enhanced activations of paracingulate, anterior and posterior cingulate and amygdala. We therefore suggest that ToM and empathy both rely on networks associated with making inferences about mental states of others. However, empathic responding requires the additional recruitment of networks involved in emotional processing. These results have implications for our understanding of disorders characterized by impairments of social cognition, such as autism and psychopathy.}
}
@article{VondricketalARXIV-12,
        title = {Inverting and Visualizing Features for Object Detection},
       author = {Carl Vondrick and Aditya Khosla and Tomasz Malisiewicz and Antonio Torralba},
      journal = {CoRR},
       volume = {arXiv:1212.2278},
         year = 2012
}
@techreport{VrbskyetalTR-90,
       author = {Vrbsky, S. V. and Liu, J. W. S. and Smith, K. P.},
        title = {An Object-Oriented Query Processor that Returns Monotonically Improving Approximate Answers},
        issue = {UIUCDCS-R-90-1568},
  institution = {Department of Computer Science, {UIUC}},
         year = 1990,
}
@book{Vygotsky1987cognition,
        title = {{The Collected Works of L.S. Vygotsky: Volume 1: Problems of General Psychology, Including the Volume Thinking and Speech}},
       author = {Rieber, R.W. and Carton, A.S. and Vygotski, L.S. and Minick, N. and Wollock, J. and Knox, J.E. and Bruner, J.S.},
       series = {Cognition and Language: A Series in Psycholinguistics},
         year = {1987},
    publisher = {Plenum Press}
}
@article{VyletaetalELIFE-16,
        title = {Plasticity-dependent, full detonation at hippocampal mossy fiber-{CA3} pyramidal neuron synapses},
       author = {Vyleta, Nicholas P and Borges-Merjane, Carolina and Jonas, Peter},
       editor = {Raman, Indira M},
      journal = {eLife},
    publisher = {eLife Sciences Publications, Ltd},
       volume = {5},
         year = {2016},
        pages = {e17977},
     abstract = {Mossy fiber synapses on CA3 pyramidal cells are 'conditional detonators' that reliably discharge postsynaptic targets. The 'conditional' nature implies that burst activity in dentate gyrus granule cells is required for detonation. Whether single unitary excitatory postsynaptic potentials (EPSPs) trigger spikes in CA3 neurons remains unknown. Mossy fiber synapses exhibit both pronounced short-term facilitation and uniquely large post-tetanic potentiation (PTP). We tested whether PTP could convert mossy fiber synapses from subdetonator into detonator mode, using a recently developed method to selectively and noninvasively stimulate individual presynaptic terminals in rat brain slices. Unitary EPSPs failed to initiate a spike in CA3 neurons under control conditions, but reliably discharged them after induction of presynaptic short-term plasticity. Remarkably, PTP switched mossy fiber synapses into full detonators for tens of seconds. Plasticity-dependent detonation may be critical for efficient coding, storage, and recall of information in the granule cell-CA3 cell network.},
}
@misc{WT10g,
       author = {{CSIRO}},
        title = {{TREC} Web Corpus: {WT10g}},
 howpublished = {\url{http://www.ted.cmis.csiro.au/TRECWeb/wt10g.html}},
         year = 2001,
}
@article{WacongneetalJN-12,
       author = {Wacongne, Catherine and Changeux, Jean-Pierre and Dehaene, Stanislas},
        title = {A Neuronal Model of Predictive Coding Accounting for the Mismatch Negativity},
      journal = {Journal of Neuroscience},
       volume = {32},
       number = {11},
         year = {2012},
        pages = {3665-3678},
    publisher = {Society for Neuroscience},
     abstract = {The mismatch negativity (MMN) is thought to index the activation of specialized neural networks for active prediction and deviance detection. However, a detailed neuronal model of the neurobiological mechanisms underlying the MMN is still lacking, and its computational foundations remain debated. We propose here a detailed neuronal model of auditory cortex, based on predictive coding, that accounts for the critical features of MMN. The model is entirely composed of spiking excitatory and inhibitory neurons interconnected in a layered cortical architecture with distinct input, predictive, and prediction error units. A spike-timing dependent learning rule, relying upon NMDA receptor synaptic transmission, allows the network to adjust its internal predictions and use a memory of the recent past inputs to anticipate on future stimuli based on transition statistics. We demonstrate that this simple architecture can account for the major empirical properties of the MMN. These include a frequency-dependent response to rare deviants, a response to unexpected repeats in alternating sequences (ABABAA{\textellipsis}), a lack of consideration of the global sequence context, a response to sound omission, and a sensitivity of the MMN to NMDA receptor antagonists. Novel predictions are presented, and a new magnetoencephalography experiment in healthy human subjects is presented that validates our key hypothesis: the MMN results from active cortical prediction rather than passive synaptic habituation.},
}
@article{WagneretalNATURE-17,
       author = {Wagner, Mark J. and Kim, Tony Hyun and Savall, Joan and Schnitzer, Mark J. and Luo, Liqun},
        title = {Cerebellar granule cells encode the expectation of reward},
      journal = {Nature},
         year = {2017},
    publisher = {Macmillan Publishers Limited},
       volume = {544},
        issue = {7648},
        pages = {96-100},
     abstract = {The human brain contains approximately 60 billion cerebellar granule cells, which outnumber all other brain neurons combined. Classical theories posit that a large, diverse population of granule cells allows for highly detailed representations of sensorimotor context, enabling downstream Purkinje cells to sense fine contextual changes. Although evidence suggests a role for the cerebellum in cognition, granule cells are known to encode only sensory and motor context. Here, using two-photon calcium imaging in behaving mice, we show that granule cells convey information about the expectation of reward. Mice initiated voluntary forelimb movements for delayed sugar-water reward. Some granule cells responded preferentially to reward or reward omission, whereas others selectively encoded reward anticipation. Reward responses were not restricted to forelimb movement, as a Pavlovian task evoked similar responses. Compared to predictable rewards, unexpected rewards elicited markedly different granule cell activity despite identical stimuli and licking responses. In both tasks, reward signals were widespread throughout multiple cerebellar lobules. Tracking the same granule cells over several days of learning revealed that cells with reward-anticipating responses emerged from those that responded at the start of learning to reward delivery, whereas reward-omission responses grew stronger as learning progressed. The discovery of predictive, non-sensorimotor encoding in granule cells is a major departure from the current understanding of these neurons and markedly enriches the contextual information available to postsynaptic Purkinje cells, with important implications for cognitive processing in the cerebellum.}
}
@article{WahlstrometalCoRR-14,
       author = {Niklas Wahlstr\"{o}m, Thomas B. Sch\"{o}n, Marc Peter Deisenroth},
        title = {Learning deep dynamical models from image pixels},
      journal = {CoRR},
       volume = {arXiv:1410.7550},
         year = {2014},
     abstract = {Modeling dynamical systems is important in many disciplines, e.g., control, robotics, or neurotechnology. Commonly the state of these systems is not directly observed, but only available through noisy and potentially high-dimensional observations. In these cases, system identification, i.e., finding the measurement mapping and the transition mapping (system dynamics) in latent space can be challenging. For linear system dynamics and measurement mappings efficient solutions for system identification are available. However, in practical applications, the linearity assumptions does not hold, requiring non-linear system identification techniques. If additionally the observations are high-dimensional (e.g., images), non-linear system identification is inherently hard. To address the problem of non-linear system identification from high-dimensional observations, we combine recent advances in deep learning and system identification. In particular, we jointly learn a low-dimensional embedding of the observation by means of deep auto-encoders and a predictive transition model in this low-dimensional space. We demonstrate that our model enables learning good predictive models of dynamical systems from pixel information only.}
}
@incollection{Wainwrightetal00,
        title = {Natural image statistics and divisive normalization: {Modeling} nonlinearity and adaptation in cortical neurons},
       author = {M. J. Wainwright and O. Schwartz and E. P. Simoncelli},
    booktitle = {Probabilistic Models of the Brain: {Perception} and Neural Function},
       editor = {R. Rao and B. Olshausen and M. Lewicki},
    publisher = {MIT Press},
        pages = {203-222},
         year = 2002,
          url = {http://www.cns.nyu.edu/pub/eero/wainwright00c.pdf}
}
@article{WakanaetalRADIOLOGY-04,
       author = {Wakana, S.  and Jiang, H.  and Nagae-Poetscher, L. M.  and van Zijl, P. C. and Mori, S.},
        title = {Fiber tract-based atlas of human white matter anatomy},
      journal = {Radiology},
         year = {2004},
       volume = {230},
       number = {1},
        pages = {77-87},
     abstract = {Two- and three-dimensional (3D) white matter atlases were created on the basis of high-spatial-resolution diffusion tensor magnetic resonance (MR) imaging and 3D tract reconstruction. The 3D trajectories of 17 prominent white matter tracts could be reconstructed and depicted. Tracts were superimposed on coregistered anatomic MR images to parcel the white matter. These parcellation maps were then compared with coregistered diffusion tensor imaging color maps to assign visible structures. The results showed (a). which anatomic structures can be identified on diffusion tensor images and (b). where these anatomic units are located at each section level and orientation. The atlas may prove useful for educational and clinical purposes.}
}
@book{Waldrop01,
       author = {M. Mitchell Waldrop},
        title = {The Dream Machine: J.\ C.\ R.\ Licklider and the Revolution That Made Computing Personal},
    publisher = {Viking},
      address = {New York, NY},
         year = 2001,
}
@article{WaldropNATURE-12,
        title = {Computer modelling: Brain in a box},
       author = {Waldrop, M.M.},
      journal = {Nature},
       volume = 482,
        issue = 7386,
        pages = {456-8},
         year = 2012,
}
@article{WalesARPC-18,
       author = {Wales, David J.},
        title = {Exploring Energy Landscapes},
      journal = {Annual Review of Physical Chemistry},
       volume = {69},
       number = {1},
         year = {2018},
        pages = {401-425},
     abstract = { Recent advances in the potential energy landscapes approach are highlighted, including both theoretical and computational contributions. Treating the high dimensionality of molecular and condensed matter systems of contemporary interest is important for understanding how emergent properties are encoded in the landscape and for calculating these properties while faithfully representing barriers between different morphologies. The pathways characterized in full dimensionality, which are used to construct kinetic transition networks, may prove useful in guiding such calculations. The energy landscape perspective has also produced new procedures for structure prediction and analysis of thermodynamic properties. Basin-hopping global optimization, with alternative acceptance criteria and generalizations to multiple metric spaces, has been used to treat systems ranging from biomolecules to nanoalloy clusters and condensed matter. This review also illustrates how all this methodology, developed in the context of chemical physics, can be transferred to landscapes defined by cost functions associated with machine learning. }
}
@book{Walker2017,
       author = {Matthew Walker},
        title = {Why We Sleep - Unlocking the Power of Sleep and Dreams},
    publisher = {Simon \& Schuster},
      address = {New York, NY},
         year = 2017
}
@article{WalkerandvanderHelmPSYCHOLOGICAL-BULLETIN-09,
       author = {Walker, Matthew P. and van der Helm, Els},
        title = {Overnight therapy?: {T}he role of sleep in emotional brain processing},
      journal = {Psychological Bulletin},
       volume = {135},
        issue = {5},
         year = {2009},
        pages = {731-748},
     abstract = {Cognitive neuroscience continues to build meaningful connections between affective behavior and human brain function. Within the biological sciences, a similar renaissance has taken place, focusing on the role of sleep in various neurocognitive processes and, most recently, on the interaction between sleep and emotional regulation. This review surveys an array of diverse findings across basic and clinical research domains, resulting in a convergent view of sleep-dependent emotional brain processing. On the basis of the unique neurobiology of sleep, the authors outline a model describing the overnight modulation of affective neural systems and the (re)processing of recent emotional experiences, both of which appear to redress the appropriate next-day reactivity of limbic and associated autonomic networks. Furthermore, a rapid eye movement (REM) sleep hypothesis of emotional-memory processing is proposed, the implications of which may provide brain-based insights into the association between sleep abnormalities and the initiation and maintenance of mood disturbances.}
}
@inproceedings{WalkeretalACL-97,
       author = {Marilyn A. Walker and Diane J. Litman and Candace A. Kamm and Alicia Abella},
        title = {{PARADISE:} {A} Framework for Evaluating Spoken Dialogue Agents},
    booktitle = {35th Annual Meeting of the Association for Computational Linguistics},
         year = 1997,
        pages = {271-280},
}
@article{WalkeretalCBR-02,
       author = {Walker, M. P. and Liston, C. and Hobson, J. A. and Stickgold, R.},
        title = {{C}ognitive flexibility across the sleep-wake cycle: {REM}-sleep enhancement of anagram problem solving},
      journal = {Cognitive Brain Research},
         year = {2002},
       volume = {14},
       number = {3},
        pages = {317-324},
     abstract = {Flexible or 'fluid' cognitive processes are regarded as fundamental to problem solving and creative ability, requiring a specific neurophysiological milieu. REM-sleep dreaming is associated with creative processes and abstract reasoning with increased strength of weak associations in cognitive networks. REM sleep is also mediated by a distinctive neurophysiological profile, different to that of wake and NREM sleep. This study compared the performance of 16 subjects on a test of cognitive flexibility using anagram word puzzles following REM and NREM awakenings across the night, and waking performances during the day. REM awakenings provided a significant 32\% advantage in the number of anagrams solved compared with NREM awakenings and was equal to that of wake time trials. Correlations of individual performance profiles suggest that REM sleep may offer a different mode of problem solving compared with wake and NREM. When early and late REM and NREM awakening data were separated, a dissociation was evident, with NREM task performance becoming more REM-like later in the night, while REM performance remained constant. These data suggest that the neurophysiology of REM sleep represents a brain state more amenable to flexible cognitive processing than NREM and different from that in wake, and may offer insights into the neurocognitive properties of REM-sleep dreaming.}
}
@techreport{WallachTR-04,
       author = {Hanna M. Wallach},
        title = {Conditional Random Fields: An Introduction},
        issue = {MS-CIS-04-21},
  institution = {Department of Computer and Information Science, University of Pennsylvania},
         year = 2004,
}
@book{Walletal00,
       author = {Larry Wall and Tom Christiansen and Jon Orwant},
        title = {Programming Perl},
    publisher = {Third edition, O'Reilly},
      address = {Sebastopol, CA},
         year = 2000,
}
@article{WallingetalMOLSCI-09,
        title = {Quantum dots for live cell and in vivo imaging},
       author = {Walling, M.A. and Novak, J.A. and Shepard, J.R.E.},
      journal = {International Journal Molecular Science},
       volume = 10,
        issue = 2,
        pages = {441-491},
         year = 2009,
     abstract = {In the past few decades, technology has made immeasurable strides to enable visualization, identification, and quantitation in biological systems. Many of these technological advancements are occurring on the nanometer scale, where multiple scientific disciplines are combining to create new materials with enhanced properties. The integration of inorganic synthetic methods with a size reduction to the nano-scale has lead to the creation of a new class of optical reporters, called quantum dots. These semiconductor quantum dot nanocrystals have emerged as an alternative to organic dyes and fluorescent proteins, and are brighter and more stable against photobleaching than standard fluorescent indicators. Quantum dots have tunable optical properties that have proved useful in a wide range of applications from multiplexed analysis such as DNA detection and cell sorting and tracking, to most recently demonstrating promise for in vivo imaging and diagnostics. This review provides an in-depth discussion of past, present, and future trends in quantum dot use with an emphasis on in vivo imaging and its related applications.}
}
@article{WallravenetalNETWORK-05,
        title = {Learning from humans: Computational modeling of face recognition},
       author = {Christian Wallraven and Adrian Schwaninger and Heinrich B\"{u}thoff},
      journal = {Network: Computation in Neural Systems},
       volume = 16,
        issue = 4,
        pages = {401-418},
         year = 2005,
}
@article{WalshetalCELL-08,
       author = {Walsh, Christopher A. and Morrow, Eric M. and Rubenstein, John L. R.},
        title = {Autism and brain development},
      journal = {Cell},
       volume = {135},
        issue = {3},
         year = {2008},
        pages = {396-400},
     abstract = {Genetic studies are refining our understanding of neurodevelopmental mechanisms in autism. Some autism-related mutations appear to disrupt genes regulated by neuronal activity, which are especially important in development of the postnatal nervous system. Gene replacement studies in mice indicate that the developmental window to ameliorate symptoms may be wider than previously anticipated.},
}
@article{WalterSCIAM-50,
       author = {William Grey Walter},
        title = {An imitation of life},
      journal = {Scientific American},
         year = 1950,
       volume = 182,
        issue = 5,
        pages = {42-45},
}
@article{WaltherandKochNN-06,
       author = {Dirk Walther and Christof Koch},
        title = {Modeling attention to salient proto-objects},
      journal = {Neural Networks},
       volume = 19,
        issue = 9,
         year = 2006,
        pages = {1395-1407},
}
@article{WaltheretalCVIU-05,
       author = {Dirk Walther and Ueli Rutishauser and Christof Koch and Pietro Perona},
        title = {Selective visual attention enables learning and recognition of multiple objects in cluttered scenes},
      journal = {Computer Vision and Image Understanding},
       volume = 100,
        issue = {1-2},
        pages = {41-63},
         year = 2005,
}
@article{WandellARN-16,
       author = {Wandell, Brian A.},
        title = {Clarifying Human White Matter},
      journal = {Annual Review of Neuroscience},
       volume = {39},
       number = {1},
         year = {2016},
        pages = {103-128},
     abstract = {Progress in magnetic resonance imaging (MRI) now makes it possible to identify the major white matter tracts in the living human brain. These tracts are important because they carry many of the signals communicated between different brain regions. MRI methods coupled with biophysical modeling can measure the tissue properties and structural features of the tracts that impact our ability to think, feel, and perceive. This review describes the fundamental ideas of the MRI methods used to identify the major white matter tracts in the living human brain.}
}
@article{WandellandYeatmanCOIN-13,
       author = {Wandell, B.A. and Yeatman, J.D.},
        title = {Biological development of reading circuits},
      journal = {Current Opinion Neurobiology},
       volume = 23,
        issue = 2,
         year = 2013,
        pages = {261-268},
}
@article{WandelletalNEURON-07,
        title = {Visual Field Maps in Human Cortex},
       author = {Brian A. Wandell and Serge O. Dumoulin and Alyssa A. Brewer},
      journal = {Neuron},
       volume = {56},
        issue = {2},
         year = {2007},
        pages = {366-383},
     abstract = {Much of the visual cortex is organized into visual field maps: nearby neurons have receptive fields at nearby locations in the image. Mammalian species generally have multiple visual field maps with each species having similar, but not identical, maps. The introduction of functional magnetic resonance imaging made it possible to identify visual field maps in human cortex, including several near (1) medial occipital (V1,V2,V3), (2) lateral occipital (LO-1,LO-2, hMT+), (3) ventral occipital (hV4, VO-1, VO-2), (4) dorsal occipital (V3A, V3B), and (5) posterior parietal cortex (IPS-0 to IPS-4). Evidence is accumulating for additional maps, including some in the frontal lobe. Cortical maps are arranged into clusters in which several maps have parallel eccentricity representations, while the angular representations within a cluster alternate in visual field sign. Visual field maps have been linked to functional and perceptual properties of the visual system at various spatial scales, ranging from the level of individual maps to map clusters to dorsal-ventral streams. We survey recent measurements of human visual field maps, describe hypotheses about the function and relationships between maps, and consider methods to improve map measurements and characterize the response properties of neurons comprising these maps.}
}
@article{WandelletalPTRS-B-05,
       author = {Wandell, Brian A. and Brewer, Alyssa A. and Dougherty, Robert F.},
        title = {Visual field map clusters in human cortex},
      journal = {Philosophical transactions of the Royal Society of London. Series B, Biological sciences},
    publisher = {The Royal Society},
       volume = {360},
        issue = {1456},
         year = {2005},
        pages = {693-707},
     abstract = {We describe the location and general properties of nine human visual field maps. The cortical location of each map, as well as many examples of the eccentricity and angular representations within these maps, are shown in a series of images that summarize a large set of functional MRI data. The organization and properties of these maps are compared and contrasted with descriptions by other investigators. We hypothesize that the human visual field maps are arranged in several clusters, each comprising a group of maps that share a common foveal representation and semicircular eccentricity map. The spatial organization of these clusters suggests that the perceptual processing within each cluster serves related functions.},
}
@article{WangCOIN-12,
       author = {Wang, X. J.},
        title = {Neural dynamics and circuit mechanisms of decision-making},
      journal = {Current Opinion in Neurobiology},
       volume = {22},
       number = {6},
         year = {2012},
        pages = {1039-1046},
     abstract = {In this review, I briefly summarize current neurobiological studies of decision-making that bear on two general themes. The first focuses on the nature of neural representation and dynamics in a decision circuit. Experimental and computational results suggest that ramping-to-threshold in the temporal domain and trajectory of population activity in the state space represent a duality of perspectives on a decision process. Moreover, a decision circuit can display several different dynamical regimes, such as the ramping mode and the jumping mode with distinct defining properties. The second is concerned with the relationship between biologically-based mechanistic models and normative-type models. A fruitful interplay between experiments and these models at different levels of abstraction have enabled investigators to pose increasingly refined questions and gain new insights into the neural basis of decision-making. In particular, recent work on multi-alternative decisions suggests that deviations from rational models of choice behavior can be explained by established neural mechanisms.}
}
@article{WangCoRR-19,
       author = {Ke Wang},
        title = {Learning Scalable and Precise Representation of Program Semantics},
      journal = {CoRR},
       volume = {arXiv:1905.05251},
         year = {2019},
     abstract = {Neural program embedding has shown potential in aiding the analysis of large-scale, complicated software. Newly proposed deep neural architectures pride themselves on learning program semantics rather than superficial syntactic features. However, by considering the source code only, the vast majority of neural networks do not capture a deep, precise representation of program semantics. In this paper, we present dypro, a novel deep neural network that learns from program execution traces. Compared to the prior dynamic models, not only is dypro capable of generalizing across multiple executions for learning a program's dynamic semantics in its entirety, but dypro is also more efficient when dealing with programs yielding long execution traces. For evaluation, we task dypro with semantic classification (i.e. categorizing programs based on their semantics) and compared it against two prominent static models: Gated Graph Neural Network and TreeLSTM. We find that dypro achieves the highest prediction accuracy among all models. To further reveal the capacity of all aforementioned deep neural architectures, we examine if the models can learn to detect deeper semantic properties of a program. In particular given a task of recognizing loop invariants, we show dypro beats all static models by a wide margin.}
}
@article{WangNATURE-PHOTONICS-09,
       author = {Wang, Lihong V.},
        title = {Multiscale photoacoustic microscopy and computed tomography},
      journal = {Nature Photonics},
    publisher = {Nature Publishing Group},
       volume = {3},
        issue = {9},
         year = {2009},
        pages = {503-509},
     abstract = {Photoacoustic tomography (PAT) is probably the fastest-growing area of biomedical imaging technology, owing to its capacity for high-resolution sensing of rich optical contrast in vivo at depths beyond the optical transport mean free path (\approx{}1 mm in human skin). Existing high-resolution optical imaging technologies, such as confocal microscopy and two-photon microscopy, have had a fundamental impact on biomedicine but cannot reach the penetration depths of PAT. By utilizing low ultrasonic scattering, PAT indirectly improves tissue transparency up to 1000-fold and consequently enables deeply penetrating functional and molecular imaging at high spatial resolution. Furthermore, PAT promises in vivo imaging at multiple length-scales; it can image subcellular organelles to organs with the same contrast origin\emdash{}an important application in multiscale systems biology research.}
}
@article{WangPHYSIOLOGICAL_REVIEWS-10,
       author = {Wang, X. J.},
        title = {Neurophysiological and computational principles of cortical rhythms in cognition},
      journal = {Physiological Reviews},
         year = 2010,
       volume = 90,
        issue = 3,
        pages = {1195-1268},
     abstract = {Synchronous rhythms represent a core mechanism for sculpting temporal coordination of neural activity in the brain-wide network. This review focuses on oscillations in the cerebral cortex that occur during cognition, in alert behaving conditions. Over the last two decades, experimental and modeling work has made great strides in elucidating the detailed cellular and circuit basis of these rhythms, particularly gamma and theta rhythms. The underlying physiological mechanisms are diverse (ranging from resonance and pacemaker properties of single cells to multiple scenarios for population synchronization and wave propagation), but also exhibit unifying principles. A major conceptual advance was the realization that synaptic inhibition plays a fundamental role in rhythmogenesis, either in an interneuronal network or in a reciprocal excitatory-inhibitory loop. Computational functions of synchronous oscillations in cognition are still a matter of debate among systems neuroscientists, in part because the notion of regular oscillation seems to contradict the common observation that spiking discharges of individual neurons in the cortex are highly stochastic and far from being clocklike. However, recent findings have led to a framework that goes beyond the conventional theory of coupled oscillators and reconciles the apparent dichotomy between irregular single neuron activity and field potential oscillations. From this perspective, a plethora of studies will be reviewed on the involvement of long-distance neuronal coherence in cognitive functions such as multisensory integration, working memory, and selective attention. Finally, implications of abnormal neural synchronization are discussed as they relate to mental disorders like schizophrenia and autism.}
}
@article{WangWetalNSR-16,
       author = {Wang, Wei and Panin, Gennady N. and Fu, Xiao and Zhang, Lei and Ilanchezhiyan, P. and Pelenovich, Vasiliy O. and Fu, Dejun and Kang, Tae Won},
        title = {{MoS2} memristor with photoresistive switching},
      journal = {Scientific Reports},
         year = {2016},
       volume = {6},
        pages = {31224},
     abstract = {A MoS2 nanosphere memristor with lateral gold electrodes was found to show photoresistive switching. The new device can be controlled by the polarization of nanospheres, which causes resistance switching in an electric field in the dark or under white light illumination. The polarization charge allows to change the switching voltage of the photomemristor, providing its multi-level operation. The device, polarized at a voltage 6 V, switches abruptly from a high resistance state (HRSL6) to a low resistance state (LRSL6) with the On/Off resistance ratio of about 10 under white light and smooth in the dark. Analysis of device conductivity in different resistive states indicates that its resistive state could be changed by the modulation of the charge in an electric field in the dark or under light, resulting in the formation/disruption of filaments with high conductivity. A MoS2 photomemristor has great potential as a multifunctional device designed by using cost-effective fabrication techniques.}
}
@inproceedings{WangandAdelsonSPIE-94,
       author = {J. Y. A. Wang and E. H. Adelson},
        title = {Spatio-Temporal Segmentation of Video Sequence},
    booktitle = {Proceedings of {SPIE} on Image and Video Processing {II}},
        pages = {120-131},
       volume = 2182,
         year = 1994,
}
@article{WangandAdelsonTIP-94,
       author = {J. Y. A. Wang and E. H. Adelson},
        title = {Representing Moving Images With Layers},
      journal = {The {IEEE} Transactions on Image Processing Special Issue: Image Sequence Compression},
       volume = 3,
        issue = 5,
        pages = {625-638},
         year = 1994,
}
@article{WangandHuSCIENCE-12,
        title = {Photoacoustic Tomography: In Vivo Imaging from Organelles to Organs},
       author = {Lihong V. Wang and Song Hu},
      journal = {Science},
       volume = 335,
        issue = 6075,
         year = 2012,
        pages = {1458-1462},
     abstract = {Photoacoustic tomography (PAT) can create multiscale multicontrast images of living biological structures ranging from organelles to organs. This emerging technology overcomes the high degree of scattering of optical photons in biological tissue by making use of the photoacoustic effect. Light absorption by molecules creates a thermally induced pressure jump that launches ultrasonic waves, which are received by acoustic detectors to form images. Different implementations of PAT allow the spatial resolution to be scaled with the desired imaging depth in tissue while a high depth-to-resolution ratio is maintained. As a rule of thumb, the achievable spatial resolution is on the order of 1/200 of the desired imaging depth, which can reach up to 7 centimeters. PAT provides anatomical, functional, metabolic, molecular, and genetic contrasts of vasculature, hemodynamics, oxygen metabolism, biomarkers, and gene expression. We review the state of the art of PAT for both biological and clinical studies and discuss future prospects.}
}
@inproceedings{WangandJiangICLR-17,
       author = {Shuohang Wand and Jing Jiang},
        title = {Machine Comprehension Using Match-{LSTM} and Answer Pointer},
    booktitle = {International Conference on Learning Representations},
       volume = {arXiv:1608.07905},
         year = {2017},
     abstract = {Machine comprehension of text is an important problem in natural language processing. A recently released dataset, the Stanford Question Answering Dataset (SQuAD), offers a large number of real questions and their answers created by humans through crowdsourcing. SQuAD provides a challenging testbed for eval- uating machine comprehension algorithms, partly because compared with previ- ous datasets, in SQuAD the answers do not come from a small set of candidate answers and they have variable lengths. We propose an end-to-end neural architec- ture for the task. The architecture is based on match-LSTM, a model we proposed previously for textual entailment, and Pointer Net, a sequence-to-sequence model proposed by Vinyals et al. (2015) to constrain the output tokens to be from the input sequences. We propose two ways of using Pointer Net for our tasks. Our experiments show that both of our two models substantially outperform the best results obtained by Rajpurkar et al. (2016) using logistic regression and manually crafted features. Besides, our boundary model also achieves the best performance on the MSMARCO dataset (Nguyen et al., 2016).}
}
@article{WangandKennedyCOIN-16,
       author = {Xiao-Jing Wang and Henry Kennedy},
        title = {Brain structure and dynamics across scales: in search of rules},
      journal = {Current Opinion in Neurobiology},
       volume = {37},
        pages = {92-98},
         year = {2016},
     abstract = {Louis Henry Sullivan, the father of skyscrapers, famously stated 'Form ever follows function'. In this short review, we will focus on the relationship between form (structure) and function (dynamics) in the brain. We summarize recent advances on the quantification of directed- and weighted-mesoscopic connectivity of mammalian cortex, the exponential distance rule for mesoscopic and microscopic circuit wiring, a spatially embedded random model of inter-areal cortical networks, and a large-scale dynamical circuit model of money's cortex that gives rise to a hierarchy of timescales. These findings demonstrate that inter-areal cortical networks are dense (hence such concepts as 'small-world' need to be refined when applied to the brain), spatially dependent (therefore purely topological approach of graph theory has limited applicability) and heterogeneous (consequently cortical areas cannot be treated as identical 'nodes').}
}
@inproceedings{WangandLiIJCAI-09,
       author = {Zheshen Wang and Baoxin Li},
        title = {Human Activity Encoding and Recognition Using Low-level Visual Features},
    booktitle = {Proceedings of the 21th International Joint Conference on Artificial Intelligence},
         year = 2009,
}
@incollection{WangandLingNanotechnology-Encyclopedia-12,
       author = {Wang, ZhongLin and Liu, Ying},
        title = {Piezoelectric Effect at Nanoscale},
    booktitle = {Encyclopedia of Nanotechnology},
       editor = {Bhushan, Bharat},
    publisher = {Springer Netherlands},
         year = 2012,
        pages = {2085-2099},
}
@article{WangandSongSCIENCE-06,
       author = {Zhong Lin Wang and Jinhui Song},
        title = {Piezoelectric nanogenerators based on zinc oxide nanowire arrays},
      journal = {Science},
       volume = 312,
        issue = 5771,
         year = 2006,
        pages = {242-246},
     abstract = {We have converted nanoscale mechanical energy into electrical energy by means of piezoelectric zinc oxide nanowire (NW) arrays. The aligned NWs are deflected with a conductive atomic force microscope tip in contact mode. The coupling of piezoelectric and semiconducting properties in zinc oxide creates a strain field and charge separation across the NW as a result of its bending. The rectifying characteristic of the Schottky barrier formed between the metal tip and the NW leads to electrical current generation. The efficiency of the NW-based piezoelectric power generator is estimated to be 17 to 30\%. This approach has the potential of converting mechanical, vibrational, and/or hydraulic energy into electricity for powering nanodevices.}
}
@article{WangandSuPOPL-19,
       author = {Ke Wang and Zhendong Su},
        title = {Learning Blended, Precise Semantic Program Embeddings},
      journal = {CoRR},
       volume = {arXiv:1907.02136},
         year = {2019},
     abstract = {Learning neural program embeddings is key to utilizing deep neural networks in program languages research --- precise and efficient program representations enable the application of deep models to a wide range of program analysis tasks. Existing approaches predominately learn to embed programs from their source code, and, as a result, they do not capture deep, precise program semantics. On the other hand, models learned from runtime information critically depend on the quality of program executions, thus leading to trained models with highly variant quality. This paper tackles these inherent weaknesses of prior approaches by introducing a new deep neural network, liger, which learns program representations from a mixture of symbolic and concrete execution traces. We have evaluated liger on coset, a recently proposed benchmark suite for evaluating neural program embeddings. Results show liger (1) is significantly more accurate than the state-of-the-art syntax-based models Gated Graph Neural Network and code2vec in classifying program semantics, and (2) requires on average 10x fewer executions covering 74\% fewer paths than the state-of-the-art dynamic model dypro. Furthermore, we extend liger to predict the name for a method from its body's vector representation. Learning on the same set of functions (more than 170K in total), liger significantly outperforms code2seq, the previous state-of-the-art for method name prediction.}
}
@article{WangandWangFiP-19,
       author = {Wang, Wei and Wang, Guang-Zhong},
        title = {Understanding Molecular Mechanisms of the Brain Through Transcriptomics},
      journal = {Frontiers in physiology},
    publisher = {Frontiers Media S.A.},
       volume = {10},
         year = {2019},
        pages = {214-214},
     abstract = {The brain is the most complicated organ in the human body with more than ten thousand genes expressed in each region. The molecular activity of the brain is divergent in various brain regions, both spatially and temporally. The function of each brain region lies in the fact that each region has different gene expression profiles, the possibility of differential RNA splicing, as well as various post-transcriptional and translational modification processes. Understanding the overall activity of the brain at the molecular level is essential for a comprehensive understanding of how the brain works. Fortunately, the development of next generation sequencing technology has made it possible to measure the molecular activity of a specific tissue as a daily routine approach of research. Therefore, at the molecular level, the application of sequencing technology to investigate the molecular organization of the brain has become a novel field, and significant progress has been made recently in this field. In this paper, we reviewed the major computational methods used in the analysis of brain transcriptome, including the application of these methods to the research of human and non-human mammal brains. Finally, we discussed the utilization of transcriptome methods in neurological diseases.},
}
@inproceedings{WangandYangAAAI-11,
       author = {Huayan Wang and Qiang Yang},
        title = {Transfer Learning by Structural Analogy},
    booktitle = {Proceedings of the Twenty-Sixth {AAAI} Conference on Artificial Intelligence},
         year = {2011},
     abstract = {Transfer learning allows knowledge to be extracted from auxiliary domains and be used to enhance learning in a target domain. For transfer learning to be successful, it is critical to find the similarity between auxiliary and target domains, even when such mappings are not obvious. In this paper, we present a novel algorithm for finding the structural similarity between two domains, to enable transfer learning at a structured knowledge level. In particular, we address the problem of how to learn a non-trivial structural similarity mapping between two different domains when they are completely different on the representation level. This problem is challenging because we cannot directly compare features across domains. Our algorithm extracts the structural features within each domain and then maps the features into the Reproducing Kernel Hilbert Space (RKHS), such that the "structural dependencies" of features across domains can be estimated by kernel matrices of the features within each domain. By treating the analogues from both domains as equivalent, we can transfer knowledge to achieve a better understanding of the domains and improved performance for learning. We validate our approach on synthetic and real-world datasets.},
}
@book{Wangetal2006subthreshold,
       author = {Wang, Alice and Calhoun, Benton H. and Chandrakasan, Anantha P.},
        title = {Sub-Threshold Design for Ultra Low-Power Systems (Series on Integrated Circuits and Systems)},
         year = {2006},
    publisher = {Springer-Verlag},
      address = {Berlin, Heidelberg}
}
@inproceedings{WangetalACCV-07,
       author = {Liming Wang and Jianbo Shi and Gang Song and I-Fan Shen},
        title = {Object Detection Combining Recognition and Segmentation},
    booktitle = {Proceedings of the 8th Asian Conference on Computer Vision},
       editor = {Yasushi Yagi and Sing Bing Kang and In-So Kweon and Hongbin Zha},
     location = {Tokyo, Japan},
    publisher = {Springer},
       series = {Lecture Notes in Computer Science},
       volume = 4843,
        pages = {189-199},
         year = 2007,
}
@article{WangetalBIORXIV-18,
       author = {Wang, Jane X and Kurth-Nelson, Zeb and Kumaran, Dharshan and Tirumala, Dhruva and Soyer, Hubert and Leibo, Joel Z and Hassabis, Demis and Botvinick, Matthew},
        title = {Prefrontal cortex as a meta-reinforcement learning system},
      journal = {bioRxiv},
         year = {2018},
    publisher = {Cold Spring Harbor Laboratory},
     abstract = {Over the past twenty years, neuroscience research on reward-based learning has converged on a canonical model, under which the neurotransmitter dopamine {\textquoteright}stamps in{\textquoteright} associations between situations, actions and rewards by modulating the strength of synaptic connections between neurons. However, a growing number of recent findings have placed this standard model under strain. In the present work, we draw on recent advances in artificial intelligence to introduce a new theory of reward-based learning. Here, the dopamine system trains another part of the brain, the prefrontal cortex, to operate as its own free-standing learning system. This new perspective accommodates the findings that motivated the standard model, but also deals gracefully with a wider range of observations, providing a fresh foundation for future research.},
          url = {https://www.biorxiv.org/content/early/2018/04/06/295964},
}
@inproceedings{WangetalBMVC-09,
        title = {Evaluation of local spatio-temporal features for action recognition},
       author = {H. Wang and M. M. Ullah and A. Kl\"{a}ser and I. Laptev and C. Schmid},
    booktitle = {Proceedings of the {B}ritish Machine Vision Conference},
     location = {London},
         year = 2009,
}
@inproceedings{WangetalBMVC-10,
       author = {Heng Wang and Muhammad Muneeb Ullah and Alexander Kl{\"a}ser and Ivan Laptev and Cordelia Schmid},
        title = {Evaluation of local spatio-temporal features for action recognition},
    booktitle = {British Machine Vision Conference},
        pages = 127,
         year = 2010,
}
@inproceedings{WangetalCOLING-16,
       author = {Wang, Xingyou and Jiang, Weijie and Luo, Zhiyong},
        title = {Combination of Convolutional and Recurrent Neural Network for Sentiment Analysis of Short Texts},
    booktitle = {Proceedings of {COLING} 2016, the 26th International Conference on Computational Linguistics: Technical Papers},
    publisher = {The COLING 2016 Organizing Committee},
         year = {2016},
        pages = {2428-2437},
     abstract = {Sentiment analysis of short texts is challenging because of the limited contextual information they usually contain. In recent years, deep learning models such as convolutional neural networks (CNNs) and recurrent neural networks (RNNs) have been applied to text sentiment analysis with comparatively remarkable results. In this paper, we describe a jointed CNN and RNN architecture, taking advantage of the coarse-grained local features generated by CNN and long-distance dependencies learned via RNN for sentiment analysis of short texts. Experimental results show an obvious improvement upon the state-of-the-art on three benchmark corpora, MR, SST1 and SST2, with 82.28{\%}, 51.50{\%} and 89.95{\%} accuracy, respectively.},
}
@article{WangetalCOMPLEXITY-08,
       author = {Wang, Jun and Dam, Gregory and Yildirim, Sule and Rand, William and Wilensky, Uri and Houk, James C.},
        title = {Reciprocity between the cerebellum and the cerebral cortex: Nonlinear dynamics in microscopic modules for generating voluntary motor commands},
      journal = {Complexity},
       volume = {14},
       number = {2},
         year = {2008},
        pages = {29-45},
     abstract = {The cerebellum and basal ganglia are reciprocally connected with the cerebral cortex, forming many loops that function as distributed processing modules. Here we present a detailed model of one microscopic loop between the motor cortex and the cerebellum, and we show how small arrays of these microscopic loops (CB modules) can be used to generate biologically plausible motor commands for controlling movement. A fundamental feature of CB modules is the presence of positive feedback loops between the cerebellar nucleus and the motor cortex. We use nonlinear dynamics to model one microscopic loop and to investigate its bistable properties. Simulations demonstrate an ability to program a motor command well in advance of command generation and an ability to vary command duration. However, control of command intensity is minimal, which could interfere with the control of movement velocity. To assess these hypotheses, we use a minimal nonlinear model of the neuromuscular (NM) system that translates motor commands into actual movements. Simulations of the combined CB-NM modular model indicate that movement duration is readily controlled, whereas velocity is poorly controlled. We then explore how an array of eight CB-NM modules can be used to control the direction and endpoint of a planar movement. In actuality, thousands of such microscopic loops function together as an array of adjustable pattern generators for programming and regulating the composite motor commands that control limb movements. We discuss the biological plausibility and limitations of the model. We also discuss ways in which an agent-based representation can take advantage of the modularity in order to model this complex system.},
}
@article{WangetalCoRR-16,
       author = {Sida I. Wang and Percy Liang and Christopher D. Manning},
        title = {Learning Language Games through Interaction},
      journal = {CoRR},
       volume = {arXiv:1606.02447},
         year = {2016},
 abstract = {We introduce a new language learning setting relevant to building adaptive natural language interfaces. It is inspired by Wittgenstein's language games: a human wishes to accomplish some task (e.g., achieving a certain configuration of blocks), but can only communicate with a computer, who performs the actual actions (e.g., removing all red blocks). The computer initially knows nothing about language and therefore must learn it from scratch through interaction, while the human adapts to the computer's capabilities. We created a game called SHRDLURN in a blocks world and collected interactions from 100 people playing it. First, we analyze the humans' strategies, showing that using compositionality and avoiding synonyms correlates positively with task performance. Second, we compare computer strategies, showing that modeling pragmatics on a semantic parsing model accelerates learning for more strategic players.}
}
@article{WangetalCoRR-17,
       author = {Ke Wang and Rishabh Singh and Zhendong Su},
        title = {Dynamic Neural Program Embedding for Program Repair},
      journal = {CoRR},
       volume = {arXiv:1711.07163},
         year = {2017},
     abstract = {Neural program embeddings have shown much promise recently for a variety of program analysis tasks, including program synthesis, program repair, fault localization, etc. However, most existing program embeddings are based on syntactic features of programs, such as raw token sequences or abstract syntax trees. Unlike images and text, a program has an unambiguous semantic meaning that can be difficult to capture by only considering its syntax (i.e. syntactically similar programs can exhibit vastly different run-time behavior), which makes syntax-based program embeddings fundamentally limited. This paper proposes a novel semantic program embedding that is learned from program execution traces. Our key insight is that program states expressed as sequential tuples of live variable values not only captures program semantics more precisely, but also offer a more natural fit for Recurrent Neural Networks to model. We evaluate different syntactic and semantic program embeddings on predicting the types of errors that students make in their submissions to an introductory programming class and two exercises on the CodeHunt education platform. Evaluation results show that our new semantic program embedding significantly outperforms the syntactic program embeddings based on token sequences and abstract syntax trees. In addition, we augment a search-based program repair system with the predictions obtained from our semantic embedding, and show that search efficiency is also significantly improved.}
} 
@article{WangetalCoRR-18,
       author = {Chenglong Wang and Po-Sen Huang and Alex Polozov and Marc Brockschmidt and Rishabh Singh},
        title = {Execution-Guided Neural Program Decoding},
      journal = {CoRR},
       volume = {arXiv:1807.03100},
         year = {2018},
     abstract = {We present a neural semantic parser that translates natural language questions into executable SQL queries with two key ideas. First, we develop an encoder-decoder model, where the decoder uses a simple type system of SQL to constraint the output prediction, and propose a value-based loss when copying from input tokens. Second, we explore using the execution semantics of SQL to repair decoded programs that result in runtime error or return empty result. We propose two model-agnostic repair approaches, an ensemble model and a local program repair, and demonstrate their effectiveness over the original model. We evaluate our model on the WikiSQL dataset and show that our model achieves close to state-of-the-art results with lesser model complexity.}
}
@article{WangetalCoRR-18b,
       author = {Linnan Wang and Yiyang Zhao and Yuu Jinnai},
        title = {{AlphaX: eXploring Neural Architectures with Deep Neural Networks and Monte Carlo Tree Search}},
      journal = {CoRR},
       volume = {arXiv:1805.07440},
         year = {2018},
     abstract = {Neural Architecture Search (NAS) has shown great success in automating the design of neural networks, but the prohibitive amount of computations behind current NAS methods requires further investigations in improving the sample efficiency and the network evaluation cost to get better results in a shorter time. In this paper, we present a novel scalable Monte Carlo Tree Search (MCTS) based NAS agent, named AlphaX, to tackle these two aspects. AlphaX improves the search efficiency by adaptively balancing the exploration and exploitation at the state level, and by a Meta-Deep Neural Network (DNN) to predict network accuracies for biasing the search toward a promising region. To amortize the network evaluation cost, AlphaX accelerates MCTS rollouts with a distributed design and reduces the number of epochs in evaluating a network by transfer learning, which is guided with the tree structure in MCTS. In 12 GPU days and 1000 samples, AlphaX found an architecture that reaches 97.84\% top-1 accuracy on CIFAR-10, and 75.5\% top-1 accuracy on ImageNet, exceeding SOTA NAS methods in both the accuracy and sampling efficiency. Particularly, we also evaluate AlphaX on NASBench-101, a large scale NAS dataset; AlphaX is 3x and 2.8x more sample efficient than Random Search and Regularized Evolution in finding the global optimum. Finally, we show the searched architecture improves a variety of vision applications from Neural Style Transfer, to Image Captioning and Object Detection.}
}
@article {WangetalELIFE-19,
        title = {Anterior insular cortex plays a critical role in interoceptive attention},
       author = {Wang, Xingchao and Wu, Qiong and Egan, Laura and Gu, Xiaosi and Liu, Pinan and Gu, Hong and Yang, Yihong and Luo, Jing and Wu, Yanhong and Gao, Zhixian and Fan, Jin},
       editor = {Stephan, Klaas Enno and Frank, Michael J. and Stephan, Klaas Enno and Faull, Olivia K. and Allen, Micah},
      journal = {eLife},
    publisher = {eLife Sciences Publications, Ltd},
       volume = 8,
         year = 2019,
        pages = {e42265},
     abstract = {Accumulating evidence indicates that the anterior insular cortex (AIC) mediates interoceptive attention which refers to attention towards physiological signals arising from the body. However, the necessity of the AIC in this process has not been demonstrated. Using a novel task that directs attention toward breathing rhythm, we assessed the involvement of the AIC in interoceptive attention in healthy participants using functional magnetic resonance imaging and examined the necessity of the AIC in interoceptive attention in patients with AIC lesions. Results showed that interoceptive attention was associated with increased AIC activation, as well as enhanced coupling between the AIC and somatosensory areas along with reduced coupling between the AIC and visual sensory areas. In addition, AIC activation was predictive of individual differences in interoceptive accuracy. Importantly, AIC lesion patients showed disrupted interoceptive discrimination accuracy and sensitivity. These results provide compelling evidence that the AIC plays a critical role in interoceptive attention.},
}
@inproceedings{WangetalICCV-11,
        title = {Action Recognition by Dense Trajectories},
       author = {Heng Wang and Alexander Kl\"{a}ser and Cordelia Schmid and Cheng-Lin Li},
    booktitle = {Proceedings of the International Conference on Computer Vision},
    publisher = {IEEE Computer Society},
         year = 2011,
        pages = {3169-3176},
}
@article{WangetalICLR-18,
       author = {Ke Wang and Rishabh Singh and Zhendong Su},
        title = {Dynamic Neural Program Embedding for Program Repair},
      journal = {International Conference on Learning Representations},
         year = {2018},
      abstract = {Neural program embeddings have shown much promise recently for a variety of program analysis tasks, including program synthesis, program repair, code- completion, and fault localization. However, most existing program embeddings are based on syntactic features of programs, such as token sequences or abstract syntax trees. Unlike images and text, a program has well-defined semantics that can be difficult to capture by only considering its syntax (i.e. syntactically similar programs can exhibit vastly different run-time behavior), which makes syntax- based program embeddings fundamentally limited. We propose a novel semantic program embedding that is learned from program execution traces. Our key in- sight is that program states expressed as sequential tuples of live variable values not only capture program semantics more precisely, but also offer a more natural fit for Recurrent Neural Networks to model. We evaluate different syntactic and semantic program embeddings on the task of classifying the types of errors that students make in their submissions to an introductory programming class and on the CodeHunt education platform. Our evaluation results show that the semantic program embeddings significantly outperform the syntactic program embeddings based on token sequences and abstract syntax trees. In addition, we augment a search-based program repair system with predictions made from our semantic embedding and demonstrate significantly improved search efficiency.},
}
@article{WangetalJBO-12,
       author = {Po-Hsun Wang and Hao-Li Liu and Po-Hung Hsu and Chia-Yu Lin and Churng-Ren Chris Wang and Pin-Yuan Chen and Kuo-Chen Wei and Tzu-Chen Yen and Meng-Lin Li},
        title = {Gold-nanorod contrast-enhanced photoacoustic micro-imaging of focused-ultrasound induced blood-brain-barrier opening in a rat model},
      journal = {Journal of Biomedical Optics},
       volume = 17,
        issue = 6,
         year = 2012,
        pages = 061222,
}
@article{WangetalJoNM-13,
       author = {Shuxia Wang and Dai Shan and Jiankun Dai and Haichen Niu and Yuanye Ma and Fuchun Lin and Hao Lei},
        title = {Anatomical MRI templates of tree shrew brain for volumetric analysis and voxel-based morphometry},
      journal = {Journal of Neuroscience Methods},
       volume = {220},
       number = {1},
         year = {2013},
        pages = {9-17},
     abstract = {Background: Tree shrews are close relatives of primates, and are increasingly used as models in the research of vision, social stress and neurological/psychiatric diseases. However, neuroimaging techniques, for example magnetic resonance (MR) imaging, are only rarely applied to this species to study the structure and function of the brain. A template MR image set, which is essential for morphometry/volumetric analysis, of tree shrew brain has been lacking in the literature.

New method: High-resolution anatomical MR images and diffusion tensor images of the brain were acquired from male Chinese tree shrews (Tupaia belangeri chinensis), and resampled to an isotropic resolution of 200$\mu$m\%200$\mu$m\%200$\mu$m. Population-based image templates of tree shrew brain, including gray matter/white matter/cerebrospinal fluid probability maps and a fractional anisotropy template, were constructed at this spatial resolution, all in a reference space. Digital masks of representative anatomical structures, including hippocampus, amygdala and cingulum bundle, were created.

Result: With the templates constructed, the volumes of bilateral hippocampus and amygdala were measured using a template-facilitated semi-automated approach to be 59.8±8.3 and 64.3±3.4mm3, respectively.

Comparison: For the first time, high-resolution MR image templates of tree shrew brain were reported. The average volume of bilateral hippocampus measured with the template-facilitated semi-automated approach was found to be similar to the result obtained by the much more labor-intensive manual approach.

Conclusions: The MR image templates obtained in this study are useful for analyzing neuroimage data of tree shrew brain. The templates are freely available to the scientific community upon request.}
}
@article{WangetalMTA-20,
       author = {Wang, Xin and Jiang, Xiaotao and Regedzai, Gloria Rumbidzai and Meng, Haohao and Sun, Lingyun},
        title = {Gated neural network framework for interactive character control},
      journal = {Multimedia Tools and Applications},
         year = {2020},
     abstract = {Recently, interactive character control models based on neural network have become a hot research topic in computer graphics and motion synthesis. A real-time interactive character control model with two substructures called gated neural network is proposed in this paper. In the first part of the model, a gated network is used to calculate the expert weights based on the user's input parameters and the character's current pose dynamically. Mixture of experts is used in choosing different control strategies for user input. The character posture is controlled by selecting mode-adjustment or phase-adjustment. In the second part, a simple neural network is used to adjust the character's state through additional input parameters on different landscapes and to synthesize the final character motion. Experimental results show that the model can improve the performance of character control.},
}
@article{WangetalNATURE-METHODS-15,
       author = {Yao, Junjie and Wang, Lidai and Yang, Joon-Mo and Maslov, Konstantin I. and Wong, Terence T. W. and Li, Lei and Huang, Chih-Hsien and Zou, Jun and Wang, Lihong V.},
        title = {High-speed label-free functional photoacoustic microscopy of mouse brain in action},
      journal = {Nature Methods},
         year = 2015,
    publisher = {Nature Publishing Group},
       volume = {advance online publication},
     abstract = {We present fast functional photoacoustic microscopy (PAM) for three-dimensional high-resolution, high-speed imaging of the mouse brain, complementary to other imaging modalities. We implemented a single-wavelength pulse-width-based method with a one-dimensional imaging rate of 100 khz to image blood oxygenation with capillary-level resolution. We applied PAM to image the vascular morphology, blood oxygenation, blood flow and oxygen metabolism in both resting and stimulated states in the mouse brain.}
}
@article{WangetalNATURE-NEUROSCIENCE-18,
       author = {Wang, Jane X. and Kurth-Nelson, Zeb and Kumaran, Dharshan and Tirumala, Dhruva and Soyer, Hubert and Leibo, Joel Z. and Hassabis, Demis and Botvinick, Matthew},
        title = {Prefrontal cortex as a meta-reinforcement learning system},
      journal = {Nature Neuroscience},
         year = {2018},
       volume = {21},
        issue = {6},
        pages = {860-868},
     abstract = {Over the past 20 years, neuroscience research on reward-based learning has converged on a canonical model, under which the neurotransmitter dopamine 'stamps in' associations between situations, actions and rewards by modulating the strength of synaptic connections between neurons. However, a growing number of recent findings have placed this standard model under strain. We now draw on recent advances in artificial intelligence to introduce a new theory of reward-based learning. Here, the dopamine system trains another part of the brain, the prefrontal cortex, to operate as its own free-standing learning system. This new perspective accommodates the findings that motivated the standard model, but also deals gracefully with a wider range of observations, providing a fresh foundation for future research.},
}
@inproceedings{WangetalNIPS-05,
       author = {Wang, J. M. and Fleet, D. J. and Hertzmann, A},
        title = {Gaussian Process Dynamical Models},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 18,
     location = {Vancouver, Canada},
         year = 2005,
}
@incollection{WangetalNIPS-14,
        title = {Attentional Neural Network: Feature Selection Using Cognitive Feedback},
       author = {Wang, Qian and Zhang, Jiaxing and Song, Sen and Zhang, Zheng},
    booktitle = {Advances in Neural Information Processing Systems 27},
       editor = {Z. Ghahramani and M. Welling and C. Cortes and N. D. Lawrence and K. Q. Weinberger},
    publisher = {Curran Associates, Inc.},
         year = {2014},
        pages = {2033-2041},
     abstract = {Attentional Neural Network is a new framework that integrates top-down cognitive bias and bottom-up feature extraction in one coherent architecture. The top-down influence is especially effective when dealing with high noise or dif- ficult segmentation problems. Our system is modular and extensible. It is also easy to train and cheap to run, and yet can accommodate complex behaviors. We obtain classification accuracy better than or competitive with state of art results on the MNIST variation dataset, and successfully disentangle overlaid digits with high success rates. We view such a general purpose framework as an essential foundation for a larger system emulating the cognitive abilities of the whole brain.}
}
@incollection{WangetalNIPS-17,
       author = {Ziyu Wang and Josh Merel and Scott Reed and Greg Wayne and Nando de Freitas, Nicolas Hees},
        title = {Robust Imitation of Diverse Behaviors},
    booktitle = {Advances in Neural Information Processing Systems 30},
       editor = {I. Guyon and U. V. Luxburg and S. Bengio and H. Wallach and R. Fergus and S. Vishwanathan and R. Garnett},
    publisher = {Curran Associates, Inc.},
        pages = {5320-5329},
         year = {2017},
     abstract = {Deep generative models have recently shown great promise in imitation learning for motor control. Given enough data, even supervised approaches can do one-shot imitation learning; however, they are vulnerable to cascading failures when the agent trajectory diverges from the demonstrations. Compared to purely supervised methods, Generative Adversarial Imitation Learning (GAIL) can learn more robust controllers from fewer demonstrations, but is inherently mode-seeking and more difficult to train. In this paper, we show how to combine the favourable aspects of these two approaches. The base of our model is a new type of variational autoencoder on demonstration trajectories that learns semantic policy embeddings. We show that these embeddings can be learned on a 9 DoF Jaco robot arm in reaching tasks, and then smoothly interpolated with a resulting smooth interpolation of reaching behavior. Leveraging these policy representations, we develop a new version of GAIL that (1) is much more robust than the purely-supervised controller, especially with few demonstrations, and (2) avoids mode collapse, capturing many diverse behaviors when GAIL on its own does not. We demonstrate our approach on learning diverse gaits from demonstration on a 2D biped and a 62 DoF 3D humanoid in the MuJoCo physics environment.}
}
@article{WangetalNSR-16,
       author = {Wang, Sheng and Peng, Jian and Ma, Jianzhu and Xu, Jinbo},
        title = {Protein Secondary Structure Prediction Using Deep Convolutional Neural Fields},
      journal = {Scientific Reports},
    publisher = {Nature Publishing Group},
       volume = {6},
       number = {18962},
         year = {2016},
     abstract = {Protein secondary structure (SS) prediction is important for studying protein structure and function. When only the sequence (profile) information is used as input feature, currently the best predictors can obtain ~80\% Q3 accuracy, which has not been improved in the past decade. Here we present DeepCNF (Deep Convolutional Neural Fields) for protein SS prediction. DeepCNF is a Deep Learning extension of Conditional Neural Fields (CNF), which is an integration of Conditional Random Fields (CRF) and shallow neural networks. DeepCNF can model not only complex sequence-structure relationship by a deep hierarchical architecture, but also interdependency between adjacent SS labels, so it is much more powerful than CNF. Experimental results show that DeepCNF can obtain ~84\% Q3 accuracy, ~85\% SOV score, and ~72\% Q8 accuracy, respectively, on the CASP and CAMEO test proteins, greatly outperforming currently popular predictors. As a general framework, DeepCNF can be used to predict other protein structure properties such as contact number, disorder regions, and solvent accessibility.},
}
@article{WangetalOP-13,
        title = {Going Deep: Brain Imaging with Multi-Photon Microscopy},
       author = {Ke Wang, Nicholas G. Horton and Chris Xu},
      journal = {Optics \& Photonics},
         year = 2013,
}
@article{WangetalPAMI-14,
       author = {J. Wang and N. Wang and Y. Jia and J. Li and G. Zeng and H. Zha and X. S. Hua},
      journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
        title = {Trinary-Projection Trees for Approximate Nearest Neighbor Search},
         year = {2014},
       volume = {36},
       number = {2},
        pages = {388-403},
     abstract = {We address the problem of approximate nearest neighbor (ANN) search for visual descriptor indexing. Most spatial partition trees, such as KD trees, VP trees, and so on, follow the hierarchical binary space partitioning framework. The key effort is to design different partition functions (hyperplane or hypersphere) to divide the points so that 1) the data points can be well grouped to support effective NN candidate location and 2) the partition functions can be quickly evaluated to support efficient NN candidate location. We design a trinary-projection direction-based partition function. The trinary-projection direction is defined as a combination of a few coordinate axes with the weights being 1 or -1. We pursue the projection direction using the widely adopted maximum variance criterion to guarantee good space partitioning and find fewer coordinate axes to guarantee efficient partition function evaluation. We present a coordinate-wise enumeration algorithm to find the principal trinary-projection direction. In addition, we provide an extension using multiple randomized trees for improved performance. We justify our approach on large-scale local patch indexing and similar image search.},
}
@misc{WangetalPFC-META-RL-18,
        title = {Prefrontal Cortex as a Meta-Reinforcement Learning System},
       author = {Jane X. Wang and Zeb Kurth-Nelson and Dharshan Kumaran and Dhruva Tirumala and Hubert Soyer and Joel Z. Leibo and Demis Hassabis and Matthew Botvinick},
 howpublished = {https://talks.ox.ac.uk/talks/id/f8d80f87-fb39-4473-af48-8403b9a355c8/},
         year = {2018},
     abstract = {Over the past twenty years, neuroscience research on reward-based learning has converged on a canonical model, under which the neurotransmitter dopamine 'stamps in' associations between situations, actions and rewards by modulating the strength of synaptic connections between neurons. However, a growing number of recent findings have placed this standard model under strain. In the present work, we draw on recent advances in artificial intelligence to introduce a new theory of reward-based learning. Here, the dopamine system trains another part of the brain, the prefrontal cortex, to operate as its own free-standing learning system. This new perspective accommodates the findings that motivated the standard model, but also deals gracefully with a wider range of observations, providing a fresh foundation for future research.}
}
@article{WangetalPLoS-17,
       author = {Wang, Sheng and Sun, Siqi and Li, Zhen and Zhang, Renyu and Xu, Jinbo},
      journal = {PLOS Computational Biology},
    publisher = {Public Library of Science},
        title = {Accurate De Novo Prediction of Protein Contact Map by Ultra-Deep Learning Model},
       volume = {13},
       number = {1},
         year = {2017},
        pages = {1-34},
     abstract = {Protein contact prediction and contact-assisted folding has made good progress due to direct evolutionary coupling analysis (DCA). However, DCA is effective on only some proteins with a very large number of sequence homologs. To further improve contact prediction, we borrow ideas from deep learning, which has recently revolutionized object recognition, speech recognition and the GO game. Our deep learning method can model complex sequence-structure relationship and high-order correlation (i.e., contact occurrence patterns) and thus, improve contact prediction accuracy greatly. Our test results show that our method greatly outperforms the state-of-the-art methods regardless how many sequence homologs are available for a protein in question. Ab initio folding guided by our predicted contacts may fold many more test proteins than the other contact predictors. Our contact-assisted 3D models also have much better quality than homology models built from the training proteins, especially for membrane proteins. One interesting finding is that even trained mostly with soluble proteins, our method performs very well on membrane proteins. Recent blind CAMEO test confirms that our method can fold large proteins with a new fold and only a small number of sequence homologs.},
}
@article{WangetalTIP-04,
  author = {Z. Wang and A. Bovik and H. Sheikh and E. Simoncelli},
   title = {Image quality assessment: From error visibility to structural similarity},
 journal = {{IEEE} Transactions on Image Processing},
    year = 2004,
  volume = 13,
   issue = 4,
   pages = {600-612},
}
@article{WangetalTNN-10,
     title = {Linear time maximum margin clustering},
    author = {Fei Wang and Bin Zhao and Changshui Zhang},
   journal = {{IEEE} Transactions on Neural Networks},
    volume = 21,
     issue = 2,
      year = 2010,
     pages = {319-332},
}
@misc{WanneretalLARVAL_ZEBRAFISH_OLFACTORY_BULB-16,
       author = {Wanner, Adrian and Genoud, Christel and Friedrich, Rainer},
        title = {Dense skeleton reconstructions of neurons in the larval zebrafish olfactory bulb},
 howpublished = {Zenodo (Data set)},
         year = 2016,
          url = {https://doi.org/10.5281/zenodo.58985}
}
@article{WanneretalNATURE-NEUROSCIENCE-16,
       author = {Wanner, Adrian and Genoud, Christel and Masudi, Tafheem and Siksou, Lea and Friedrich, Rainer},
        title = {Dense {EM}-based reconstruction of the interglomerular projectome in the zebrafish olfactory bulb},
      journal = {Nature Neuroscience},
    publisher = {Nature Publishing Group},
       volume = {19},
        issue = {6},
         year = 2016,
        pages = {816-825},
     abstract = {The dense reconstruction of neuronal circuits from volumetric electron microscopy (EM) data has the potential to uncover fundamental structure-function relationships in the brain. To address bottlenecks in the workflow of this emerging methodology, we developed a procedure for conductive sample embedding and a pipeline for neuron reconstruction. We reconstructed [sim]98\% of all neurons (>1,000) in the olfactory bulb of a zebrafish larva with high accuracy and annotated all synapses on subsets of neurons representing different types. The organization of the larval olfactory bulb showed marked differences from that of the adult but similarities to that of the insect antennal lobe. Interneurons comprised multiple types but granule cells were rare. Interglomerular projections of interneurons were complex and bidirectional. Projections were not random but biased toward glomerular groups receiving input from common types of sensory neurons. Hence, the interneuron network in the olfactory bulb exhibits a specific topological organization that is governed by glomerular identity.},
}
@article{WardenetalNATURE-12,
        title = {A prefrontal cortex-brainstem neuronal projection that controls response to behavioural challenge},
       author = {Warden, M.R. and Selimbeyoglu, A. and Mirzabekov, J.J. and Lo, M. and Thompson, K.R. and Kim, S. and Adhikari, A. and Tye, K.M. and Frank, L.M. and Deisseroth, K.},
      journal = {Nature},
         year = 2012,
     abstract = {The prefrontal cortex (PFC) is thought to participate in high-level control of the generation of behaviours (including the decision to execute actions); indeed, imaging and lesion studies in human beings have revealed that PFC dysfunction can lead to either impulsive states with increased tendency to initiate action, or to amotivational states characterized by symptoms such as reduced activity, hopelessness and depressed mood. Considering the opposite valence of these two phenotypes as well as the broad complexity of other tasks attributed to PFC, we sought to elucidate the PFC circuitry that favours effortful behavioural responses to challenging situations. Here we develop and use a quantitative method for the continuous assessment and control of active response to a behavioural challenge, synchronized with single-unit electrophysiology and optogenetics in freely moving rats. In recording from the medial PFC (mPFC), we observed that many neurons were not simply movement-related in their spike-firing patterns but instead were selectively modulated from moment to moment, according to the animal's decision to act in a challenging situation. Surprisingly, we next found that direct activation of principal neurons in the mPFC had no detectable causal effect on this behaviour. We tested whether this behaviour could be causally mediated by only a subclass of mPFC cells defined by specific downstream wiring. Indeed, by leveraging optogenetic projection-targeting to control cells with specific efferent wiring patterns, we found that selective activation of those mPFC cells projecting to the brainstem dorsal raphe nucleus (DRN), a serotonergic nucleus implicated in major depressive disorder, induced a profound, rapid and reversible effect on selection of the active behavioural state. These results may be of importance in understanding the neural circuitry underlying normal and pathological patterns of action selection and motivation in behaviour.}
}
@techreport{WarrenTR-74,
       author = {Warren, David},
        title = {Warplan: A System for Generating Plans},
        issue = {Memo 76},
  institution = {Department of Computational Logic, University of Edinburgh},
         year = 1974,
}
@article{WatanabeetalNATURE-04,
        title = {{DNA} sequence and comparative analysis of chimpanzee chromosome 22},
       author = {Watanabe, H. {\it{etal}}},
      comment = {Watanabe, H. and Fujiyama, A. and Hattori, M. and Taylor . D. and Toyoda, A. and Kuroki, Y. and Noguchi, H. and BenKahla . and Lehrach, H. and Sudbrak, R. and Kube, M. and Taenzer . and Galgoczy, P. and Platzer, M. and Scharfe, M. and Nordsiek . and Blocker, H. and Hellmann, I. and Khaitovich, P. and Paabo . and Reinhardt, R. and Zheng, H. J. and Zhang, X. L. and Zhu . F. and Wang, B. F. and Fu, G. and Ren, S. X. and Zhao . P. and Chen, Z. and Lee, Y. S. and Cheong, J. E. and Choi . H. and Wu, K. M. and Liu, T. T. and Hsiao, K. J. and Tsai . F. and Kim, C. G. and OOta, S. and Kitano, T. and Kohara . and Saitou, N. and Park, H. S. and Wang, S. Y. and Yaspo, M. L. and Sakaki, Y.},
      journal = {Nature},
        issue = 6990,
        pages = {382-388},
       volume = 429,
         year = 2004,
     abstract = {Human-chimpanzee comparative genome research is ssential for narrowing down genetic changes involved in the cquisition of unique human features, such as highly developed ognitive functions, bipedalism or the use of complex anguage. Here, we report the high-quality DNA sequence of 33.3 egabases of chimpanzee chromosome 22. By comparing the whole equence with the human counterpart, chromosome 21, we found that .44\% of the chromosome consists of single-base substitutions in ddition to nearly 68,000 insertions or deletions. These ifferences are sufficient to generate changes in most of the roteins. Indeed, 83\% of the 231 coding sequences, including unctionally important genes, show differences at the amino acid equence level. Furthermore, we demonstrate different expansion of articular subfamilies of retrotransposons between the lineages suggesting different impacts of retrotranspositions on human and himpanzee evolution. The genomic changes after speciation and heir biological consequences seem more complex than originally ypothesized.},
}
@book{WatermanandHayes-Roth,
       author = {Waterman, D. A. and Hayes-Roth, Frederick},
        title = {Pattern-Directed Inference Systems},
    publisher = {Academic Press},
      address = {New York},
         year = 1978,
}
@phdthesis{WatkinsPhD,
       author = {Watkins, C. J. C. H},
        title = {Learning from Delayed Rewards},
       school = {Cambridge University},
         year = 1989,
}
@article{WatkinsetalNEUROPSYCHOLOGIA-03,
        title = {Seeing and hearing speech excites the motor system involved in speech production},
       author = {Watkins, K.E and Strafella, A.P. and Paus, T.},
      journal = {Neuropsychologia},
    publisher = {Elsevier},
       volume = {41},
       number = {8},
        pages = {989-994},
         year = {2003},
     abstract = {The perception of action is associated with increased activity in motor regions, implicating such regions in the recognition, understanding and imitation of actions. We examined the possibility that perception of speech, both auditory and visual, would also result in changes in the excitability of the motor system underlying speech production. Transcranial magnetic stimulation was applied to the face area of primary motor cortex to elicit motor-evoked potentials in the lip muscles. The size of the motor-evoked potentials was compared under the following conditions: listening to speech, listening to non-verbal sounds, viewing speech-related lip movements, and viewing eye and brow movements. Compared to control conditions, listening to and viewing speech enhanced the size of the motor-evoked potential. This effect was only seen in response to stimulation of the left hemisphere; stimulation of the right hemisphere produced no changes in motor-evoked potentials in any of the conditions. In a control experiment, the size of the motor-evoked potentials elicited in the muscles of the right hand did not differ among conditions, suggesting that speech-related changes in excitability are specific to the lip muscles. These results provide evidence that both auditory and visual speech perception facilitate the excitability of the motor system involved in speech production.}
}
@book{Watson1980,
       author = {James D. Watson},
        title = {The Double Helix: A Personal Account of the Discovery of the Structure of DNA},
    publisher = {W. W. Norton},
      address = {New York, NY},
         year = 1980,
}
@article{WatsonandAhumadaJOSA-85,
       author = {Andrew B. Watson and Albert J. Ahumada, Jr.},
        title = {Model of human visual-motion sensing},
      journal = {Journal of the Optical Society of America A},
        issue = 2,
    publisher = {Optical Society of America},
        pages = {322-341},
       volume = 2,
         year = 1985,
     abstract = {We propose a model of how humans sense the velocity of moving images. The model exploits constraints provided by human psychophysics, notably that motion-sensing elements appear tuned for two-dimensional spatial frequency, and by the frequency spectrum of a moving image, namely, that its support lies in the plane in which the temporal frequency equals the dot product of the spatial frequency and the image velocity. The first stage of the model is a set of spatial-frequency-tuned, direction-selective linear sensors. The temporal frequency of the response of each sensor is shown to encode the component of the image velocity in the sensor direction. At the second stage, these components are resolved in order to measure the velocity of image motion at each of a number of spatial locations and spatial frequencies. The model has been applied to several illustrative examples, including apparent motion, coherent gratings, and natural image sequences. The model agrees qualitatively with human perception.},
}
@article{WatsonandCrickNATURE-53a,
       author = {Watson, J. D. and Crick, F. H. C.},
        title = {Molecular Structure of Nucleic Acids: A Structure for Deoxyribose Nucleic Acid},
      journal = {Nature},
         year = {1953},
       volume = {171},
        issue = {4356},
        pages = {737-738},
}
@article{WatsonandCrickNATURE-53b,
       author = {Watson, J. D. and Crick, F. H. C.},
        title = {Genetical Implications of the Structure of Deoxyribonucleic Acid},
      journal = {Nature},
         year = {1953},
       volume = {171},
        issue = {4361},
        pages = {964-967},
}
@article{WatsonetalFiSN-14,
       author = {Watson, Thomas C. and Becker, Nadine and Apps, Richard and Jones, Matthew W.},
        title = {Back to front: cerebellar connections and interactions with the prefrontal cortex},
      journal = {Frontiers in Systems Neuroscience},
         year = {2014},
    publisher = {Frontiers Media S.A.},
       volume = {8},
        pages = {4},
     abstract = {Although recent neuroanatomical evidence has demonstrated closed-loop connectivity between prefrontal cortex and the cerebellum, the physiology of cerebello-cerebral circuits and the extent to which cerebellar output modulates neuronal activity in neocortex during behavior remain relatively unexplored. We show that electrical stimulation of the contralateral cerebellar fastigial nucleus (FN) in awake, behaving rats evokes distinct local field potential (LFP) responses (onset latency {\textasciitilde}13 ms) in the prelimbic (PrL) subdivision of the medial prefrontal cortex. Trains of FN stimulation evoke heterogeneous patterns of response in putative pyramidal cells in frontal and prefrontal regions in both urethane-anesthetized and awake, behaving rats. However, the majority of cells showed decreased firing rates during stimulation and subsequent rebound increases; more than 90\% of cells showed significant changes in response. Simultaneous recording of on-going LFP activity from FN and PrL while rats were at rest or actively exploring an open field arena revealed significant network coherence restricted to the theta frequency range (5-10 Hz). Granger causality analysis indicated that this coherence was significantly directed from cerebellum to PrL during active locomotion. Our results demonstrate the presence of a cerebello-prefrontal pathway in rat and reveal behaviorally dependent coordinated network activity between the two structures, which could facilitate transfer of sensorimotor information into ongoing neocortical processing during goal directed behaviors.},
}
@article{WattersetalCoRR-19,
       author = {Nicholas Watters and  Loic Matthey and  Matko Bosnjak and  Christopher P. Burgess and  Alexander Lerchner},
        title = {{COBRA:} Data-Efficient Model-Based {RL} through Unsupervised Object  Discovery and Curiosity-Driven Exploration},
      journal = {CoRR},
       volume = {arXiv:1905.09275},
         year = {2019},
     abstract = {Data efficiency and robustness to task-irrelevant perturbations are long-standing challenges for deep reinforcement learning algorithms. Here we introduce a modular approach to addressing these challenges in a continuous control environment, without using hand-crafted or supervised information. Our Curious Object-Based seaRch Agent (COBRA) uses task-free intrinsically motivated exploration and unsupervised learning to build object-based models of its environment and action space. Subsequently, it can learn a variety of tasks through model-based search in very few steps and excel on structured hold-out tests of policy robustness}
}
@incollection{WattersetalNIPS-17,
        title = {Visual Interaction Networks: Learning a Physics Simulator from Video},
       author = {Watters, Nicholas and Zoran, Daniel and Weber, Theophane and Battaglia, Peter and Pascanu, Razvan and Tacchetti, Andrea},
    booktitle = {Advances in Neural Information Processing Systems 30},
       editor = {I. Guyon and U. V. Luxburg and S. Bengio and H. Wallach and R. Fergus and S. Vishwanathan and R. Garnett},
    publisher = {Curran Associates, Inc.},
         year = {2017},
        pages = {4539-4547},
     abstract = {From just a glance, humans can make rich predictions about the future of a wide range of physical systems. On the other hand, modern approaches from engineering, robotics, and graphics are often restricted to narrow domains or require information about the underlying state. We introduce the Visual Interaction Network, a general-purpose model for learning the dynamics of a physical system from raw visual observations. Our model consists of a perceptual front-end based on convolutional neural networks and a dynamics predictor based on interaction networks. Through joint training, the perceptual front-end learns to parse a dynamic visual scene into a set of factored latent object representations. The dynamics predictor learns to roll these states forward in time by computing their interactions, producing a predicted physical trajectory of arbitrary length. We found that from just six input video frames the Visual Interaction Network can generate accurate future trajectories of hundreds of time steps on a wide range of physical systems. Our model can also be applied to scenes with invisible objects, inferring their future states from their effects on the visible objects, and can implicitly infer the unknown mass of objects. This work opens new opportunities for model-based decision-making and planning from raw sensory observations in complex physical environments.},
}
@book{Watts2003,
    author = {Duncan J. Watts},
     title = {Six Degrees: The Science of a Connected Age},
 publisher = {W. W. Norton},
   address = {New York, NY},
      year = 2003,
}
@book{Watts2011,
    author = {Duncan J. Watts},
     title = {Everything is Obvious: Once You Know the Answer},
 publisher = {Random House},
   address = {New York, NY},
      year = 2003,
}
@article{WattsetalSCI-02,
   author = {Duncan J. Watts and Peter Sheridan Dodds and M. E. J. Newman},
    title = {Identity and Search in Social Networks},
  journal = {Science},
   volume = 296,
    pages = {1302-1305},
     year = 2002,
}
@article{WayneetalCoRR-18,
       author = {Greg Wayne and Chia-Chun Hung and David Amos and Mehdi Mirza and Arun Ahuja and Agnieszka Grabska-Barwinska and Jack Rae and Piotr Mirowski and Joel Z. Leibo and Adam Santoro and Mevlana Gemici and Malcolm Reynolds and Tim Harley and Josh Abramson and Shakir Mohamed and Danilo Rezende and David Saxton and Adam Cain and Chloe Hillier and David Silver and Koray Kavukcuoglu and Matt Botvinick and Demis Hassabis and Timothy Lillicrap},
        title = {Unsupervised Predictive Memory in a Goal-Directed Agent},
      journal = {CoRR},
       volume = {arXiv:1803.10760},
         year = {2018},
     abstract = {Animals execute goal-directed behaviours despite the limited range and scope of their sensors. To cope, they explore environments and store memories maintaining estimates of important information that is not presently available. Recently, progress has been made with artificial intelligence (AI) agents that learn to perform tasks from sensory input, even at a human level, by merging reinforcement learning (RL) algorithms with deep neural networks, and the excitement surrounding these results has led to the pursuit of related ideas as explanations of non-human animal learning. However, we demonstrate that contemporary RL algorithms struggle to solve simple tasks when enough information is concealed from the sensors of the agent, a property called "partial observability". An obvious requirement for handling partially observed tasks is access to extensive memory, but we show memory is not enough; it is critical that the right information be stored in the right format. We develop a model, the Memory, RL, and Inference Network (MERLIN), in which memory formation is guided by a process of predictive modeling. MERLIN facilitates the solution of tasks in 3D virtual reality environments for which partial observability is severe and memories must be maintained over long durations. Our model demonstrates a single learning agent architecture that can solve canonical behavioural tasks in psychology and neurobiology without strong simplifying assumptions about the dimensionality of sensory input or the duration of experiences.},
}
@book{Weatherall2013,
        title = {The Physics of Wall Street: A Brief History of Predicting the Unpredictable},
       author = {Weatherall, J.O.},
    publisher = {Houghton Mifflin Harcourt},
         year = {2013},
}
@article{WebbetalPNAS-16,
       author = {Webb, T. W. and Igelstrom, K. M. and Schurger, A. and Graziano, M. S.},
        title = {Cortical networks involved in visual awareness independent of visual attention},
      journal = {Proceedings National Academy of Science U.S.A.},
         year = {2016},
       volume = {113},
       number = {48},
        pages = {13923-13928},
     abstract = {It is now well established that visual attention, as measured with standard spatial attention tasks, and visual awareness, as measured by report, can be dissociated. It is possible to attend to a stimulus with no reported awareness of the stimulus. We used a behavioral paradigm in which people were aware of a stimulus in one condition and unaware of it in another condition, but the stimulus drew a similar amount of spatial attention in both conditions. The paradigm allowed us to test for brain regions active in association with awareness independent of level of attention. Participants performed the task in an MRI scanner. We looked for brain regions that were more active in the aware than the unaware trials. The largest cluster of activity was obtained in the temporoparietal junction (TPJ) bilaterally. Local independent component analysis (ICA) revealed that this activity contained three distinct, but overlapping, components: a bilateral, anterior component; a left dorsal component; and a right dorsal component. These components had brain-wide functional connectivity that partially overlapped the ventral attention network and the frontoparietal control network. In contrast, no significant activity in association with awareness was found in the banks of the intraparietal sulcus, a region connected to the dorsal attention network and traditionally associated with attention control. These results show the importance of separating awareness and attention when testing for cortical substrates. They are also consistent with a recent proposal that awareness is associated with ventral attention areas, especially in the TPJ.}
}
@inproceedings{WeberIJCAI-89,
       author = {Weber, Jay C.},
        title = {A Parallel Algorithm for Statistical Belief Refinement and its use in Causal Reasoning},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
         year = 1989,
}
@article{WeberetalCoRR-17,
       author = {Theophane Weber and  S{\'{e}}bastien Racani{\`{e}}re and David P. Reichert and Lars Buesing and Arthur Guez and Danilo Jimenez Rezende and Adri{\`{a}} Puigdom{\`{e}}nech Badia and Oriol Vinyals and Nicolas Heess and Yujia Li and Razvan Pascanu and Peter Battaglia and David Silver and Daan Wierstra},
        title = {Imagination-Augmented Agents for Deep Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1707.06203},
         year = {2017},
     abstract = {We introduce Imagination-Augmented Agents (I2As), a novel architecture for deep reinforcement learning combining model-free and model-based aspects. In contrast to most existing model-based reinforcement learning and planning methods, which prescribe how a model should be used to arrive at a policy, I2As learn to interpret predictions from a learned environment model to construct implicit plans in arbitrary ways, by using the predictions as additional context in deep policy networks. I2As show improved data efficiency, performance, and robustness to model misspecification compared to several baselines.}
}
@inproceedings{WeberetalECCC-2000,
       author = {Weber, M. and Welling, M. and Perona, P.},
        title = {Unsupervised Learning of Models for Recognition},
    booktitle = {European Conference on Computer Vision: Volume {I}},
        pages = {18-32},
         year = 2000,
}
@inproceedings{WeberetalECCV-00,
       author = {Markus Weber and Max Welling and Pietro Perona},
        title = {Unsupervised Learning of Models for Recognition},
    booktitle = {Proceedings of the European Conference on Computer Vision},
        pages = {18-32},
         year = 2000,
}
@article{WeedenetalSCIENCE-12,
        title = {The Geometric Structure of the Brain Fiber Pathways},
       author = {Wedeen, Van J. and Rosene, Douglas L. and Wang, Ruopeng and Dai, Guangping and Mortazavi, Farzad and Hagmann, Patric and Kaas, Jon H. and Tseng, Wen-Yih I.},
      journal = {Science},
       volume = 335,
        issue = 6076,
         year = 2012,
        pages = {1628-1634},
     abstract = {The structure of the brain as a product of morphogenesis is difficult to reconcile with the observed complexity of cerebral connectivity. We therefore analyzed relationships of adjacency and crossing between cerebral fiber pathways in four nonhuman primate species and in humans by using diffusion magnetic resonance imaging.  The cerebral fiber pathways formed a rectilinear three-dimensional  grid continuous with the three principal axes of development.  Cortico-cortical pathways formed parallel sheets of interwoven paths  in the longitudinal and medio-lateral axes, in which major pathways  were local condensations. Cross-species homology was strong and  showed emergence of complex gyral connectivity by continuous  elaboration of this grid structure. This architecture naturally  supports functional spatio-temporal coherence, developmental  path-finding, and incremental rewiring with correlated adaptation  of structure and function in cerebral plasticity and evolution.},
}
@misc{WeerakoonPhD-16,
        title = {Artificial Potential Field and Feature Extraction Method for Mobile Robot Path Planning in Structured Environments},
       author = {W.M.M. Tharindu Weerakoon},
         year = 2016,
 howpublished = {Department of Brain Science and Engineering. Graduate School of Life Science and Systems Engineering Kyushu Institute of Technology},
}
@book{Wegner2002,
       author = {Daniel M. Wegner},
        title = {The Illusion of Conscious Will},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2002,
      comment = {The concept of transactive memory was proposed by Wegner in 1985. A transactive memory system is a system through which groups  collectively encode, store, and retrieve knowledge. Transactive  memory suggests an analysis not only of how couples and families  in close relationships coordinate memory and tasks at home, but  how teams, larger groups and organizations come to develop a "group  mind", a memory system that is more complex and potentially more effective than that of any of the individuals that comprise it. --- \urlh{http://en.wikipedia.org/wiki/Daniel_Wegner}{Wikipedia}},
}
@article{WegnerBBS-04,
       author = {Daniel M. Wegner},
      journal = {Behavioural and Brain Sciences},
        issue = 5,
        title = {{Pr\'{e}cis of The Illusion of Conscious Will}},
       volume = 27,
         year = 2004,
     abstract = {The experience of conscious will is the feeling that we are doing things. This feeling occurs for many things we do, conveying to us again and again the sense that we consciously cause our actions. But the feeling may not be a true reading of what is happening in our minds, brains, and bodies as our actions are produced. The feeling of conscious will can be fooled. This happens in clinical disorders such as alien hand syndrome, dissociative identity disorder, and schizophrenic auditory hallucinations. And in people without disorders, phenomena such as hypnosis, automatic writing, Ouija board spelling, water dowsing, facilitated communication, speaking in tongues, spirit possession, and trance channeling also illustrate anomalies of will--cases when actions occur without will or will occurs without action. This book brings these cases together with research evidence from laboratories in psychology to explore a theory of apparent mental causation. According to this theory, when a thought appears in consciousness just prior to an action, is consistent with the action, and appears exclusive of salient alternative causes of the action, we experience conscious will and ascribe authorship to ourselves for the action. Experiences of conscious will thus arise from processes whereby the mind interprets itself --- not from processes whereby mind creates action. Conscious will, in this view, is an indication that we think we have caused an action, not a revelation of the causal sequence by which the action was produced.},
}
@article{WegnerIEEE-92,
       author = {Wegner, Peter},
        title = {Dimensions of object-oriented modeling},
      journal = {{IEEE} Computer},
       volume = 25,
        issue = 10,
        pages = {12-21},
         year = 1992,
}
@inproceedings{WeiandCaines,
       author = {Y.J. Wei and P. E. Caines},
        title = {Hierarchical {COCOLOG} for Finite Machines},
    booktitle = {Proceedings of 11th {INRIA} Conference on Analysis and Optimization of Systems},
      address = {Sophia-Antipolis, France},
         year = 1994,
}
@article{WeibeletalPNAS-05,
        title = {Microoxen: microorganisms to move microscale loads},
       author = {Weibel, D.B. and Garstecki, P. and Ryan, D. and DiLuzio, W.R. and Mayer, M. and Seto, J.E. and Whitesides, G.M.},
      journal = {Proceedings of the National Academy of Science},
       volume = 102,
        issue = 34,
         year = 2005,
        pages = {11963-11967},
     abstract = {It is difficult to harness the power generated by biological motors to carry out mechanical work in systems outside the cell. Efforts to capture the mechanical energy of nanomotors ex vivo require in vitro reconstitution of motor proteins and, often, protein engineering. This study presents a method for harnessing the power produced by biological motors that uses intact cells. The unicellular, biflagellated algae Chlamydomonas reinhardtii serve as "microoxen." This method uses surface chemistry to attach loads (1- to 6-microm-diameter polystyrene beads) to cells, phototaxis to steer swimming cells, and photochemistry to release loads. These motile microorganisms can transport microscale loads (3-microm-diameter beads) at velocities of approximately 100-200 microm.sec(-1) and over distances as large as 20 cm.}
}
@inproceedings{Weickert-97,
       author = {Joachim Weickert},
        title = {A Review of Nonlinear Diffusion Filtering},
    booktitle = {Scale-Space Theories in Computer Vision},
        pages = {3-28},
         year = 1997,
}
@inproceedings{WeietalNIPS-19,
        title = {Code Generation as a Dual Task of Code Summarization},
       author = {Wei, Bolin and Li, Ge and Xia, Xin and Fu, Zhiyi and Jin, Zhi},
    booktitle = {Advances in Neural Information Processing Systems 32},
       editor = {H. Wallach and H. Larochelle and A. Beygelzimer and F. d' Alch\'{e}-Buc and E. Fox and R. Garnett},
    publisher = {Curran Associates, Inc.},
         year = {2019},
        pages = {6563-6573},
     abstract = {Code summarization (CS) and code generation (CG) are two crucial tasks in the field of automatic software development. Various neural network-based approaches are proposed to solve these two tasks separately. However, there exists a specific intuitive correlation between CS and CG, which has not been exploited in previous work. In this paper, we apply the relations between two tasks to improve the performance of both tasks. In other words, exploiting the duality between the two tasks, we propose a dual training framework to train the two tasks simultaneously. In this framework, we consider the dualities on probability and attention weights, and design corresponding regularization terms to constrain the duality. We evaluate our approach on two datasets collected from GitHub, and experimental results show that our dual framework can improve the performance of CS and CG tasks over baselines.}
}
@book{WeigendandGershenfeld94,
       editor = {Weigend, Andreas S. and Gershenfeld, Neil A.},
        title = {Time Series Prediction: Forecasting the Future and Understanding the Past},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1994,
}
@incollection{WeilleretalNL-16,
       author = {Cornelius Weiller and Tobias Bormann and Dorothee Kuemmerer and Mariachristina Musso and Michel Rijntjes},
        title = {The Dual Loop Model in Language},
       editor = {Hickok, Gregory  and Small, Steven L.},
    booktitle = {Neurobiology of Language},
    publisher = {Academic Press},
      address = {San Diego},
         year = {2016},
        pages = {325-337},
     abstract = {A dual loop model for the processing of language in the brain was proposed early in history and has also formed the basis for many neuropsychological models. These models incorporate a (direct) route for sensorimotor mapping and a (indirect) route for "semantic" processing. Dual loop models also emerged in the field of visual processing, motor control, or spatial attention. Thus, a general dual loop system may provide the framework for the interpretation of cognition in human and primate brains independent of the modality and species. Modern imaging techniques like diffusion tensor imaging (DTI)-based fiber tracking identified long human association tracts for ventral and dorsal pathways. The extreme capsule (EmC) and uncinate fascicle (UF) are part of the ventral system, and the superior longitudinal fasciculi (SLF I, II, III) and the arcuate fasciculus (AF) are all dorsal pathways. This chapter reviews language processing in the brain in the context of a domain general dual loop system.}
}
@inproceedings{WeinbergeretalICML-04,
       author = {Kilian Q. Weinberger and Fei Sha and Lawrence K. Saul},
        title = {Learning a kernel matrix for nonlinear dimensionality reduction},
    booktitle = {Proceedings of the twenty-first international conference on Machine learning},
        pages = 106,
     location = {Banff, Alberta, Canada},
         year = 2004,
}
@article{WeinbergeretalIEEE-82,
       author = {Weinberger, M. J. and Lempel, A. and Ziv, J.},
        title = {A Sequential Algorithm for the Universal Coding of Finite Memory Sources},
      journal = {{IEEE} Transactions on Information Theory},
       volume = 38,
         year = 1982,
}
@inproceedings{WeinbergeretalNIPS-05,
       author = {Kilian Weinberger and John Blitzer and Lawrence Saul},
        title = {Distance Metric Learning for Large Margin Nearest Neighbor Classification},
    booktitle = {Advances in Neural Information Processing Systems 18},
         year = 2005,
}
@Article{WeinerGrill-SpectorPSYCHOLOGICAL-RESEARCH-11,
       author = {Weiner, K. S. and Grill-Spector, K.},
        title = {{{N}eural representations of faces and limbs neighbor in human high-level visual cortex: evidence for a new organization principle}},
      journal = {Psychological Research},
         year = {2011},
       volume = {77},
       number = {1},
        pages = {74-97},
     abstract = {Neurophysiology and optical imaging studies in monkeys and functional magnetic resonance imaging (fMRI) studies in both monkeys and humans have localized clustered neural responses in inferotemporal cortex selective for images of biologically relevant categories, such as faces and limbs. Using higher resolution (1.5 mm voxels) fMRI scanning methods than past studies (3-5 mm voxels), we recently reported a network of multiple face- and limb-selective regions that neighbor one another in human ventral temporal cortex (Weiner and Grill-Spector, Neuroimage, 52(4):1559-1573, 2010) and lateral occipitotemporal cortex (Weiner and Grill-Spector, Neuroimage, 56(4):2183-2199, 2011). Here, we expand on three basic organization principles of high-level visual cortex revealed by these findings: (1) consistency in the anatomical location of functional regions, (2) preserved spatial relationship among functional regions, and (3) a topographic organization of face- and limb-selective regions in adjacent and alternating clusters. We highlight the implications of this structure in comparing functional brain organization between typical and atypical populations. We conclude with a new model of high-level visual cortex consisting of ventral, lateral, and dorsal components, where multimodal processing related to vision, action, haptics, and language converges in the lateral pathway.}
}
@article{WeinhardetalNATURE-COMMUNICATIONS-18,
       author = {Weinhard, Laetitia and di Bartolomei, Giulia and Bolasco, Giulia and Machado, Pedro and Schieber, Nicole L. and Neniskyte, Urte and Exiga, Melanie and Vadisiute, Auguste and Raggioli, Angelo and Schertel, Andreas and Schwab, Yannick and Gross, Cornelius T.},
        title = {Microglia remodel synapses by presynaptic trogocytosis and spine head filopodia induction},
      journal = {Nature Communications},
         year = {2018},
       volume = {9},
        issue = {1},
        pages = {1228},
     abstract = {Microglia are highly motile glial cells that are proposed to mediate synaptic pruning during neuronal circuit formation. Disruption of signaling between microglia and neurons leads to an excess of immature synaptic connections, thought to be the result of impaired phagocytosis of synapses by microglia. However, until now the direct phagocytosis of synapses by microglia has not been reported and fundamental questions remain about the precise synaptic structures and phagocytic mechanisms involved. Here we used light sheet fluorescence microscopy to follow microglia-synapse interactions in developing organotypic hippocampal cultures, complemented by a 3D ultrastructural characterization using correlative light and electron microscopy (CLEM). Our findings define a set of dynamic microglia-synapse interactions, including the selective partial phagocytosis, or trogocytosis (trogo-: nibble), of presynaptic structures and the induction of postsynaptic spine head filopodia by microglia. These findings allow us to propose a mechanism for the facilitatory role of microglia in synaptic circuit remodeling and maturation.},
}
@book{WeinsteinandFineberg80,
       author = {Weinstein, Milton C. and Fineberg, Harvey V.},
        title = {Clinical Decision Analysis},
    publisher = {W. B. Saunders Co.},
      address = {Philadelphia},
         year = 1980,
}
@article{WeintraubSCIAM-19,
       author = {Karen Weintraub},
        title = {The Adult Brain Does Grow New Neurons After All, Study Says},
      journal = {Scientific American},
       volume = {March 25},
         year = 2019,
}
@inproceedings{WeissICCV-99,
       author = {Y. Weiss},
    booktitle = {Proceedings of the Seventh IEEE International Conference on Computer Vision},
        title = {Segmentation using eigenvectors: a unifying view},
         year = {1999},
       volume = {2},
        pages = {975-982},
     abstract = {Automatic grouping and segmentation of images remains a challenging problem in computer vision. Recently, a number of authors have demonstrated good performance on this task using methods that are based on eigenvectors of the affinity matrix. These approaches are extremely attractive in that they are based on simple eigen-decomposition algorithms whose stability is well understood. Nevertheless, the use of eigendecompositions in the context of segmentation is far from well understood. In this paper we give a unified treatment of these algorithms, and show the close connections between them while highlighting their distinguishing features. We then prove results on eigenvectors of block matrices that allow us to analyze the performance of these algorithms in simple grouping settings. Finally, we use our analysis to motivate a variation on the existing methods that combines aspects from different eigenvector segmentation algorithms. We illustrate our analysis with results on real and synthetic images},
}
@article{WeissandFreemanNC-02,
       author = {Jair Weiss and William T. Freeman},
        title = {Correctness of Belief Propagation in Gaussian Graphical Models of Arbitrary Topology},
      journal = {Neural Computation},
    publisher = {MIT Press},
      address = {Cambridge, MA},
       volume = 13,
        issue = 10,
        pages = {2173-2200},
         year = 2001,
}
@inproceedings{WeissetalNIPS-08,
       author = {Yair Weiss and Antonio B. Torralba and Robert Fergus},
        title = {Spectral Hashing},
       editor = {Daphne Koller and Dale Schuurmans and Yoshua Bengio and L{\'e}on Bottou},
    booktitle = {Advances in Neural Information Processing Systems 21, Proceedings of the Twenty-Second Annual Conference on Neural Information Processing Systems},
         year = 2008,
        pages = {1753-1760},
}
@article{Weizenbaum66,
       author = {Joseph Weizenbaum},
        title = {{ELIZA} -- A Computer Program For the Study of Natural Language Communication Between Man and Machine},
      journal = {Communications of the {ACM}},
       volume = 9,
        issue = 1,
         year = 1966,
        pages = {36-45},
}
@article{WeldAIM-94,
       author = {Weld, Daniel S.},
        title = {An Introduction to Partial-Order Planning},
      journal = {AI Magazine},
       volume = {Winter},
         year = 1994,
}
@incollection{WellingetalNIPS-03,
       author = {Max Welling and Geoffrey Hinton and Simon Osindero},
        title = {Learning Sparse Topographic Representations with Products of {S}tudent-t Distributions},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = {15},
       editor = {S. Becker, S. Thrun and K. Obermayer},
    publisher = {MIT Press},
      address = {Cambridge, MA},
        pages = {1359-1366},
         year = 2003,
}
@inproceedings{WellmanAAAI-90,
       author = {Wellman, Michael P.},
        title = {The {STRIPS} Assumption for Planning Under Uncertainty},
    booktitle = {Proceedings AAAI-90},
     location = {Boston, Massachusetts},
 organization = {AAAI},
         year = 1990,
        pages = {198-203},
}
@inproceedings{WellmanAAAI-91,
       author = {Wellman, Michael P. and Doyle, Jon},
        title = {Preferential semantics for goals},
    booktitle = {Proceedings AAAI-91},
     location = {Anaheim, California},
 organization = {AAAI},
         year = 1991,
}
@article{WellmanAIJ-90,
       author = {Wellman, Michael P.},
        title = {Fundamental Concepts of Qualitative Probabilistic Networks},
      journal = AIJ,
       volume = 44,
        pages = {257-303},
         year = 1990,
}
@inproceedings{WellmanIJCAI-87,
       author = {Wellman, Michael P.},
        title = {Dominance and Subsumption in Constraint-Posting Planning},
    booktitle = {Proceedings {IJCAI} 10},
     location = {Milan, Italy},
 organization = {IJCAII},
         year = 1987,
}
@techreport{WellmanPhD,
       author = {Wellman, Michael P.},
        title = {Formulation of Tradeoffs in Planning Under Uncertainty},
        issue = {MIT/LCS/TR-427},
  institution = {{MIT} {AI} Laboratory},
         year = 1988,
}
@article{WellmanandWahRSF-17,
       author = {Michael P. Wellman and Elaine Wah},
        title = {Strategic Agent-Based Modeling of Financial Markets},
      journal = {The Russell Sage Foundation Journal of the Social Sciences},
       volume = {3},
       number = {1},
        pages = {104-119},
         year = {2017},
     abstract = {Understanding the implications of algorithmic trading calls for modeling financial markets at a level of fidelity that often precludes analytic solution. We describe how agent-based simulation modeling can be combined with game-theoretic reasoning to examine the effects of market variables on outcomes of interest. The approach is illustrated in a basic model where investors trade a single security through a continuous double auction mechanism. Our results demonstrate the feasibility of the approach, and raise questions about the use of spreads as a proxy for trading cost and welfare.}
}
@article{WenetalNEURON-12,
       author = {Wen, Q. and Po, M. D. and Hulme, E. and Chen, S. and Liu, X. and Kwok, S. W. and Gershow, M. and Leifer, A. M. and Butler, V. and Fang-Yen, C. and Kawano, T. and Schafer, W. R. and Whitesides, G. and Wyart, M. and Chklovskii, D. B. and Zhen, M. and Samuel, A. D.},
        title = {Proprioceptive coupling within motor neurons drives {\it{Caenorhabditis elegans}} forward locomotion},
      journal = {Neuron},
         year = {2012},
       volume = {76},
       number = {4},
        pages = {750-761},
     abstract = {Locomotion requires coordinated motor activity throughout an animal's body. In both vertebrates and invertebrates, chains of coupled central pattern generators (CPGs) are commonly evoked to explain local rhythmic behaviors. In C. elegans, we report that proprioception within the motor circuit is responsible for propagating and coordinating rhythmic undulatory waves from head to tail during forward movement. Proprioceptive coupling between adjacent body regions transduces rhythmic movement initiated near the head into bending waves driven along the body by a chain of reflexes. Using optogenetics and calcium imaging to manipulate and monitor motor circuit activity of moving C. elegans held in microfluidic devices, we found that the B-type cholinergic motor neurons transduce the proprioceptive signal. In C. elegans, a sensorimotor feedback loop operating within a specific type of motor neuron both drives and organizes body movement.}
}
@article{WengandChenIS-00,
       author = {Juyang Weng and Shaoyun Chen},
        title = {Visual Learning with Navigation as an Example},
      journal = {{IEEE} Intelligent Systems},
       volume = 15,
        issue = 5,
          url = {http://www.computer.org/intelligent/ex2000/x5063abs.htm},
         year = 2000,
        pages = {63-71},
}
@article{WennekersetalNC-99,
       author = {Thomas Wennekers and Friedrich T. Sommer and Gunther Palm},
        title = {Iterative retrieval in associative memories by threshold control of different neural models},
      journal = {Neural Computation},
       volume = 11,
         year = 1999,
        pages = {21-66},
     abstract = {We investigate the retrieval properties of Hebbian auto-associative memories in the limit of sparse coding. Appropriately chosen threshold control strategies in finite size associative memories increase the completion capacity for iterative retrieval (and even for the very fast two-step retrieval) above the asymptotic capacity for extremely large networks. We relate these results to a biologically motivated network consisting of excitatorily coupled cells which are controlled by a globally acting inhibitory interneuron. Choosing a homogenous coupling matrix and different excitatory single neuron types, we find in an explicit numerical comparison, that the global behavior of spiking neurons in general is different from rate-function and probabilistic binary units. We also show that a network of spiking neurons with Hebbian coupling matrix is able to complete and segregate several distorted patterns that are simultaneously presented at the input. In this case the global dynamics falls into rhythmic activity and processes one input pattern per period, a behavior that might be related to rhythmic cortical activity as found, for example, in the visual cortices of cats and monkeys.}
}
@techreport{WerbosPhD,
       author = {Werbos, P.J.},
        title = {Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Sciences},
         type = {{Ph.D.} Thesis},
  institution = {Harvard University},
         year = 1974,
}
@article{WersingandKoernerNC-03,
        title = {Learning Optimized Features for Hierarchical Models of Invariant Object Recognition},
       author = {Heiko Wersing and Edgar K\"{o}rner},
      journal = {Neural Computation},
       volume = 15,
        issue = 7,
        pages = {1559-1588},
         year = 2003,
}
@inproceedings{WersingetalICANN-03,
       author = {Heiko Wersing and Julian Eggert and Edgar K\"{o}rner},
        title = {Sparse Coding with Invariance Constraints},
    booktitle = ICANN,
        pages = {385-392},
         year = 2003,
}
@article{Wertheimer23,
       author = {Wertheimer, M.},
        title = {Untersuchungen zur Lehre von der Gestalt, {II} ({I}nvestigations of the principles of Gestalt, {II})},
      journal = {Psychologische Forschung},
       volume = 4,
        pages = {301-350},
         year = 1923,
}
@article{WestbrookandBraverNEURON-16,
       author = {Westbrook, A. and Braver, T. S.},
        title = {Dopamine Does Double Duty in Motivating Cognitive Effort},
      journal = {Neuron},
         year = {2016},
       volume = {89},
       number = {4},
        pages = {695-710},
     abstract = {Cognitive control is subjectively costly, suggesting that engagement is modulated in relationship to incentive state. Dopamine appears to play key roles. In particular, dopamine may mediate cognitive effort by two broad classes of functions: (1) modulating the functional parameters of working memory circuits subserving effortful cognition, and (2) mediating value-learning and decision-making about effortful cognitive action. Here, we tie together these two lines of research, proposing how dopamine serves "double duty", translating incentive information into cognitive motivation.}
}
@article{WestbrookandFrankCOiBS-18,
       author = {Westbrook, A. and Frank, M.},
        title = {Dopamine and Proximity in Motivation and Cognitive Control},
      journal = {Current Opinion in Behavioral Science},
       volume = {22},
         year = {2018},
        pages = {28-34},
     abstract = {Cognitive control - the ability to override a salient or prepotent action to execute a more deliberate one - is required for flexible, goal-directed behavior, and yet it is subjectively costly: decision-makers avoid allocating control resources, even when doing so affords more valuable outcomes. Dopamine likely offsets effort costs just as it does for physical effort. And yet, dopamine can also promote impulsive action, undermining control. We propose a novel hypothesis that reconciles opposing effects of dopamine on cognitive control: during action selection, striatal dopamine biases benefits relative to costs, but does so preferentially for "proximal" motor and cognitive actions. Considering the nature of instrumental affordances and their dynamics during action selection facilitates a parsimonious interpretation and conserved corticostriatal mechanisms across physical and cognitive domains.}
}
@article{WestonCoRR-16,
       author = {Jason Weston},
        title = {Dialog-based Language Learning},
      journal = {CoRR},
       volume = {arXiv:1604.06045},
         year = {2016},
     abstract = {A long-term goal of machine learning research is to build an intelligent dialog agent. Most research in natural language understanding has focused on learning from fixed training sets of labeled data, with supervision either at the word level (tagging, parsing tasks) or sentence level (question answering, machine translation). This kind of supervision is not realistic of how humans learn, where language is both learned by, and used for, communication. In this work, we study dialog-based language learning, where supervision is given naturally and implicitly in the response of the dialog partner during the conversation. We study this setup in two domains: the bAbI dataset of (Weston et al., 2015) and large-scale question answering from (Dodge et al., 2015). We evaluate a set of baseline learning strategies on these tasks, and show that a novel model incorporating predictive lookahead is a promising approach for learning from a teacher's response. In particular, a surprising result is that it can learn to answer questions correctly without any reward-based supervision at all.}
}
@article{WestonetalCoRR-14,
       author = {Jason Weston and Sumit Chopra and Antoine Bordes},
        title = {Memory Networks},
      journal = {CoRR},
       volume = {arXiv:1410.3916},
         year = 2014,
          url = {http://arxiv.org/abs/1410.3916},
     abstract = {We describe a new class of learning models called memory networks. Memory networks reason with inference components combined with a long-term memory component; they learn how to use these jointly. The long-term memory can be read and written to, with the goal of using it for prediction. We investigate these models in the context of question answering (QA) where the long-term memory effectively acts as a (dynamic) knowledge base, and the output is a textual response. We evaluate them on a large-scale QA task, and a smaller, but more complex, toy task generated from a simulated world. In the latter, we show the reasoning power of such models by chaining multiple supporting sentences to answer questions that require understanding the intension of verbs.}
}
@article{WestonetalCoRR-15,
       author = {Jason Weston and Antoine Bordes and Sumit Chopra and Tomas Mikolov},
        title = {Towards AI-Complete Question Answering: {A} Set of Prerequisite Toy Tasks},
      journal = {CoRR},
       volume = {arXiv:1502.05698},
         year = {2015},
     abstract = {One long-term goal of machine learning research is to produce methods that are applicable to reasoning and natural language, in particular building an intelligent dialogue agent. To measure progress towards that goal, we argue for the usefulness of a set of proxy tasks that evaluate reading comprehension via question answering. Our tasks measure understanding in several ways: whether a system is able to answer questions via chaining facts, simple induction, deduction and many more. The tasks are designed to be prerequisites for any system that aims to be capable of conversing with a human. We believe many existing learning systems can currently not solve them, and hence our aim is to classify these tasks into skill sets, so that researchers can identify (and then rectify) the failings of their systems. We also extend and improve the recently introduced Memory Networks model, and show it is able to solve some, but not all, of the tasks.}
}
@inproceedings{WestonetalICML-08,
       author = {Weston, Jason and Ratle, Fr{\'e}d{\'e}ric and Collobert, Ronan},
        title = {Deep learning via semi-supervised embedding},
    booktitle = {Proceedings of the 25th International Conference on Machine Learning},
    publisher = {ACM},
      address = {New York, NY, USA},
     location = {Helsinki, Finland},
         year = 2008,
        pages = {1168-1175},
}
@incollection{WestonetalToT-12,
        title = {Deep Learning via Semi-supervised Embedding},
       author = {Weston, Jason and Ratle, Frederic and Mobahi, Hossein and Collobert, Ronan},
    booktitle = {Neural Networks: Tricks of the Trade},
       volume = 7700,
       series = {Lecture Notes in Computer Science},
       editor = {Montavon, Gregoire and Orr, Genevieve. and Muller, Klaus-Robert},
    publisher = {Springer Berlin Heidelberg},
         year = 2012,
        pages = {639-655}
}
@article{WestphaletalSCIENCE-08,
       author = {Westphal, Volker and Rizzoli, Silvio O. and Lauterbach, Marcel A. and Kamin, Dirk and Jahn, Reinhard and Hell, Stefan W.},
        title = {Video-Rate Far-Field Optical Nanoscopy Dissects Synaptic Vesicle Movement},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {320},
       number = {5873},
         year = {2008},
        pages = {246-249},
     abstract = {We present video-rate (28 frames per second) far-field optical imaging with a focal spot size of 62 nanometers in living cells. Fluorescently labeled synaptic vesicles inside the axons of cultured neurons were recorded with stimulated emission depletion (STED) microscopy in a 2.5-micrometer by 1.8-micrometer field of view. By reducing the cross-sectional area of the focal spot by about a factor of 18 below the diffraction limit (260 nanometers), STED allowed us to map and describe the vesicle mobility within the highly confined space of synaptic boutons. Although restricted within boutons, the vesicle movement was substantially faster in nonbouton areas, consistent with the observation that a sizable vesicle pool continuously transits through the axons. Our study demonstrates the emerging ability of optical microscopy to investigate intracellular physiological processes on the nanoscale in real time.},
}
@book{Westraetal2007,
        title = {Oscillators and Oscillator Systems: Classification, Analysis and Synthesis},
       author = {Jan R. Westra and Chris J.M. Verhoeven and Arthur van Roermund},
    publisher = {Springer-Verlag},
      address = {Berlin, Heidelberg},
         year = 2007,
}
@article{WeyhrauchAIJ-80,
       author = {Weyhrauch, Richard W.},
        title = {Prolegomena to a Theory of Mechanized Formal Reasoning},
      journal = AIJ,
       volume = 13,
         year = 1980,
        pages = {133-170},
}
@article{WheeleretalNATURE-NEUROSCIENCE-16,
       author = {Wheeler, Michael A. and Smith, Cody J.and Ottolini, Matteo and Barker, Bryan S. and Purohit, Aarti M. and Grippo, Ryan M. and Gaykema, Ronald P. and Spano, Anthony J. and Beenhakker, Mark P. and Kucenas, Sarahand Patel, Manoj K. and Deppmann, Christopher D. and Guler, Ali D.},
        title = {Genetically targeted magnetic control of the nervous system},
      journal = {Nature Neuroscience},
         year = {2016},
    publisher = {Nature Publishing Group},
       volume = {Advance Online Publication},
     abstract = {Optogenetic and chemogenetic actuators are critical for deconstructing the neural correlates of behavior. However, these tools have several limitations, including invasive modes of stimulation or slow on/off kinetics. We have overcome these disadvantages by synthesizing a single-component, magnetically sensitive actuator, Magneto, comprising the cation channel TRPV4 fused to the paramagnetic protein ferritin. We validated noninvasive magnetic control over neuronal activity by demonstrating remote stimulation of cells using in vitro calcium imaging assays, electrophysiological recordings in brain slices, in vivo electrophysiological recordings in the brains of freely moving mice, and behavioral outputs in zebrafish and mice. As proof of concept, we used Magneto to delineate a causal role of striatal dopamine receptor 1 neurons in mediating reward behavior in mice. Together our results present Magneto as an actuator capable of remotely controlling circuits associated with complex animal behaviors.},
}
@article{WhelanDECEREBRATED-CAT-96,
       author = {Whelan, P. J.},
        title = {{C}ontrol of locomotion in the decerebrate cat},
      journal = {Progress in Neurobiology},
       volume = {49},
       number = {5},
         year = {1996},
        pages = {481-515},
     abstract = {Many of the general concepts regarding the control of walking were described years ago by: Sherrington (1906) Integrative Actions of the Nervous System. Yale University Press: New Haven, CT; Sherrington (1910a) Remarks on the reflex mechanism of the step, Brain 33, 1-25; Sherrington (1910b) Flexor-reflex of the limb, crossed extension reflex, and reflex stepping and standing (cat and dog), J. Physiol. (Lond.) 40, 28-121; Sherrington (1931) Quantitative management of contraction in lowest level coordination, Brain 54, 1-28; Graham-Brown (1912) The intrinsic factors in the act of progression in the mammal, Proc. R. Soc. Lond. 84, 308-319; Graham-Brown (1914) On the nature of the fundamental activity of the nervous centres; together with an analysis of the conditioning of rhythmic activity in progression, and a theory of the evolution of function in the nervous system, J. Physiol. 49, 18-46; Graham-Brown (1915) On the activities of the central nervous system of the unborn foetus of the cat, with a discussion of the question whether progression (walking, etc.) is a 'learnt' complex, J. Physiol. 49, 208-215; Graham-Brown (1922) The physiology of stepping, J. Neur. Psychopathol. 3, 112-116. Only in recent years, however, have the mechanisms been analyzed in detail. Quite a few of these mechanisms have been described using the decerebrate cat. Locomotion is initiated in decerebrate cats by activation of the mesencephalic locomotor region (MLR) that activates the medial medullary reticular formation (MRF) which in turn projects axons to the spinal cord which descend within the ventrolateral funiculus (VLF). The MRF region regulates as well as initiates the stepping pattern and is thought to be involved in interlimb coordination. Afferent feedback from proprioceptors and exteroceptors can modify the ongoing locomotor pattern. Recently, the types of afferents responsible for signaling the stance to swing transition have been identified. A general rule states that if the limb is unloaded and the leg is extended, then swing will occur. The afferents that detect unloading of the limb are the Golgi tendon organs and stimulation of these afferents (at group I strengths) prolongs the stance phase in walking cats. The afferents that detect the extension of the leg have been found to be the length- and velocity-sensitive muscle afferents located in flexor muscles. Plasticity of locomotor systems is discussed briefly in this article. Descerebrate animals can adapt locomotor behaviors to respond to new environmental conditions. Oligosynaptic reflex pathways that control locomotion can be recalibrated after injury in a manner that appears to be functionally related to the recovery of the animal.}
}
@article{WhitakeretalDEVELOPMENTAL-SCIENCE-18,
       author = {Whitaker, Kirstie J. and Vendetti, Michael S. and Wendelken, Carter and Bunge, Silvia A.},
        title = {Neuroscientific insights into the development of analogical reasoning},
      journal = {Developmental science},
         year = {2018},
    publisher = {John Wiley and Sons Inc.},
       volume = {21},
        issue = {2},
        pages = {e12531},
     abstract = {Analogical reasoning, or the ability to find correspondences between entities based on shared relationships, supports knowledge acquisition. As such, the development of this ability during childhood is thought to promote learning. Here, we sought to better understand the mechanisms by which analogical reasoning about semantic relations improves over childhood and adolescence (e.g. chalk is to chalkboard as pen is to ?). We hypothesized that age-related differences would manifest as differences in the brain regions associated with one or more of the following cognitive functions: (1) controlled semantic retrieval, or the ability to retrieve task-relevant semantic associations; (2) response control, or the ability to override the tendency to respond to a salient distractor; and/or (3) relational integration, or the ability to consider jointly two mental relations. In order to test these hypotheses, we analyzed patterns of fMRI activation during performance of a pictorial propositional analogy task across 95 typically developing children between the ages of 6 and 18 years old. Despite large age-related differences in task performance, particularly over ages 6-10 but through to around age 14, participants across the whole age range recruited a common network of frontal, parietal and temporal regions. However, activation in a brain region that has been implicated in controlled semantic retrieval - left anterior prefrontal cortex (BA 47/45) - was positively correlated with age, and also with performance after controlling for age. This finding indicates that improved performance over middle childhood and early adolescence on this analogical reasoning task is driven largely by improvements in the ability to selectively retrieve task-relevant semantic relationships.},
}
@article{WhiteandEldeibOR-86,
       author = {White, C. C. and Eldeib, H. K.},
        title = {Parameter Imprecision in Finite State, Finite Action Dynamic Programs},
      journal = {Operations Research},
       volume = 34,
         year = 1986,
        pages = {120-129},
}
@article{WhiteandEldeibOR-94,
       author = {White, C. C. and Eldeib, H. K.},
        title = {{M}arkov Decision Processes with Imprecise Transition Probabilities},
      journal = {Operations Research},
       volume = 42,
         year = 1994,
        pages = {739-749},
}
@article{WhiteandSchererOR-89,
       author = {White, Chelsea C. {III} and Scherer, William T.},
        title = {Solution Procedures for Partially Observed {M}arkov Decision Processes},
      journal = {Operations Research},
       volume = 37,
        issue = 5,
         year = 1989,
        pages = {791-797},
}
@article{WhiteetalCELL-08,
       author = {White, R.M. and Sessa, A. and Burke, C. and Bowman, T. and LeBlanc, J. and Ceol, C. and Bourque, C. and Dovey, M. Goessling, W. and Burns, C.E. and Zon, L.I.},
        title = {Transparent adult zebrafish as a tool for in vivo transplantation analysis.},
      journal = {Cell Stem Cell},
       volume = 2,
        issue = 2,
         year = 2008,
        pages = {183-189},
}
@article{WhiteetalPTRSB-86,
       author = {White, J. G. and Southgate, E. and Thomson, J. N. and Brenner, S.},
        title = {The Structure of the Nervous System of the Nematode Caenorhabditis elegans},
      journal = {Philosophical Transactions of the Royal Society B: Biological Sciences},
       volume = 314,
        issue = 1165,
         year = 1986,
        pages = {1-340},
}
@inproceedings{WhiteheadAAAI-91,
       author = {Whitehead, Steven D.},
        title = {A Complexity Analysis of Cooperative Mechanisms in Reinforcement Learning},
    booktitle = {Proceedings AAAI-91},
     location = {Anaheim, California},
 organization = {AAAI},
         year = 1991,
}
@inproceedings{WhiteheadandBallardML-90,
       author = {Whitehead, Steven D. and Ballard, Dana H.},
        title = {Active Perception and Reinforcement Learning},
    booktitle = {Proceedings 7th International Conference on Machine Learning},
     location = {Austin, Texas},
         year = 1990,
}
@article{WhiteheadandBallardML-91,
       author = {Whitehead, Steven D. and Ballard, Dana H.},
        title = {Learning to Perceive and Act by Trial and Error},
      journal = {Machine Learning},
       volume = 7,
         year = 1991,
}
@article{WhiteheadandBallardNC-90,
       author = {Whitehead, Steven D. and Ballard, Dana H.},
        title = {Active Perception and Reinforcement Learning},
      journal = {Neural Computation},
       volume = {2},
       number = {4},
        pages = {409-419},
         year = {1990},
     abstract = {This paper considers adaptive control architectures that integrate active sensorimotor systems with decision systems based on reinforcement learning. One unavoidable consequence of active perception is that the agent's internal representation often confounds external world states. We call this phenomenon perceptual aliasing and show that it destabilizes existing reinforcement learning algorithms with respect to the optimal decision policy. A new decision system that overcomes these difficulties is described. The system incorporates a perceptual subcycle within the overall decision cycle and uses a modified learning algorithm to suppress the effects of perceptual aliasing. The result is a control architecture that learns not only how to solve a task but also where to focus its attention in order to collect necessary sensory information. }
}
@book{Whitmore,
       editor = {Whitmore, G. A. and Findlay, M. C.},
        title = {Stochastic Dominance: An Approach to Decision Making Under Risk},
    booktitle = {Stochastic Dominance: An Approach to Decision Making Under Risk},
    publisher = {D. C. Heath and Company},
      address = {Lexington, Massachusetts},
         year = 1978,
}
@article{Whitneyetal03,
       author = {Whitney, David and Goltz, Herbert C. and Thomas, Christopher G. and  Gati, Joseph S. and Menon, Ravi S. and Goodale, Melvyn A.},
        title = {Flexible Retinotopy: Motion-Dependent Position Coding in the Visual Cortex},
      journal = {Science},
       volume = 302,
        issue = 5646,
        pages = {878-881},
         year = 2003,
}
@book{Whittaker90,
       author = {Whittaker, J.},
        title = {Graphical Models in Applied Multivariate Statistics},
    publisher = {Wiley},
      address = {New York},
         year = 1990,
}
@article{WiddowsandCohenJIGPL-14,
       author = {Dominic Widdows and Trevor Cohen},
        title = {Reasoning with Vectors: A Continuous Model for Fast Robust Inference},
      journal = {Logic Journal of the IGPL},
       volume = {10.1093/jigpal/jzu028},
        issue = {0},
         year = {2014},
        pages = {000-000},
     abstract = {This article describes the use of continuous vector space models for reasoning with a formal knowledge base. The practical significance of these models is that they support fast, approximate but robust inference and hypothesis generation, which is complementary to the slow, exact, but sometimes brittle behaviour of more traditional deduction engines such as theorem provers.<p>The article explains the way logical connectives can be used in semantic vector models, and summarizes the development of Predication-based Semantic Indexing, which involves the use of Vector Symbolic Architectures to represent the concepts and relationships from a knowledge base of subject-predicate-object triples. Experiments show that the use of continuous models for formal reasoning is not only possible, but already demonstrably effective for some recognized informatics tasks, and showing promise in other traditional problem areas. Examples described in this article include: predicting new uses for existing drugs in biomedical informatics; removing unwanted meanings from search results in information retrieval and concept navigation; type inference from attributes; comparing words based on their orthography; and representing tabular data, including modelling numerical values.<p>The algorithms and techniques described in this article are all publicly released and freely available in the Semantic Vectors open-source software package.}
}
@inproceedings{WidrowandHoff,
       author = {Widrow, B. and Hoff, M. E.},
        title = {Adaptive Switching Circuits},
    booktitle = {1960 WESCON Convention Record Part IV, (Reprinted in J. A. Anderson and E. Rosenfeld, {\em Neurocomputing: Foundations of Research}, The {MIT} Press, Cambridge, MA, 1988)},
        pages = {96-104},
         year = 1960,
}
@book{WidrowandStearns85,
       author = {Widrow, B. and Stearns, S. D.},
        title = {Adaptive Signal Processing},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, N.J.},
         year = 1985,
}
@article{WieckiandFrankPSYCHOLOGY-REVIEW-13,
       author = {Wiecki, T. V. and Frank, M. J.},
        title = {A computational model of inhibitory control in frontal cortex and basal ganglia},
      journal = {Psychology Review},
         year = {2013},
       volume = {120},
       number = {2},
        pages = {329-355},
     abstract = {Planning and executing volitional actions in the face of conflicting habitual responses is a critical aspect of human behavior. At the core of the interplay between these 2 control systems lies an override mechanism that can suppress the habitual action selection process and allow executive control to take over. Here, we construct a neural circuit model informed by behavioral and electrophysiological data collected on various response inhibition paradigms. This model extends a well-established model of action selection in the basal ganglia by including a frontal executive control network that integrates information about sensory input and task rules to facilitate well-informed decision making via the oculomotor system. Our simulations of the anti-saccade, Simon, and saccade-override tasks ensue in conflict between a prepotent and controlled response that causes the network to pause action selection via projections to the subthalamic nucleus. Our model reproduces key behavioral and electrophysiological patterns and their sensitivity to lesions and pharmacological manipulations. Finally, we show how this network can be extended to include the inferior frontal cortex to simulate key qualitative patterns of global response inhibition demands as required in the stop-signal task.}
}
@book{Wiener1994,
       author = {Jonathan Wiener},
        title = {The Beak of the Finch: A Story of Evolution in Our Time},
    publisher = {Alfred A. Knoph},
      address = {New York},
         year = 1994,
}
@book{Wiener1999,
       author = {Jonathan Wiener},
        title = {Time, Love, Memory: A Great Biologist and His Quest for the Origins of Behavior},
    publisher = {Alfred A. Knoph},
      address = {New York},
         year = 1994,
}
@article {WieraandMozrzymasFiCN-17,
        title = {Extracellular proteolysis in structural and functional plasticity of mossy fiber synapses in hippocampus},
       author = {Wiera, Grzegorz and Mozrzymas, Jerzy W},
      journal = {Frontiers in Cellular Neuroscience},
    publisher = {Frontiers Media S.A.},
       volume = {9},
         year = {2015},
        pages = {427},
     abstract = {Brain is continuously altered in response to experience and environmental changes. One of the underlying mechanisms is synaptic plasticity, which is manifested by modification of synapse structure and function. It is becoming clear that regulated extracellular proteolysis plays a pivotal role in the structural and functional remodeling of synapses during brain development, learning and memory formation. Clearly, plasticity mechanisms may substantially differ between projections. Mossy fiber synapses onto CA3 pyramidal cells display several unique functional features, including pronounced short-term facilitation, a presynaptically expressed long-term potentiation (LTP) that is independent of NMDAR activation, and NMDA-dependent metaplasticity. Moreover, structural plasticity at mossy fiber synapses ranges from the reorganization of projection topology after hippocampus-dependent learning, through intrinsically different dynamic properties of synaptic boutons to pre- and postsynaptic structural changes accompanying LTP induction. Although concomitant functional and structural plasticity in this pathway strongly suggests a role of extracellular proteolysis, its impact only starts to be investigated in this projection. In the present report, we review the role of extracellular proteolysis in various aspects of synaptic plasticity in hippocampal mossy fiber synapses. A growing body of evidence demonstrates that among perisynaptic proteases, tissue plasminogen activator (tPA)/plasmin system, β-site amyloid precursor protein-cleaving enzyme 1 (BACE1) and metalloproteinases play a crucial role in shaping plastic changes in this projection. We discuss recent advances and emerging hypotheses on the roles of proteases in mechanisms underlying mossy fiber target specific synaptic plasticity and memory formation.}
}
@article{WieringandSchmidhuberAB-87,
       author = {Wiering, M. and Schmidhuber, J.},
        title = {{HQ} Learning},
      journal = {Adaptive Behavior},
       volume = 6,
        issue = 2,
        pages = {219-246},
         year = 1997,
}
@article{WiestandNicolelisNN-03,
       author = {Wiest, M.C. and Nicolelis, M.A.},
        title = {Behavioral detection of tactile stimuli during 7-12 Hz cortical oscillations in awake rats.},
      journal = {Nature Neuroscience},
       volume = 6,
        issue = 9,
         year = 2003,
        pages = {913-914},
     abstract = {Prominent 7--12 Hz oscillations in the primary somatosensory cortex (S1) of awake but immobile rats might represent a seizure-like state in which neuronal burst firing renders animals unresponsive to incoming tactile stimuli; others have proposed that these oscillations are analogous to human mu rhythm. To test whether rats can respond to tactile stimuli during 7--12 Hz oscillatory activity, we trained head-immobilized awake animals to indicate whether they could detect the occurrence of transient whisker deflections while we recorded local field potentials (LFPs) from microelectrode arrays implanted bilaterally in the S1 whisker representation area. They responded rapidly and reliably, suggesting that this brain rhythm represents normal physiological activity that does not preclude perception.},
}
@article{WignerCPAM-60,
       author = {Wigner, Eugene~P.},
        title = {{The Unreasonable Effectiveness of Mathematics in the Natural Sciences}},
      journal = {Communications in Pure Applied Mathematics},
         year = 1960,
       volume = {13},
        issue = {1},
        pages = {1-14},
}
@article {WildangeretalJoM-09,
       author = {Wildanger, D. and Medda, R. and Kastrup, L. and Hell, S.W.},
        title = {A compact {STED} microscope providing {3D} nanoscale resolution},
      journal = {Journal of Microscopy},
    publisher = {Blackwell Publishing Ltd},
       volume = {236},
       number = {1},
        pages = {35--43},
         year = {2009},
     abstract = {The advent of supercontinuum laser sources has enabled the implementation of compact and tunable stimulated emission depletion fluorescence microscopes for imaging far below the diffraction barrier. Here we report on an enhanced version of this approach displaying an all-physics based resolution down to (19 ± 3) nm in the focal plane. Alternatively, this single objective lens system can be configured for 3D imaging with resolution down to 45 \% 45 \% 108 nm in a cell. The obtained results can be further improved by mathematical restoration algorithms. The far-field optical nanoscale resolution is attained in a variety of biological samples featuring strong variations in the local density of features.},
}
@book{Wilensky83,
       author = {Wilensky, Robert},
        title = {Planning and Understanding},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1983,
}
@book{Wilensky86,
       author = {Wilensky, Robert},
        title = {Common Lisp Craft},
    publisher = {Norton},
      address = {New York},
         year = 1986,
}
@article{WilhelmetalNEUROSCIENCE-19,
       author = {Ricardo A. Wilhelm and Matthew W. Miller and Philip A. Gable},
        title = {Neural and Attentional Correlates of Intrinsic Motivation Resulting from Social Performance Expectancy},
      journal = {Neuroscience},
       volume = {416},
         year = {2019},
        pages = {137-146},
     abstract = {Some models of motivation distinguish between intrinsic and extrinsic motivation. While past work has examined the neural and cognitive correlates of extrinsic motivation, research on intrinsic motivation has relied primarily on behavioral measures of performance and learning. In particular, no past work has examined the neural and cognitive correlates of social performance expectancy, which is linked to intrinsic motivation. The current study manipulated expectancy of difficult (vs. easy) trials on a cued flanker task and assessed attentional scope and performance. EEG was used to examine motor-action preparation as measured by suppression of beta band activity over the motor cortex and feedback processing as measured by the Reward Positivity (RewP). Results revealed expectancy of difficult (vs. easy) trials narrowed attentional scope, reduced beta activity over the motor cortex, and enhanced RewP amplitudes to win feedback. These findings suggest that enhancing intrinsic motivation through expectancies of positive social comparison engages similar neural and cognitive correlates as extrinsic motivators high in motivational intensity.}
}
@book{Wilkins88,
       author = {Wilkins, David E.},
        title = {Practical Planning: Extending the Classical {AI} Planning Paradigm},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1988,
}
@article{WilkinsAIJ-84,
       author = {Wilkins, David E.},
        title = {Domain Independent Planning: Representation and Plan Generation},
      journal = AIJ,
      comment = {SIPE},
       volume = 22,
         year = 1984,
        pages = {269-302},
}
@article{WilkinsCIJ-85,
       author = {Wilkins, David E.},
        title = {Recovering from Execution Errors in {SIPE}},
      journal = CIJ,
       volume = 1,
         year = 1985,
        pages = {33-45},
}
@article{WilkinsCIJ-88,
       author = {Wilkins, David E.},
        title = {Causal Reasoning in Planning},
      journal = CIJ,
         year = 1988,
}
@inproceedings{WilkinsECAI-86,
       author = {Wilkins, David E.},
        title = {Hierarchical Planning: Definition and Implementation},
    booktitle = {Proceedings of ECAI-86},
 organization = {ECAI},
         year = 1986,
}
@book{Wilkinson63,
       author = {Wilkinson, J. H.},
        title = {Rounding Errors in Algebraic Processes},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, N.J.},
         year = 1963,
}
@article{WillemetBRAIN-SCIENCE-12,
       author = {Willemet, Romain},
        title = {Understanding the evolution of Mammalian brain structures; the need for a (new) cerebrotype approach},
      journal = {Brain sciences},
       volume = {2},
        issue = {2},
         year = {2012},
        pages = {203-224},
     abstract = {The mammalian brain varies in size by a factor of 100,000 and is composed of anatomically and functionally distinct structures. Theoretically, the manner in which brain composition can evolve is limited, ranging from highly modular (``mosaic evolution'') to coordinated changes in brain structure size (``concerted evolution'') or anything between these two extremes. There is a debate about the relative importance of these distinct evolutionary trends. It is shown here that the presence of taxa-specific allometric relationships between brain structures makes a taxa-specific approach obligatory. In some taxa, the evolution of the size of brain structures follows a unique, coordinated pattern, which, in addition to other characteristics at different anatomical levels, defines what has been called here a ``taxon cerebrotype''. In other taxa, no clear pattern is found, reflecting heterogeneity of the species' lifestyles. These results suggest that the evolution of brain size and composition depends on the complex interplay between selection pressures and constraints that have changed constantly during mammalian evolution. Therefore the variability in brain composition between species should not be considered as deviations from the normal, concerted mammalian trend, but in taxa and species-specific versions of the mammalian brain. Because it forms homogenous groups of species within this complex ``space'' of constraints and selection pressures, the cerebrotype approach developed here could constitute an adequate level of analysis for evo-devo studies, and by extension, for a wide range of disciplines related to brain evolution.},
}
@article{WillemsetalTIT-95,
       author = {F. M. J. Willems and Y. M. Shtarkov and Tj. J. Tjalkens},
        title = {The Context-tree Weighting Method: Basic Properties},
      journal = {{IEEE} Transactions on Information Theory},
        pages = {653-664},
         year = 1995,
}
@article{WilliamsBRAIN-68,
       author = {Williams, Denis},
        title = {{Man's Temporal Lobe}},
      journal = {Brain},
       volume = {91},
       number = {4},
        pages = {639-654},
         year = {1968},
      abtract = {Study of the temporal lobe of the brain remained neglected mainly because animal experimentation, clinical observation and surgical extirpation failed to give results which could be recorded upon a chart, or photographed, or subjected to statistical enquiry. Like the frontal lobe it was a silent area of the brain, for the only function served by the cortex below the fissure of Sylvius which could be quantitatively studied in this way was hearing, and even that sense continued when Heschl's gyrus was removed, or its connections were cut on one side, because of its bilateral receptor mechanism. Today, man's temporal lobe interests all those who study disturbed feeling and behaviour, and it has become common ground for such diverse specialists as psychiatrists, neurological surgeons, criminologists, psychologists, sociologists, and of course experimental physiologists, but clinical neurologists, whose study after all is man, have been slow to share this interest - except in the field of the epilepsies in which they have contributed so much.}
}
@article{WilliamsML-92,
       author = {Ronald J. Williams},
        title = {Simple Statistical Gradient-Following Algorithms for Connectionist Reinforcement Learning},
      journal = {Machine Learning},
       volume = 8,
        pages = {229-256},
         year = 1992,
}
@article{WilliamsPNAS-13,
       author = {Williams, Sarah C. P.},
        title = {The other microbiome},
      journal = {Proceedings of the National Academy of Science},
    publisher = {National Academy of Sciences},
         year = {2013},
       volume = {110},
        issue = {8},
          url = {http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3581985/},
        pages = {2682-2684},
}
@techreport{WilliamsTR-93,
       author = {Ronald J. Williams and Baird, Leemon C. {III}},
        title = {Tight Performance Bounds on Greedy Policies Based on Imperfect Value Functions},
  institution = {Northeastern University, College of Computer Science},
        issue = {NU-CCS-93-13},
      address = {Boston, Massachusetts},
         year = 1993,
}
@techreport{WilliamsTR-99,
       author = {Ronald J. Williams},
        title = {Towards a Theory of Reinforcement-Learning Connectionist Systems},
  institution = {Northeastern University},
        issue = {NU-CCS-88-3},
      address = {Boston, Massachusetts},
         year = 1988,
}
@inproceedings{WilliamsandBaird,
       author = {Williams, Ronald J. and Baird, Leemon C.},
        title = {A Mathematical Analysis of Actor-Critic Architectures for Learning Optimal Controls Through Incremental Dynamic Programming},
    booktitle = {Proceedings of the Sixth Yale Workshop on Adaptive and Learning Systems},
     location = {New Haven, Connecticut},
 organization = {Yale University},
         year = 1990,
        pages = {96-101},
}
@inproceedings{WilliamsandMillarAAAI-98,
       author = {Brian C Williams and Bill Millar},
        title = {Decompositional, Model-based Learning and its Analogy to Diagnosis},
    booktitle = {Proceedings {AAAI}-98},
     location = {Madison, Wisconsin},
 organization = {AAAI},
         year = 1998,
}
@article{WilliamsetalCoRR-17,
       author = {Jason D. Williams and Kavosh Asadi and Geoffrey Zweig},
        title = {Hybrid Code Networks: practical and efficient end-to-end dialog control with supervised and reinforcement learning},
      journal = {CoRR},
       volume = {arXiv:1702.03274},
         year = {2017},
     abstract = {End-to-end learning of recurrent neural networks (RNNs) is an attractive solution for dialog systems; however, current techniques are data-intensive and require thousands of dialogs to learn simple behaviors. We introduce Hybrid Code Networks (HCNs), which combine an RNN with domain-specific knowledge encoded as software and system action templates. Compared to existing end-to-end approaches, HCNs considerably reduce the amount of training data required, while retaining the key benefit of inferring a latent representation of dialog state. In addition, HCNs can be optimized with supervised learning, reinforcement learning, or a mixture of both. HCNs attain state-of-the-art performance on the bAbI dialog dataset, and outperform two commercially deployed customer-facing dialog systems.}
}
@inproceedings{WilliamsonHanksAIPS-94,
       author = {Williamson, Mike and Steve Hanks},
        title = {Optimal Planning with a Goal-Directed Utility Model},
    booktitle = {Second International Conference on {AI} Planning Systems},
         year = 1994,
}
@article{WilliamsonetalJoN-11,
       author = {Williamson, L. L. and Sholar, P. W. and Mistry, R. S. and Smith, S. H. and Bilbo, S. D.},
        title = {Microglia and memory: modulation by early-life infection},
      journal = {Journal of Neuroscience},
       volume = {31},
       number = {43},
         year = {2011},
        pages = {15511-15521},
     abstract = {The proinflammatory cytokine interleukin-1\^{I}$^2$ (IL-1\^{I}$^2$) is critical for normal hippocampus (HP)-dependent cognition, whereas high levels can disrupt memory and are implicated in neurodegeneration. However, the cellular source of IL-1\^{I}$^2$ during learning has not been shown, and little is known about the risk factors leading to cytokine dysregulation within the HP. We have reported that neonatal bacterial infection in rats leads to marked HP-dependent memory deficits in adulthood. However, deficits are only observed if unmasked by a subsequent immune challenge [lipopolysaccharide (LPS)] around the time of learning. These data implicate a long-term change within the immune system that, upon activation with the "second hit," LPS, acutely impacts the neural processes underlying memory. Indeed, inhibiting brain IL-1\^{I}$^2$ before the LPS challenge prevents memory impairment in neonatally infected (NI) rats. We aimed to determine the cellular source of IL-1\^{I}$^2$ during normal learning and thereby lend insight into the mechanism by which this cytokine is enduringly altered by early-life infection. We show for the first time that CD11b(+) enriched cells are the source of IL-1\^{I}$^2$ during normal HP-dependent learning. CD11b(+) cells from NI rats are functionally sensitized within the adult HP and produce exaggerated IL-1\^{I}$^2$ ex vivo compared with controls. However, an exaggerated IL-1\^{I}$^2$ response in vivo requires LPS before learning. Moreover, preventing microglial activation during learning prevents memory impairment in NI rats, even following an LPS challenge. Thus, early-life events can significantly modulate normal learning-dependent cytokine activity within the HP, via a specific, enduring impact on brain microglial function.}
}
@book{Wills91,
       author = {Christopher Wills},
        title = {Exons, Introns, and Talking Genes: The Science Behind the Human Genome Project},
    publisher = {Basic Books},
      address = {New York},
         year = 1991,
}
@article{WillshawetalPTRS_B-15,
       author = {D. J. Willshaw  and P. Dayan  and R. G. M. Morris },
        title = {Memory, modelling and Marr: a commentary on Marr (1971) Simple memory: a theory of archicortex},
      journal = {Philosophical Transactions of the Royal Society B: Biological Sciences},
       volume = {370},
       number = {1666},
        pages = {20140383},
         year = {2015},
     abstract = {David Marr's theory of the archicortex, a brain structure now more commonly known as the hippocampus and hippocampal formation, is an epochal contribution to theoretical neuroscience. Addressing the problem of how information about 10 000 events could be stored in the archicortex during the day so that they can be retrieved using partial information and then transferred to the neocortex overnight, the paper presages a whole wealth of later empirical and theoretical work, proving impressively prescient. Despite this impending success, Marr later apparently grew dissatisfied with this style of modelling, but he went on to make seminal suggestions that continue to resonate loudly throughout the field of theoretical neuroscience. We describe Marr's theory of the archicortex and his theory of theories, setting them into their original and a contemporary context, and assessing their impact. This commentary was written to celebrate the 350th anniversary of the journal Philosophical Transactions of the Royal Society.}
}
@book{Wilson1978,
        title = {On Human Nature},
       author = {Wilson, E.O.},
         year = {1978},
    publisher = {Harvard University Press}
}
@misc{WilsonTHE-NEW-SCIENTIST-18,
        title = {Memory special: Can you trust your memories?},
       author = {Clare Wilson},
 howpublished = {The New Sciencist},
         year = {2018},
}
@article{WilsonandCowanBPJ-72,
       author = {Wilson, Hugh R. and Cowan, Jack D.},
        title = {Excitatory and Inhibitory Interactions in Localized Populations of Model Neurons},
      journal = {Biophysical Journal},
         year = {1972},
    publisher = {Elsevier},
       volume = {12},
        issue = {1},
        pages = {1-24},
     abstract = {Coupled nonlinear differential equations are derived for the dynamics of spatially localized populations containing both excitatory and inhibitory model neurons. Phase plane methods and numerical solutions are then used to investigate population responses to various types of stimuli. The results obtained show simple and multiple hysteresis phenomena and limit cycle activity. The latter is particularly interesting since the frequency of the limit cycle oscillation is found to be a monotonic function of stimulus intensity. Finally, it is proved that the existence of limit cycle dynamics in response to one class of stimuli implies the existence of multiple stable states and hysteresis in response to a different class of stimuli. The relation between these findings and a number of experiments is discussed.},
}
@article{WilsonandMcNaughtonSCIENCE-94,
       author = {Wilson, M. A. and McNaughton, B.L.},
        title = {Reactivation of hippocampal ensemble memories during sleep},
      journal = {Science},
    publisher = {American Association for the Advancement of Science},
       volume = {265},
       number = {5172},
        pages = {676-679},
         year = {1994},
     abstract = {Simultaneous recordings were made from large ensembles of hippocampal "place cells" in three rats during spatial behavioral tasks and in slow-wave sleep preceding and following these behaviors. Cells that fired together when the animal occupied particular locations in the environment exhibited an increased tendency to fire together during subsequent sleep, in comparison to sleep episodes preceding the behavioral tasks. Cells that were inactive during behavior, or that were active but had non-overlapping spatial firing, did not show this increase. This effect, which declined gradually during each post-behavior sleep session, may result from synaptic modification during waking experience. Information acquired during active behavior is thus re-expressed in hippocampal circuits during sleep, as postulated by some theories of memory consolidation.},
}
@article{WilsonandNicollNATURE-01,
       author = {Wilson, Rachel I. and Nicoll, Roger A.},
        title = {Endogenous cannabinoids mediate retrograde signalling at hippocampal synapses},
      journal = {Nature},
         year = 2001,
    publisher = {Macmillian Magazines Ltd.},
       volume = 410,
        issue = 6828,
        pages = {588-592},
      abtract = {Marijuana affects brain function primarily by activating the G-protein-coupled cannabinoid receptor-1 (CB1), which is expressed throughout the brain at high levels. Two endogenous lipids, anandamide and 2-arachidonylglycerol (2-AG), have been identified as CB1 ligands. Depolarized hippocampal neurons rapidly release both anandamide and 2-AG in a Ca2+-dependent manner. In the hippocampus, CB1 is expressed mainly by GABA (gamma-aminobutyric acid)-mediated inhibitory interneurons, where CB1 clusters on the axon terminal. A synthetic CB1 agonist depresses GABA release from hippocampal slices. These findings indicate that the function of endogenous cannabinoids released by depolarized hippocampal neurons might be to downregulate GABA release. Here we show that the transient suppression of GABA-mediated transmission that follows depolarization of hippocampal pyramidal neurons is mediated by retrograde signalling through release of endogenous cannabinoids. Signalling by the endocannabinoid system thus represents a mechanism by which neurons can communicate backwards across synapses to modulate their inputs.}
}
@inproceedings{WinderandBrownCVPR-07,
       author = {Winder, Simon  A. and Brown, Matthew},
        title = {Learning Local Image Descriptors},
    booktitle = {Proceedings 2007 of the {IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE Computer Society},
        pages = {1-8},
         year = 2007,
          url = {http://research.microsoft.com/~brown/papers/cvpr2007b.pdf},
}
@article{WindridgeFiRAI-17,
       author = {Windridge, David},
        title = {Emergent Intentionality in Perception-Action Subsumption Hierarchies},
      journal = {Frontiers in Robotics and AI},
       volume = {4},
        pages = {38},
         year = {2017},
     abstract = {A cognitively autonomous artificial agent may be defined as one able to modify both its external world-model and the framework by which it represents the world, requiring two simultaneous optimization objectives. This presents deep epistemological issues centered on the question of how a framework for representation (as opposed to the entities it represents) may be objectively validated. In this article, formalizing previous work in this field, it is argued that subsumptive perception-action learning has the capacity to resolve these issues by (a) building the perceptual hierarchy from the bottom up so as to ground all proposed representations and (b) maintaining a bijective coupling between proposed percepts and projected action possibilities to ensure empirical falsifiability of these grounded representations. In doing so, we will show that such subsumptive perception-action learners intrinsically incorporate a model for how intentionality emerges from randomized exploratory activity in the form of "motor babbling." Moreover, such a model of intentionality also naturally translates into a model for human-computer interfacing that makes minimal assumptions as to cognitive states.}
}
@inproceedings{WinnandJojicICCV-05,
    author = {John Winn and Nebojsa Jojic},
     title = {{LOCUS}: Learning Object Classes with Unsupervised Segmentation},
 booktitle = {Proceedings of the 10th {IEEE} International Conference on Computer Vision},
 publisher = {IEEE Computer Society},
    volume = 1,
     pages = {756-763},
      year = 2005,
}
@inproceedings{WinnandShottonCVPR-06,
    author = {Winn, John and Shotton, Jamie},
     title = {The Layout Consistent Random Field for Recognizing and Segmenting Partially Occluded Objects},
 booktitle = {Proceedings of IEEE Computer Vision and Pattern Recognition},
 publisher = {IEEE Computer Society},
    volume = 1,
     pages = {37-44},
      year = 2006
}
@book{Winograd83,
       author = {Winograd, Terry},
        title = {Language as a Cognitive Process, Volume 1: Syntax},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1983,
}
@article{WinogradARAC-18,
       author = {Winograd, Nicholas},
        title = {Gas Cluster Ion Beams for Secondary Ion Mass Spectrometry},
      journal = {Annual Review of Analytical Chemistry},
       volume = {11},
       number = {1},
        pages = {29-48},
         year = {2018},
     abstract = {Gas cluster ion beams (GCIBs) provide new opportunities for bioimaging and molecular depth profiling with secondary ion mass spectrometry (SIMS). These beams, consisting of clusters containing thousands of particles, initiate desorption of target molecules with high yield and minimal fragmentation. This review emphasizes the unique opportunities for implementing these sources, especially for bioimaging applications. Theoretical aspects of the cluster ion/solid interaction are developed to maximize conditions for successful mass spectrometry. In addition, the history of how GCIBs have become practical laboratory tools is reviewed. Special emphasis is placed on the versatility of these sources, as size, kinetic energy, and chemical composition can be varied easily to maximize lateral resolution, hopefully to less than 1 micron, and to maximize ionization efficiency. Recent examples of bioimaging applications are also presented.}
}
@book{WinogradandFLores1986,
        title = {Understanding Computers and Cognition: A New Foundation for Design},
       author = {Winograd, T. and Flores, F. and Flores, F.F.},
       series = {Language and Being},
         year = {1986},
    publisher = {Ablex Publishing Corporation}
}
@inproceedings{WinslettAAAI-88,
       author = {Winslett, Marianne},
        title = {Reasoning About Action Using a Possible Models Approach},
    booktitle = {Proceedings AAAI-88},
     location = {St. Paul, Minnesota},
 organization = {AAAI},
         year = 1988,
        pages = {89-93},
}
@book{Winston91,
       author = {Winston, Wayne L.},
        title = {Operations Research: Applications and Algorithms, Second Edition},
    publisher = {PWS-Kent Publishing Company},
      address = {Boston, Massachusetts},
         year = 1991,
}
@book{Winston92,
       author = {Winston, Patrick H.},
        title = {Artificial Intelligence, Third Edition},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1992,
}
@incollection{WiskottPSN-03,
       author = {Laurenz Wiskott},
        title = {How Does Our Visual System Achieve Shift and Size Invariance?},
    booktitle = {Problems in Systems Neuroscience},
       editor = {J. L. {van Hemmen} and T. J. Sejnowski},
    publisher = {Oxford University Press},
          url = {http://cogprints.org/3321/02/Wiskott2003.ps},
         year = 2003,
}
@article{WiskottandSejnowskiNC-02,
       author = {Laurenz Wiskott and Terrence Sejnowski},
        title = {Slow Feature Analysis: Unsupervised Learning of Invariances},
      journal = {Neural Computation},
       volume = 14,
        issue = 4,
        pages = {715-770},
          url = {http://itb.biologie.hu-berlin.de/~wiskott/},
         year = 2002,
}
@article{WiskottetalHIPPOCAMPUS-06,
       author = {Wiskott, L. and Rasch, M. J. and Kempermann, G.},
        title = {A functional hypothesis for adult hippocampal neurogenesis: avoidance of catastrophic interference in the dentate gyrus},
      journal = {Hippocampus},
         year = {2006},
       volume = {16},
       number = {3},
        pages = {329-343},
     abstract = {The dentate gyrus is part of the hippocampal memory system and special in that it generates new neurons throughout life. Here we discuss the question of what the functional role of these new neurons might be. Our hypothesis is that they help the dentate gyrus to avoid the problem of catastrophic interference when adapting to new environments. We assume that old neurons are rather stable and preserve an optimal encoding learned for known environments while new neurons are plastic to adapt to those features that are qualitatively new in a new environment. A simple network simulation demonstrates that adding new plastic neurons is indeed a successful strategy for adaptation without catastrophic interference.}
}
@inproceedings{WitkinIJCAI-83,
       author = {Andrew P. Witkin},
        title = {Scale-Space Filtering},
    booktitle = {Proceedings {IJCAI} 8},
     location = {Karlsruhe, West Germany},
        pages = {1019-1022},
         year = 1983,
}
@inproceedings{WitkinandTenenbaumIJCAI-83,
       author = {Andrew P. Witkin and Jay M. Tenenbaum},
        title = {What Is Perceptual Organization For?},
    booktitle = {Proceedings {IJCAI} 8},
     location = {Karlsruhe, West Germany},
        pages = {1023-1026},
         year = 1983,
}
@book{Wittgenstein1953philosophical,
        title = {Philosophical Investigations},
       author = {Wittgenstein, L. and Anscombe, G.E.M.},
    publisher = {Blackwell},
         year = {1968},
}
@book{Wittgenstein2009philosophical,
        title = {Philosophical Investigations},
       author = {Wittgenstein, Ludwig},
         year = {2009},
    publisher = {John Wiley \& Sons}
}
@article{WixtedetalPNAS-18,
       author = {Wixted, John T. and Goldinger, Stephen D. and Squire, Larry R. and Kuhn, Joel R. and Papesh, Megan H. and Smith, Kris A. and Treiman, David M. and Steinmetz, Peter N.},
        title = {Coding of episodic memory in the human hippocampus},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {115},
       number = {5},
        pages = {1093-1098},
         year = {2018},
    publisher = {National Academy of Sciences},
     abstract = {Neurocomputational models hold that episodic memories are represented by sparse, stimulus-specific neural codes. In tests of episodic memory, single-unit recording studies of the human hippocampus have found neurons that operate as general novelty detectors or general familiarity detectors. Here, we investigated whether neurons can be found that sparsely code some recently studied items and not others. In the left hippocampus, but not the amygdala, we found that small fractions of neurons exhibited strong responses to specific repeated words. The remaining large fractions of neurons exhibited a concurrent reduction in firing rates relative to novel words. Both findings are consistent with predictions made by neurocomputational models of how episodic memory is coded in the hippocampus.Neurocomputational models have long posited that episodic memories in the human hippocampus are represented by sparse, stimulus-specific neural codes. A concomitant proposal is that when sparse-distributed neural assemblies become active, they suppress the activity of competing neurons (neural sharpening). We investigated episodic memory coding in the hippocampus and amygdala by measuring single-neuron responses from 20 epilepsy patients (12 female) undergoing intracranial monitoring while they completed a continuous recognition memory task. In the left hippocampus, the distribution of single-neuron activity indicated that only a small fraction of neurons exhibited strong responding to a given repeated word and that each repeated word elicited strong responding in a different small fraction of neurons. This finding reflects sparse distributed coding. The remaining large fraction of neurons exhibited a concurrent reduction in firing rates relative to novel words. The observed pattern accords with longstanding predictions that have previously received scant support from single-cell recordings from human hippocampus.},
}
@book{Wolfe00,
       author = {Tom Wolfe},
        title = {Hooking Up},
    publisher = {Farrar, Straus \& Giroux},
      address = {New York, NY},
         year = 2000,
}
@book{Wolfe79,
       author = {Tom Wolfe},
        title = {The Right Stuff},
    publisher = {Farrar, Straus \& Giroux},
      address = {New York, NY},
         year = 1979,
}
@article{WolffetalJCM-15,
       author = {Wolff, Tanya and Iyer, Nirmala A. and Rubin, Gerald M.},
        title = {Neuroarchitecture and neuroanatomy of the Drosophila central complex: A {GAL4}-based dissection of protocerebral bridge neurons and circuits},
      journal = {Journal Comparative Neurology},
         year = {2015},
    publisher = {BlackWell Publishing Ltd},
      address = {Oxford, UK},
       volume = {523},
        issue = {7},
        pages = {997-1037},
     abstract = {Insects exhibit an elaborate repertoire of behaviors in response to environmental stimuli. The central complex plays a key role in combining various modalities of sensory information with an insect's internal state and past experience to select appropriate responses. Progress has been made in understanding the broad spectrum of outputs from the central complex neuropils and circuits involved in numerous behaviors. Many resident neurons have also been identified. However, the specific roles of these intricate structures and the functional connections between them remain largely obscure. Significant gains rely on obtaining a comprehensive catalog of the neurons and associated GAL4 lines that arborize within these brain regions, and on mapping neuronal pathways connecting these structures. To this end, small populations of neurons in the Drosophila melanogaster central complex were stochastically labeled using the multicolor flip-out technique and a catalog was created of the neurons, their morphologies, trajectories, relative arrangements, and corresponding GAL4 lines. This report focuses on one structure of the central complex, the protocerebral bridge, and identifies just 17 morphologically distinct cell types that arborize in this structure. This work also provides new insights into the anatomical structure of the four components of the central complex and its accessory neuropils. Most strikingly, we found that the protocerebral bridge contains 18 glomeruli, not 16, as previously believed. Revised wiring diagrams that take into account this updated architectural design are presented. This updated map of the Drosophila central complex will facilitate a deeper behavioral and physiological dissection of this sophisticated set of structures.},
}
@book{Wolfram91,
       author = {Wolfram, Steven},
        title = {Mathematica: A System for Doing Mathematics by Computer, Second Edition},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1991,
}
@book{Wolovich74,
       author = {Wolovich, William A.},
        title = {Linear Multivariable Systems},
    publisher = {Springer-Verlag},
      address = {New York},
         year = 1974,
}
@book{Wolovich87,
       author = {Wolovich, William A.},
        title = {Robotics: Basic Analysis and Design},
    publisher = {Holt, Rinehart and Winston},
         year = 1987,
}
@book{WolpawaandWolpaw2012,
        title = {Brain-Computer Interfaces: Principles and Practice},
       author = {Wolpaw, J. and Wolpaw, E.W.},
         year = {2012},
    publisher = {Oxford University Press, USA}
}
@book{Wolpert95,
       editor = {Wolpert, David H.},
        title = {The Mathematics of Generalization},
    publisher = {Addison-Wesley},
      address = {Reading, Massachusetts},
         year = 1995,
}
@article{Wolpert97,
       author = {David H. Wolpert and William G. Macready},
        title = {No Free Lunch Theorems for Optimization},
      journal = {{IEEE} Transactions on Evolutionary Computation},
       volume = 1,
        issue = 1,
        pages = {67-82},
         year = 1997,
}
@incollection{WolpertSFI-95a,
       author = {Wolpert, David H.},
        title = {The Relationship Between {PAC}, the Statistical Physics Framework, the {B}ayesian Framework, and the {VC} Framework},
    booktitle = {The Mathematics of Generalization},
       editor = {Wolpert, David H.},
    publisher = {Addison-Wesley},
         year = 1995,
}
@incollection{WolpertSFI-95b,
       author = {Wolpert, David H.},
        title = {The Status of Supervised Learning Science Circa 1994: The Search for Consensus},
    booktitle = {The Mathematics of Generalization},
       editor = {Wolpert, David H.},
    publisher = {Addison-Wesley},
         year = 1995,
}
@article{WolterandYaoNIPS-18,
       author = {Moritz Wolter and Angela Yao},
        title = {Gated Complex Recurrent Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1806.08267},
         year = {2018},
     abstract = {Complex numbers have long been favoured for digital signal processing, yet complex representations rarely appear in deep learning architectures. RNNs, widely used to process time series and sequence information, could greatly benefit from complex representations. We present a novel complex gated recurrent cell, which is a hybrid cell combining complex-valued and norm-preserving state transitions with a gating mechanism. The resulting RNN exhibits excellent stability and convergence properties and performs competitively on the synthetic memory and adding task, as well as on the real-world tasks of human motion prediction.}
}
@misc{Woman_Born_Without_Cerebellum,
       author = {IFLScience_Kean_Street_London},
        title = {{24-Year-Old Woman Born Without Cerebellum}},
 howpublished = {https://www.iflscience.com/brain/24-year-old-woman-born-without-cerebellum-her-brain/},
         year = {2014}
}
@article{WondersandAndersonNATURE-06,
       author = {Wonders, Carl P. and Anderson, Stewart A.},
        title = {The origin and specification of cortical interneurons},
      journal = {Nature Reviews Neuroscience},
         year = 2006,
       volume = 7,
        issue = 9,
        pages = {687-696},
     abstract = {GABA-containing interneurons are crucial to both the development and function of the cerebral cortex. Unlike cortical projection neurons, which have a relatively conserved set of characteristics, interneurons include multiple phenotypes that vary on morphological, physiological and neurochemical axes. This diversity, and the relatively late, context-dependent maturation of defining features, has challenged efforts to uncover the transcriptional control of cortical interneuron development. Here, we discuss recent data that are beginning to illuminate the origins and specification of distinct subgroups of cortical interneurons.},
}
@article{WonetalNATURE-14,
        title = {A highly stretchable, helical copper nanowire conductor exhibiting a stretchability of 700\%},
       author = {Yulim Won and Areum Kim and Wooseok Yang and Sunho Jeong  and Jooho Moon},
      journal = {NPG Asia Materials},
    publisher = {Nature Japan KK},
       volume = 6,
         year = 2014,
        pages = {e132},
     abstract = {Recent advances in unconventional foldable and stretchable electronics have forged a new field in electronics. However, traditional conducting metal oxides and metal thin films are inappropriate as electrodes for stretchable devices because they are vulnerable to tensile strain as well as bending strain. In this study, we describe the fabrication of annealing-free, copper nanowire (CuNW)-based stretchable electrodes using an inexpensive metal source through a simple and scalable process at low temperature without a vacuum. We also introduce a reversible and extremely stretchable (up to 700\% of strain) helical, CuNW-based conducting spring, which has not been previously used for stretchable electrodes.},
}
@inproceedings{WongandCipollaICCV-07,
       author = {Shu-Fai Wong and Roberto Cipolla},
        title = {Extracting Spatiotemporal Interest Points using Global Information},
    booktitle = {Proceedings of the 11th {IEEE} International Conference on Computer Vision},
         year = 2007,
}
@article{WongandWangJoN-06,
       author = {Wong, K. F. and Wang, X. J.},
        title = {A recurrent network mechanism of time integration in perceptual decisions},
      journal = {Journal of Neuroscience},
       volume = {26},
       number = {4},
         year = {2006},
        pages = {1314-1328},
     abstract = {Recent physiological studies using behaving monkeys revealed that, in a two-alternative forced-choice visual motion discrimination task, reaction time was correlated with ramping of spike activity of lateral intraparietal cortical neurons. The ramping activity appears to reflect temporal accumulation, on a timescale of hundreds of milliseconds, of sensory evidence before a decision is reached. To elucidate the cellular and circuit basis of such integration times, we developed and investigated a simplified two-variable version of a biophysically realistic cortical network model of decision making. In this model, slow time integration can be achieved robustly if excitatory reverberation is primarily mediated by NMDA receptors; our model with only fast AMPA receptors at recurrent synapses produces decision times that are not comparable with experimental observations. Moreover, we found two distinct modes of network behavior, in which decision computation by winner-take-all competition is instantiated with or without attractor states for working memory. Decision process is closely linked to the local dynamics, in the "decision space" of the system, in the vicinity of an unstable saddle steady state that separates the basins of attraction for the two alternative choices. This picture provides a rigorous and quantitative explanation for the dependence of performance and response time on the degree of task difficulty, and the reason for which reaction times are longer in error trials than in correct trials as observed in the monkey experiment. Our reduced two-variable neural model offers a simple yet biophysically plausible framework for studying perceptual decision making in general.}
}
@article{WongetalNATURE-95,
        title = {Early functional neural networks in the developing retina},
       author = {Wong, R. O. L. and Chernjavsky, A. and Smith, S. J. and Shatz, C. J.},
      journal = {Nature},
       volume = {374},
       number = {6524},
         year = {1995},
        pages = {716-718},
     abstract = {In the adult mammalian retina, the principal direction of information flow is along a vertical pathway from photoreceptors to retinal interneurons to ganglion cells, the output neurons of the retina. We report here, however, that initially in development, at a time when the photoreceptors are not yet even present, there are already functionally defined networks within the retina. These networks are spontaneously active rather than visually driven, and they involve horizontal rather than vertical pathways. By means of optical recording using the calcium-sensitive dye Fura-2, we have found that sets of retinal ganglion cells and amacrine cells, a type of retinal interneuron, undergo synchronized oscillations in intracellular calcium concentration. These oscillations are highly correlated among subgroups of neighbouring cells, and spread in a wave-like fashion tangentially across the retina. Thus, in development of retinal circuitry, the initial patterning of neuronal function occurs in the horizontal domain before the adult pattern of vertical information transfer emerges.},
}
@article{WoodetalJoN-88,
       author = {Charles C. Wood and Dennis D. Spencer and Truett Allison and Gregory McCarthy and Peter D. Williamson and William R. Goff},
        title = {Localization of human sensorimotor cortex during surgery by cortical surface recording of somatosensory evoked potentials},
      journal = {Journal of Neurosurgery},
         year = {1988},
    publisher = {Journal of Neurosurgery Publishing Group},
       volume = {68},
       number = {1},
        pages = {99-111},
     abstract = {The traditional means of localizing sensorimotor cortex during surgery is Penfield's procedure of mapping sensory and motor responses elicited by electrical stimulation of the cortical surface. This procedure can accurately localize sensorimotor cortex but is time-consuming and best carried out in awake, cooperative patients. An alternative localization procedure is presented that involves cortical surface recordings of somatosensory evoked potentials (SEP's), providing accurate and rapid localization in patients under either local or general anesthesia.

The morphology and amplitude of median nerve SEP's recorded from the cortical surface varied systematically as a function of spatial location relative to the sensorimotor hand representation area. These results were validated in 18 patients operated on under local anesthesia in whom the sensorimotor cortex was independently localized by electrical stimulation mapping; the two procedures were in agreement in all cases. Similar SEP results were demonstrated in an additional 27 patients operated on under general anesthesia without electrical stimulation mapping.

The following three spatial relationships between SEP's and the anatomy of the sensorimotor cortex permit rapid and accurate localization of the sensorimotor hand area: (1) SEP's with approximately mirror-image waveforms are recorded at electrode sites in the hand area on opposite sides of the central sulcus (P20-N30 precentrally (for consistency) and N20-P30 postcentrally); (2) the P25-N35 is recorded from the postcentral gyrus as well as a small region of the precentral gyrus in the immediate vicinity of the central sulcus: this waveform is largest on the postcentral gyrus about 1 cm medial to the focus of the 20- and 30-msec potentials; and (3) regardless of component identification, maximum SEP amplitudes are recorded from the hand representation area on the precentral and postcentral gyri.},
}
@article{WoodetalNEUROSCIENCE-11,
       author = {Wood, William E. and Lovell, Peter V. and Mello, Claudio V. and Perkel, David J.},
        title = {Serotonin, via {HTR2} Receptors, Excites Neurons in a Cortical-like Premotor Nucleus Necessary for Song Learning and Production},
      journal = {The Journal of Neuroscience},
       volume = 31,
        issue = 39,
        pages = {13808-13815},
         year = 2011,
     abstract = {Serotonin (5-HT) is a neuromodulator that is important for neural development, learning and memory, mood, and perception. Dysfunction of the serotonin system is central to depression and other clinically important mood disorders and has been linked with learning deficits. In mammals, 5-HT release from the raphe nuclei in the brainstem can modulate the functional properties of cortical neurons, influencing sensory and motor processing. Birds also have serotonergic neurons in the dorsal raphe, suggesting that 5-HT plays similar roles in sensory and motor processing, perhaps modulating brain circuitry underlying birdsong. To investigate this possibility, we measured the effects of 5-HT on spontaneous firing of projection neurons in the premotor robust nucleus of the arcopallium in brain slices from male zebra finches. These neurons are thought be akin to cortical layer V pyramidal neurons. 5-HT dramatically and reversibly enhanced the endogenous firing of RA neurons. Using pharmacological agonists and antagonists in vitro, we determined this action is mediated via HTR2 receptors, which we verified are expressed by in situ hybridization. Finally, focal administration of the serotonin selective reuptake inhibitor fluvoxamine revealed that endogenous 5-HT is sufficient to mediate this effect in vivo. These findings reveal a modulatory action of serotonin on the physiology of the song system circuitry and suggest a novel role of serotonin in regulating song production and/or learning; further understanding of the role of 5-HT in this system may help illuminate the complex role of this neuromodulator in social interactions and motor plasticity in humans.},
}
@article{Woods1978Semantics,
        title = {Semantics and Quantification in Natural Language Question Answering},
       author = {William A. Woods},
      journal = {Advances in Computers},
       volume = {17},
         year = {1978},
        pages = {1-87}
}
@book{Wosetal84,
       author = {Wos, Larry and Overbeek, Ross and Lusk, Ewing and Boyle, Jim},
        title = {Automated Reasoning: Introduction and Applications},
    publisher = {Prentice-Hall},
      address = {Englewood Cliffs, N.J.},
         year = 1984,
}
@book{Wrangham2019goodness,
        title = {The Goodness Paradox: {H}ow Evolution Made Us Both More and Less Violent},
       author = {Wrangham, Richard},
         year = {2019},
    publisher = {Profile}
}
@article{WuandFosterJoN-14,
       author = {Wu, X. and Foster, D. J.},
        title = {{{H}ippocampal replay captures the unique topological structure of a novel environment}},
      journal = {Journal of Neuroscience},
       volume = {34},
       number = {19},
         year = {2014},
        pages = {6459-6469},
     abstract = {Hippocampal place-cell replay has been proposed as a fundamental mechanism of learning and memory, which might support navigational learning and planning. An important hypothesis of relevance to these proposed functions is that the information encoded in replay should reflect the topological structure of experienced environments; that is, which places in the environment are connected with which others. Here we report several attributes of replay observed in rats exploring a novel forked environment that support the hypothesis. First, we observed that overlapping replays depicting divergent trajectories through the fork recruited the same population of cells with the same firing rates to represent the common portion of the trajectories. Second, replay tended to be directional and to flip the represented direction at the fork. Third, replay-associated sharp-wave-ripple events in the local field potential exhibited substructure that mapped onto the maze topology. Thus, the spatial complexity of our recording environment was accurately captured by replay: the underlying neuronal activities reflected the bifurcating shape, and both directionality and associated ripple structure reflected the segmentation of the maze. Finally, we observed that replays occurred rapidly after small numbers of experiences. Our results suggest that hippocampal replay captures learned information about environmental topology to support a role in navigation.}
}
@article{WuetalCoRR-19,
       author = {Zonghan Wu and Shirui Pan and Fengwen Chen and Guodong Long and Chengqi Zhang and Philip S. Yu},
        title = {A Comprehensive Survey on Graph Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1901.00596},
         year = {2019},
          url = {http://arxiv.org/abs/1901.00596},
     abstract = {Deep learning has revolutionized many machine learning tasks in recent years, ranging from image classification and video processing to speech recognition and natural language understanding. The data in these tasks are typically represented in the Euclidean space. However, there is an increasing number of applications where data are generated from non-Euclidean domains and are represented as graphs with complex relationships and interdependency between objects. The complexity of graph data has imposed significant challenges on existing machine learning algorithms. Recently, many studies on extending deep learning approaches for graph data have emerged. In this survey, we provide a comprehensive overview of graph neural networks (GNNs) in data mining and machine learning fields. We propose a new taxonomy to divide the state-of-the-art graph neural networks into different categories. With a focus on graph convolutional networks, we review alternative architectures that have recently been developed; these learning paradigms include graph attention networks, graph autoencoders, graph generative networks, and graph spatial-temporal networks. We further discuss the applications of graph neural networks across various domains and summarize the open source codes and benchmarks of the existing algorithms on different learning tasks. Finally, we propose potential research directions in this fast-growing field.},
}
@article{WuetalCoRR-20,
        title = {Visual Transformers: Token-based Image Representation and Processing for Computer Vision},
       author = {Bichen Wu and Chenfeng Xu and Xiaoliang Dai and Alvin Wan and Peizhao Zhang and Zhicheng Yan and Masayoshi Tomizuka and Joseph Gonzalez and Kurt Keutzer and Peter Vajda},
      journal = {CoRR},
       volume = {arXiv:2006.03677},
         year = {2020},
     abstract = {Computer vision has achieved remarkable success by (a) representing images as uniformly-arranged pixel arrays and (b) convolving highly-localized features. However, convolutions treat all image pixels equally regardless of importance; explicitly model all concepts across all images, regardless of content; and struggle to relate spatially-distant concepts. In this work, we challenge this paradigm by (a) representing images as semantic visual tokens and (b) running transformers to densely model token relationships. Critically, our Visual Transformer operates in a semantic token space, judiciously attending to different image parts based on context. This is in sharp contrast to pixel-space transformers that require orders-of-magnitude more compute. Using an advanced training recipe, our VTs significantly outperform their convolutional counterparts, raising ResNet accuracy on ImageNet top-1 by 4.6 to 7 points while using fewer FLOPs and parameters. For semantic segmentation on LIP and COCO-stuff, VT-based feature pyramid networks (FPN) achieve 0.35 points higher mIoU while reducing the FPN module's FLOPs by 6.5x.}
}
@article{WuetalIEEE-15,
       author = {Xinyu Wu and Vishal Saxena and Kehan Zhu},
        title = {Homogeneous Spiking Neuromorphic System for Real-World Pattern Recognition},
      journal = {IEEE Journal on Emerging and Selected Topics in Circuits and Systems},
       volume = 5,
        issue = 2,
         year = 2015,
     abstract = {A neuromorphic chip that combines CMOS analog spiking neurons and memristive synapses offers a promising solution to brain-inspired computing, as it can provide massive neural network parallelism and density. Previous hybrid analog CMOS-memristor approaches required extensive CMOS circuitry for training, and thus eliminated most of the density advantages gained by the adoption of memristor synapses. Further, they used different waveforms for pre and post-synaptic spikes that added undesirable circuit overhead. Here we describe a hardware architecture that can feature a large number of memristor synapses to learn real-world patterns. We present a versatile CMOS neuron that combines integrate-and-fire behavior, drives passive memristors and implements competitive learning in a compact circuit module, and enables in-situ plasticity in the memristor synapses. We demonstrate handwritten-digits recognition using the proposed architecture using transistor-level circuit simulations. As the described neuromorphic architecture is homogeneous, it realizes a fundamental building block for large-scale energy-efficient brain-inspired silicon chips that could lead to next-generation cognitive computing.}
}
@inproceedings{WuetalISPAN-09,
     author = {Di Wu and Fan Zhang and Naiyong Ao and Fang Wang and Xiaoguang Liu and Gang Wang}, 
  booktitle = {Proceedings of the 10th International Symposium on, Pervasive Systems, Algorithms, and Networks (ISPAN)}, 
      title = {A Batched GPU Algorithm for Set Intersection}, 
      pages = {752-756}, 
       year = 2009, 
}
@article{WuetalMICRO-NANO-15,
       author = {Wu, Sung-Yueh and Yang, Chen and Hsu, Wensyang and Lin, Liwei},
        title = {{3D} printed microelectronics for integrated circuitry and passive wireless sensors},
      journal = {Microsystems \& Nanoengineering},
         year = {2015},
    publisher = {Nature Publishing Group},
       volume = {1},
        pages = {15013},
     abstract = {Three-dimensional (3D) additive manufacturing techniques have been utilized to make {3D} electrical components, such as resistors, capacitors, and inductors, as well as circuits and passive wireless sensors. Using the fused deposition modeling technology and a multiple-nozzle system with a printing resolution of 30 \hmmu{}m, {3D} structures with both supporting and sacrificial structures are constructed. After removing the sacrificial materials, suspensions with silver particles are injected subsequently solidified to form metallic elements/interconnects. The prototype results show good characteristics of fabricated {3D} microelectronics components, including an inductor-capacitor-resonant tank circuitry with a resonance frequency at 0.53 GHz. A {3D} 'smart cap' with an embedded inductor-capacitor tank as the wireless passive sensor was demonstrated to monitor the quality of liquid food (e.g., milk and juice) wirelessly. The result shows a 4.3\% resonance frequency shift from milk stored in the room temperature environment for 36 h. This work establishes an innovative approach to construct arbitrary {3D} systems with embedded electrical structures as integrated circuitry for various applications, including the demonstrated passive wireless sensors.},
}
@article{WuetalNANO-08,
        title = {Carbon nanotubes protect DNA strands during cellular delivery},
       author = {Wu, Y. and Phillips, J.A. and Liu, H. and Yang, R. and Tan, W.},
      journal = {ACS Nano},
       volume = 2,
        issue = 10,
        pages = {2023-2028},
         year = 2008,
     abstract = {To protect against nuclease digestion, or single-strand binding protein interactions, oligonucleotides for targeted delivery into intracellular systems must be stable. To accomplish this, we have developed single-walled carbon nanotubes as a carrier for single-stranded DNA probe delivery. This has resulted in superior biostability for intracellular application and, hence, has achieved the desired protective attributes, which are particularly important when DNA probes are used for intracellular measurements. Specifically, when bound to single-walled carbon nanotubes, DNA probes are protected from enzymatic cleavage and interference from nucleic acid binding proteins. Moreover, and equally important, our study shows that a single-walled carbon nanotube-modified DNA probe, which targets a specific mRNA inside living cells, has increased self-delivery capability and intracellular biostability when compared to free DNA probes. Therefore, this new conjugate provides significant advantages for basic genomic studies in which DNA probes are used to monitor intracellular levels of molecules.}
}
@inproceedings{WuetalSPIE-14,
       author = {Xinyu Wu and Vishal Saxena and Kristy A. Campbell},
        title = {Energy-efficient {STDP}-based Learning Circuits with Memristor Synapses},
    booktitle = {Machine Intelligence and Bio-inspired Computation: Theory and Applications VIII},
       editor = {Misty Blowers and Jonathan Williams},
       volume = 9119,
         year = 2014,
     abstract = {It is now accepted that the traditional von Neumann architecture, with processor and memory separation, is ill suited to process parallel data streams which a mammalian brain can efficiently handle. Moreover, researchers now envision computing architectures which enable cognitive processing of massive amounts of data by identifying spatio-temporal relationships in real-time and solving complex pattern recognition problems. Memristor cross-point arrays, integrated with standard CMOS technology, are expected to result in massively parallel and low-power Neuromorphic computing architectures. Recently, significant progress has been made in spiking neural networks (SNN) which emulate data processing in the cortical brain. These architectures comprise of a dense network of neurons and the synapses formed between the axons and dendrites. Further, unsupervised or supervised competitive learning schemes are being investigated for global training of the network. In contrast to a software implementation, hardware realization of these networks requires massive circuit overhead for addressing and individually updating network weights. Instead, we employ bio-inspired learning rules such as the spike-timing-dependent plasticity (STOP) to efficiently update the network weights locally. To realize SNNs on a chip, we propose to use densely integrating mixed-signal integrate-and- fire neurons (IFNs) and cross-point arrays of memristors in back-end-of-the-line (BEOL) of CMOS chips. Novel IFN circuits have been designed to drive memristive synapses in parallel while maintaining overall power efficiency (<I pJ/spike/synapse), even at spike rate greater than 10 MHz. We present circuit design details and simulation results of the IFN with memristor synapses, its response to incoming spike trains and STOP learning characterization.}
}
@inproceedings{WuetalWWW-10,
       author = {Shoamei Wu and Jake Hofman and Winter A. Mason and Duncan J. Watts},
        title = {Who Says What to Whom on Twitter},
    booktitle = {Proceedings of the 20th International World Wide Web Conference},
     location = {Hyderabad, India},
         year = 2010,
}
@inproceedings{XavieretalNABIC-11,
       author = {R. S. Xavier and N. Omar and L. N. de Castro},
    booktitle = {Proceedings of the World Congress on Nature and Biologically Inspired Computing},
        title = {Bacterial colony: Information processing and computational behavior},
         year = {2011},
        pages = {439-443},
     abstract = {A bacterium presents rudimentary, though powerful, cognitive functions, such as sensing the environment, processing information and contextual interpretation. When part of a colony, bacteria show collective memory, distributed information storage and processing, collective decision making abilities, cooperation and competition, various communication methods and learning at different levels. These features and capabilities allow us to view bacteria colonies as complex adaptive systems, from which much inspiration can be taken to the design of bioinspired tools and algorithms for problem solving and the modeling of biological phenomena. Despite some effort has been taken in this direction, emphasizing applications in optimization, the present paper brings a discussion of bacteria as useful information processing systems under an intelligent agent framework. The main benefits of this discussion, thus, are the presentation of an intelligent agent-based perspective of bacteria, the stress of the main communication skills and mechanisms of bacteria, and a structured discussion of their main information processing and learning. Our hope is that this paper paves the conceptual ground for the design of novel bacteria inspired algorithms with applications over several areas, from the mining of large databases, to the design of sophisticated decision-making mechanisms for collective robotics.},
}
@inproceedings{XiaetalMICCAI-01,
        title = {Methodology of Precise Skull Model Creation},
       author = {James Xia and Jaime Gateno and John Teichgraeber and Andrew Rosen},
    booktitle = {Medical Image Computing and Computer-Assisted Intervention},
       editor = {Wiro J. Niessen and Max A. Viergever},
         year = 2001,
        pages = {434-440},
}
@article{XiaetalNANO-LETTERS-10,
       author = {Xia, Qiangfei and Yang, J. Joshua and Wu, Wei and Li, Xuema and Williams, R. Stanley},
        title = {Self-Aligned Memristor Cross-Point Arrays Fabricated with One Nanoimprint Lithography Step},
      journal = {Nano Letters},
       volume = 10,
        issue = 8,
         year = 2010,
        pages = {2909-2914},
     abstract = {We demonstrate a technique to fabricate memristor cross-point arrays using a self-aligned, one step nanoimprint lithography process that simultaneously patterns the bottom electrode, switching material film and the top electrode. Since this process does not require overlay alignment, the fabrication complexity is greatly reduced and the throughput is significantly increased. The critical interfaces are exposed to much less contamination and thus under better chemical control. With this technique, we fabricated arrays of TiO2-based memristive devices (junction area 100 nm by 100 nm) that did not require electrical forming and were operated with nanoampere currents.}
}
@article{XiangetalCEREBRAL-CORTEX-09,
       author = {Xiang, Hua-Dong and Fonteijn, Hubert M. and Norris, David G. and Hagoort, Peter},
        title = {{Topographical Functional Connectivity Pattern in the Perisylvian Language Networks}},
      journal = {Cerebral Cortex},
       volume = {20},
       number = {3},
         year = {2009},
        pages = {549-560},
     abstract = {We performed a resting-state functional connectivity study to investigate directly the functional correlations within the perisylvian language networks by seeding from 3 subregions of Broca's complex (pars opercularis, pars triangularis, and pars orbitalis) and their right hemisphere homologues. A clear topographical functional connectivity pattern in the left middle frontal, parietal, and temporal areas was revealed for the 3 left seeds. This is the first demonstration that a functional connectivity topology can be observed in the perisylvian language networks. The results support the assumption of the functional division for phonology, syntax, and semantics of Broca's complex as proposed by the memory, unification, and control (MUC) model and indicated a topographical functional organization in the perisylvian language networks, which suggests a possible division of labor for phonological, syntactic, and semantic function in the left frontal, parietal, and temporal areas.},
}
@inproceedings{XiangetalUAI-96,
    author = {Xiang, Y. and Wong, S. K. M. and Cercone N.},
     title = {Critical Remarks on Single Link Search in Learning Belief Networks},
 booktitle = {Twelfth Conference on Uncertainty in Artificial Intelligence},
  location = {Portland, Oregon},
      year = 1996,
}
@inproceedings{XiaoetalCVPR-10,
      title = {{SUN} Database: Large-scale Scene Recognition from {Abbey} to {Zoo}},
     author = {Jianxiong Xiao and James Hays and Krista A. Ehinger and Aude Oliva and Antonio Torralba},
  booktitle = {Proceedings of International Conference on Computer Vision and Pattern Recognition},
  publisher = {IEEE Computer Society},
       year = 2010,
}
@article{XiaoetalICLR-18,
       author = {Da Xiao and Jo-Yu Liao and Xingyuan Yuan},
        title = {Improving the Universality and Learnability of Neural Programmer-Interpreters with Combinator Abstraction},
      journal = {CoRR},
       volume = {arXiv:1802.02696},
         year = {2018},
     abstract = {To overcome the limitations of Neural Programmer-Interpreters (NPI) in its universality and learnability, we propose the incorporation of combinator abstraction into neural programing and a new NPI architecture to support this abstraction, which we call Combinatory Neural Programmer-Interpreter (CNPI). Combinator abstraction dramatically reduces the number and complexity of programs that need to be interpreted by the core controller of CNPI, while still allowing the CNPI to represent and interpret arbitrary complex programs by the collaboration of the core with the other components. We propose a small set of four combinators to capture the most pervasive programming patterns. Due to the finiteness and simplicity of this combinator set and the offloading of some burden of interpretation from the core, we are able construct a CNPI that is universal with respect to the set of all combinatorizable programs, which is adequate for solving most algorithmic tasks. Moreover, besides supervised training on execution traces, CNPI can be trained by policy gradient reinforcement learning with appropriately designed curricula.},
}
@article{XiaojieetalNATURE-NANO-12,
    author = {Duan, Xiaojie and Gao, Ruixuan and Xie, Ping and Cohen-Karni, Tzahi and Qing, Quan 
             and Choe, Hwan Sung and Tian, Bozhi and Jiang, Xiaocheng and Lieber, Charles M},
     title = {Intracellular recordings of action potentials by an extracellular nanoscale field-effect transistor},
   journal = {Nature Nanotechnology},
    volume = 7,
     issue = 3,
      year = 2012,
     pages = {174-179},
}
@article{XieandSeungNC-03,
       author = {Xie, X.  and Seung, H. S.},
        title = {{{E}quivalence of backpropagation and contrastive {H}ebbian learning in a layered network}},
      journal = {Neural Computation},
       volume = {15},
       number = {2},
         year = {2003},
        pages = {441-454},
     abstract = {Backpropagation and contrastive Hebbian learning are two methods of training networks with hidden neurons. Backpropagation computes an error signal for the output neurons and spreads it over the hidden neurons. Contrastive Hebbian learning involves clamping the output neurons at desired values and letting the effect spread through feedback connections over the entire network. To investigate the relationship between these two forms of learning, we consider a special case in which they are identical: a multilayer perceptron with linear output units, to which weak feedback connections have been added. In this case, the change in network state caused by clamping the output neurons turns out to be the same as the error signal spread by backpropagation, except for a scalar prefactor. This suggests that the functionality of backpropagation can be realized alternatively by a Hebbian-type learning algorithm, which is suitable for implementation in biological networks.}
}
@article{XieetalNATURE-NANO-12,
       author = {Xie, Chong and Lin, Ziliang and Hanson, Lindsey and Cui, Yi and Cui, Bianxiao},
        title = {Intracellular recording of action potentials by nanopillar electroporation},
      journal = {Nature Nanotechnology},
    publisher = {Nature Publishing Group},
       volume = 7,
        issue = 3,
         year = 2012,
        pages = {185-190},
}
@article{XieetalNEUROTV-14,
       author = {L, Xie Diana and L, Miller Steven and Leanne, Boucher and L, Kubie John and Fran{\c{c}}ois, Gari{\`{e}}py Jean-},
        title = {{NEURO.TV}: Neuroscience Education on the Internet},
      journal = {Malaysian Journal of Medical Science},
         year = {2014},
    publisher = {Penerbit Universiti Sains Malaysia},
       volume = {21},
        issue = {3},
        pages = {78-80},
     abstract = {NEURO.tv is a new educational project that seeks to bring advanced concepts in neuroscience to the general public. We film one-hour discussions with leading neuroscientists, philosophers, and psychologists who have had significant impact on our current understanding of brain function, and we publish these discussions on YouTube, iTunes, and other social media outlets. Here, we explain the motivations behind this new program.},
}
@inproceedings{XingetalNIPS-03,
    author = {E. Xing and A. Ng and M. Jordan and S. Russell},
     title = {Distance metric learning, with application to clustering with side-information},
 booktitle = {Advances in Neural Information Processing Systems},
    editor = {Sebastian Thrun and Lawrence Saul},
      year = 2003,
}
@article{XinuanetalCoRR-18,
       author = {Xinyuan Zhang and Yitong Li and Dinghan Shen and Lawrence Carin},
        title = {Diffusion Maps for Textual Network Embedding},
      journal = {CoRR},
       volume = {arXiv:1805.09906},
         year = {2018},
     abstract = {Textual network embedding leverages rich text information associated with the network to learn low-dimensional vectorial representations of vertices. Rather than using typical natural language processing (NLP) approaches, recent research exploits the relationship of texts on the same edge to graphically embed text. However, these models neglect to measure the complete level of connectivity between any two texts in the graph. We present diffusion maps for textual network embedding (DMTE), integrating global structural information of the graph to capture the semantic relatedness between texts, with a diffusion-convolution operation applied on the text inputs. In addition, a new objective function is designed to efficiently preserve the high-order proximity using the graph diffusion. Experimental results show that the proposed approach outperforms state-of-the-art methods on the vertex-classification and link-prediction tasks.}
}
@article{XiongetalCoRR-16,
       author = {Caiming Xiong and Stephen Merity and Richard Socher},
        title = {Dynamic Memory Networks for Visual and Textual Question Answering},
      journal = {CoRR},
       volume = {arXiv:1603.01417},
         year = {2016},
     abstract = {Neural network architectures with memory and attention mechanisms exhibit certain reasoning capabilities required for question answering. One such architecture, the dynamic memory network (DMN), obtained high accuracy on a variety of language tasks. However, it was not shown whether the architecture achieves strong results for question answering when supporting facts are not marked during training or whether it could be applied to other modalities such as images. Based on an analysis of the DMN, we propose several improvements to its memory and input modules. Together with these changes we introduce a novel input module for images in order to be able to answer visual questions. Our new DMN+ model improves the state of the art on both the Visual Question Answering dataset and the \babi-10k text question-answering dataset without supporting fact supervision.}
}
@article{XueetalPMC-10,
       author = {Xue, Gui and Chen, Chuansheng and Lu, Zhong-Lin and Dong, Qi},
        title = {Brain Imaging Techniques and Their Applications in Decision-Making Research},
      journal = {Acta Psychologica Sinica},
         year = {2010},
       volume = {42},
        issue = {1},
        pages = {120-137},
     abstract = {Advanced noninvasive neuroimaging techniques such as EEG and fMRI allow researchers to directly observe brain activities while subjects perform various perceptual, motor, and/or cognitive tasks. By combining functional brain imaging with sophisticated experimental designs and data analysis methods, functions of brain regions and their interactions can be examined. A nascent field called neuroeconomics has recently emerged as a result of the enormous success of applications of functional brain imaging techniques in the study of human decision-making. In this article, we first provide an overview of brain imaging techniques, focusing on the recent developments in multivariate analysis and multi-modal data integration. We then present several studies on risky decision making, intertemporal choice, and social decision making, to illustrate how neuroimaging techniques can be used to advance our knowledge on decision making. Finally, we discuss challenges and future directions in neuroeconomics.}
}
@inproceedings{XuetalAAAI-15,
        title = {Jointly Modeling Deep Video and Compositional Text to Bridge Vision and Language in a Unified Framework},
       author = {Ran Xu and Caiming Xiong and Wei Chen and Jason J. Corso},
    booktitle = {Proceedings of the Twenty-ninth {AAAI} Conference on Artificial Intelligence},
         year = {2015},
     abstract = {Recently, joint video-language modeling has been attracting more and more attention. However, most existing approaches focus on exploring the language model upon on a fixed visual model. In this paper, we propose a unified framework that jointly models video and the corresponding text sentences. The framework consists of three parts: a compositional semantics language model, a deep video model and a joint embedding model. In our language model, we propose a dependency-tree structure model that embeds sentence into a continuous vector space, which preserves visually grounded meanings and word order. In the visual model, we leverage deep neural networks to capture essential semantic information from videos. In the joint embedding model, we minimize the distance of the outputs of the deep video model and compositional language model in the joint space, and update these two models jointly. Based on these three parts, our system is able to accomplish three tasks: 1) natural language generation, and 2) video retrieval and 3) language retrieval. In the experiments, the results show our approach outperforms SVM, CRF and CCA baselines in predicting Subject-Verb-Object triplet and natural sentence generation, and is better than CCA in video retrieval and language retrieval tasks.}
}
@article {XuetalBioRxiv-20,
       author = {Xu, C. Shan and Januszewski, Michal and Lu, Zhiyuan and Takemura, Shin-ya and Hayworth, Kenneth J. and Huang, Gary and Shinomiya, Kazunori and Maitin-Shepard, Jeremy and Ackerman, David and Berg, Stuart and Blakely, Tim and Bogovic, John and Clements, Jody and Dolafi, Tom and Hubbard, Philip and Kainmueller, Dagmar and Katz, William and Kawase, Takashi and Khairy, Khaled A. and Leavitt, Laramie and Li, Peter H. and Lindsey, Larry and Neubarth, Nicole and Olbris, Donald J. and Otsuna, Hideo and Troutman, Eric T. and Umayam, Lowell and Zhao, Ting and Ito, Masayoshi and Goldammer, Jens and Wolff, Tanya and Svirskas, Robert and Schlegel, Philipp and Neace, Erika R. and Knecht, Christopher J. and Alvarado, Chelsea X. and Bailey, Dennis A. and Ballinger, Samantha and Borycz, Jolanta A and Canino, Brandon S. and Cheatham, Natasha and Cook, Michael and Dreher, Marisa and Duclos, Octave and Eubanks, Bryon and Fairbanks, Kelli and Finley, Samantha and Forknall, Nora and Francis, Audrey and Hopkins, Gary Patrick and Joyce, Emily M. and Kim, SungJin and Kirk, Nicole A. and Kovalyak, Julie and Lauchie, Shirley A. and Lohff, Alanna and Maldonado, Charli and Manley, Emily A. and McLin, Sari and Mooney, Caroline and Ndama, Miatta and Ogundeyi, Omotara and Okeoma, Nneoma and Ordish, Christopher and Padilla, Nicholas and Patrick, Christopher and Paterson, Tyler and Phillips, Elliott E. and Phillips, Emily M. and Rampally, Neha and Ribeiro, Caitlin and Robertson, Madelaine K and Rymer, Jon Thomson and Ryan, Sean M. and Sammons, Megan and Scott, Anne K. and Scott, Ashley L. and Shinomiya, Aya and Smith, Claire and Smith, Kelsey and Smith, Natalie L. and Sobeski, Margaret A. and Suleiman, Alia and Swift, Jackie and Takemura, Satoko and Talebi, Iris and Tarnogorska, Dorota and Tenshaw, Emily and Tokhi, Temour and Walsh, John J. and Yang, Tansy and Horne, Jane Anne and Li, Feng and Parekh, Ruchi and Rivlin, Patricia K. and Jayaraman, Vivek and Ito, Kei and Saalfeld, Stephan and George, Reed and Meinertzhagen, Ian and Rubin, Gerald M. and Hess, Harald F. and Scheffer, Louis K. and Jain, Viren and Plaza, Stephen M.},
        title = {A Connectome of the Adult Drosophila Central Brain},
      journal = {bioRxiv},
         year = {2020},
    publisher = {Cold Spring Harbor Laboratory},
     abstract = {The neural circuits responsible for behavior remain largely unknown. Previous efforts have reconstructed the complete circuits of small animals, with hundreds of neurons, and selected circuits for larger animals. Here we (the FlyEM project at Janelia and collaborators at Google) summarize new methods and present the complete circuitry of a large fraction of the brain of a much more complex animal, the fruit fly Drosophila melanogaster. Improved methods include new procedures to prepare, image, align, segment, find synapses, and proofread such large data sets; new methods that define cell types based on connectivity in addition to morphology; and new methods to simplify access to a large and evolving data set. From the resulting data we derive a better definition of computational compartments and their connections; an exhaustive atlas of cell examples and types, many of them novel; detailed circuits for most of the central brain; and exploration of the statistics and structure of different brain compartments, and the brain as a whole. We make the data public, with a web site and resources specifically designed to make it easy to explore, for all levels of expertise from the expert to the merely curious. The public availability of these data, and the simplified means to access it, dramatically reduces the effort needed to answer typical circuit questions, such as the identity of upstream and downstream neural partners, the circuitry of brain regions, and to link the neurons defined by our analysis with genetic reagents that can be used to study their functions.Note: In the next few weeks, we will release a series of papers with more involved discussions. One paper will detail the hemibrain reconstruction with more extensive analysis and interpretation made possible by this dense connectome. Another paper will explore the central complex, a brain region involved in navigation, motor control, and sleep. A final paper will present insights from the mushroom body, a center of multimodal associative learning in the fly brain.},
}
@article{XuetalCEREBRAL-CORTEX-09,
       author = {Xue, G. and Lu, Z. and Levin, I. P. and Weller, J. A. and Li, X. and Bechara, A.},
        title = {{F}unctional dissociations of risk and reward processing in the medial prefrontal cortex},
      journal = {Cerebral Cortex},
         year = {2009},
       volume = {19},
       number = {5},
        pages = {1019-1027},
     abstract = {Making a risky decision is a complex process that involves evaluation of both the value of the options and the associated risk level. Yet the neural processes underlying these processes have not so far been clearly identified. Using functional magnetic resonance imaging and a task that simulates risky decisions, we found that the dorsal region of the medial prefrontal cortex (MPFC) was activated whenever a risky decision was made, but the degree of this activity across subjects was negatively correlated with their risk preference. In contrast, the ventral MPFC was parametrically modulated by the received gain/loss, and the activation in this region was positively correlated with an individual's risk preference. These results extend existing neurological evidence by showing that the dorsal and ventral MPFC convey different decision signals (i.e., aversion to uncertainty vs. approach to rewarding outcomes), where the relative strengths of these signals determine behavioral decisions involving risk and uncertainty.}
}
@article{XuetalCoRR-15,
        title = {Show, Attend and Tell: Neural Image Caption Generation with Visual Attention},
       author = {Kelvin Xu and Jimmy Ba and Ryan Kiros and Kyunghyun Cho and Aaron Courville and Ruslan Salakhutdinov and Richard Zemel and Yoshua Bengio},
      journal = {CoRR},
       volume = {arXiv:1502.03044},
         year = 2015,
     abstract = {Inspired by recent work in machine translation and object detection, we introduce an attention based model that automatically learns to describe the content of images. We describe how we can train this model in a deterministic manner using standard backpropagation techniques and stochastically by maximizing a variational lower bound. We also show through visualization how the model is able to automatically learn to fix its gaze on salient objects while generating the corresponding words in the output sequence. We validate the use of attention with state-of-the-art performance on three benchmark datasets: Flickr8k, Flickr30k and MS COCO.},
}
@article{XuetalCoRR-17,
       author = {Xiaojun Xu and Chang Liu and Qian Feng and Heng Yin and Le Song and Dawn Song},
        title = {Neural Network-based Graph Embedding for Cross-Platform Binary Code Similarity Detection},
      journal = {CoRR},
       volume = {arXiv:1708.06525},
         year = {2017},
     abstract = {The problem of cross-platform binary code similarity detection aims at detecting whether two binary functions coming from different platforms are similar or not. It has many security applications, including plagiarism detection, malware detection, vulnerability search, etc. Existing approaches rely on approximate graph matching algorithms, which are inevitably slow and sometimes inaccurate, and hard to adapt to a new task. To address these issues, in this work, we propose a novel neural network-based approach to compute the embedding, i.e., a numeric vector, based on the control flow graph of each binary function, then the similarity detection can be done efficiently by measuring the distance between the embeddings for two functions. We implement a prototype called Gemini. Our extensive evaluation shows that Gemini outperforms the state-of-the-art approaches by large margins with respect to similarity detection accuracy. Further, Gemini can speed up prior art's embedding generation time by 3 to 4 orders of magnitude and reduce the required training time from more than 1 week down to 30 minutes to 10 hours. Our real world case studies demonstrate that Gemini can identify significantly more vulnerable firmware images than the state-of-the-art, i.e., Genius. Our research showcases a successful application of deep learning on computer security problems.}
}
@inproceedings{XuetalICML-18,
        title = {Representation Learning on Graphs with Jumping Knowledge Networks},
       author = {Xu, Keyulu and Li, Chengtao and Tian, Yonglong and Sonobe, Tomohiro and Kawarabayashi, Ken-ichi and Jegelka, Stefanie},
    booktitle = {Proceedings of the 35th International Conference on Machine Learning},
       editor = {Jennifer Dy and Andreas Krause},
        pages = {5453-5462},
         year = {2018},
       volume = {80},
     abstract = {Recent deep learning approaches for representation learning on graphs follow a neighborhood aggregation procedure. We analyze some important properties of these models, and propose a strategy to overcome those. In particular, the range of "neighboring" nodes that a node's representation draws from strongly depends on the graph structure, analogous to the spread of a random walk. To adapt to local neighborhood properties and tasks, we explore an architecture - jumping knowledge (JK) networks - that flexibly leverages, for each node, different neighborhood ranges to enable better structure-aware representation. In a number of experiments on social, bioinformatics and citation networks, we demonstrate that our model achieves state-of-the-art performance. Furthermore, combining the JK framework with models like Graph Convolutional Networks, GraphSAGE and Graph Attention Networks consistently improves those models' performance.}
}
@article{XuetalIEEE-16,
       author = {C. Xu and C. Lu and X. Liang and J. Gao and W. Zheng and T. Wang and S. Yan},
      journal = {IEEE Transactions on Circuits and Systems for Video Technology},
        title = {Multi-loss Regularized Deep Neural Network},
         year = {2016},
       volume = {26},
       number = {12},
        pages = {2273-2283},
     abstract = {A proper strategy to alleviate overfitting is critical to a deep neural network (DNN). In this paper, we introduce the cross-loss-function regularization for boosting the generalization capability of the DNN, which results in the multi-loss regularized DNN (ML-DNN) framework. For a particular learning task, e.g., image classification, only a single-loss function is used for all previous DNNs, and the intuition behind the multiloss framework is that the extra loss functions with different theoretical motivations (e.g., pairwise loss and LambdaRank loss) may drag the algorithm away from overfitting to one particular single-loss function (e.g., softmax loss). In the training stage, we pretrain the model with the single-core-loss function and then warm start the whole ML-DNN with the convolutional parameters transferred from the pretrained model. In the testing stage, the outputs by the ML-DNN from different loss functions are fused with average pooling to produce the ultimate prediction. The experiments conducted on several benchmark datasets (CIFAR-10, CIFAR-100, MNIST, and SVHN) demonstrate that the proposed ML-DNN framework, instantiated by the recently proposed network in network, considerably outperforms all other state-of-the-art methods.},
}
@article{XuetalSCIENCE-13,
       author = {Xu, Ke and Zhong, Guisheng and Zhuang, Xiaowei},
        title = {Actin, Spectrin, and Associated Proteins Form a Periodic Cytoskeletal Structure in Axons},
      journal = {Science},
       volume = 339,
        issue = 6118,
         year = 2013,
        pages = {452-456},
     abstract = {Actin and spectrin play important roles in neurons, but their organization in axons and dendrites remains unclear. We used stochastic optical reconstruction microscopy to study the organization of actin, spectrin, and associated proteins in neurons. Actin formed ringlike structures that wrapped around the circumference of axons and were evenly spaced along axonal shafts with a periodicity of \approx{}180 to 190 nanometers. This periodic structure was not observed in dendrites, which instead contained long actin filaments running along dendritic shafts. Adducin, an actin-capping protein, colocalized with the actin rings. Spectrin exhibited periodic structures alternating with those of actin and adducin, and the distance between adjacent actin-adducin rings was comparable to the length of a spectrin tetramer. Sodium channels in axons were distributed in a periodic pattern coordinated with the underlying actin-spectrin-based cytoskeleton.},
}
@misc{XuetalSCIENCE-13.supplement,
        Title = {Actin, Spectrin, and Associated Proteins Form a Periodic Cytoskeletal Structure in Axons: Supplementary Materials},
       author = {Xu, Ke and Zhong, Guisheng and Zhuang, Xiaowei},
         year = 2013,
 howpublished = {Available from the pages of Science Magazine},
}
@article{XuetalSCIENCE-ADVANCES-16,
       author = {Xu, Wentao and Min, Sung-Yong and Hwang, Hyunsang and Lee, Tae-Woo},
        title = {Organic core-sheath nanowire artificial synapses with femtojoule energy consumption},
      journal = {Science Advances},
    publisher = {American Association for the Advancement of Science},
       volume = {2},
       number = {6},
         year = {2016},
     abstract = {Emulation of biological synapses is an important step toward construction of large-scale brain-inspired electronics. Despite remarkable progress in emulating synaptic functions, current synaptic devices still consume energy that is orders of magnitude greater than do biological synapses (~10 fJ per synaptic event). Reduction of energy consumption of artificial synapses remains a difficult challenge. We report organic nanowire (ONW) synaptic transistors (STs) that emulate the important working principles of a biological synapse. The ONWs emulate the morphology of nerve fibers. With a core-sheath{\textendash}structured ONW active channel and a well-confined 300-nm channel length obtained using ONW lithography, ~1.23 fJ per synaptic event for individual ONW was attained, which rivals that of biological synapses. The ONW STs provide a significant step toward realizing low-energy{\textendash}consuming artificial intelligent electronics and open new approaches to assembling soft neuromorphic systems with nanometer feature size.},
}
@article{XuetalTNNLS-17,
       author = {David Xu and Andrew Clappison and Cameron Seth and Jeff Orchard},
        title = {Symmetric Predictive Estimator for Biologically Plausible Neural Learning},
      journal = {IEEE Transactions on Neural Networks and Learning Systems},
       volume = {28},
        issue = {10},
         year = {2017},
     abstract = {In a real brain, the act of perception is a bidirectional process, depending on both feedforward sensory pathways and feedback pathways that carry expectations. We are interested in how such a neural network might emerge from a biologically plausible learning rule. Other neural network learning methods either only apply to feedforward networks, or employ assumptions (such as weight copying) that render them unlikely in a real brain. Predictive estimators (PEs) offer a better solution to this bidirectional learning scenario. However, PEs also depend on weight copying. In this paper, we propose the symmetric PE (SPE), an architecture that can learn both feedforward and feedback connection weights individually using only locally available information. We demonstrate that the SPE can learn complicated mappings without the use of weight copying. The SPE networks also show promise in deeper architectures.},
}
@misc{YadavVQ-VAE-19,
        title = {Understanding Vector Quantized Variational Autoencoders},
       author = {Sasshank Kwon},
 howpublished = {\url{https://blog.usejournal.com/understanding-vector-quantized-variational-autoencoders-vq-vae-323d710a888a}},
         year = {2019},
}
@book{YadavetalNEURAL-NET-HISTORY-15,
        title = {An Introduction to Neural Network Methods for Differential Equations},
       author = {Yadav, N. and Yadav, A. and Kumar, M.},
         year = {2015},
    publisher = {Springer Netherlands}
}
@inproceedings{YagniketalICCV-11,
        title = {The Power of Comparative Reasoning},
       author = {Yagnik, Jay and Strelow, Dennis and Ross, David A. and Lin, Ruei-sung},
    booktitle = {Proceedings of the International Conference on Computer Vision},
    publisher = {{IEEE} Computer Society},
         year = 2011,
     abstract = {Rank correlation measures are known for their resilience to perturbations in numeric values and are widely used in many evaluation metrics. Such ordinal measures have rarely been applied in treatment of numeric features as a representational transformation. We emphasize the benefits of ordinal representations of input features both theoretically and empirically. We present a family of algorithms for computing ordinal embeddings based on partial order statistics. Apart from having the stability benefits of ordinal measures, these embeddings are highly nonlinear, giving rise to sparse feature spaces highly favored by several machine learning methods. These embeddings are deterministic, data independent and by virtue of being based on partial order statistics, add another degree of resilience to noise. These machine-learning-free methods when applied to the task of fast similarity search outperform state-of-the-art machine learning methods with complex optimization setups. For solving classification problems, the embeddings provide a nonlinear transformation resulting in sparse binary codes that are well-suited for a large class of machine learning algorithms.}
}
@article{YakovlevetalIEEE-12,
       author = {Yakovlev, A. and Sanghoek Kim and Poon, A.},
      journal = {Communications Magazine, IEEE},
        title = {Implantable biomedical devices: Wireless powering and communication},
         year = 2012,
       volume = 50,
        issue = 4,
        pages = {152-159},
     abstract = {In recent years, there has been major progress on implantable biomedical systems that support most of the functionalities of wireless implantable devices. Nevertheless, these devices remain mostly restricted to research, in part due to limited miniaturization, power supply constraints, and lack of a reliable interface between implants and external devices. This article provides a tutorial on the design of implantable biomedical devices that addresses these limitations. Specifically, it presents analysis and techniques for wireless power transfer and efficient data transfer from both theoretical and practical standpoints. Their potential implementations are also discussed.},
}
@article{YamadaASS-14,
        title = {Historical milestones and future prospects of cluster ion beam technology},
       author = {Isao Yamada},
      journal = {Applied Surface Science},
       volume = {310},
        pages = {77-88},
         year = {2014},
     abstract = {Development of technology for processing of surfaces by means of gas cluster ion beams began only about a quarter century ago even though fundamental research related to generation of gas clusters began much earlier. Industrial applications of cluster ion beams did not start to be explored until commercial equipment was first introduced to the ion beam community in around 2000. The technology is now evolving rapidly with industrial equipment being engineered for many diverse surface processing applications which are made possible by the unique characteristics of cluster-ion/solid-surface interactions. In this paper, important historical milestones in cluster ion beam development are described. Present activities related to a wide range of industrial applications in semiconductors, magnetic and optical devices, and bio-medical devices are reviewed. Several emerging new advances in cluster beam applications for the future are also discussed.}
}
@article{YamamotoandLopez-BenditoEJoN-12,
        title = {Shaping brain connections through spontaneous neural activity},
       author = {Nobuhiko Yamamoto and Guillermina L{\'o}pez-Bendito},
      journal = {The European Journal of Neuroscience},
         year = {2012},
       volume = {35},
       number = {10},
        pages = {1595-604},
     abstract = {An overwhelming number of observations demonstrate that neural activity and genetic programs interact to specify the composition and organization of neural circuits during all stages of development. Spontaneous neuronal activities have been documented in several developing neural regions in both invertebrates and vertebrates, and their roles are mostly conserved among species. Among these roles, Ca(2+) spikes and levels of electrical activity have been shown to regulate neurite growth, axon extension and axon branching. Here, we review selected findings concerning the role of spontaneous activity on circuit development.}
}
@incollection{YamaneetalICNIP-15,
       author = {Yamane, Toshiyuki and Katayama, Yasunao and Nakane, Ryosho and Tanaka, Gouhei and Nakano, Daiju},
       editor = {Arik, Sabri and Huang, Tingwen and Lai, Kin Weng and Liu, Qingshan},
        title = {Wave-Based Reservoir Computing by Synchronization of Coupled Oscillators},
    booktitle = {22nd International Conference on Neural Information Processing},
    publisher = {Springer International Publishing},
         year = {2015},
        pages = {198-205},
}
@article{YamaneetalJN-06,
        title = {Representation of the Spatial Relationship Among Object Parts by Neurons in Macaque Inferotemporal Cortex},
       author = {Yamane, Yukako and Tsunoda, Kazushige and Matsumoto, Madoka and Phillips, Adam N. and Tanifuji, Manabu},
      journal = {Journal of Neurophysiology},
       volume = 96,
        issue = 6,
         year = 2006,
        pages = {3147-3156},
     abstract = {We investigated object representation in area TE, the  anterior part of monkey inferotemporal (IT) cortex, with a combination of optical and extracellular recordings in anesthetized monkeys. We found neurons that respond to visual stimuli composed of naturally distinguishable parts. These neurons were sensitive to a particular spatial arrangement of parts but less sensitive to differences in local features within individual parts. Thus these neurons were activated when arbitrary local features were arranged in a particular spatial configuration, suggesting that they may be responsible for representing the spatial configuration of object images. Previously it has been reported that many neurons in area TE respond to visual features less complex than natural objects, but it has remained unclear whether these features are related to local features of object images or to more global features. These results indicate that TE neurons represent not only local features but also global features such as the spatial relationship among object parts.},
}
@article{YamaneetalNATURE-08,
       author = {Yamane, Yukako and Carlson, Eric T. and Bowman, Katherine C. and Wang, Zhihong and Connor, Charles E.},
        title = {A neural code for three-dimensional object shape in macaque inferotemporal cortex},
      journal = {Nature Neuroscience},
       volume = 11,
        issue = 11,
         year = 2008,
        pages = {1352-1360},
}
@article{YamashitaetalNN-08,
        title = {Developmental learning of complex syntactical song in the Bengalese finch: A neural network model},
       author = {Yuichi Yamashita and Miki Takahasi and Tetsu Okumura and Maki Ikebuchi and Hiroko Yamada and Madoka Suzuki and Kazuo Okanoya and Jun Tani},
      journal = {Neural Networks},
       volume = 21,
        issue = 9,
         year = 2008,
        pages = {1224-1231},
     abstract = {We developed a neural network model for studying neural mechanisms underlying complex syntactical songs of the Bengalese finch, which result from interactions between sensori-motor nuclei, the nucleus \{HVC\} (HVC) and the nucleus interfacialis (NIf). Results of simulations are tested by comparison with the song development of real young birds learning the same songs from their fathers. The model shows that complex syntactical songs can be reproduced from the simple interaction between the deterministic dynamics of a recurrent neural network and random noise. Features of the learning process in the simulations show similar trends to those observed in empirical data on the song development of real birds. These observations suggest that the temporal note sequences of songs take the form of a dynamical process involving recurrent connections in the network of the HVC, as opposed to feedforward activities, the mechanism proposed in the previous model.}
}
@article{YamazakiandTanakaNN-07,
        title = {The cerebellum as a liquid state machine},
       author = {Tadashi Yamazaki and Shigeru Tanaka},
      journal = {Neural Networks},
       volume = 20,
        issue = 3,
         year = 2007,
        pages = {290-297},
     abstract = {We examined closely the cerebellar circuit model that we have proposed previously. The model granular layer generates a finite but very long sequence of active neuron populations without recurrence, which is able to represent the passage of time. For all the possible binary patterns fed into mossy fibres, the circuit generates the same number of different sequences of active neuron populations. Model Purkinje cells that receive parallel fiber inputs from neurons in the granular layer learn to stop eliciting spikes at the timing instructed by the arrival of signals from the inferior olive. These functional roles of the granular layer and Purkinje cells are regarded as a liquid state generator and readout neurons, respectively. Thus, the cerebellum that has been considered to date as a biological counterpart of a perceptron is reinterpreted to be a liquid state machine that possesses powerful information processing capability more than a perceptron.}
}
@inproceedings{YaminsIJCAAMS-05,
       author = {Yamins, Daniel},
        title = {Towards a Theory of "Local to Global" in Distributed Multi-agent Systems (I)},
    booktitle = {Proceedings of the Fourth International Joint Conference on Autonomous Agents and Multiagent Systems},
     location = {The Netherlands},
    publisher = {ACM},
      address = {New York, NY, USA},
         year = {2005},
        pages = {183-190},
}
@techreport{YaminsPhD-07,
       author = {Yamins, Daniel},
        title = {A theory of local-to-global algorithms for one-dimensional spatial multi-agent systems},
  institution = {Harvard University},
      address = {Cambridge, Massachusetts},
         year = 2007,
}
@article{YaminsandDiCarloCOIN-16,
        title = {Eight open problems in the computational modeling of higher sensory cortex},
       author = {Yamins, D.~L. and DiCarlo, J.~J.},
      journal = {Current Opinion in Neurobiology},
       volume = {37},
         year = {2016},
        pages = {144-120},
     abstract = {Propelled by recent advances in biologically-inspired computer vision and artificial intelligence, the past five years have seen significant progress in using deep neural networks to model response patterns of neurons in higher visual cortical areas. In this paper, we briefly review this progress and then discuss eight key "open questions" that we believe will drive research in computational models of sensory systems over the next five years, both in visual cortex and beyond. Throughout, our focus is on challenges that will require both cutting-edge algorithmic developments as well as next-generation neuroscience and cognitive science experiments.}
}
@article{YaminsetalCoRR-19,
       author = {Chengxu Zhuang and Alex Lin Zhai and Daniel Yamins},
        title = {Local Aggregation for Unsupervised Learning of Visual Embeddings},
      journal = {CoRR},
       volume = {arXiv:1903.12355},
         year = {2019},
     abstract = {Unsupervised approaches to learning in neural networks are of substantial interest for furthering artificial intelligence, both because they would enable the training of networks without the need for large numbers of expensive annotations, and because they would be better models of the kind of general-purpose learning deployed by humans. However, unsupervised networks have long lagged behind the performance of their supervised counterparts, especially in the domain of large-scale visual recognition. Recent developments in training deep convolutional embeddings to maximize non-parametric instance separation and clustering objectives have shown promise in closing this gap. Here, we describe a method that trains an embedding function to maximize a metric of local aggregation, causing similar data instances to move together in the embedding space, while allowing dissimilar instances to separate. This aggregation metric is dynamic, allowing soft clusters of different scales to emerge. We evaluate our procedure on several large-scale visual recognition datasets, achieving state-of-the-art unsupervised transfer learning performance on object recognition in ImageNet, scene recognition in Places 205, and object detection in PASCAL VOC.},
}
@inproceedings{YaminsetalNIPS-13,
        title = {Hierarchical Modular Optimization of Convolutional Networks Achieves Representations Similar to Macaque IT and Human Ventral Stream},
       author = {Yamins, D.~L. and Hong, H. and Cadieu, C. and J.~J. {DiCarlo}},
    booktitle = {Advances in Neural Information Processing Systems 26},
      address = {Tahoe, CA},
         year = 2013,
        pages = {3093-3101},
     abstract = {Humans recognize visually-presented objects rapidly and accurately. To understand this ability, we seek to construct models of the ventral stream, the series of cortical areas thought to subserve object recognition. One tool to assess the quality of a model of the ventral stream is the Representation Dissimilarity Matrix (RDM), which uses a set of visual stimuli and measures the distances produced in either the brain (i.e. fMRI voxel responses, neural firing rates) or in models (features). Previous work has shown that all known models of the ventral stream fail to capture the RDM pattern observed in either IT cortex, the highest ventral area, or in the human ventral stream. In this work, we construct models of the ventral stream using a novel optimization procedure for category-level object recognition problems, and produce RDMs resembling both macaque IT and human ventral stream. The model, while novel in the optimization procedure, further develops a long-standing functional hypothesis that the ventral visual stream is a hierarchically arranged series of processing stages optimized for visual object recognition.}
}
@article{YaminsetalPNAS-14,
       author = {Yamins, Daniel L. K. and Hong, Ha and Cadieu, Charles F. and Solomon, Ethan A. and Seibert, Darren and DiCarlo, James J.},
        title = {Performance-optimized hierarchical models predict neural responses in higher visual cortex},
      journal = {Proceedings of the National Academy of Sciences},
       volume = {111},
       number = {23},
         year = {2014},
        pages = {8619-8624},
     abstract = {The ventral visual stream underlies key human visual object recognition abilities. However, neural encoding in the higher areas of the ventral stream remains poorly understood. Here, we describe a modeling approach that yields a quantitatively accurate model of inferior temporal (IT) cortex, the highest ventral cortical area. Using high-throughput computational techniques, we discovered that, within a class of biologically plausible hierarchical neural network models, there is a strong correlation between a model's categorization performance and its ability to predict individual IT neural unit response data. To pursue this idea, we then identified a high-performing neural network that matches human performance on a range of recognition tasks. Critically, even though we did not constrain this model to match neural data, its top output layer turns out to be highly predictive of IT spiking responses to complex naturalistic images at both the single site and population levels. Moreover, the model's intermediate layers are highly predictive of neural responses in the V4 cortex, a midlevel visual area that provides the dominant cortical input to IT. These results show that performance optimization--applied in a biologically appropriate model class--can be used to build quantitative predictive models of neural processing.},
}
@article {YaminsetalPNAS-2014,
       author = {Yamins, Daniel L. K. and Hong, Ha and Cadieu, Charles F. and Solomon, Ethan A. and Seibert, Darren and DiCarlo, James J.},
        title = {Performance-optimized hierarchical models predict neural responses in higher visual cortex},
      journal = {Proceedings of the National Academy of Sciences},
    publisher = {National Academy of Sciences},
       volume = {111},
       number = {23},
         year = {2014},
        pages = {8619-8624},
     abstract = {Humans and monkeys easily recognize objects in scenes. This ability is known to be supported by a network of hierarchically interconnected brain areas. However, understanding neurons in higher levels of this hierarchy has long remained a major challenge in visual systems neuroscience. We use computational techniques to identify a neural network model that matches human performance on challenging object categorization tasks. Although not explicitly constrained to match neural data, this model turns out to be highly predictive of neural responses in both the V4 and inferior temporal cortex, the top two layers of the ventral visual hierarchy. In addition to yielding greatly improved models of visual cortex, these results suggest that a process of biological performance optimization directly shaped neural mechanisms.The ventral visual stream underlies key human visual object recognition abilities. However, neural encoding in the higher areas of the ventral stream remains poorly understood. Here, we describe a modeling approach that yields a quantitatively accurate model of inferior temporal (IT) cortex, the highest ventral cortical area. Using high-throughput computational techniques, we discovered that, within a class of biologically plausible hierarchical neural network models, there is a strong correlation between a model{\textquoteright}s categorization performance and its ability to predict individual IT neural unit response data. To pursue this idea, we then identified a high-performing neural network that matches human performance on a range of recognition tasks. Critically, even though we did not constrain this model to match neural data, its top output layer turns out to be highly predictive of IT spiking responses to complex naturalistic images at both the single site and population levels. Moreover, the model{\textquoteright}s intermediate layers are highly predictive of neural responses in the V4 cortex, a midlevel visual area that provides the dominant cortical input to IT. These results show that performance optimization{\textemdash}applied in a biologically appropriate model class{\textemdash}can be used to build quantitative predictive models of neural processing.},
}
@article{YanetalCoRR-20,
        title = {Neural Execution Engines: Learning to Execute Subroutines},
       author = {Yujun Yan and Kevin Swersky and Danai Koutra and Parthasarathy Ranganathan and Milad Hashemi},
      journal = {CoRR},
       volume = {arXiv:2006.08084},
         year = {2020},
}
@article{YanetalJoP-12,
       author = {Xiaogang Yan and Ankit Khambhati and Lei Liu and Tai Sing Lee},
        title = {Neural dynamics of image representation in the primary visual cortex},
      journal = {Journal of Physiology Paris},
       volume = 106,
        issue = {5-6},
         year = 2012,
        pages = {250-256},
     abstract = {Horizontal connections in the primary visual cortex have been hypothesized to play a number of computational roles: association field for contour completion, surface interpolation, surround suppression, and saliency computation. Here, we argue that horizontal connections might also serve a critical role of computing the appropriate codes for image representation. That the early visual cortex or V1 explicitly represents the image we perceive has been a common assumption on computational theories of efficient coding (Olshausen and Field 1996), yet such a framework for understanding the circuitry in V1 has not been seriously entertained in the neurophysiological community. In fact, a number of recent fMRI and neurophysiological studies cast doubt on the neural validity of such an isomorphic representation (Cornelissen et al. 2006, von der Heydt et al. 2003). In this study, we investigated, neurophysiologically, how V1 neurons respond to uniform color surfaces and show that spiking activities of neurons can be decomposed into three components: a bottom-up feedforward input, an articulation of color tuning and a contextual modulation signal that is inversely proportional to the distance away from the bounding contrast border. We demonstrate through computational simulations that the behaviors of a model for image representation are consistent with many aspects of our neural observations. We conclude that the hypothesis of isomorphic representation of images in V1 remains viable and this hypothesis suggests an additional new interpretation of the functional roles of horizontal connections in the primary visual cortex.}
}
@inproceedings{YanetalNIPS-20,
        title = {Neural Execution Engines: Learning to Execute Subroutines},
       author = {Yujun Yan and Kevin Swersky and Danai Koutra and Parthasarathy Ranganathan and Milad Hashemi},
       editor = {H. Larochelle and M. Ranzato and R. Hadsell and M.F. Balcan and H. Lin},
    booktitle = {Advances in Neural Information Processing Systems 33},
         year = {2020},
     abstract = {A significant effort has been made to train neural networks that replicate algorithmic reasoning, but they often fail to learn the abstract concepts underlying these algorithms. This is evidenced by their inability to generalize to data distributions that are outside of their restricted training sets, namely larger inputs and unseen data. We study these generalization issues at the level of numerical subroutines that comprise common algorithms like sorting, shortest paths, and minimum spanning trees. First, we observe that transformer-based sequence-to-sequence models can learn subroutines like sorting a list of numbers, but their performance rapidly degrades as the length of lists grows beyond those found in the training set. We demonstrate that this is due to attention weights that lose fidelity with longer sequences, particularly when the input numbers are numerically similar. To address the issue, we propose a learned conditional masking mechanism, which enables the model to strongly generalize far outside of its training range with near-perfect accuracy on a variety of algorithms. Second, to generalize to unseen data, we show that encoding numbers with a binary representation leads to embeddings with rich structure once trained on downstream tasks like addition or multiplication. This allows the embedding to handle missing data by faithfully interpolating numbers not seen during training.}
}
@article{YangandHospedalesCoRR-16,
       author = {Yongxin Yang and Timothy M. Hospedales},
        title = {Deep Multi-task Representation Learning: {A} Tensor Factorisation Approach},
      journal = {CoRR},
       volume = {arXiv:605.06391},
         year = {2016},
     abstract = {Most contemporary multi-task learning methods assume linear models. This setting is considered shallow in the era of deep learning. In this paper, we present a new deep multi-task representation learning framework that learns cross-task sharing structure at every layer in a deep network. Our approach is based on generalising the matrix factorisation techniques explicitly or implicitly used by many conventional MTL algorithms to tensor factorisation, to realise automatic learning of end-to-end knowledge sharing in deep networks. This is in contrast to existing deep learning approaches that need a user-defined multi-task sharing strategy. Our approach applies to both homogeneous and heterogeneous MTL. Experiments demonstrate the efficacy of our deep multi-task representation learning in terms of both higher accuracy and fewer design choices.}
}
@inproceedings{YangandToderciCVPR-11, 
      title = {Discriminative Tag Learning on YouTube Videos with Latent Sub-tags},
     author = {Weilong Yang and George Toderci},
  booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
       year = 2011
}
@inproceedings{YangetalCVPR-09,
     author = {Jianchao Yang and Kai Yu and Yihong Gong and Thomas Huang},
      title = {Linear Spatial Pyramid Matching Using Sparse Coding for Image Classification},
  booktitle = {{IEEE} Computer Society Conference on Computer Vision and Pattern Recognition},
  publisher = {IEEE Computer Society},
       year = 2009,
}
@inproceedings{YangetalNAACL-16,
       author = {Yang, Zichao  and  Yang, Diyi  and  Dyer, Chris  and  He, Xiaodong  and  Smola, Alex  and  Hovy, Eduard},
        title = {Hierarchical Attention Networks for Document Classification},
    booktitle = {Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics},
    publisher = {Association for Computational Linguistics},
      address = {San Diego, California},
         year = {2016},
        pages = {1480-1489},
     abstract = {We propose a hierarchical attention network for document classification. Our model has two distinctive characteristics: (i) it has a hierarchical structure that mirrors the hierarchical structure of documents; (ii) it has two levels of attention mechanisms applied at the wordand sentence-level, enabling it to attend differentially to more and less important content when constructing the document representation. Experiments conducted on six large scale text classification tasks demonstrate that the proposed architecture outperform previous methods by a substantial margin. Visualization of the attention layers illustrates that the model selects qualitatively informative words and sentences.},
}
@article{YangetalNATURE-MATERIALS-19,
       author = {Yang, Xiao and Zhou, Tao and Zwang, Theodore J. and Hong, Guosong and Zhao, Yunlong and Viveros, Robert D. and Fu, Tian-Ming and Gao, Teng and Lieber, Charles M.},
        title = {Bioinspired neuron-like electronics},
      journal = {Nature Materials},
         year = {2019},
     abstract = {As an important application of functional biomaterials, neural probes have contributed substantially to studying the brain. Bioinspired and biomimetic strategies have begun to be applied to the development of neural probes, although these and previous generations of probes have had structural and mechanical dissimilarities from their neuron targets that lead to neuronal loss, neuroinflammatory responses and measurement instabilities. Here, we present a bioinspired design for neural probes neuron-like electronics (NeuE) where the key building blocks mimic the subcellular structural features and mechanical properties of neurons. Full three-dimensional mapping of implanted NeuE brain interfaces highlights the structural indistinguishability and intimate interpenetration of NeuE and neurons. Time-dependent histology and electrophysiology studies further reveal a structurally and functionally stable interface with the neuronal and glial networks shortly following implantation, thus opening opportunities for next-generation brain machine interfaces. Finally, the NeuE subcellular structural features are shown to facilitate migration of endogenous neural progenitor cells, thus holding promise as an electrically active platform for transplantation-free regenerative medicine.},
}
@article{YangetalNATURE-NANO-13,
       author = {Yang, J. J. and Strukov, D. B. and Stewart, D. R.},
        title = {Memristive devices for computing},
      journal = {Nature Nanotechnology},
         year = 2013,
       volume = 8,
        issue = 1,
        pages = {13-24},
}
@article{YangetalTIPS-10,
       author = {Yang, Jianchao and Wright, John and Huang, Thomas S. and Ma, Yi},
        title = {Image Super-resolution via Sparse Representation},
      journal = {Transactions Image Processing},
    publisher = {IEEE Press},
       volume = {19},
       number = {11},
         year = {2010},
        pages = {2861-2873},
     abstract = {This paper presents a new approach to single-image superresolution, based upon sparse signal representation. Research on image statistics suggests that image patches can be well-represented as a sparse linear combination of elements from an appropriately chosen over-complete dictionary. Inspired by this observation, we seek a sparse representation for each patch of the low-resolution input, and then use the coefficients of this representation to generate the high-resolution output. Theoretical results from compressed sensing suggest that under mild conditions, the sparse representation can be correctly recovered from the downsampled signals. By jointly training two dictionaries for the low- and high-resolution image patches, we can enforce the similarity of sparse representations between the low-resolution and high-resolution image patch pair with respect to their own dictionaries. Therefore, the sparse representation of a low-resolution image patch can be applied with the high-resolution image patch dictionary to generate a high-resolution image patch. The learned dictionary pair is a more compact representation of the patch pairs, compared to previous approaches, which simply sample a large amount of image patch pairs [1], reducing the computational cost substantially. The effectiveness of such a sparsity prior is demonstrated for both general image super-resolution (SR) and the special case of face hallucination. In both cases, our algorithm generates high-resolution images that are competitive or even superior in quality to images produced by other similar SR methods. In addition, the local sparse modeling of our approach is naturally robust to noise, and therefore the proposed algorithm can handle SR with noisy inputs in a more unified framework.},
}
@inproceedings{YannakakisSTOC-91,
    author = {Yannakakis, Mihalis and Lee, David},
     title = {Testing Finite State Machines},
 booktitle = STOC,
 publisher = {{ACM} Press},
      year = 1991,
     pages = {476-485},
}
@inproceedings{YaoetalCVPR-11,
    author = {Bangpeng Yao and Aditya Khosla and Li Fei-Fei},
     title = {Combining Randomization and Discrimination for Fine-Grained Image Categorization},
 booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
      year = 2011,
     pages = {1577-1584},
}
@inproceedings{YaoetalCVPR-12,
    author = {Bangpeng Yao and Gary Bradski and Li Fei-Fei},
     title = {A Codebook-Free and Annotation-Free Approach for Fine-Grained Image Categorization},
 booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
      year = 2012,
   address = {Providence, RI},
}
@article{YapicietalEMBO-14,
       author = {Yapici, N. and Zimmer, M. and Domingos, A. I.},
        title = {Cellular and molecular basis of decision-making},
      journal = {EMBO Reports},
         year = {2014},
       volume = {15},
       number = {10},
        pages = {1023-1035},
     abstract = {People think they are in control of their own decisions: what to eat or drink, whom to marry or pick a fight with, where to live, what to buy. Behavioural economists and neurophysiologists have long studied decision-making behaviours. However, these behaviours have only recently been studied through the light of molecular genetics. Here, we review recent research in mice, Drosophila melanogaster and Caenorhabditis elegans, that analyses the molecular and cellular mechanisms underlying decision-making. These studies interrogate decision-making about food, sexual behaviour, aggression or foraging strategies, and add molecular and cell biology understanding onto the consilience of brain and decision.}
}
@book{Yarbus1967,
       author = {Yarbus, A. L.},
        title = {Eye Movements and Vision},
         year = {Plenum Press},
    publisher = {New York},
      address = 1967
}
@article{YasudaetalCELL-98,
       author = {R. Yasuda and H. Noji and K. Kinosita Jr. and M. Yoshida},
        title = {{F}$_1$-{ATP}ase is a highly efficient molecular motor that rotates with discrete $120^\circ$ steps},
      journal = {Cell},
       volume = 93,
         year = 1998,
        pages = {1117-1124},
}
@article{YeatmanetalPLoS-12,
       author = {Yeatman, J. D. and Dougherty, R. F. and Myall, N. J. and Wandell, B. A. and Feldman, H. M.},
        title = {{{T}ract profiles of white matter properties: automating fiber-tract quantification}},
      journal = {PLoS ONE},
         year = {2012},
       volume = {7},
       number = {11},
        pages = {e49790},
     abstract = {Tractography based on diffusion weighted imaging (DWI) data is a method for identifying the major white matter fascicles (tracts) in the living human brain. The health of these tracts is an important factor underlying many cognitive and neurological disorders. In vivo, tissue properties may vary systematically along each tract for several reasons: different populations of axons enter and exit the tract, and disease can strike at local positions within the tract. Hence quantifying and understanding diffusion measures along each fiber tract (Tract Profile) may reveal new insights into white matter development, function, and disease that are not obvious from mean measures of that tract. We demonstrate several novel findings related to Tract Profiles in the brains of typically developing children and children at risk for white matter injury secondary to preterm birth. First, fractional anisotropy (FA) values vary substantially within a tract but the Tract FA Profile is consistent across subjects. Thus, Tract Profiles contain far more information than mean diffusion measures. Second, developmental changes in FA occur at specific positions within the Tract Profile, rather than along the entire tract. Third, Tract Profiles can be used to compare white matter properties of individual patients to standardized Tract Profiles of a healthy population to elucidate unique features of that patient's clinical condition. Fourth, Tract Profiles can be used to evaluate the association between white matter properties and behavioral outcomes. Specifically, in the preterm group reading ability is positively correlated with FA measured at specific locations on the left arcuate and left superior longitudinal fasciculus and the magnitude of the correlation varies significantly along the Tract Profiles. We introduce open source software for automated fiber-tract quantification (AFQ) that measures Tract Profiles of MRI parameters for 18 white matter tracts. With further validation, AFQ Tract Profiles have potential for informing clinical management and decision-making.}
}
@incollection{Yedidiaetal03,
       author = {Jonathan S. Yedidia and William T. Freeman and Yair Weiss},
        title = {Understanding belief propagation and its generalizations},
    booktitle = {Exploring artificial intelligence in the new millennium},
         year = 2003,
        pages = {239-269},
       editor = {Gerhard Lakemeyer and Bernhard Nebel},
    publisher = {Morgan Kaufmann Publishers Inc.},
      address = {San Francisco},
}
@inproceedings{YeetalCVPR-10,
       author = {Guangnan Ye and Dong Liu and I-Hong Jhuo and Shih-Fu Chang},
        title = {Robust Late Fusion with Rank Minimization},
    booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
         year = 2012,
}
@inproceedings{YellinSODA-92,
       author = {Daniel Yellin},
        title = {Algorithms for subset testing and finding maximal sets},
    booktitle = {Proceedings of the 3rd Annual ACM-SIAM Symposium on Discrete Algorithms},
        pages = {386-392},
     location = {Orlando, Florida},
         year = 1992,
}
@inproceedings{YeminiFOCS-79,
       author = {Yemini, Yechiam},
        title = {Some Theoretical Aspects of Position-Location Problems},
    booktitle = {Proceedings of the 20th Symposium on the Foundations of Computer Science},
         year = 1979,
        pages = {1-7},
}
@article{YenetalJN-97,
        title = {Heterogeneity in the Responses of Adjacent Neurons to Natural Stimuli in Cat Striate Cortex},
       author = {Yen, S.C. and Baker, J. and C.M. Gray},
      journal = {Journal of Neurophysiology},
       volume = 97,
        pages = {1326-1341},
         year = 2007,
}
@article{YildizetalPLoS-13,
       author = {Yildiz, Izzet B. and von Kriegstein, Katharina and Kiebel, Stefan J.},
      journal = {{PLoS} Compututational Biology},
    publisher = {Public Library of Science},
        title = {From Birdsong to Human Speech Recognition: Bayesian Inference on a Hierarchy of Nonlinear Dynamical Systems},
         year = 2013,
       volume = 9,
        pages = {e10013219},
     abstract = {Neuroscience still lacks a concrete explanation of how humans recognize speech. Even though neuroimaging techniques are helpful in determining the brain areas involved in speech recognition, there are rarely mechanistic explanations at a neuronal level. Here, we assume that songbirds and humans solve a very similar task: extracting information from sound wave modulations produced by a singing bird or a speaking human. Given strong evidence that both humans and songbirds, although genetically very distant, converged to a similar solution, we combined the vast amount of neurobiological findings for songbirds with nonlinear dynamical systems theory to develop a hierarchical, Bayesian model which explains fundamental functions in recognition of sound sequences. We found that the resulting model is good at learning and recognizing human speech. We suggest that this translated model can be used to qualitatively explain or predict experimental data, and the underlying mechanism can be used to construct improved automatic speech recognition algorithms.},
}
@inproceedings{YilmazetalACM-14,
       author = {Emre Yilmaz and Konstantinos Rematas and Tinne Tuytelaars and Hugo Van hamme},
        title = {Learning Like a Toddler: Watching Television Series to Learn Vocabulary from Images and Audio},
    booktitle = {ACM Multimedia},
         year = {2014},
     abstract = {This paper presents the initial findings of the efforts to build an unsupervised multimodal learning scheme in a realistic scenario. For this purpose, a new multimodal dataset, namely Musti3D, containing episodes from an animation series for toddlers has been created. Annotated with audiovisual information, this database is used for the investigation of a non-negative matrix factorization (NMF)-based audiovisual learning technique. The performance of the technique, i.e. correctly matching the audio and visual representations of the objects from Musti3D, has been evaluated by gradually reducing the level of supervision starting from the ground truth transcriptions. Moreover, we have performed experiments using different visual representations and time spans for combining the audiovisual information. The preliminary results have shown the feasibility of the proposed audiovisual learning framework.}
}
@inproceedings{YinandNeubigACL-17,
        title = {A Syntactic Neural Model for General-Purpose Code Generation},
       author = {Pengcheng Yin and Graham Neubig},
    booktitle = {The 55th Annual Meeting of the Association for Computational Linguistics (ACL)},
      address = {Vancouver, Canada},
         year = {2017},
     abstract = {We consider the problem of parsing natural language descriptions into source code written in a general-purpose programming language like Python. Existing data-driven methods treat this problem as a language generation task without considering the underlying syntax of the target programming language. Informed by previous work in semantic parsing, in this paper we propose a novel neural architecture powered by a grammar model to explicitly capture the target syntax as prior knowledge. Experiments find this an effective way to scale up to generation of complex programs from natural language descriptions, achieving state-of-the-art results that well outperform previous code generation and semantic parsing approaches.}
}
@inproceedings{YingetalICLR-18,
        title = {Interactive Grounded Language Acquisition and Generalization in a 2D World},
       author = {Haonan Yu and Haichao Zhang and Wei Xu},
    booktitle = {International Conference on Learning Representations},
         year = {2018},
     abstract = {We build a virtual agent for learning language in a 2D maze-like world. The agent sees images of the surrounding environment, listens to a virtual teacher, and takes actions to receive rewards. It interactively learns the teacher's language from scratch based on two language use cases: sentence-directed navigation and question answering. It learns simultaneously the visual representations of the world, the language, and the action control. By disentangling language grounding from other computational routines and sharing a concept detection function between language grounding and prediction, the agent reliably interpolates and extrapolates to interpret sentences that contain new word combinations or new words missing from training sentences. The new words are transferred from the answers of language prediction. Such a language ability is trained and evaluated on a popu- lation of over 1.6 million distinct sentences consisting of 119 object words, 8 color words, 9 spatial-relation words, and 50 grammatical words. The proposed model significantly outperforms five comparison methods for interpreting zero-shot sentences. In addition, we demonstrate human-interpretable intermediate outputs of the model in the appendix.},
}
@incollection{YitongetalNIPS-18,
        title = {Extracting Relationships by Multi-Domain Matching},
       author = {Li, Yitong and Murias, Michael and Dawson, Geraldine and Carlson, David E},
    booktitle = {Advances in Neural Information Processing Systems 31},
       editor = {S. Bengio and H. Wallach and H. Larochelle and K. Grauman and N. Cesa-Bianchi and R. Garnett},
        pages = {6799-6810},
         year = {2018},
    publisher = {Curran Associates, Inc.},
     abstract = {In many biological and medical contexts, we construct a large labeled corpus by aggregating many sources to use in target prediction tasks. Unfortunately, many of the sources may be irrelevant to our target task, so ignoring the structure of the dataset is detrimental. This work proposes a novel approach, the Multiple Domain Matching Network (MDMN), to exploit this structure. MDMN embeds all data into a shared feature space while learning which domains share strong statistical relationships. These relationships are often insightful in their own right, and they allow domains to share strength without interference from irrelevant data. This methodology builds on existing distribution-matching approaches by assuming that source domains are varied and outcomes multi-factorial. Therefore, each domain should only match a relevant subset. Theoretical analysis shows that the proposed approach can have a tighter generalization bound than existing multiple-domain adaptation approaches. Empirically, we show that the proposed methodology handles higher numbers of source domains (up to 21 empirically), and provides state-of-the-art performance on image, text, and multi-channel time series classification, including clinical outcome data in an open label trial evaluating a novel treatment for Autism Spectrum Disorder.} 
}
@article{YizharetalNEURON-11,
        title = {Optogenetics in neural systems},
       author = {Yizhar, O. and Fenno, L.E. and Davidson, T.J. and Mogri, M. and Deisseroth, K.},
      journal = {Neuron},
       volume = 71,
        issue = 1,
        pages = {9-34},
         year = 2011,
     abstract = {Both observational and perturbational technologies are essential for advancing the understanding of brain function and dysfunction. But while observational techniques have greatly advanced in the last century, techniques for perturbation that are matched to the speed and heterogeneity of neural systems have lagged behind. The technology of optogenetics represents a step toward addressing this disparity. Reliable and targetable single-component tools (which encompass both light sensation and effector function within a single protein) have enabled versatile new classes of investigation in the study of neural systems. Here we provide a primer on the application of optogenetics in neuroscience, focusing on the single-component tools and highlighting important problems, challenges, and technical considerations.}
}
@article{YogatamaetalICLR-18,
        title = {Memory Architectures in Recurrent Neural Network Language Models},
       author = {Dani Yogatama and Yishu Miao and Gabor Melis and Wang Ling and Adhiguna Kuncoro and Chris Dyer and Phil Blunsom},
      journal = {International Conference on Learning Representations (Accepted)},
         year = {2018},
          url = {https://openreview.net/forum?id=SkFqf0lAZ},
     abstract = {We compare and analyze sequential, random access, and stack memory architectures for recurrent neural network language models. Our experiments on the Penn Treebank and Wikitext-2 datasets show that stack-based memory architectures consistently achieve the best performance in terms of held out perplexity. We also propose a generalization to existing continuous stack models (Joulin \& Mikolov,2015; Grefenstette et al., 2015)  to allow a variable number of pop operations more naturally that further improves performance. We further evaluate these language models in terms of their ability to capture non-local syntactic dependencies on a subject-verb agreement dataset  (Linzen et al., 2016) and establish new state of the art results using memory augmented language models. Our results demonstrate the value of stack-structured memory for explaining the distribution of words in natural language, in line with linguistic theories claiming a context-free backbone for natural language.}
}
@book{YongMULTITUDES-16,
        title = {I Contain Multitudes: The Microbes Within Us and a Grander View of Life},
       author = {Yong, Ed},
    publisher = {HarperCollins Publishers},
         year = {2016},
}
@inproceedings{YoonetalICLR-18,
        title = {Lifelong Learning with Dynamically Expandable Networks},
       author = {Jaehong Yoon and Eunho Yang and Jeongtae Lee and Sung Ju Hwang},
    booktitle = {International Conference on Learning Representations},
         year = {2018},
      comment = {We propose a novel deep network architecture for lifelong learning which we refer to as Dynamically Expandable Network (DEN), that can dynamically decide its network capacity as it trains on a sequence of tasks, to learn a compact overlapping knowledge sharing structure among tasks. DEN is efficiently trained in an online manner by performing selective retraining, dynamically expands network capacity upon arrival of each task with only the necessary number of units, and effectively prevents semantic drift by splitting/duplicating units and timestamping them. We validate DEN on multiple public datasets in lifelong learning scenarios on multiple public datasets, on which it not only significantly outperforms existing lifelong learning methods for deep networks, but also achieves the same level of performance as the batch model with substantially fewer number of parameters.}
}
@misc{Yost98,
       author = {Kirk A. Yost and Alan Washburn},
         year = 1998,
        title = {Briefing at {DARPA/Rome Laboratory} meeting},
      address = {Monterey, California},
}
@article{YouetalACS-NANO-14,
       author = {You, Byoung Kuk and Park, Woon Ik and Kim, Jong Min and Park, Kwi-Il and Seo, Hyeon Kook and Lee, Jeong Yong and Jung, Yeon Sik and Lee, Keon Jae},
        title = {Reliable Control of Filament Formation in Resistive Memories by Self-Assembled Nanoinsulators Derived from a Block Copolymer},
      journal = {ACS Nano},
       volume = 8,
        issue = 9,
         year = 2014,
        pages = {9492-9502},
     abstract = {Resistive random access memory (ReRAM) is a promising candidate for future nonvolatile memories. Resistive switching in a metal-insulator-metal structure is generally assumed to be caused by the formation/rupture of nanoscale conductive filaments (CFs) under an applied electric field. The critical issue of ReRAM for practical memory applications, however, is insufficient repeatability of the operating voltage and resistance ratio. Here, we present an innovative approach to reliably and reproducibly control the CF growth in unipolar NiO resistive memory by exploiting uniform formation of insulating SiOx nanostructures from the self-assembly of a Si-containing block copolymer. In this way, the standard deviation (SD) of set and reset voltages was markedly reduced by 76.9\% and 59.4\%, respectively. The SD of high resistance state also decreased significantly, from 6.3 \% 107 \ohm to 5.4 \% 104 \ohm. Moreover, we report direct observations of localized metallic Ni CF formation and their controllable growth using electron microscopy and discuss electrothermal simulation results based on the finite element method supporting our analysis results.}
}
@phdthesis{YoungPhD,
       author = {Joel Young},
        title = {Exploiting the Structure of the {W}eb for Spidering. Forthcoming},
       school = {Brown University},
      address = {Providence, RI},
         year = 2003,
}
@inproceedings{YoungandDeanUAI-03,
        title = {Exploiting Locality in Searching the Web},
       author = {Joel Young and Thomas Dean},
    booktitle = {Proceedings of the 19th Annual Conference on Uncertainty in Artificial Intelligence ({UAI}-03)},
 organization = {AUAI},
    publisher = {Morgan Kaufmann Publishers},
      address = {San Francisco, California},
     location = {Acapulco, Mexico},
         year = 2003,
        pages = {608-615},
}
@article{YoungetalCoRR-18,
        title = {Integrating Episodic Memory into a Reinforcement Learning Agent Using Reservoir Sampling},
       author = {Kenny J. Young and Richard S. Sutton and Shuo Yang},
      journal = {CoRR},
       volume = {arXiv:1806.00540},
         year = {2018},
     abstract = {Episodic memory is a psychology term which refers to the ability to recall specific events from the past. We suggest one advantage of this particular type of memory is the ability to easily assign credit to a specific state when remembered information is found to be useful. Inspired by this idea, and the increasing popularity of external memory mechanisms to handle long-term dependencies in deep learning systems, we propose a novel algorithm which uses a reservoir sampling procedure to maintain an external memory consisting of a fixed number of past states. The algorithm allows a deep reinforcement learning agent to learn online to preferentially remember those states which are found to be useful to recall later on. Critically this method allows for efficient online computation of gradient estimates with respect to the write process of the external memory. Thus unlike most prior mechanisms for external memory it is feasible to use in an online reinforcement learning setting.}
}
@article{YoungetalICLR-18,
        title = {Integrating Episodic Memory into a Reinforcement Learning Agent Using Reservoir Sampling},
       author = {Kenny J. Young and Shuo Yang and Richard S. Sutton},
      journal = {Submitted to International Conference on Learning Representations},
         year = {2018},
          url = {https://openreview.net/forum?id=ByJDAIe0b},
     abstract = {Episodic memory is a psychology term which refers to the ability to recall specific events from the past. We suggest one advantage of this particular type of memory is the ability to easily assign credit to a specific state when remembered information is found to be useful. Inspired by this idea, and the increasing popularity of external memory mechanisms to handle long-term dependencies in deep learning systems, we propose a novel algorithm which uses a reservoir sampling procedure to maintain an external memory consisting of a fixed number of past states. The algorithm allows a deep reinforcement learning agent to learn online to preferentially remember those states which are found to be useful to recall later on. Critically this method allows for efficient online computation of gradient estimates with respect to the write process of the external memory. Thus unlike most prior mechanisms for external memory it is feasible to use in an online reinforcement learning setting.}
}
@article{YoungetalIEEE-13,
       author = {Young, S. and Gasic, M. and Thomson, B. and Williams, J.D.},
      journal = {Proceedings of the {IEEE}},
        title = {{POMDP}-Based Statistical Spoken Dialog Systems: A Review},
         year = 2013,
       volume = 101,
        issue = 5,
        pages = {1160-1179},
     abstract = {Statistical dialog systems (SDSs) are motivated by the need for a data-driven framework that reduces the cost of laboriously handcrafting complex dialog managers and that provides robustness against the errors created by speech recognizers operating in noisy environments. By including an explicit Bayesian model of uncertainty and by optimizing the policy via a reward-driven process, partially observable Markov decision processes (POMDPs) provide such a framework. However, exact model representation and optimization is computationally intractable. Hence, the practical application of POMDP-based systems requires efficient algorithms and carefully constructed approximations. This review article provides an overview of the current state of the art in the development of POMDP-based spoken dialog systems.},
}
@inproceedings{YuanandDruzdzelAAAI-07,
       author = {Changhe Yuan and Marek Druzdzel},
        title = {Generalized Evidence Pre-propagated Importance Sampling for Hybrid Bayesian Networks},
    booktitle = {Proceedings of the Twenty-Second AAAI Conference on Artificial Intelligence},
        pages = {1290-1295},
         year = 2007,
}
@article{YuanandDruzdzelMCM-05,
       author = {Changhe Yuan and Marek Druzdzel},
        title = {Importance sampling algorithms for Bayesian networks: Principles and performance},
      journal = {Mathematical and Computer Modeling},
       volume = 43,
        issue = {9-10},
        pages = {1189-1207},
         year = 2005,
}
@incollection{YuanandMcClellandCSS-19,
        title = {A model of structure learning, inference, and generation for scene understanding},
       author = {Yuan, Arianna and McClelland, Jay},
    booktitle = {Proceedings of the 41st Annual Meeting of the Cognitive Science Society},
    publisher = {Cognitive Science Society},
         year = {2019},
     abstract = {Previous studies including Ramani and Siegler (2008) have shown that playing a number board game improved students performance on several numerical tasks, including numeral identification, magnitude comparison, counting and number line estimation. However, the computational mechanism underlying such number sense acquisition remains unclear. Here, we aim to fill this gap by building a model that simulates play of the game as well as the basic numerical tasks. We hypothesize that cognitive components that are used in the basic tasks are recruited to work together when children play the game, so that the learning induced by playing the game also manifests itself in those tasks. We reproduced the empirical findings with a neural network model implementing our hypothesis. This computational approach demonstrates how a single model that coordinates components of number processing in different modalities (visual, language and spatially-guided action) can explain the number sense acquisition in number board game playing.}
}
@incollection{YuandDayanNIPS-05,
       author = {Yu, Angela J. and Dayan, Peter},
        title = {Inference, Attention, and Decision in a {B}ayesian Neural Architecture},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = 17,
       editor = {Lawrence K. Saul and Yair Weiss and {L\'{e}on} Bottou},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = 2005,
}
@article{YuandFersterJoN-13,
       author = {Yu, J. and Ferster, D.},
        title = {{F}unctional coupling from simple to complex cells in the visually driven cortical circuit},
      journal = {Journal of Neuroscience},
         year = 2013,
       volume = 33,
        issue = 48,
        pages = {18855-18866},
     abstract = {In the classic model of the primary visual cortex, upper-layer complex cells are driven by feedforward inputs from layer 4 simple cells. Based on spike cross-correlation, previousin vivo work has suggested that this connection is strong and dense, with a high probability of connection (50\%) and significant strength in connected pairs. A much sparser projection has been found in brain slices, however, with the probability of layer 4 cells connecting to layer 2/3 cells being relatively low (10\%). Here, we explore this connection in vivo in the cat primary visual cortex by recording simultaneously spikes of layer 4 simple cells and the membrane potential (Vm) of layer 2/3 complex cells. By triggering the average of the complex cell's Vm on the spikes of the simple cell (Vm-STA), we found functional coupling to be very common during visual stimulation: the simple cell's spikes tended to occur near the troughs of the complex cell's Vm fluctuations and were, on average, followed by a significant (1 mV) fast-rising (10 ms) depolarization in the complex cell. In the absence of visual stimulation, however, when single simple cells were activated electrically through the recording electrode, no significant depolarization, or at most a very weak input (0.1-0.2 mV), was detected in the complex cell. We suggest the functional coupling observed during visual stimulation arises from coordinated or nearly synchronous activity among a large population of simple cells, only a small fraction of which are presynaptic to the recorded complex cell.},
}
@inproceedings{YuandSiskindACL-13,
        title = {Grounded Language Learning from Video Described with Sentences},
       author = {Haonan Yu and Jeffrey Mark Siskind},
    booktitle = {ACL},
         year = {2013},
     abstract = {We present a method that learns representations for word meanings from short video clips paired with sentences. Unlike prior work on learning language from symbolic input, our input consists of video of people interacting with multiple complex objects in outdoor environments. Unlike prior computer-vision approaches that learn from videos with verb labels or im- ages with noun labels, our labels are sentences containing nouns, verbs, preposi- tions, adjectives, and adverbs. The correspondence between words and concepts in the video is learned in an unsupervised fashion, even when the video depicts simultaneous events described by multiple sentences or when different aspects of a single event are described with multiple sentences. The learned word meanings can be subsequently used to automatically generate description of new video.}
}
@article{YuetalAM-15,
       author = {Hyeon Gyun Yoo and Myunghwan Byun and Chang Kyu Jeong and Keon Jae Lee},
        title = {Performance Enhancement of Electronic and Energy Devices via Block Copolymer Self-Assembly},
      journal = {Advanced Materials},
         year = 2015,
     abstract = {The use of self-assembled block copolymers (BCPs) for the fabrication of electronic and energy devices has received a tremendous amount of attention as a non-traditional approach to patterning integrated circuit elements at nanometer dimensions and densities inaccessible to traditional lithography techniques. The exquisite control over the dimensional features of the self-assembled nanostructures (i.e., shape, size, and periodicity) is one of the most attractive properties of BCP self-assembly. Harmonic spatial arrangement of the self-assembled nanoelements at desired positions on the chip may offer a new strategy for the fabrication of electronic and energy devices. Several recent reports show the great promise in using BCP self-assembly for practical applications of electronic and energy devices, leading to substantial enhancements of the device performance. Recent progress is summarized here, with regard to the performance enhancements of non-volatile memory, electrical sensor, and energy devices enabled by directed BCP self-assembly.}
}
@article{YuetalBRAIN-14,
       author = {Yu, Feng and Jiang, Qing-jun and Sun, Xi-yan and Zhang, Rong-wei},
        title = {A new case of complete primary cerebellar agenesis: clinical and imaging findings in a living patient},
      journal = {Brain},
       volume = {138},
       number = {6},
        pages = {e353-e353},
         year = {2014},
     abstract = {We read with great interest the articles previously published in Brain by Glickstein (1994) and Boyd (2010) describing the clinical evidence and review of cerebellar agenesis. Here we report a new case of a living patient whom we believe to be another patient with cerebellar agenesis. Cerebellar agenesis is an extremely rare condition implying complete absence of the cerebellum. The pathogenesis and molecular basis of this disease remain unknown. There are very few reported cases of complete cerebellar agenesis, making it challenging and controversial to understand the degree of cerebellum development necessary to avoid deficits in motor and non-motor functions. Further, a detailed description of neurological findings in a living adult with cerebellar agenesis is almost non-existent; most cases are reported based on autopsy reports. The patient presents with mild mental impairment and medium motor deficits. CT and MRI scans revealed no remnants of any cerebellar tissues, verifying complete absence of the cerebellum. A comprehensive literature review reveals novel implications for differential diagnosis. We conclude that primary cerebellar agenesis is actually the same diagnosis as Chiari malformation IV.},
}
@article{YuetalBRAIN-15,
       author = {Yu, Feng and Jiang, Qing-jun and Sun, Xi-yan and Zhang, Rong-wei},
        title = {A new case of complete primary cerebellar agenesis: clinical and imaging findings in a living patient},
      journal = {Brain},
       volume = {138},
       number = {6},
        pages = {e353},
         year = {2015},
     abstract = {Previously published in Brain by Glickstein (1994) and Boyd (2010) describes the clinical evidence and review of cerebellar agenesis. Here we report a new case of a living patient whom we believe to be another patient with cerebellar agenesis. Cerebellar agenesis is an extremely rare condition implying complete absence of the cerebellum. The pathogenesis and molecular basis of this disease remain unknown. There are very few reported cases of complete cerebellar agenesis, making it challenging and controversial to understand the degree of cerebellum development neces- sary to avoid deficits in motor and non-motor functions. Further, a detailed description of neurological findings in a living adult with cerebellar agenesis is almost non-existent; most cases are reported based on autopsy reports. The patient presents with mild mental impairment and medium motor deficits. CT and MRI scans revealed no remnants of any cerebellar tissues, verifying complete absence of the cerebellum. A comprehensive literature review reveals novel implications for differential diagnosis. We conclude that primary cerebellar agenesis is actually the same diagnosis as Chiari malformation IV.}
}
@article{YuetalCoRR-17,
       author = {Haonan Yu and Haichao Zhang and Wei Xu},
        title = {A Deep Compositional Framework for Human-like Language Acquisition in Virtual Environment},
      journal = {CoRR},
       volume = {arXiv:1703.09831},
         year = {2017},
     abstract = {We tackle a task where an agent learns to navigate in a 2D maze-like environment called XWORLD. In each session, the agent perceives a sequence of raw-pixel frames, a natural language command issued by a teacher, and a set of rewards. The agent learns the teacher's language from scratch in a grounded and compositional manner, such that after training it is able to correctly execute zero-shot commands: 1) the combination of words in the command never appeared before, and/or 2) the command contains new object concepts that are learned from another task but never learned from navigation. Our deep framework for the agent is trained end to end: it learns simultaneously the visual representations of the environment, the syntax and semantics of the language, and the action module that outputs actions. The zero-shot learning capability of our framework results from its compositionality and modularity with parameter tying. We visualize the intermediate outputs of the framework, demonstrating that the agent truly understands how to solve the problem. We believe that our results provide some preliminary insights on how to train an agent with similar abilities in a 3D environment.}
}
@article{YuetalJAIR-15,
        title = {A Compositional Framework for Grounding Language Inference, Generation, and Acquisition in Video},
       author = {Haonan Yu and N. Siddharth and Andrei Barbu and Jeffrey Mark Siskind},
      journal = {Journal of Artificial Intelligence Research},
         year = {2015},
       volume = {52},
        pages = {601-713},
     abstract = {We present an approach to simultaneously reasoning about a video clip and an entire natural-language sentence. The compositional nature of language is exploited to construct models which represent the meanings of entire sentences composed out of the meanings of the words in those sentences mediated by a grammar that encodes the predicate-argument relations. We demonstrate that these models faithfully represent the meanings of sentences and are sensitive to how the roles played by participants (nouns), their characteristics (adjectives), the actions performed (verbs), the manner of such actions (adverbs), and changing spatial relations between participants (prepositions) affect the meaning of a sentence and how it is grounded in video. We exploit this methodology in three ways. In the first, a video clip along with a sentence are taken as input and the participants in the event described by the sentence are highlighted, even when the clip depicts multiple similar simultaneous events. In the second, a video clip is taken as input without a sentence and a sentence is generated that describes an event in that clip. In the third, a corpus of video clips is paired with sentences which describe some of the events in those clips and the meanings of the words in those sentences are learned. We learn these meanings without needing to specify which attribute of the video clips each word in a given sentence refers to. The learned
meaning representations are shown to be intelligible to humans.}
}
@article{YujiaetalCoRR-15,
       author = {Yujia Li and Daniel Tarlow and Marc Brockschmidt and Richard S. Zemel},
        title = {Gated Graph Sequence Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1511.05493},
         year = {2015},
      abstract = {Graph-structured data appears frequently in domains including chemistry, natural language semantics, social networks, and knowledge bases. In this work, we study feature learning techniques for graph-structured inputs. Our starting point is previous work on Graph Neural Networks (Scarselli et al., 2009), which we modify to use gated recurrent units and modern optimization techniques and then extend to output sequences. The result is a flexible and broadly useful class of neural network models that has favorable inductive biases relative to purely sequence-based models (e.g., LSTMs) when the problem is graph-structured. We demonstrate the capabilities on some simple AI (bAbI) and graph algorithm learning tasks. We then show it achieves state-of-the-art performance on a problem from program verification, in which subgraphs need to be matched to abstract data structures.}
}
@article {YukselCGF-15,
       author = {Yuksel, Cem},
        title = {Sample Elimination for Generating {P}oisson {D}isk {S}ample Sets},
      journal = {Computer Graphics Forum},
       volume = 34,
        issue = 2,
         year = 2015,
        pages = {25-32},
     abstract = {In this paper we describe sample elimination for generating Poisson disk sample sets with a desired size. We introduce a greedy sample elimination algorithm that assigns a weight to each sample in a given set and eliminates the ones with greater weights in order to pick a subset of a desired size with Poisson disk property without having to specify a Poisson disk radius. This new algorithm is simple, computationally efficient, and it can work in any sampling domain, producing sample sets with more pronounced blue noise characteristics than dart throwing. Most importantly, it allows unbiased progressive (adaptive) sampling and it scales better to high dimensions than previous methods. However, it cannot guarantee maximal coverage. We provide a statistical analysis of our algorithm in 2D and higher dimensions as well as results from our tests with different example applications.},
}
@article{YusteNATURE_REVIEWS_NEUROSCIENCE-15,
       author = {Yuste, Raphael},
        title = {From the neuron doctrine to neural networks},
      journal = {Nature Reviews Neuroscience},
       volume = {16},
       number = {8},
         year = {2015},
        pages = {487-497},
     abstract = {For over a century, the neuron doctrine--which states that the neuron is the structural and functional unit of the nervous system--has provided a conceptual foundation for neuroscience. This viewpoint reflects its origins in a time when the use of single-neuron anatomical and physiological techniques was prominent. However, newer multineuronal recording methods have revealed that ensembles of neurons, rather than individual cells, can form physiological units and generate emergent functional properties and states. As a new paradigm for neuroscience, neural network models have the potential to incorporate knowledge acquired with single-neuron approaches to help us understand how emergent functional states generate behaviour, cognition and mental disease.}
}
@article{YusteNEURON-05,
     title = {Origin and classification of neocortical interneurons},
    author = {Yuste, Rafael},
   journal = {Neuron},
    volume = 48,
     issue = 4,
      year = 2005,
     pages = {524-527},
}
@article{YusteNEURON-11,
       author = {Yuste, R.},
        title = {Dendritic spines and distributed circuits},
      journal = {Neuron},
         year = {2011},
       volume = {71},
       number = {5},
        pages = {772-781},
     abstract = {Dendritic spines receive most excitatory connections in pyramidal cells and many other principal neurons. But why do neurons use spines, when they could accommodate excitatory contacts directly on their dendritic shafts? One suggestion is that spines serve to connect with passing axons, thus increasing the connectivity of the dendrites. Another hypothesis is that spines are biochemical compartments that enable input-specific synaptic plasticity. A third possibility is that spines have an electrical role, filtering synaptic potentials and electrically isolating inputs from each other. In this review, I argue that, when viewed from the perspective of the circuit function, these three functions dovetail with one another to achieve a single overarching goal: to implement a distributed circuit with widespread connectivity. Spines would endow these circuits with nonsaturating, linear integration and input-specific learning rules, which would enable them to function as neural networks, with emergent encoding and processing of information.}
}
@article{YusteetalNATURE_REVIEWS_NEUROSCIENCE-05,
       author = {Yuste, R. and MacLean, J. N. and Smith, J. and Lansner, A.},
        title = {The cortex as a central pattern generator},
      journal = {Nature Reviews Neuroscience},
         year = {2005},
       volume = {6},
       number = {6},
        pages = {477-483},
     abstract = {Vertebrate spinal cord and brainstem central pattern generator (CPG) circuits share profound similarities with neocortical circuits. CPGs can produce meaningful functional output in the absence of sensory inputs. Neocortical circuits could be considered analogous to CPGs as they have rich spontaneous dynamics that, similar to CPGs, are powerfully modulated or engaged by sensory inputs, but can also generate output in their absence. We find compelling evidence for this argument at the anatomical, biophysical, developmental, dynamic and pathological levels of analysis. Although it is possible that cortical circuits are particularly plastic types of CPG ('learning CPGs'), we argue that present knowledge about CPGs is likely to foretell the basic principles of the organization and dynamic function of cortical circuits.}
}
@book{Yuval_Noah_HarariDEUS-17,
        title = {Homo Deus: A Brief History of Tomorrow},
       author = {Yuval Noah Harari},
         year = {2017},
    publisher = {HarperCollins Publishers}
}
@book{Yuval_Noah_HarariSAPIENS-15,
        title = {Sapiens: A Brief History of Humankind},
       author = {Yuval Noah Harari},
         year = {2015},
    publisher = {HarperCollins Publishers}
}
@book{Zadeh,
       author = {Zadeh, Lotfi A.},
        title = {Fuzzy Sets and Applications},
    publisher = {Wiley},
      address = {New York},
         year = 1987,
}
@article{ZadehIC-65,
       author = {Zadeh, Lotfi A.},
        title = {Fuzzy Sets},
      journal = {Information and Control},
       volume = 8,
        pages = {338-353},
         year = 1965,
}
@article{ZadehIC-68,
       author = {Zadeh, Lotfi A.},
        title = {Fuzzy Algorithms},
      journal = {Information and Control},
       volume = 12,
        pages = {99-102},
         year = 1968,
} 
@article{ZadehSMC-73,
       author = {Zadeh, Lotfi A.},
        title = {Outline of a New Approach to the Analysis of Complex Systems and Decision Processes},
      journal = {{IEEE} Transactions on Systems, Man and Cybernetics},
       volume = 3,
        pages = {28-44},
         year = 1973,
}
@article{ZadorBIORXIV-19,
       author = {Zador, Anthony},
        title = {A Critique of Pure Learning: What Artificial Neural Networks can Learn from Animal Brains},
      journal = {bioRxiv},
         year = {2019},
    publisher = {Cold Spring Harbor Laboratory},
     abstract = {Over the last decade, artificial neural networks (ANNs), have undergone a revolution, catalyzed in large part by better tools for supervised learning. However, training such networks requires enormous data sets of labeled examples, whereas young animals (including humans) typically learn with few or no labeled examples. This stark contrast with biological learning has led many in the ANN community posit that instead of supervised paradigms, animals must rely instead primarily on unsupervised learning, leading the search for better unsupervised algorithms. Here we argue that much of an animal's behavioral repertoire is not the result of clever learning algorithms--supervised or unsupervised--but arises instead from behavior programs already present at birth. These programs arise through evolution, are encoded in the genome, and emerge as a consequence of wiring up the brain. Specifically, animals are born with highly structured brain connectivity, which enables them learn very rapidly. Recognizing the importance of the highly structured connectivity suggests a path toward building ANNs capable of rapid learning.},
}
@article{ZadorandDobrunzNEURON-97,
       author = {Anthony M Zador and Lynn E Dobrunz},
        title = {Dynamic Synapses in the Cortex},
      journal = {Neuron},
       volume = {19},
       number = {1},
         year = {1997},
        pages = {1-4},
}
@article{ZadoretalPLoS-12,
        title = {Sequencing the Connectome},
       author = {Zador, Anthony M. and Dubnau, Joshua and Oyibo, Hassana K. and Zhan, Huiqing and Cao, Gang and Peikon, Ian D.},
      journal = {{PLoS} Biology},
    publisher = {Public Library of Science},
       volume = 10,
        issue = 10,
         year = 2012,
        pages = {e1001411},
     abstract = {Recasting the study of neural circuitry as a problem of high-throughput DNA sequencing instead of microscopy holds the potential to increase efficiency by orders of magnitude.},
}
@incollection{ZaghloulandBoahenVPOD-07,
        title = {Visual Prosthesis and Ophthalmic Devices: New Hope in Sight},
       author = {Kareem A. Zaghloul and Kwabena Boahen},
    booktitle = {Circuit Designs That Model the Properties of the Outer and Inner Retina},
       editor = {Joyce Tombran-Tink and Colin J. Barnstable and, Joseph F. Rizzo III},
    publisher = {Springer-Verlag},
      address = {Berlin, Heidelberg},
         year = 2007,
        pages = {135-158},
}
@article{ZambaldietalCoRR-18,
       author = {Vin\'{i}cius Flores Zambaldi and David Raposo and Adam Santoro and Victor Bapst and Yujia Li and Igor Babuschkin and Karl Tuyls and David P. Reichert and Timothy P. Lillicrap and Edward Lockhart and Murray Shanahan and Victoria Langston and Razvan Pascanu and Matthew Botvinick and Oriol Vinyals and Peter Battaglia},
        title = {Relational Deep Reinforcement Learning},
      journal = {CoRR},
       volume = {arXiv:1806.01830},
         year = {2018},
     abstract = {We introduce an approach for deep reinforcement learning (RL) that improves upon the efficiency, generalization capacity, and interpretability of conventional approaches through structured perception and relational reasoning. It uses self-attention to iteratively reason about the relations between entities in a scene and to guide a model-free policy. Our results show that in a novel navigation and planning task called Box-World, our agent finds interpretable solutions that improve upon baselines in terms of sample complexity, ability to generalize to more complex scenes than experienced during training, and overall performance. In the StarCraft II Learning Environment, our agent achieves state-of-the-art performance on six mini-games -- surpassing human grandmaster performance on four. By considering architectural inductive biases, our work opens new directions for overcoming important, but stubborn, challenges in deep RL.}
}
@article{ZamftetalPLoS-12,
        title = {Measuring Cation Dependent {DNA} Polymerase Fidelity Landscapes by Deep Sequencing},
       author = {Zamft, Bradley Michael and Marblestone, Adam H. and Kording, Konrad and Schmidt, Daniel and Martin-Alarcon, Daniel and Tyo, Keith and Boyden, Edward S. and Church, George},
      journal = {{PLoS} ONE},
    publisher = {Public Library of Science},
         year = 2012,
       volume = 7,
        issue = 8,
        pages = {e43876},
     abstract = {High-throughput recording of signals embedded within inaccessible micro-environments is a technological challenge. The ideal recording device would be a nanoscale machine capable of quantitatively transducing a wide range of variables into a molecular recording medium suitable for long-term storage and facile readout in the form of digital data. We have recently proposed such a device, in which cation concentrations modulate the misincorporation rate of a DNA polymerase (DNAP) on a known template, allowing DNA sequences to encode information about the local cation concentration. In this work we quantify the cation sensitivity of DNAP misincorporation rates, making possible the indirect readout of cation concentration by DNA sequencing. Using multiplexed deep sequencing, we quantify the misincorporation properties of two DNA polymerases - Dpo4 and Klenow exo<sup>−</sup> - obtaining the probability and base selectivity of misincorporation at all positions within the template. We find that Dpo4 acts as a DNA recording device for Mn$^{2+}$ with a misincorporation rate gain of \textasciitilde{}2\%/mM. This modulation of misincorporation rate is selective to the template base: the probability of misincorporation on template T by Dpo4 increases more than 50-fold over the range tested, while the other template bases are affected less strongly. Furthermore, cation concentrations act as scaling factors for misincorporation: on a given template base, Mn$^{2+}$ and Mg$^{2+}$ change the overall misincorporation rate but do not alter the relative frequencies of incoming misincorporated nucleotides. Characterization of the ion dependence of DNAP misincorporation serves as the first step towards repurposing it as a molecular recording device.}
}
@article{ZamiretalCoRR-16,
       author = {Amir Roshan Zamir and Te{-}Lin Wu and Lin Sun and William B. Shen and Jitendra Malik and Silvio Savarese},
        title = {Feedback Networks},
      journal = {CoRR},
       volume = {arXiv:1612.09508},
         year = {2016},
     abstract = {Currently, the most successful learning models in computer vision are based on learning successive representations followed by a decision layer. This is usually actualized through feedforward multilayer neural networks, e.g. ConvNets, where each layer forms one of such successive representations. However, an alternative that can achieve the same goal is a feedback based approach in which the representation is formed in an iterative manner based on a feedback received from previous iteration's output.
We establish that a feedback based approach has several fundamental advantages over feedforward: it enables making early predictions at the query time, its output naturally conforms to a hierarchical structure in the label space (e.g. a taxonomy), and it provides a new basis for Curriculum Learning. We observe that feedback networks develop a considerably different representation compared to feedforward counterparts, in line with the aforementioned advantages. We put forth a general feedback based learning architecture with the endpoint results on par or better than existing feedforward networks with the addition of the above advantages. We also investigate several mechanisms in feedback architectures (e.g. skip connections in time) and design choices (e.g. feedback length). We hope this study offers new perspectives in quest for more natural and practical learning models.}
}
@article{ZarembaandSutskeverCoRR-15,
       author = {Wojciech Zaremba and Ilya Sutskever},
        title = {Reinforcement Learning Neural Turing Machines},
      journal = {CoRR},
       volume = {arXiv:1505.00521},
         year = {2015},
     abstract = {The Neural Turing Machine (NTM) is more expressive than all previously considered models because of its external memory. It can be viewed as a broader effort to use abstract external Interfaces and to learn a parametric model that interacts with them. The capabilities of a model can be extended by providing it with proper Interfaces that interact with the world. These external Interfaces include memory, a database, a search engine, or a piece of software such as a theorem verifier. Some of these Interfaces are provided by the developers of the model. However, many important existing Interfaces, such as databases and search engines, are discrete. We examine feasibility of learning models to interact with discrete Interfaces. We investigate the following discrete Interfaces: a memory Tape, an input Tape, and an output Tape. We use a Reinforcement Learning algorithm to train a neural network that interacts with such Interfaces to solve simple algorithmic tasks. Our Interfaces are expressive enough to make our model Turing complete.}
}
@article{ZavershynskyietalCoRR-18,
       author = {{NAPS:} Natural Program Synthesis Dataset},
        title = {Maksym Zavershynskyi and Alex Skidanov and Illia Polosukhin},
      journal = {CoRR},
       volume = {arXiv:1807.03168},
         year = {2018},
     abstract = {We present a program synthesis-oriented dataset consisting of human written problem statements and solutions for these problems. The problem statements were collected via crowdsourcing and the program solutions were extracted from human-written solutions in programming competitions, accompanied by input/output examples. We propose using this dataset for the program synthesis tasks aimed for working with real user-generated data. As a baseline we present few models, with the best model achieving 8.8\% accuracy, showcasing both the complexity of the dataset and large room for future research.}
}
@article{Zeguraetal97,
       author = {E. W. Zegura and K. L. Calvert and M. J. Donahoo},
        title = {A quantitative comparison of graph-based model for internetworks},
      journal = {IEEE/ACM Transactions on Networking},
       volume = {5:6},
        pages = {770-783},
         year = 1997,
}
@article{ZeidmanandMaguireNATURE-REVIEWS-NEUROSCIENCE-16,
       author = {Zeidman, Peter and Maguire, Eleanor A.},
        title = {Anterior hippocampus: the anatomy of perception, imagination and episodic memory},
      journal = {Nature Reviews Neuroscience},
         year = {2016},
       volume = {17},
        issue = {3},
        pages = {173-182},
     abstract = {The brain creates a model of the world around us. We can use this representation to perceive and comprehend what we see at any given moment, but also to vividly re-experience scenes from our past and imagine future (or even fanciful) scenarios. Recent work has shown that these cognitive functions--perception, imagination and recall of scenes and events--all engage the anterior hippocampus. In this Opinion article, we capitalize on new findings from functional neuroimaging to propose a model that links high-level cognitive functions to specific structures within the anterior hippocampus.},
}
@article{ZeilerandFergusCoRR-13,
       author = {Matthew D. Zeiler and Rob Fergus},
        title = {Visualizing and Understanding Convolutional Networks},
      journal = {CoRR},
       volume = {arXiv:1311.290},
         year = 2013,
     abstract = {Large Convolutional Network models have recently demonstrated impressive classification performance on the ImageNet benchmark. However there is no clear understanding of why they perform so well, or how they might be improved. In this paper we address both issues. We introduce a novel visualization technique that gives insight into the function of intermediate feature layers and the operation of the classifier. We also perform an ablation study to discover the performance contribution from different model layers. This enables us to find model architectures that outperform Krizhevsky \etal on the ImageNet classification benchmark. We show our ImageNet model generalizes well to other datasets: when the softmax classifier is retrained, it convincingly beats the current state-of-the-art results on Caltech-101 and Caltech-256 datasets.}
}
@inproceedings{ZeileretalCVPR-10,
        title = {Deconvolutional Networks},
       author = {Matthew D. Zeiler and Dilip Kirshnan and Graham W. Taylor and Rob Fergus},
    booktitle = {{IEEE} International Conference on Computer Vision and Pattern Recognition},
         year = 2010,
        pages = {2528-2535},
     abstract = {Building robust low and mid-level image representations, beyond edge primitives, is a long-standing goal in vision. Many existing feature detectors spatially pool edge information which destroys cues such as edge intersections, parallelism and symmetry. We present a learning framework where features that capture these mid-level cues spontaneously emerge from image data. Our approach is based on the convolutional decomposition of images under a sparsity constraint and is totally unsupervised. By building a hierarchy of such decompositions we can learn rich feature sets that are a robust image representation for both the analysis and synthesis of images.},
}
@inproceedings{ZeileretalICCV-11,
       author = {Zeiler, M.D. and Taylor, G.W. and Fergus, R.},
    booktitle = {{IEEE} International Conference on Computer Vision},
        title = {Adaptive deconvolutional networks for mid and high level feature learning},
         year = 2011,
        pages = {2018-2025},
     abstract = {We present a hierarchical model that learns image decompositions via alternating layers of convolutional sparse coding and max pooling. When trained on natural images, the layers of our model capture image information in a variety of forms: low-level edges, mid-level edge junctions, high-level object parts and complete objects. To build our model we rely on a novel inference scheme that ensures each layer reconstructs the input, rather than just the output of the layer directly beneath, as is common with existing hierarchical approaches. This makes it possible to learn multiple layers of representation and we show models with 4 layers, trained on images from the Caltech-101 and 256 datasets. When combined with a standard classifier, features extracted from these models outperform SIFT, as well as representations from other feature learning methods.},
}
@article{ZeiseletalSCIENCE-15,
       author = {Zeisel, Amit and Munoz-Manchado, Ana B. and Codeluppi, Simone and Lönnerberg, Peter and La Manno, Gioele and Jur\'{e}us, Anna and Marques, Sueli and Munguba, Hermany and He, Liqun and Betsholtz, Christer and Rolny, Charlotte and Castelo-Branco, Goncalo and Hjerling-Leffler, Jens and Linnarsson, Sten},
        title = {Cell types in the mouse cortex and hippocampus revealed by single-cell {RNA}-seq},
      journal = {Science},
       volume = 347,
        issue = 6226,
         year = 2015,
        pages = {1138-1142},
     abstract = {The mammalian cerebral cortex supports cognitive functions such as sensorimotor integration, memory, and social behaviors. Normal brain function relies on a diverse set of differentiated cell types, including neurons, glia, and vasculature. Here, we have used large-scale single-cell RNA sequencing (RNA-seq) to classify cells in the mouse somatosensory cortex and hippocampal CA1 region. We found 47 molecularly distinct subclasses, comprising all known major cell types in the cortex. We identified numerous marker genes, which allowed alignment with known cell types, morphology, and location. We found a layer I interneuron expressing Pax6 and a distinct postmitotic oligodendrocyte subclass marked by Itpr2. Across the diversity of cortical cell types, transcription factors formed a complex, layered regulatory code, suggesting a mechanism for the maintenance of adult cell type identity.},
}
@misc{Zelikman2020seniorthesis,
        title = {Learning is its Own Reward
Exploring Worlds with Curiosity-driven Spiking Neural Networks},
       author = {Eric Zelikman},
         year = 2016,
 howpublished = {Symbolic Systems, Thesis for Bachelor of Science with Honors},
}
@misc{ZelikmanYinWangSPIKES-19,
        title = {Learning as Reinforcement: Applying Principles of Neuroscience for More General Reinforcement Learning Agents},
       author = {Eric Zelikman and William Yin and Kenneth Wang},
         year = {2019},
 howpublished = {A significant challenge in developing AI that can generalize well is designing agents that learn about their world without being told what to learn, and apply that learning to challenges with sparse rewards. Moreover, most traditional rein- forcement learning approaches explicitly separate learning and decision making in a way that does not correspond to biological learning. We im- plement an architecture founded in principles of experimental neuroscience, by combining compu- tationally efficient abstractions of biological algo- rithms. Our approach is inspired by research on spike-timing dependent plasticity, the transition between short and long term memory, and the role of various neurotransmitters in rewarding curios- ity. The Neurons-in-a-Box architecture can learn in a wholly generalizable manner, and demon- strates an efficient way to build and apply repre- sentations without explicitly optimizing over a set of criteria or actions.}
}
@incollection{ZemelTCC-00,
       author = {Richard Zemel},
        title = {Cortical Belief Networks},
       editor = {R. Hecht-Neilsen},
    publisher = {Springer-Verlag},
      address = {New York, NY},
    booktitle = {Theories of the Cerebral Cortex},
          url = {http://www.cs.toronto.edu/~zemel/papers.html},
         year = 2000,
}
@incollection{ZemeletalNIPS-05,
       author = {Zemel, Richard S. and Huys, Quentin J. M. and Natarajan, Rama and Dayan, Peter},
        title = {Probabilistic Computation in Spiking Neurons},
    booktitle = {Advances in Neural Information Processing Systems},
       volume = {17},
       editor = {Lawrence K. Saul and Yair Weiss and {L\'{e}on} Bottou},
    publisher = {MIT Press},
      address = {Cambridge, MA},
         year = {2005},
}
@article{ZengetalCOIN-18,
       author = {Hongkui Zeng},
        title = {Mesoscale connectomics},
      journal = {Current Opinion in Neurobiology},
       volume = {50},
        pages = {154-162},
         year = {2018},
     abstract = {Brain cells communicate with one another via local and long-range synaptic connections. Structural connectivity is the foundation for neural function. Brain-wide connectivity can be described at macroscopic, mesoscopic and microscopic levels. The mesoscale connectome represents connections between neuronal types across different brain regions. Building a mesoscale connectome requires a detailed understanding of the cell type composition of different brain regions and the patterns of inputs and outputs that each of these cell types receives and forms, respectively. In this review, I discuss historical and contemporary tracing techniques in both anterograde and retrograde directions to map the input and output connections at population and individual cell levels, as well as imaging and network analysis approaches to build mesoscale connectomes for mammalian brains.}
}
@article{ZenkeandGerstnerFiN-14,
       author = {Zenke, F. and Gerstner, W.},
        title = {Limits to high-speed simulations of spiking neural networks using general-purpose computers},
      journal = {Frontiers in Neuroinformatics},
       volume = {8},
        issue = {76},
         year = {2014},
     abstract = {To understand how the central nervous system performs computations using recurrent neuronal circuitry, simulations have become an indispensable tool for theoretical neuroscience. To study neuronal circuits and their ability to self-organize, increasing attention has been directed toward synaptic plasticity. In particular spike-timing-dependent plasticity (STDP) creates specific demands for simulations of spiking neural networks. On the one hand a high temporal resolution is required to capture the millisecond timescale of typical STDP windows. On the other hand network simulations have to evolve over hours up to days, to capture the timescale of long-term plasticity. To do this efficiently, fast simulation speed is the crucial ingredient rather than large neuron numbers. Using different medium-sized network models consisting of several thousands of neurons and off-the-shelf hardware, we compare the simulation speed of the simulators: Brian, NEST and Neuron as well as our own simulator Auryn. Our results show that real-time simulations of different plastic network models are possible in parallel simulations in which numerical precision is not a primary concern. Even so, the speed-up margin of parallelism is limited and boosting simulation speeds beyond one tenth of real-time is difficult. By profiling simulation code we show that the run times of typical plastic network simulations encounter a hard boundary. This limit is partly due to latencies in the inter-process communications and thus cannot be overcome by increased parallelism. Overall, these results show that to study plasticity in medium-sized spiking neural networks, adequate simulation tools are readily available which run efficiently on small clusters. However, to run simulations substantially faster than real-time, special hardware is a prerequisite.}
}
@article{ZetzscheetalJOSA-99,
       author = {Christoph Zetzsche and Gerhard Krieger and Bernhard Wegmann},
        title = {The atoms of vision: Cartesian or polar?},
      journal = {Journal of the Optical Society of America},
    publisher = {Optical Society of America},
        issue = {7},
       volume = {16},
        pages = {1554-1565},
         year = {1999},
}
@article{ZhanetalNATURE-NEUROSCIENCE-14,
       author = {Zhan, Y. and Paolicelli, R. C. and Sforazzini, F. and Weinhard, L. and Bolasco, G. and Pagani, F. and Vyssotski, A. L. and Bifone, A. and Gozzi, A. and Ragozzino, D. and Gross, C. T.},
        title = {Deficient neuron-microglia signaling results in impaired functional brain connectivity and social behavior},
      journal = {Nature Neuroscience},
       volume = {17},
       number = {3},
         year = {2014},
        pages = {400-406},
     abstract = {Microglia are phagocytic cells that infiltrate the brain during development and have a role in the elimination of synapses during brain maturation. Changes in microglial morphology and gene expression have been associated with neurodevelopmental disorders. However, it remains unknown whether these changes are a primary cause or a secondary consequence of neuronal deficits. Here we tested whether a primary deficit in microglia was sufficient to induce some autism-related behavioral and functional connectivity deficits. Mice lacking the chemokine receptor Cx3cr1 exhibit a transient reduction of microglia during the early postnatal period and a consequent deficit in synaptic pruning. We show that deficient synaptic pruning is associated with weak synaptic transmission, decreased functional brain connectivity, deficits in social interaction and increased repetitive-behavior phenotypes that have been previously associated with autism and other neurodevelopmental and neuropsychiatric disorders. These findings open the possibility that disruptions in microglia-mediated synaptic pruning could contribute to neurodevelopmental and neuropsychiatric disorders.}
}
@inproceedings{ZhangandGoldmanNIPS-01,
       author = {Qi Zhang and Sally A. Goldman},
        title = {{EM-DD}: An Improved Multiple-Instance Learning Technique},
    booktitle = {In Advances in Neural Information Processing Systems},
    publisher = {MIT Press},
        pages = {1073-1080},
         year = {2001},
}
@article{ZhangandLeCunCoRR-15,
       author = {Zhang, Xiang and LeCun, Yann},
        title = {Text Understanding from Scratch},
      journal = {CoRR},
       volume = {arXiv:1502.01710},
         year = 2015,
          url = {http://arxiv.org/abs/1502.01710},
     abstract = {This article demonstrates that we can apply deep learning to text understanding from character-level inputs all the way up to abstract text concepts, using temporal convolutional networks (ConvNets). We apply ConvNets to various large-scale datasets, including ontology classification, sentiment analysis and text categorization. We show that temporal ConvNets can achieve astonishing performance without the knowledge of words, phrases, sentences and any other syntactic or semantic structures with regards to a human language. Evidence shows that our models can work for both English and Chinese.}
}
@article{ZhangandWangTHERANOSTICS-12,
     title = {Quantum Dot Enabled Molecular Sensing and Diagnostics},
    author = {Zhang, Y. and Wang, T.},
   journal = {Theranostics},
    volume = 2,
     issue = 7,
     pages = {631-654},
      year = 2012,
  abstract = {Since its emergence, semiconductor nanoparticles known as quantum dots (QDs) have drawn considerable attention and have quickly extended their applicability to numerous fields within the life sciences. This is largely due to their unique optical properties such as high brightness and narrow emission band as well as other advantages over traditional organic fluorophores. New molecular sensing strategies based on QDs have been developed in pursuit of high sensitivity, high throughput, and multiplexing capabilities. For traditional biological applications, QDs have already begun to replace traditional organic fluorophores to serve as simple fluorescent reporters in immunoassays, microarrays, fluorescent imaging applications, and other assay platforms. In addition, smarter, more advanced QD probes such as quantum dot fluorescence resonance energy transfer (QD-FRET) sensors, quenching sensors, and barcoding systems are paving the way for highly-sensitive genetic and epigenetic detection of diseases, multiplexed identification of infectious pathogens, and tracking of intracellular drug and gene delivery. When combined with microfluidics and confocal fluorescence spectroscopy, the detection limit is further enhanced to single molecule level. Recently, investigations have revealed that QDs participate in series of new phenomena and exhibit interesting non-photoluminescent properties. Some of these new findings are now being incorporated into novel assays for gene copy number variation (CNV) studies and DNA methylation analysis with improved quantification resolution. Herein, we provide a comprehensive review on the latest developments of QD based molecular diagnostic platforms in which QD plays a versatile and essential role.}
}
@inproceedings{ZhangetalCASA-16,
       author = {Zhang, Juzheng and Thalmann, Nadia Magnenat and Zheng, Jianmin},
        title = {Combining Memory and Emotion With Dialog on Social Companion: A Review},
    booktitle = {Proceedings of the 29th International Conference on Computer Animation and Social Agents},
    publisher = {ACM},
     location = {Geneva, Switzerland},
         year = {2016},
        pages = {1-9},
     abstract = {In the coming era of social companions, many researchers have been pursuing natural dialog interactions and long-term relations between social companions and users. With respect to the quick decrease of user interest after the first few interactions, various emotion and memory models have been developed and integrated with social companions for better user engagement. This paper reviews related works in the effort of combining memory and emotion with natural language dialog on social companions. We separate these works into three categories: (1) Affective system with dialog, (2) Task-driven memory with dialog, (3) Chat-driven memory with dialog. In addition, we discussed limitations and challenging issues to be solved. Finally, we also introduced our framework of social companions.}
}
@article{ZhangetalCB-14,
       author = {Xilin Zhang and Jiang Qiu and Yanyu Zhang and Shihui Han and Fang Fang},
        title = {Misbinding of Color and Motion in Human Visual Cortex},
      journal = {Current Biology},
       volume = 24,
        issue = 12,
        pages = {1354-1360},
         year = 2014,
     abstract = {Summary A fundamental challenge for the visual system is to integrate visual features into a coherent scene, known as the binding problem. The neural mechanisms of feature binding are hard to identify because of difficulties in separating active feature binding from feature co-occurrence. In previous studies on feature binding [1-5], visual features were superimposed and presented simultaneously. Neurons throughout the visual cortex are known to code multiple features [6]. Therefore, the observed binding effects could be due to the physical co-occurrence of features and the sensory representation of feature pairings. It is uncertain whether the mechanisms responsible for perceptual binding were actually recruited [7, 8]. To address this issue, we performed psychophysical and fMRI experiments to investigate the neural mechanisms of a steady-state misbinding of color and motion [9], because feature misbinding is probably the most striking evidence for the active existence of the binding mechanisms [10]. We found that adapting to the color-motion misbinding generated the color-contingent motion aftereffect, as well as the color-contingent motion adaptation effect in visual cortex. Notably, \{V2\} exhibited the strongest adaptation effect, which significantly correlated with the aftereffect across subjects. Furthermore, effective connectivity analysis using dynamic causal modeling showed that the misbinding was closely associated with enhanced feedback from \{V4\} and \{V5\} to V2. These findings provide strong evidence for active feature binding in early visual cortex and suggest a critical role of reentrant connections from specialized intermediate areas to early visual cortex in this process.}
}
@article{ZhangetalCOGNITIVE-NEURODYNAMICS-16,
       author = {Zhang, Yan and Pan, Xiaochuan and Wang, Rubin and Sakagami, Masamichi},
        title = {Functional connectivity between prefrontal cortex and striatum estimated by phase locking value},
      journal = {Cognitive Neurodynamics},
    publisher = {Springer Netherlands},
         year = {2016},
       volume = {10},
        issue = {3},
        pages = {245-254},
     abstract = {The interplay between the prefrontal cortex (PFC) and striatum has an important role in cognitive processes. To investigate interactive functions between the two areas in reward processing, we recorded local field potentials (LFPs) simultaneously from the two areas of two monkeys performing a reward prediction task (large reward vs small reward). The power of the LFPs was calculated in three frequency bands: the beta band (15-29Hz), the low gamma band (30-49Hz), and the high gamma band (50-100Hz). We found that both the PFC and striatum encoded the reward information in the beta band. The reward information was also found in the high gamma band in the PFC, not in the striatum. We further calculated the phase-locking value (PLV) between two LFP signals to measure the phase synchrony between the PFC and striatum. It was found that significant differences occurred between PLVs in different task periods and in different frequency bands. The PLVs in small reward condition were significant higher than that in large reward condition in the beta band. In contrast, the PLVs in the high gamma band were stronger in large reward trials than in small trials. These results suggested that the functional connectivity between the PFC and striatum depended on the task periods and reward conditions. The beta synchrony between the PFC and striatum may regulate behavioral outputs of the monkeys in the small reward condition.},
}
@article {ZhangetalJoN-16,
       author = {Zhang, Wen-hao and Chen, Aihua and Rasch, Malte J. and Wu, Si},
        title = {Decentralized Multisensory Information Integration in Neural Systems},
      journal = {Journal of Neuroscience},
    publisher = {Society for Neuroscience},
       volume = {36},
       number = {2},
         year = {2016},
        pages = {532-547},
     abstract = {How multiple sensory cues are integrated in neural circuitry remains a challenge. The common hypothesis is that information integration might be accomplished in a dedicated multisensory integration area receiving feedforward inputs from the modalities. However, recent experimental evidence suggests that it is not a single multisensory brain area, but rather many multisensory brain areas that are simultaneously involved in the integration of information. Why many mutually connected areas should be needed for information integration is puzzling. Here, we investigated theoretically how information integration could be achieved in a distributed fashion within a network of interconnected multisensory areas. Using biologically realistic neural network models, we developed a decentralized information integration system that comprises multiple interconnected integration areas. Studying an example of combining visual and vestibular cues to infer heading direction, we show that such a decentralized system is in good agreement with anatomical evidence and experimental observations. In particular, we show that this decentralized system can integrate information optimally. The decentralized system predicts that optimally integrated information should emerge locally from the dynamics of the communication between brain areas and sheds new light on the interpretation of the connectivity between multisensory brain areas. Significance Statement: To extract information reliably from ambiguous environments, the brain integrates multiple sensory cues, which provide different aspects of information about the same entity of interest. Here, we propose a decentralized architecture for multisensory integration. In such a system, no processor is in the center of the network topology and information integration is achieved in a distributed manner through reciprocally connected local processors. Through studying the inference of heading direction with visual and vestibular cues, we show that the decentralized system can integrate information optimally, with the reciprocal connections between processers determining the extent of cue integration. Our model reproduces known multisensory integration behaviors observed in experiments and sheds new light on our understanding of how information is integrated in the brain.},
}
@article{ZhangetalNATURE-10,
     title = {Optogenetic interrogation of neural circuits: technology for probing mammalian brain structures},
    author = {Zhang, F. and Gradinaru, V. and Adamantidis, A.R. and Durand, R. and Airan, R.D. and de Lecea, L. and Deisseroth, K.},
   journal = {Nature Protocols},
    volume = 5,
     issue = 3,
     pages = {439-56},
      year = 2010,
  abstract = {Elucidation of the neural substrates underlying complex animal behaviors depends on precise activity control tools, as well as compatible readout methods. Recent developments in optogenetics have addressed this need, opening up new possibilities for systems neuroscience. Interrogation of even deep neural circuits can be conducted by directly probing the necessity and sufficiency of defined circuit elements with millisecond-scale, cell type-specific optical perturbations, coupled with suitable readouts such as electrophysiology, optical circuit dynamics measures and freely moving behavior in mammals. Here we collect in detail our strategies for delivering microbial opsin genes to deep mammalian brain structures in vivo, along with protocols for integrating the resulting optical control with compatible readouts (electrophysiological, optical and behavioral). The procedures described here, from initial virus preparation to systems-level functional readout, can be completed within 4-5 weeks. Together, these methods may help in providing circuit-level insight into the dynamics underlying complex mammalian behaviors in health and disease.}
}
@article{ZhangetalNATURE-13,
    author = {Zhang, Guo and Li, Juxue and Purkayastha, Sudarshana and Tang, Yizhe and Zhang, Hai and Yin, Ye and Li, Bo and Liu, Gang and Cai, Dongsheng},
     title = {Hypothalamic programming of systemic ageing involving IKK-[bgr], NF-[kgr]B and GnRH},
   journal = {Nature},
 publisher = {Nature Publishing Group},
    volume = 497,
     issue = 7448,
      year = 2013,
     pages = {211-216},
  abstract = {Ageing is a result of gradual and overall functional deteriorations across the body; however, it is unknown whether an individual tissue primarily works to mediate the ageing progress and control lifespan. Here we show that the hypothalamus is important for the development of whole-body ageing in mice, and that the underlying basis involves hypothalamic immunity mediated by I$\kappa$B kinase-$\beta$ (IKK-$\beta$), nuclear factor $\kappa$B (NF-$\kappa$B) and related microglia-neuron immune crosstalk. Several interventional models were developed showing that ageing retardation and lifespan extension are achieved in mice by preventing ageing-related hypothalamic or brain IKK-$\beta$ and NF-$\kappa$B activation. Mechanistic studies further revealed that IKK-$\beta$ and NF-$\kappa$B inhibit gonadotropin-releasing hormone (GnRH) to mediate ageing-related hypothalamic GnRH decline, and GnRH treatment amends ageing-impaired neurogenesis and decelerates ageing. In conclusion, the hypothalamus has a programmatic role in ageing development via immune-neuroendocrine integration, and immune inhibition or GnRH restoration in the hypothalamus/brain represent two potential strategies for optimizing lifespan and combating ageing-related health problems.},
}
@article{ZhangetalNEUROIMAGE-15,
       author = {Zhang, W. and Li, R. and Deng, H. and Wang, L. and Lin, W. and Ji, S. and Shen, D.},
        title = {Deep convolutional neural networks for multi-modality isointense infant brain image segmentation},
      journal = {Neuroimage},
         year = {2015},
       volume = {108},
        pages = {214-224},
     abstract = {The segmentation of infant brain tissue images into white matter (WM), gray matter (GM), and cerebrospinal fluid (CSF) plays an important role in studying early brain development in health and disease. In the isointense stage (approximately 6-8 months of age), WM and GM exhibit similar levels of intensity in both T1 and T2 MR images, making the tissue segmentation very challenging. Only a small number of existing methods have been designed for tissue segmentation in this isointense stage; however, they only used a single T1 or T2 images, or the combination of T1 and T2 images. In this paper, we propose to use deep convolutional neural networks (CNNs) for segmenting isointense stage brain tissues using multi-modality MR images. CNNs are a type of deep models in which trainable filters and local neighborhood pooling operations are applied alternatingly on the raw input images, resulting in a hierarchy of increasingly complex features. Specifically, we used multi-modality information from T1, T2, and fractional anisotropy (FA) images as inputs and then generated the segmentation maps as outputs. The multiple intermediate layers applied convolution, pooling, normalization, and other operations to capture the highly nonlinear mappings between inputs and outputs. We compared the performance of our approach with that of the commonly used segmentation methods on a set of manually segmented isointense stage brain images. Results showed that our proposed model significantly outperformed prior methods on infant brain tissue segmentation. In addition, our results indicated that integration of multi-modality images led to significant performance improvement.}
}
@article{ZhangetalPLoS-11,
       author = {Zhang, Zi-Wei and Kang, Jun Il and Vaucher, Elvire},
      journal = {PLoS ONE},
    publisher = {Public Library of Science},
        title = {Axonal Varicosity Density as an Index of Local Neuronal Interactions},
       volume = 6,
       number = 7,
         year = 2011,
        pages = {e22543},
     abstract = {Diffuse transmission is an important non-synaptic communication mode in the cerebral neocortex, in which neurotransmitters released from en passant varicosities interact with surrounding cells. In a previous study we have shown that the cholinergic axonal segments which were in the microproximity with dopaminergic fibers possessed a greater density of en passant varicosities compared to more distant segments, suggesting an activity-dependent level of en passant varicosities in the axonal zone of interaction. To further evaluate this plastic relationship, the density of cholinergic varicosities was quantified on fiber segments within the microproximity of activated or non-activated pyramidal cells of the prefrontal cortex (mPFC). Repetitive 14 days patterned visual stimulation paired with an electrical stimulation of the cholinergic fibers projecting to the mPFC from the HDB was performed to induce persistent axonal plastic changes. The c-Fos early gene immunoreactivity was used as a neuronal activity marker of layer V pyramidal cells, labelled with anti-glutamate transporter EAAC1. Cholinergic fibers were labeled with anti-ChAT (choline acetyltransferase) immunostaining. The density of ChAT+ varicosities on and the length of fiber segments within the 3 \hmmu{}m microproximity of c-Fos positive/negative pyramidal cells were evaluated on confocal images. More than 50\% of the pyramidal cells in the mPFC were c-Fos immunoreactive. Density of ChAT+ varicosities was significantly increased within 3 \hmmu{}m vicinity of activated pyramidal cells (0.50\hmplusminus{}0.01 per \hmmu{}m of ChAT+ fiber length) compared to non-activated cells in this group (0.34\hmplusminus{}0.001; p\hmlessthanorequal{}0.05) or control rats (0.32\hmplusminus{}0.02; p\hmlessthanorequal{}0.05). Different types of stimulation (visual, HDB or visual/HDB) induced similar increase of the density of ChAT+ varicosities within microproximity of activated pyramidal cells. This study demonstrated at the subcellular level an activity-dependent enrichment of ChAT+ varicosities in the axonal zone of interaction with other neuronal elements.},
}
@article{ZhangetalSENSORS-13,
    author = {Nan Zhang and Xiaodi Su and Paul Free and Xiaodong Zhou and Koon Gee Neoh and Jinghua Teng and Wolfgang Knoll},
     title = {Plasmonic metal nanostructure array by glancing angle deposition for biosensing application},
   journal = {{Sensors and Actuators B: Chemical}},
    volume = 183,
     issue = 0,
     pages = {310-318},
      year = 2013,
  abstract = {This paper investigates the application of gold nanostructure array fabricated on a layer of closely packed polystyrene (PS) nanospheres as a plasmonic chemical or biological sensor. The sensing chip was fabricated by first functionalizing a layer of closely packed \{PS\} nanospheres as a colloidal template, and then depositing the gold at an oblique angle, where the nanospheres shaded each other and formed a gold nanostructure on the top of the \{PS\} nanospheres. The surface profiles of the gold nanostructures on \{PS\} nanospheres were investigated by Field Emission Scanning Electron Microscopy (FESEM) and Atomic Force Microscopy (AFM). A chip fabricated at the deposition angle of 30° was found to have a high consistency for keeping the shape of the nanostructures on a large fabrication area. In the study of biotin-streptavidin binding, the limit of detection (LOD) for streptavidin was [...], and the surface-confined thermodynamic binding constant is [...]. The plasmonic signal is amplified by using biotinylated Au colloids as labels. To further demonstrate the sensing capability, biotin-antibiotin IgG binding with a lower affinity constant was tested. The obtained \{LSPR\} peak shift caused by IgG (molecular weight [...]) was about 2 times larger than that by streptavidin (molecular weight [...]). The correlation of peak shift to the molecular weight further confirms the sensing capability of the chip for quantitative detection of proteins.}
}
@article{ZhangetalTBE-19,
       author = {Zhang, S.  and Liu, H.  and Huang, H.  and Zhao, Y.  and Jiang, X.  and Bowers, B.  and Guo, L.  and Hu, X.  and Sanchez, M.  and Liu, T. },
        title = {Deep Learning Models Unveiled Functional Difference Between Cortical Gyri and Sulci},
      journal = {IEEE Transactions on Biomedical Engineering},
       volume = {66},
       number = {5},
         year = {2019},
        pages = {1297-1308},
     abstract = {It is largely unknown whether there is functional role difference between cortical gyral and sulcal regions. Recent advancements in neuroimaging studies demonstrate clear difference of structural connection profiles in gyral and sulcal areas, suggesting possible functional role difference in these convex and concave cortical regions. To explore and confirm such possible functional difference, we design and apply a powerful deep learning model of convolutional neural networks (CNN) that has been proven to be superior in learning discriminative and meaningful patterns on fMRI. By using the CNN model, gyral and sulcal fMRI signals are learned and predicted, and the prediction performance is adopted to demonstrate the functional difference between gyri and sulci. By using the Human Connectome Project (HCP) fMRI data and macaque brain fMRI data, an average of 83\% and 90\% classification accuracy has been achieved to separate gyral/sulcal HCP task fMRI signals at the population and individual subject level, respectively; 81\% and 86\% classification accuracy for resting state fMRI signals at the group and individual subject level, respectively. In addition, 78\% classification accuracy has been achieved to separate gyral/sulcal resting state fMRI signals in macaque brains. Importantly, further analysis reveals that the discriminative features that are learned by CNNs to differentiate gyral/sulcal fMRI signals can be meaningfully interpreted, thus unveiling the fundamental functional difference between cortical gyri and sulci. That is, gyri are more global functional integration centers with simpler lower frequency signal components, while sulci are more local processing units with more complex higher frequency signal components.}
}
@techreport{ZhaoandCanny-TR-UCB-12,
       author = {Zhao, Huasha and Canny, John F.},
        title = {Communication-Efficient Distributed Stochastic Gradient Descent with Butterfly Mixing},
  institution = {{EECS Department, University of California, Berkeley}},
         year = 2012,
        issue = {UCB/EECS-2012-96},
          url = {http://www.eecs.berkeley.edu/Pubs/TechRpts/2012/EECS-2012-96.html},
     abstract = {Stochastic gradient descent is a widely used method to find locally-optimal models in machine learning and data mining. However, it is naturally a sequential algorithm, and parallelization involves severe compromises because the cost of synchronizing across a cluster is much larger than the time required to compute an optimal-sized gradient step. Here we explore butterfly mixing, where gradient steps are interleaved with the k stages of a butterfly network on 2^k nodes. Udp based butterfly mix steps should be extremely fast and failure-tolerant, and convergence is almost as fast as a full mix (AllReduce) on every step.}
}
@article{ZhaoandPlazaCoRR-14,
       author = {Ting Zhao and Stephen M. Plaza},
        title = {Automatic Neuron Type Identification by Neurite Localization in the Drosophila Medulla},
      journal = {CoRR},
       volume = {arXiv:1409.1892},
         year = {2014},
          url = {http://arXiv:1409.1892},
     abstract = {Mapping the connectivity of neurons in the brain (i.e., connectomics) is a challenging problem due to both the number of connections in even the smallest organisms and the nanometer resolution required to resolve them. Because of this, previous connectomes contain only hundreds of neurons, such as in the C.elegans connectome. Recent technological advances will unlock the mysteries of increasingly large connectomes (or partial connectomes). However, the value of these maps is limited by our ability to reason with this data and understand any underlying motifs. To aid connectome analysis, we introduce algorithms to cluster similarly-shaped neurons, where 3D neuronal shapes are represented as skeletons. In particular, we propose a novel location-sensitive clustering algorithm. We show clustering results on neurons reconstructed from the Drosophila medulla that show high-accuracy.}
}
@article {ZhaoetalAM-10,
       author = {Zhou, Yongfeng and Huang, Wei and Liu, Jinyao and Zhu, Xinyuan and Yan, Deyue},
        title = {Self-Assembly of Hyperbranched Polymers and Its Biomedical Applications},
      journal = {Advanced Materials},
       volume = 22,
        issue = 41,
         year = 2010,
        pages = {4567-4590},
     abstract = {Hyperbranched polymers (HBPs) are highly branched macromolecules with a three-dimensional dendritic architecture. Due to their unique topological structure and interesting physical/chemical properties, HBPs have attracted wide attention from both academia and industry. In this paper, the recent developments in HBP self-assembly and their biomedical applications have been comprehensively reviewed. Many delicate supramolecular structures from zero-dimension (0D) to three-dimension (3D), such as micelles, fibers, tubes, vesicles, membranes, large compound vesicles and physical gels, have been prepared through the solution or interfacial self-assembly of amphiphilic HBPs. In addition, these supramolecular structures have shown promising applications in the biomedical areas including drug delivery, protein purifi cation/detection/delivery, gene transfection, antibacterial/antifouling materials and cytomimetic chemistry. Such developments promote the interdiscipline researches among surpramolecular chemistry, biomedical chemistry, nano-technology and functional materials.}
}
@article {ZhaoetalAM-13,
       author = {Zhao, Yu and Sakai, Fuji and Su, Lu and Liu, Yijiang and Wei, Kongchang and Chen, Guosong and Jiang, Ming},
        title = {Progressive Macromolecular Self-Assembly: From Biomimetic Chemistry to Bio-Inspired Materials},
      journal = {Advanced Materials},
       volume = 25,
        issue = 37,
         year = 2013,
        pages = {5215-5256},
     abstract = {Macromolecular self-assembly (MSA) has been an active and fruitful research fi eld since the 1980s, especially in this new century, which is promoted by the remarkable developments in controlled radical polymerization in polymer chemistry, etc. and driven by the demands in bio-related investigations and applications. In this review, we try to summarize the trends and recent pro-gress in MSA in relation to biomimetic chemistry and bio-inspired materials. Our paper covers representative achievements in the fabrication of artifi cial building blocks for life, cell-inspired biomimetic materials, and macro-molecular assemblies mimicking the functions of natural materials and their applications. It is true that the current status of the deliberately designed and obtained nano-objects based on MSA including a variety of micelles, multi-compartment vesicles, and some hybrid and complex nano-objects is at their very fi rst stage to mimic nature, but signifi cant and encouraging progress has been made in achieving a certain similarity in morphologies or properties to that of natural ones. Such achievements also demonstrate that MSA has played an important and irreplaceable role in the grand and long-standing research of biomimetic and bio-inspired materials, the future success of which depends on mutual and persistent efforts in polymer science, material science, supramolecular chemistry, and biology.}
}
@article{ZhaoetalCoRR-16,
        title = {Deep Recurrent Models with Fast-Forward Connections for Neural Machine Translation},
       author = {Jie Zhou and Ying Cao and Xuguang Wang and Peng Li and Wei Xu},
      journal = {CoRR},
       volume = {arXiv:1606.04199},
         year = {2016},
     abstract = {Neural machine translation (NMT) aims at solving machine translation (MT) problems with purely neural networks and exhibits promising results in recent years. However, most of the existing NMT models are of shallow topology and there is still a performance gap between the single NMT model and the best conventional MT system. In this work, we introduce a new type of linear connections, named fast-forward connections, based on deep Long Short-Term Memory (LSTM) network, together with the interleaved bi-directional way for stacking them. Fast-forward connections play an essential role to propagate the gradients in building the deep topology of depth 16. On WMT'14 English- to-French task, we achieved BLEU=37.7 with single attention model, which outperforms the corresponding single shallow model by 6.2 BLEU points. It is the first time that a single NMT model achieves state-of-the-art performance and outperforms the best conventional model by 0.7 BLEU points. Even without considering attention mechanism, we can still achieve BLEU=36.3. After the special handling for unknown words and the model ensemble, we obtained the best score on this task with BLEU=40.4. Our models are also verified on the more difficult WMT'14 English-to-German task.}
}
@inproceedings{ZhaoetalICML-08,
        title = {Efficient Maximum Margin Clustering via Cutting Plane Algorithm},
       author = {Bin Zhao and Fei Wang and Changshui Zhang},
    booktitle = {Proceedings of the {SIAM} International Conference on Data Mining},
         year = 2008,
        pages = {751-762},
}
@article{ZhaoetalNATURE-16,
       author = {Zhou, Xin and Zhu, Dantong and Qi, Xue-Lian and Li, Sihai and King, Samson G. and Salinas, Emilio and Stanford, Terrence R. and Constantinidis, Christos},
        title = {Neural correlates of working memory development in adolescent primates},
      journal = {Nature Communications},
    publisher = {Nature Publishing Group},
         year = {2016},
       volume = {9},
        issue = {7},
        pages = {13423},
     abstract = {Working memory ability matures after puberty, in parallel with structural changes in the prefrontal cortex, but little is known about how changes in prefrontal neuronal activity mediate this cognitive improvement in primates. To address this issue, we compare behavioural performance and neurophysiological activity in monkeys as they transitioned from puberty into adulthood. Here we report that monkeys perform working memory tasks reliably during puberty and show modest improvement in adulthood. The adult prefrontal cortex is characterized by increased activity during the delay period of the task but no change in the representation of stimuli. Activity evoked by distracting stimuli also decreases in the adult prefrontal cortex. The increase in delay period activity relative to the baseline activity of prefrontal neurons is the best correlate of maturation and is not merely a consequence of improved performance. Our results reveal neural correlates of the working memory improvement typical of primate adolescence.},
}
@incollection{ZhaoetalNIPS-09,
        title = {Non-Parametric Bayesian Dictionary Learning for Sparse Image Representations},
       author = {Mingyuan Zhou and Haojun Chen and John Paisley and Lu Ren and Guillermo Sapiro and Lawrence Carin},
    booktitle = {Advances in Neural Information Processing Systems 22},
       editor = {Y. Bengio and D. Schuurmans and J. Lafferty and C. K. I. Williams and A. Culotta},
        pages = {2295-2303},
         year = 2009
}
@article {ZhengetalCELL-18,
       author = {Zheng, Zhihao and Lauritzen, J. Scott and Perlman, Eric and Robinson, Camenzind G. and Nichols, Matthew and Milkie, Daniel and Torrens, Omar and Price, John and Fisher, Corey B. and Sharifi, Nadiya and Calle-Schuler, Steven A. and Kmecova, Lucia and Ali, Iqbal J. and Karsh, Bill and Trautman, Eric T. and Bogovic, John and Hanslovsky, Philipp and Jefferis, Gregory S. X. E. and Kazhdan, Michael and Khairy, Khaled and Saalfeld, Stephan and Fetter, Richard D. and Bock, Davi D.},
        title = {A Complete Electron Microscopy Volume Of The Brain Of Adult Drosophila melanogaster},
      journal = {Cell},
    publisher = {Cell Press},
       volume = {174},
         year = {2018},
        pages = {1-14},
     abstract = {Drosophila melanogaster has a rich repertoire of innate and learned behaviors. Its 100,000-neuron brain is a large but tractable target for comprehensive neural circuit mapping. Only electron microscopy (EM) enables complete, unbiased mapping of synaptic connectivity; however, the fly brain is too large for conventional EM. We developed a custom high-throughput EM platform and imaged the entire brain of an adult female fly. We validated the dataset by tracing brain-spanning circuitry involving the mushroom body (MB), intensively studied for its role in learning. Here we describe the complete set of olfactory inputs to the MB; find a new cell type providing driving input to Kenyon cells (the intrinsic MB neurons); identify neurons postsynaptic to Kenyon cell dendrites; and find that axonal arbors providing input to the MB calyx are more tightly clustered than previously indicated by light-level data. This freely available EM dataset will significantly accelerate Drosophila neuroscience.},
}
@incollection{ZhiyuanandBingLML-18,
       author = {Chen, Zhiyuan and Liu, Bing},
    booktitle = {Lifelong Machine Learning, Second Edition},
        title = {Continual Learning and Catastrophic Forgetting},
       series = {Synthesis Lectures on Artificial Intelligence and Machine Learning},
    publisher = {Morgan \& Claypool Publishers},
       volume = {12},
         year = {2018},
        pages = {1-207},
     abstract = {In the recent years, lifelong learning (LL) has attracted a great deal of attention in the deep learning community, where it is often called continual learning. Though it is well-known that deep neural networks (DNNs) have achieved state-of-the-art performances in many machine learning (ML) tasks, the standard multi-layer perceptron (MLP) architecture and DNNs suffer from catastrophic forgetting [McCloskey and Cohen, 1989] which makes it difficult for continual learning. The problem is that when a neural network is used to learn a sequence of tasks, the learning of the later tasks may degrade the performance of the models learned for the earlier tasks. Our human brains, however, seem to have this remarkable ability to learn a large number of different tasks without any of them negatively interfering with each other. Continual learning algorithms try to achieve this same ability for the neural networks and to solve the catastrophic forgetting problem. Thus, in essence, continual learning performs incremental learning of new tasks. Unlike many other LL techniques, the emphasis of current continual learning algorithms has not been on how to leverage the knowledge learned in previous tasks to help learn the new task better. In this chapter, we first give an overview of catastrophic forgetting and survey the proposed continual learning techniques that address the problem. We then introduce several recent continual learning methods in more detail. Two evaluation papers are also covered to evaluate the performances of some existing continual learning algorithms. Last but not least, we give a summary of the chapter and list the relevant evaluation datasets.},
}
@article{ZhongandWonhamTAC-90,
       author = {Zhong, H. and Wonham, W. M.},
        title = {On the Consistency of Hierarchical Supervision in Discrete-Event Systems},
      journal = {{IEEE} Transactions on Automatic Control},
       volume = 35,
        issue = 10,
        pages = {1125-1134},
         year = 1990,
}
@article {ZhongetalELIFE-14,
       author = {Zhong, Guisheng and He, Jiang and Zhou, Ruobo and Lorenzo, Damaris and Babcock, Hazen P. and Bennett, Vann and Zhuang, Xiaowei},
       editor = {Singer, Robert H},
        title = {Developmental mechanism of the periodic membrane skeleton in axons},
      journal = {eLife},
    publisher = {eLife Sciences Publications Limited},
       volume = 3,
         year = 2014,
     abstract = {Actin, spectrin, and associated molecules form a periodic sub-membrane lattice structure in axons. How this membrane skeleton is developed and why it preferentially forms in axons are unknown. Here, we studied the developmental mechanism of this lattice structure. We found that this structure emerged early during axon development and propagated from proximal regions to distal ends of axons. Components of the axon initial segment were recruited to the lattice late during development. Formation of the lattice was regulated by the local concentration of I$^2$II spectrin, which is higher in axons than in dendrites. Increasing the dendritic concentration of I$^2$II spectrin by overexpression or by knocking out ankyrin B induced the formation of the periodic structure in dendrites, demonstrating that the spectrin concentration is a key determinant in the preferential development of this structure in axons and that ankyrin B is critical for the polarized distribution of I$^2$II spectrin in neurites.},
}
@article{ZhouetalNATURE-COMMUNICATION-16,
       author = {Zhou, Xin and Zhu, Dantong and Qi, Xue-Lian and Li, Sihai and King, Samson G. and Salinas, Emilio and Stanford, Terrence R. and Constantinidis, Christos},
        title = {Neural correlates of working memory development in adolescent primates},
      journal = {Nature Communications},
         year = {2016},
       volume = {7},
        issue = {1},
        pages = {13423},
     abstract = {Working memory ability matures after puberty, in parallel with structural changes in the prefrontal cortex, but little is known about how changes in prefrontal neuronal activity mediate this cognitive improvement in primates. To address this issue, we compare behavioural performance and neurophysiological activity in monkeys as they transitioned from puberty into adulthood. Here we report that monkeys perform working memory tasks reliably during puberty and show modest improvement in adulthood. The adult prefrontal cortex is characterized by increased activity during the delay period of the task but no change in the representation of stimuli. Activity evoked by distracting stimuli also decreases in the adult prefrontal cortex. The increase in delay period activity relative to the baseline activity of prefrontal neurons is the best correlate of maturation and is not merely a consequence of improved performance. Our results reveal neural correlates of the working memory improvement typical of primate adolescence.},
}
@article{ZhuCAM-13,
       author = {Zhu, Yan},
        title = {The Drosophila visual system From neural circuits to behavior},
      journal = {Cell Adhesion and Migration},
    publisher = {Landes Bioscience},
       volume = {7},
        issue = {4},
         year = {2013},
        pages = {333-344},
     abstract = {A compact genome and a tiny brain make Drosophila the prime model to understand the neural substrate of behavior. The neurogenetic efforts to reveal neural circuits underlying Drosophila vision started about half a century ago, and now the field is booming with sophisticated genetic tools, rich behavioral assays, and importantly, a greater number of scientists joining from di erent backgrounds. This review will briefly cover the structural anatomy of the Drosophila visual system, the animal's visual behaviors, the genes involved in assembling these circuits, the new and powerful techniques, and the challenges ahead for ultimately identifying the general principles of biological computation in the brain.}
}
@article{ZhuCELL-ADHESION-MIGRATION-13,
       author = {Zhu, Y.},
        title = {{{T}he {D}rosophila visual system: {F}rom neural circuits to behavior}},
      journal = {Cell Adhesion and Migration},
       volume = {7},
       number = {4},
         year = {2013},
        pages = {333-344},
     abstract = {A compact genome and a tiny brain make Drosophila the prime model to understand the neural substrate of behavior. The neurogenetic efforts to reveal neural circuits underlying Drosophila vision started about half a century ago, and now the field is booming with sophisticated genetic tools, rich behavioral assays, and importantly, a greater number of scientists joining from different backgrounds. This review will briefly cover the structural anatomy of the Drosophila visual system, the animal's visual behaviors, the genes involved in assembling these circuits, the new and powerful techniques, and the challenges ahead for ultimately identifying the general principles of biological computation in the brain.}
}
@techreport{ZhuandGhahramaniTECHREPORT-02,
        title = {Learning from labeled and unlabeled data with label propagation},
       author = {Zhu, Xiaojin and Ghahramani, Zoubin},
         year = {2002},
  institution = {Technical Report CMU-CALD-02-107, Carnegie Mellon University}
}
@inproceedings{ZhuandMumfordCVPR-00,
       author = {S.C. Zhu and D. Mumford},
        title = {Learning generic prior models for visual computation},
    booktitle = CVPR,
        pages = {463-469},
        month = {June},
         year = 1997,
}
@article{ZhuandMumfordPAMI-97,
       author = {S.C. Zhu and D. Mumford},
        title = {Prior learning and {Gibbs} reaction-diffusion},
      journal = PAMI,
       volume = 19,
        issue = 11,
         year = 1997,
} 
@article{ZhuandRozellPLoS-08,
       author = {Zhu, Mengchen and Rozell, Christopher J.},
       editor = {Bethge, Matthias},
        title = {Modeling Inhibitory Interneurons in Efficient Sensory Coding Models},
      journal = {PLoS Computational Biology},
         year = {2015},
    publisher = {Public Library of Science},
       volume = {11},
        issue = {7},
        pages = {e1004353},
     abstract = {There is still much unknown regarding the computational role of inhibitory cells in the sensory cortex. While modeling studies could potentially shed light on the critical role played by inhibition in cortical computation, there is a gap between the simplicity of many models of sensory coding and the biological complexity of the inhibitory subpopulation. In particular, many models do not respect that inhibition must be implemented in a separate subpopulation, with those inhibitory interneurons having a diversity of tuning properties and characteristic E/I cell ratios. In this study we demonstrate a computational framework for implementing inhibition in dynamical systems models that better respects these biophysical observations about inhibitory interneurons. The main approach leverages recent work related to decomposing matrices into low-rank and sparse components via convex optimization, and explicitly exploits the fact that models and input statistics often have low-dimensional structure that can be exploited for efficient implementations. While this approach is applicable to a wide range of sensory coding models (including a family of models based on Bayesian inference in a linear generative model), for concreteness we demonstrate the approach on a network implementing sparse coding. We show that the resulting implementation stays faithful to the original coding goals while using inhibitory interneurons that are much more biophysically plausible.},
} 
@article{ZhuandvonderMalsburgNEUROCOMPUTING-02,
       author = {Junmei Zhu and Christoph von der Malsburg},
        title = {Synapto-synaptic interactions speed up dynamic link matching},
      journal = {Neurocomputing},
       volume = {44-46},
         year = {2002},
        pages = {721-728},
     abstract = {An extended dynamic link matching (DLM) model is presented for the fast creation of mappings invariant to position, scale, deformation and orientation, a process underlying various visual tasks that are fast and effortless for humans but are computationally very difficult. Structured interactions between synapses are represented by control units. Each control unit stands for a group of synapses that are consistent with each other in terms of relative position, scale and orientation. Specifically, it is shown in simulations how these high-order interactions between synapses lead to very fast DLM.}
}
@inproceedings{ZhuetalCVPR-10,
      title = {Latent hierarchical structural learning for object detection},
     author = {Long Zhu and Yuanhao Chen and Alan L. Yuille and William T. Freeman},
  booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
  publisher = {IEEE},
       year = 2010,
      pages = {1062-1069},
}
@inproceedings{ZhuetalCVPR-10b,
       title = {Part and appearance sharing: Recursive Compositional Models for Multi-View and Multi-Object Detection},
      author = {Long Zhu and Yuanhao Chen and Antonio Torralba and William T. Freeman and Alan L. Yuille},
   booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
   publisher = {IEEE},
        year = 2010,
       pages = {1919-1926},
}
@article{ZhuetalCoRR-15,
        title = {Long Short-Term Memory Over Tree Structures},
       author = {Xiaodan Zhu and Parinaz Sobhani and Hongyu Guo},
      journal = {CoRR},
       volume = {arXiv:1503.04881},
         year = 2015,
     abstract = {The chain-structured long short-term memory (LSTM) has showed to be effective in a wide range of problems such as speech recognition and machine translation. In this paper, we propose to extend it to tree structures, in which a memory cell can reflect the history memories of multiple child cells or multiple descendant cells in a recursive process. We call the model S-LSTM, which provides a principled way of considering long-distance interaction over hierarchies, e.g., language or image parse structures. We leverage the models for semantic composition to understand the meaning of text, a fundamental problem in natural language understanding, and show that it outperforms a state-of-the-art recursive model by replacing its composition layers with the S-LSTM memory blocks. We also show that utilizing the given structures is helpful in achieving a performance better than that without considering the structures.}
}
@article{ZhuetalCoRR-17,
       author = {Jun-Yan Zhu and Taesung Park and Phillip Isola and Alexei A. Efros},
        title = {Unpaired Image-to-Image Translation using Cycle-Consistent Adversarial Networks},
      journal = {CoRR},
       volume = {arXiv:1703.10593},
         year = {2017},
     abstract = {Image-to-image translation is a class of vision and graphics problems where the goal is to learn the mapping between an input image and an output image using a training set of aligned image pairs. However, for many tasks, paired training data will not be available. We present an approach for learning to translate an image from a source domain X to a target domain Y in the absence of paired examples. Our goal is to learn a mapping $G:X\rightarrow{}Y$ such that the distribution of images from G(X) is indistinguishable from the distribution Y using an adversarial loss. Because this mapping is highly under-constrained, we couple it with an inverse mapping $F:Y\rightarrow{}X$ and introduce a cycle consistency loss to push $F(G(X))\rightarrow{}X$ (and vice versa). Qualitative results are presented on several tasks where paired training data does not exist, including collection style transfer, object transfiguration, season transfer, photo enhancement, etc. Quantitative comparisons against several prior methods demonstrate the superiority of our approach.},
}
@article{ZhuetalCoRR-17b,
       author = {Yan Zhu and Shaoting Zhang and Dimitris N. Metaxas},
        title = {Interactive Reinforcement Learning for Object Grounding via Self-Talking},
      journal = {CoRR},
       volume = {arXiv:1712.00576},
         year = {2017},
     abstract = {Humans are able to identify a referred visual object in a complex scene via a few rounds of natural language communications. Success communication requires both parties to engage and learn to adapt for each other. In this paper, we introduce an interactive training method to improve the natural language conversation system for a visual grounding task. During interactive training, both agents are reinforced by the guidance from a common reward function. The parametrized reward function also cooperatively updates itself via interactions, and contribute to accomplishing the task. We evaluate the method on GuessWhat?! visual grounding task, and significantly improve the task success rate. However, we observe language drifting problem during training and propose to use reward engineering to improve the interpretability for the generated conversations. Our result also indicates evaluating goal-ended visual conversation tasks require semantic relevant metrics beyond task success rate.},
}
@article{ZhuetalCoRR-18,
       author = {Qi Zhu and Xiang Ren and Jingbo Shang and Yu Zhang and Frank F. Xu and Jiawei Han},
        title = {Integrating Local Context and Global Cohesiveness for Open Information Extraction},
      journal = {CoRR},
       volume = {arXiv:1804.09931},
         year = {2018},
     abstract = {Extracting entities and their relations from text is an important task for understanding massive text corpora. Open information extraction (IE) systems mine relation tuples (i.e., entity arguments and a predicate string to describe their relation) from sentences. These relation tuples are not confined to a predefined schema for the relations of interests. However, current Open IE systems focus on modeling local context information in a sentence to extract relation tuples, while ignoring the fact that global statistics in a large corpus can be collectively leveraged to identify high-quality sentence-level extractions. In this paper, we propose a novel Open IE system, called ReMine, which integrates local context signals and global structural signals in a unified, distant-supervision framework. Leveraging facts from external knowledge bases as supervision, the new system can be applied to many different domains to facilitate sentence-level tuple extractions using corpus-level statistics. Our system operates by solving a joint optimization problem to unify (1) segmenting entity/relation phrases in individual sentences based on local context; and (2) measuring the quality of tuples extracted from individual sentences with a translating-based objective. Learning the two subtasks jointly helps correct errors produced in each subtask so that they can mutually enhance each other. Experiments on two real-world corpora from different domains demonstrate the effectiveness, generality, and robustness of ReMine when compared to state-of-the-art open IE systems.}
}
@article{ZhuetalCoRR-18b,
       author = {Yao Zhu and Saksham Suri and Pranav Kulkarni and Yueru Chen and Jiali Duan and C.{-}C. Jay Kuo},
        title = {An Interpretable Generative Model for Handwritten Digit Image Synthesis},
      journal = {CoRR},
       volume = {arXiv:1811.04507},
         year = {2018}
}
@inproceedings{ZhuetalICML-03,
    author = {Xiaojin Zhu and Zoubin Ghahramani and John Lafferty},
     title = {Semi-supervised learning using gaussian fields and harmonic functions},
 booktitle = {Proceedings of the 20th International Conference on Machine Learning},
      year = 2003,
}
@article{ZhuetalNATURE-14,
       author = {Zhu, Bo and Luo, Shyh-Chyang and Zhao, Haichao and Lin, Hsing-An and Sekine, Jun and Nakao, Aiko and Chen, Chi and Yamashita, Yoshiro and Yu, Hsiao-hua},
        title = {Large enhancement in neurite outgrowth on a cell membrane-mimicking conducting polymer},
      journal = {Nature Communications},
         year = {2014},
    publisher = {Nature Publishing Group},
       volume = {5},
     abstract = {Although electrically stimulated neurite outgrowth on bioelectronic devices is a promising means of nerve regeneration, immunogenic scar formation can insulate electrodes from targeted cells and tissues, thereby reducing the lifetime of the device. Ideally, an electrode material capable of electrically interfacing with neurons selectively and efficiently would be integrated without being recognized by the immune system and minimize its response. Here we develop a cell membrane-mimicking conducting polymer possessing several attractive features. This polymer displays high resistance towards nonspecific enzyme/cell binding and recognizes targeted cells specifically to allow intimate electrical communication over long periods of time. Its low electrical impedance relays electrical signals efficiently. This material is capable to integrate biochemical and electrical stimulation to promote neural cellular behaviour. Neurite outgrowth is enhanced greatly on this new conducting polymer; in addition, electrically stimulated secretion of proteins from primary Schwann cells can also occur on it.},
}
@article{ZhuetalPAMI-98,
   author = {S. C. Zhu and Y. N. Wu and D. B. Mumford},
    title = {{FRAME}: {F}ilters, Random field And Maximum Entropy:
            {T}owards a Unified Theory for Texture Modeling,},
  journal =  PAMI,
     year = 1998,
   volume = 27,
    issue = 2,
    pages = {1-20}
}
@article{ZhuoetalAIJ-14,
        title = {Learning hierarchical task network domains from partially observed plan traces},
       author = {Hankz Hankui Zhuo and Héctor Muñoz-Avila and Qiang Yang},
      journal = {Artificial Intelligence},
       volume = {212},
        pages = {134-157},
         year = {2014},
     abstract = {Hierarchical Task Network (HTN) planning is an effective yet knowledge intensive problem-solving technique. It requires humans to encode knowledge in the form of methods and action models. Methods describe how to decompose tasks into subtasks and the preconditions under which those methods are applicable whereas action models describe how actions change the world. Encoding such knowledge is a difficult and time-consuming process, even for domain experts. In this paper, we propose a new learning algorithm, called HTNLearn, to help acquire HTN methods and action models. HTNLearn receives as input a collection of plan traces with partially annotated intermediate state information, and a set of annotated tasks that specify the conditions before and after the tasks' completion. In addition, plan traces are annotated with potentially empty partial decomposition trees that record the processes of decomposing tasks to subtasks. HTNLearn outputs are a collection of methods and action models. HTNLearn first encodes constraints about the methods and action models as a constraint satisfaction problem, and then solves the problem using a weighted MAX-SAT solver. HTNLearn can learn methods and action models simultaneously from partially observed plan traces (i.e., plan traces where the intermediate states are partially observable). We test HTNLearn in several HTN domains. The experimental results show that our algorithm HTNLearn is both effective and efficient.}
}
@article{ZilbersteinandRussellAIJ-95,
       author = {Zilberstein, S. and Russell, S. J.},
        title = {Optimal Composition of Real-Time Systems},
      journal = {Artificial Intelligence},
       volume = 79,
        issue = 2,
         year = 1995,
}
@book{Zimmer2018,
        title = {She Has Her Mother's Laugh: The Powers, Perversions, and Potential of Heredity},
       author = {Zimmer, Carl},
         year = {2018},
    publisher = {Penguin Publishing Group}
}
@article{ZimmermanandKambhampatiAIM-03,
       author = {Zimmerman, Terry and Kambhampati, Subbarao},
        title = {Learning-assisted Automated Planning: Looking Back, Taking Stock, Going Forward},
      journal = {AI Mag.},
       volume = {24},
       number = {2},
         year = {2003},
        pages = {73-96},
     abstract = {This article reports on an extensive survey and analysis of research work related to machine learning as it applies to automated planning over the past 30 years. Major research contributions are broadly characterized by learning method and then descriptive subcategories. Survey results reveal learning techniques that have extensively been applied and a number that have received scant attention. We extend the survey analysis to suggest promising avenues for future research in learning based on both previous experience and current needs in the planning community.},
}
@book{Zipf49,
       author = {G.K. Zipf},
        title = {Human Behaviour and the Principle of Least Effort},
    publisher = {Hafner},
      address = {New York},
         year = 1949,
}
@article{ZoccolanetalPNAS-09,
       author = {Zoccolan, D. and Oertelt, N. and DiCarlo, J. J. and Cox, D. D.},
        title = {A rodent model for the study of invariant visual object recognition},
      journal = {Proceedings of the National Academy of Sciences},
       volume = 106,
        issue = 21,
         year = 2009,
        pages = {8748-8753},
}
@article{ZomorodianandCarlssonDCG-04,
       author = {Zomorodian, Afra and Carlsson, Gunnar},
        title = {Computing Persistent Homology},
      journal = {Discrete \& Computational Geometry},
         year = {2004},
       volume = {33},
       number = {2},
        pages = {249-274},
     abstract = {We show that the persistent homology of a filtered d-dimensionalsimplicial complex is simply the standard homology of a particular graded module over a polynomial ring.Our analysis establishes the existence of a simple description ofpersistent homology groups over arbitrary fields.It also enables us to derive a natural algorithm for computing persistent homology of spaces in arbitrary dimension over any field.This result generalizes and extends the previously known algorithm that was restricted to subcomplexes of S3 and Z2 coefficients.Finally, our study implies the lack of a simple classification over non-fields. Instead, we give an algorithm for computing individual persistent homology groups over an arbitrary principal ideal domain in any dimension.},
}
@article{ZonouzietalCELL-19,
        title = {Individual Oligodendrocytes Show Bias for Inhibitory Axons in the Neocortex},
       author = {Marzieh Zonouzi and Daniel Berger and Vahbiz Jokhi and Amanda Kedaigle and Jeff Lichtman and Paola Arlotta},
      journal = {Cell Reports},
       volume = {27},
       number = {10},
         year = {2019},
        pages = {2799-2808},
     abstract = {Reciprocal communication between neurons and oligodendrocytes is essential for the generation and localization of myelin, a critical feature of the CNS. In the neocortex, individual oligodendrocytes can myelinate multiple axons; however, the neuronal origin of the myelinated axons has remained undefined and, while largely assumed to be from excitatory pyramidal neurons, it also includes inhibitory interneurons. This raises the question of whether individual oligodendrocytes display bias for the class of neurons that they myelinate. Here, we find that different classes of cortical interneurons show distinct patterns of myelin distribution starting from the onset of myelination, suggesting that oligodendrocytes can recognize the class identity of individual types of interneurons that they target. Notably, we show that some oligodendrocytes disproportionately myelinate the axons of inhibitory interneurons, whereas others primarily target excitatory axons or show no bias. These results point toward very specific interactions between oligodendrocytes and neurons and raise the interesting question of why myelination is differentially directed toward different neuron types.}
}
@inproceedings{ZophandLeICLR-17,
        title = {Neural architecture search with reinforcement learning},
       author = {Barret Zoph and Quoc V. Le},
    booktitle = {Proceedings of the 5th International Conference on Learning Representations},
         year = {2017},
     abstract = {Neural networks are powerful and flexible models that work well for many difficult learning tasks in image, speech and natural language understanding. Despite their success, neural networks are still hard to design. In this paper, we use a recurrent network to generate the model descriptions of neural networks and train this RNN with reinforcement learning to maximize the expected accuracy of the generated architectures on a validation set. On the CIFAR-10 dataset, our method, starting from scratch, can design a novel network architecture that rivals the best human-invented architecture in terms of test set accuracy. Our CIFAR-10 model achieves a test error rate of 3.84, which is only 0.1 percent worse and 1.2x faster than the current state-of-the-art model. On the Penn Treebank dataset, our model can compose a novel recurrent cell that outperforms the widely-used LSTM cell, and other state-of-the-art baselines. Our cell achieves a test set perplexity of 62.4 on the Penn Treebank, which is 3.6 perplexity better than the previous state-ofthe-art.}
}
@inproceedings{ZophetalCVPR-17-SUBMITTED,
        title = {Learning Transferable Architectures for Scalable Image Recognition},
       author = {Barret Zoph and Vijay Vasudevan and Jonathon Shlens and Quoc V. Le},
    booktitle = {{IEEE} Conference on Computer Vision and Pattern Recognition},
    publisher = {IEEE},
         year = 2017,
     abstract = {Developing state-of-the-art image classification models often requires significant architecture engineering and tuning. In this paper, we attempt to substantially reduce the amount of architecture engineering by using Neural Architecture Search to learn an architectural building block on a small dataset that can be transferred to a large dataset. This approach is similar to learning the structure of a recurrent cell within a recurrent network. In our experiments, we search for the best convolutional cell on the CIFAR-10 dataset and then apply this learned cell to the ImageNet dataset by stacking together more copies of this cell, each with their own parameters. Although the cell is not learned directly on ImageNet, an architecture constructed from the best learned cell achieves state-of-the-art accuracy of 82.3\% top-1 and 96.0\% top-5 on ImageNet, which is 0.8\% better in top-1 accuracy than the best human-invented architectures while having 9 billion fewer FLOPS, a reduction of 28\% from the previous state of the art model. When evaluated at particular levels of computational cost, our learned architectural building block exceeds those of the state-of-the-art human-designed models at equivalent computational cost, across two orders of magnitude in computational cost. In particular, a smaller network constructed from the best cell also achieves 74\% top-1 accuracy, which is 3.1\% better than equivalently-sized, state-of-the-art models for mobile platforms. Finally, we demonstrate that the image features learned from image classification are generically useful and transfer to other computer vision problems --- in particular, we demonstrate that naively employing these learned features in an object detection task surpasses state-of-the-art by 4.0\% achieving 43.1\% mAP on the COCO dataset.},
}
@article{ZophetalCoRR-17,
        title = {Learning Transferable Architectures for Scalable Image Recognition},
       author = {Barret Zoph and Vijay Vasudevan and Jonathon Shlens and Quoc V. Le},
      journal = {CoRR},
       volume = {arXiv:1707.07012},
         year = {2017},
         abstract = {Developing neural network image classification models often requires significant architecture engineering. In this paper, we attempt to automate this engineering process by learning the model architectures directly on the dataset of interest. As this approach is expensive when the dataset is large, we propose to search for an architectural building block on a small dataset and then transfer the block to a larger dataset. Our key contribution is the design of a new search space which enables transferability. In our experiments, we search for the best convolutional layer (or "cell") on the CIFAR-10 dataset and then apply this cell to the ImageNet dataset by stacking together more copies of this cell, each with their own parameters. Although the cell is not searched for directly on ImageNet, an architecture constructed from the best cell achieves, among the published works, state-of-the-art accuracy of 82.7\% top-1 and 96.2\% top-5 on ImageNet. Our model is 1.2\% better in top-1 accuracy than the best human-invented architectures while having 9 billion fewer FLOPS -- a reduction of 28\% in computational demand from the previous state-of-the-art model. When evaluated at different levels of computational cost, accuracies of our models exceed those of the state-of-the-art human-designed models. For instance, a smaller network constructed from the best cell also achieves 74\% top-1 accuracy, which is 3.1\% better than equivalently-sized, state-of-the-art models for mobile platforms. On CIFAR-10, an architecture constructed from the best cell achieves 2.4\% error rate, which is also state-of-the-art. Finally, the image features learned from image classification can also be transferred to other computer vision problems. On the task of object detection, the learned features used with the Faster-RCNN framework surpass state-of-the-art by 4.0\% achieving 43.1\% mAP on the COCO dataset.},
}
@article{ZorzosetalOPTICS-12,
      title = {Three-dimensional multiwaveguide probe array for light delivery to distributed brain circuits}, 
     author = {Zorzos, A. N. and Scholvin, J. and Boyden, E. S. and Fonstad, C. G.},
    journal = {Optics Letters},
  publisher = 37,
     volume = 23,
       year = 2012,
      pages = {4841-4843},
}
@article{ZuberetalPNAS-05,
       author = {Zuber, Benoît and Nikonenko, Irina and Klauser, Paul and Muller, Dominique and Dubochet, Jacques},
        title = {The mammalian central nervous synaptic cleft contains a high density of periodically organized complexes},
      journal = {Proceedings of the National Academy of Sciences of the United States of America},
       volume = {102},
        issue = {52},
         year = {2005},
        pages = {19192-19197},
     abstract = {Cryo-electron microscopy of vitreous section makes it possible to observe cells and tissues at high resolution in a close-to-native state. The specimen remains hydrated; chemical fixation and staining are fully avoided. There is minimal molecular aggregation and the density observed in the image corresponds to the density in the object. Accordingly, organotypic hippocampal rat slices were vitrified under high pressure and controlled cryoprotection conditions, cryosectioned at a final thickness of \hmapprox{}70 nm and observed below -170\hmdegree{}C in a transmission electron microscope. The general aspect of the tissue compares with previous electron microscopy observations. The detailed analysis of the synapse reveals that the density of material in the synaptic cleft is high, even higher than in the cytoplasm, and that it is organized in 8.2-nm periodic transcleft complexes. Previously undescribed structures of presynaptic and postsynaptic elements are also described.},
}
@article{ZuckerIEEE-14,
       author = {Steven W. Zucker},
        title = {Stereo, Shading, and Surfaces: Curvature Constraints Couple Neural Computations},
      journal = {Proceedings of the {IEEE}},
       volume = 102,
        issue = 5,
         year = 2014,
        pages = {812-829},
     abstract = {Vision problems are inherently ambiguous: Do abrupt brightness changes correspond to object boundaries? Are smooth intensity changes due to shading or material properties? For stereo: Which point in the left image corresponds to which point in the right one? What is the role of color in visual information processing? To answer these (seemingly different) questions we develop an analogy between the role of orientation in organizing visual cortex and tangents in differential geometry. Machine learning experiments suggest using geometry as a surrogate for high-order statistical interactions. The cortical columnar architecture becomes a bundle structure in geometry. Connection forms within these bundles suggest answers to the above questions, and curvatures emerge in key roles. More generally, our path through these questions suggests an overall strategy for solving the inverse problems of vision: decompose the global problems into networks of smaller ones and then seek constraints from these coupled problems to reduce ambiguity. Neural computations thus amount to satisfying constraints rather than seeking uniform approximations. Even when no global formulation exists one may be able to find localized structures on which ambiguity is minimal; these can then anchor an overall approximation.}
}
@article{ZuckerJoP-12,
       author = {Steven W. Zucker},
        title = {Local field potentials and border ownership: A conjecture about computation in visual cortex},
      journal = {Journal of Physiology - Paris},
       volume = 106,
         year = 2012,
        pages = {297-315},
     abstract = {Border ownership is an intermediate-level visual task: it must integrate (upward flowing) image information about edges with (downward flowing) shape information. This highlights the familiar local-to-global aspect of border formation (linking of edge elements to form contours) with the much less studied globalto-local aspect (which edge elements form part of the same shape). To address this task we show how to incorporate certain high-level notions of distance and geometric arrangement into a form that can influence image-based edge information. The center of the argument is a reaction-diffusion equation that reveals how (global) aspects of the distance map (that is, shape) can be "read out" locally, suggesting a solution to the border ownership problem. Since the reaction-diffusion equation defines a field, a possible information processing role for the local field potential can be defined. We argue that such fields also underlie the Gestalt notion of closure, especially when it is refined using modern experimental techniques. An important implication of this theoretical argument is that, if true, then network modeling must be extended to include the substrate surrounding spiking neurons, including glia.}
}
@article{Zugaroetal05,
       author = {Zugaro, M.B. and L. Monconduit and G. Buzsaki},
        title = {Spike phase precession persists after transient intrahippocampal perturbation},
      journal = {Natural Neuroscience},
       volume = 8,
        issue = 1,
        pages = {67-71},
         year = 2005,
}
@article{ZungetalCoRR-17,
       author = {Jonathan Zung and Ignacio Tartavull and H. Sebastian Seung},
        title = {An Error Detection and Correction Framework for Connectomics},
      journal = {CoRR},
       volume = {arXiv:1708.02599},
         year = {2015},
     abstract = {Significant advances have been made in recent years on the problem of neural circuit reconstruction from electron microscopic imagery. Improvements in image acquisition, image alignment, and boundary detection have greatly reduced the achievable error rate. In order to make further progress, we argue that automated error detection is essential for focusing the effort and attention of both human and machine. In this paper, we report on the use of automated error detection as an attention signal for a flood filling error correction module. We demonstrate significant improvements upon the state of the art in segmentation performance.},
}
@article{ZustetalPLoS-15,
       author = {Z\"{u}st, Marc Alain and Colella, Patrizio and Reber, Thomas Peter and Vuilleumier, Patrik and Hauf, Martinus and Ruch, Simon and Henke, Katharina},
      journal = {PLOS ONE},
    publisher = {Public Library of Science},
        title = {Hippocampus Is Place of Interaction between Unconscious and Conscious Memories},
         year = {2015},
       volume = {10},
       number = {3},
        pages = {1-19},
     abstract = {Recent evidence suggests that humans can form and later retrieve new semantic relations unconsciously by way of hippocampus-the key structure also recruited for conscious relational (episodic) memory. If the hippocampus subserves both conscious and unconscious relational encoding/retrieval, one would expect the hippocampus to be place of unconscious-conscious interactions during memory retrieval. We tested this hypothesis in an fMRI experiment probing the interaction between the unconscious and conscious retrieval of face-associated information. For the establishment of unconscious relational memories, we presented subliminal (masked) combinations of unfamiliar faces and written occupations ("actor" or "politician"). At test, we presented the former subliminal faces, but now supraliminally, as cues for the reactivation of the unconsciously associated occupations. We hypothesized that unconscious reactivation of the associated occupation-actor or politician-would facilitate or inhibit the subsequent conscious retrieval of a celebrity's occupation, which was also actor or politician. Depending on whether the reactivated unconscious occupation was congruent or incongruent to the celebrity's occupation, we expected either quicker or delayed conscious retrieval process. Conscious retrieval was quicker in the congruent relative to a neutral baseline condition but not delayed in the incongruent condition. fMRI data collected during subliminal face-occupation encoding confirmed previous evidence that the hippocampus was interacting with neocortical storage sites of semantic knowledge to support relational encoding. fMRI data collected at test revealed that the facilitated conscious retrieval was paralleled by deactivations in the hippocampus and neocortical storage sites of semantic knowledge. We assume that the unconscious reactivation has pre-activated overlapping relational representations in the hippocampus reducing the neural effort for conscious retrieval. This finding supports the notion of synergistic interactions between conscious and unconscious relational memories in a common, cohesive hippocampal-neocortical memory space.},
}
@inproceedings{ZwebenandEskeyIJCAI-89,
       author = {Zweben, Monte and Eskey, Megan},
        title = {Constraint Satisfaction with Delayed Evaluation},
    booktitle = {Proceedings {IJCAI} 11},
     location = {Detroit, Michigan},
 organization = {IJCAII},
        pages = {875-880},
         year = 1989,
}
@book{ZwebenandFox94,
       editor = {Zweben, Monte and Fox, Mark S.},
        title = {Intelligent Scheduling},
    publisher = {Morgan Kaufmann},
      address = {San Francisco, California},
         year = 1994,
}
@inproceedings{ZweigetalICASP-17,
       author = {Geoffrey Zweig and Chengzhu Yu and Jasha Droppo and Andreas Stolcke},
        title = {Advances in all-neural speech recognition},
    booktitle = {{IEEE} International Conference on Acoustics, Speech and Signal Processing},
        pages = {4805-4809},
         year = {2017},
     abstract = {This paper advances the design of CTC-based all-neural (or end-to-end) speech recognizers. We propose a novel symbol inventory, and a novel iterated-CTC method in which a second system is used to transform a noisy initial output into a cleaner version. We present a number of stabilization and initialization methods we have found useful in training these networks. We evaluate our system on the commonly used NIST 2000 conversational telephony test set, and significantly exceed the previously published performance of similar systems, both with and without the use of an external language model and decoding technology.}
}
@proceedings{conf/iwann/2003-1,
       editor = {Jos{\'e} Mira and Jos{\'e} R. {\'A}lvarez},
        title = {Artificial Neural Nets Problem Solving Methods, 7th International Work-Conference on Artificial and Natural Neural Networks, IWANN2003, Ma{\'o}, Menorca, Spain, June 3-6, 2003 Proceedings, Part I},
    booktitle = {Proceedings of the IWANN},
    publisher = {Springer},
       series = {Lecture Notes in Computer Science},
       volume = 2686,
         year = 2003,
}
@article {deHeviaetalPNAS-14,
       author = {de Hevia, Maria Dolores and Izard, V\'{e}ronique and Coubart, Aur\'{e}lie and Spelke, Elizabeth S. and Streri, Arlette},
      journal = {Proceedings of the National Academy of Sciences},
        title = {Representations of space, time, and number in neonates},
       volume = {111},
       number = {13},
         year = {2014},
    publisher = {National Academy of Sciences},
        pages = {4809-4813},
     abstract = {Space, time, and number are connected in the world and in the human mind. How do these connections arise? Do we learn to link larger numbers and durations to longer spatial extents because they are correlated in the world, or is the human mind built to capture these relations? We showed that neonates relate both number and duration to spatial length when these dimensions vary in the same direction (number or duration increases as length increases), but not in opposite directions (number or duration increases and length decreases). After being familiarized to a pairing between two magnitudes, newborns expect these dimensions to change in the same direction. At birth, humans are sensitive to the common structure of these fundamental magnitudes. A rich concept of magnitude{\textemdash}in its numerical, spatial, and temporal forms{\textemdash}is a central foundation of mathematics, science, and technology, but the origins and developmental relations among the abstract concepts of number, space, and time are debated. Are the representations of these dimensions and their links tuned by extensive experience, or are they readily available from birth? Here, we show that, at the beginning of postnatal life, 0- to 3-d-old neonates reacted to a simultaneous increase (or decrease) in spatial extent and in duration or numerical quantity, but they did not react when the magnitudes varied in opposite directions. The findings provide evidence that representations of space, time, and number are systematically interrelated at the start of postnatal life, before acquisition of language and cultural metaphors, and before extensive experience with the natural correlations between these dimensions.},
}
@inproceedings{deKleer84,
       author = {{de Kleer}, Johan},
        title = {Choices Without Backtracking},
    booktitle = {Proceedings AAAI-84},
     location = {Austin, Texas},
 organization = {AAAI},
         year = 1984,
}
@article{deKleerAIJ-84,
       author = {{de Kleer}, Johan and Brown, John Seely},
        title = {A Qualitative Physics based on Confluences},
      journal = AIJ,
       volume = 24,
         year = 1984,
        pages = {7-83},
}
@article{deKleerAIJ-87,
       author = {{de Kleer}, Johan and Williams, Brian C.},
        title = {Diagnosing Multiple Faults},
      journal = AIJ,
       volume = 32,
        issue = 1,
        pages = {97-130},
         year = 1987,
}
@techreport{deKleerAMORD,
       author = {{de Kleer}, Johan and Doyle, Jon and Rich, Charles and Steele, Guy L. and Sussman, Gerald J.},
        title = {{AMORD} A Deductive Procedure System},
        issue = {AI-TR-435},
  institution = {{MIT} {AI} Laboratory},
         year = 1978,
}
@article{deKleerATMS,
       author = {{de Kleer}, Johan},
        title = {An Assumption-based TMS},
      journal = AIJ,
       volume = 28,
         year = 1986,
        pages = {127-163},
}
@inproceedings{deKleerIJCAI-77,
       author = {{de Kleer}, Johan},
        title = {Multiple Representations of Knowledge in a Mechanics Problem Solver},
    booktitle = {Proceedings {IJCAI} 5},
 organization = {IJCAII},
     location = {Cambridge, Massachusetts},
         year = 1977,
        pages = {299-304},
}
@incollection{deKleerQR-85,
       author = {{de Kleer}, Johan and Brown, John Seely},
        title = {A Qualitative Physics based on Confluences},
     crossref = {QR-85},
        pages = {7-84},
}
@inproceedings{deKleerandBrown,
       author = {{de Kleer}, Johan and Brown, J.S.},
        title = {Foundations of Envisioning},
    booktitle = {Proceedings AAAI-82},
     location = {Pittsburgh, Pennsylvania},
 organization = {AAAI},
         year = 1982,
}
@inproceedings{deKleeretal,
       author = {{de Kleer}, Johan and Doyle, Jon and {Steele Jr.}, Guy L. and Sussman, Gerald J.},
        title = {{AMORD}: Explicit Control of Reasoning},
    booktitle = {Proceedings of the {ACM} Symposium on Artificial Intelligence and Programming Languages},
         year = 1977,
        pages = {116-125},
}
@incollection{deKleeretal-85,
       author = {{de Kleer}, Johan and Doyle, Jon and {Steele Jr.}, Guy L. and Sussman, Gerald Jay},
        title = {{\sc Amord}: Explicit Control of Reasoning},
    booktitle = {Readings in Knowledge Representation},
    publisher = {Morgan Kaufmann},
         year = 1985,
       editor = {Brachman, Ronald J. and Levesque, Hector J.},
        pages = {346-355},
      address = {San Francisco, California},
      comment = {Originally published in 1977},
}
@article{deVilliersLINGUA-07,
       author = {de Villiers, Jill},
        title = {The Interface of Language and Theory of Mind},
      journal = {Lingua: International Review of General Linguistics},
         year = {2007},
       volume = {117},
        issue = {11},
        pages = {1858-1878},
     abstract = {The proposal is made that the interface between language and theory of mind is bidirectional. It seems probable that the conceptual developments of early Theory of Mind form an essential basis for helping to fix at least word reference. In development from two to four years, no basis exists in research for conclusions about the direction of influence between language and Theory of Mind. At the stage of false belief reasoning, after age four, the role of the mastery of syntactic complementation is highlighted as a representational tool, that is, language development assists reasoning. The paper presents a brief summary of Theory of Mind, ranging from its earliest beginnings in infancy to the appreciation around age four years that others might hold false beliefs and act according to them. For each development, the parallel language developments are described, and questions are raised about the interface between the two. In particular, research that might determine the direction of influence from one to the other is discussed. More work is called for, especially with nonverbal tasks, good experimental linguistic work, and other special populations, that might allow a more precise delineation of how language and Theory of Mind interrelate at the interface.},
}
@book{duSautoy2014primes,
        title = {The Music of the Primes: Searching to Solve the Greatest Mystery in Mathematics},
       author = {du Sautoy, Marcus},
         year = {2004},
    publisher = {HarperCollins}
}
@book{jurafsky2014speech,
        title = {Speech and Language Processing},
       author = {Jurafsky, D. and Martin, J.H.},
      comment = {Second Edition},
         year = {2014},
    publisher = {Pearson}
}
@article{vanDenOordetalCoRR-16,
       author = {A{\"{a}}ron van den Oord and Nal Kalchbrenner and Koray Kavukcuoglu},
        title = {Pixel Recurrent Neural Networks},
      journal = {CoRR},
       volume = {arXiv:1601.06759},
         year = {2016},
     abstract = {Modeling the distribution of natural images is a landmark problem in unsupervised learning. This task requires an image model that is at once expressive, tractable and scalable. We present a deep neural network that sequentially predicts the pixels in an image along the two spatial dimensions. Our method models the discrete probability of the raw pixel values and encodes the complete set of dependencies in the image. Architectural novelties include fast two-dimensional recurrent layers and an effective use of residual connections in deep recurrent networks. We achieve log-likelihood scores on natural images that are considerably better than the previous state of the art. Our main results also provide benchmarks on the diverse ImageNet dataset. Samples generated from the model appear crisp, varied and globally coherent.},
}
@article{vanDenOordetalCoRR-18,
       author = {Representation Learning with Contrastive Predictive Coding},
        title = {Aaron van den Oord and Yazhe Li and Oriol Vinyals},
      journal = {CoRR},
       volume = {arXiv:1807.03748},
         year = {2018},
     abstract = {While supervised learning has enabled great progress in many applications, unsupervised learning has not seen such widespread adoption, and remains an important and challenging endeavor for artificial intelligence. In this work, we propose a universal unsupervised learning approach to extract useful representations from high-dimensional data, which we call Contrastive Predictive Coding. The key insight of our model is to learn such representations by predicting the future in latent space by using powerful autoregressive models. We use a probabilistic contrastive loss which induces the latent space to capture information that is maximally useful to predict future samples. It also makes the model tractable by using negative sampling. While most prior work has focused on evaluating representations for a particular modality, we demonstrate that our approach is able to learn useful representations achieving strong performance on four distinct domains: speech, images, text and reinforcement learning in 3D environments.}
}
@article{vanDerWorpetalPLoS-10,
       author = {van der Worp, H.B. and Howells, D.W. and Sena, E.S. and Porritt, M.J. and Rewell, S. and O'Collins, V. and Macleod, M.R.},
        title = {Can Animal Models of Disease Reliably Inform Human Studies?},
      journal = {{PLoS} Biology},
       volume = 7,
        issue = 3,
         year = 2010,
        pages = {e1000245},
}
@article {vanGervenBIORXIV-17,
       author = {van Gerven, Marcel},
        title = {Computational Foundations of Natural Intelligence},
      journal = {bioRxiv},
         year = {2017},
    publisher = {Cold Spring Harbor Labs Journals},
     abstract = {New developments in AI and neuroscience are revitalizing the quest to understanding natural intelligence, offering insight about how to equip machines with human-like capabilities. This paper reviews some of the computational principles relevant for understanding natural intelligence and, ultimately, achieving strong AI. After reviewing basic principles, a variety of computational modeling approaches is discussed. Subsequently, I concentrate on the use of artificial neural networks as a framework for modeling cognitive processes. This paper ends by outlining some of the challenges that remain to fulfill the promise of machines that show human-like intelligence.},
}
@article{vanHaterenandvanderSchaafRSL-B-98,
       author = {van Hateren, J. H. and van der Schaaf, A.},
        title = {Independent component filters of natural images compared with simple cells in primary visual cortex},
      journal = {Proceedings Royal Society London {B}},
       volume = 265,
        pages = {359-366},
         year = 1998,
}
@article{vanHeugtenandShiJASA-10,
       author = {van Heugten, Marieke and Shi, Rushen},
        title = {Infants' sensitivity to non-adjacent dependencies across phonological phrase boundaries},
      journal = {The Journal of the Acoustical Society of America},
         year = 2010,
       volume = 128,
        issue = 5,
        pages = {223-228},
     abstract = {Natural languages contain numerous non-adjacent relationships between words or morphemes in a sentence, often straddling phonological phrase boundaries (e.g., [these sheep] [have ...]). Since phonological phrases are considered the main processing unit for infants, this may cause the acquisition of cross-phrase dependencies to be challenging. This study, however, shows that by 17 months of age, French-learning infants have nonetheless gained sensitivity to remote determiner-auxiliary co-occurrences that are interceded by phonological phrase boundaries. Infants thus possess a robust mechanism for tracking non-adjacent dependencies. This ability is essential for early grammatical development.}
}
@article{vanSwinderenCSHL-11,
       author = {van Swinderen, Bruno},
        title = {Single Fly Tethered Paradigms},
       volume = {2011},
       number = {12},
         year = {2011},
      abstract = {The most successful approaches for studying visual perception and visual learning in Drosophila have been single fly paradigms in which tethered individuals respond to different visual stimuli, as described here. The equipment and protocols involved are quite sophisticated and differ depending on whether behavior or electrophysiology will be pursued. For either approach, flies must first be secured to a metal wire. This is typically performed by first cooling flies down to 4°C and then gluing them to a copper or tungsten wire with ultraviolet (UV)-activated cement. For electrophysiology, tethering requires a few extra steps to accommodate the placement of electrodes. Prepared individuals are then placed inside a cylindrical arena where images can be presented, or in front of a computer screen or even in front of a laptop. Flight dynamics or brain activity in response to visual stimuli is recorded by using a variety of specialized and/or commercially available electronic devices.},
          url = {http://cshprotocols.cshlp.org/content/2011/12/pdb.prot066910.abstract},
      journal = {Cold Spring Harbor Protocols}
}
@book{vanVlietandHendry2004,
        title = {Modeling Financial Markets},
       author = {Van Vliet, B. and Hendry, R.},
         year = {2004},
    publisher = {McGraw-Hill Education}
}
@article{vanWassenhoveFiP-13,
       author = {van Wassenhove, Virginie},
        title = {Speech through ears and eyes: interfacing the senses with the supramodal brain},
      journal = {Frontiers in Psychology},
    publisher = {Frontiers Media},
         year = {2013},
       volume = {4},
        pages = {388-405},
     abstract = {The comprehension of auditory-visual (AV) speech integration has greatly benefited from recent advances in neurosciences and multisensory research. AV speech integration raises numerous questions relevant to the computational rules needed for binding information (within and across sensory modalities), the representational format in which speech information is encoded in the brain (e.g., auditory vs. articulatory), or how AV speech ultimately interfaces with the linguistic system. The following non-exhaustive review provides a set of empirical findings and theoretical questions that have fed the original proposal for predictive coding in AV speech processing. More recently, predictive coding has pervaded many fields of inquiries and positively reinforced the need to refine the notion of internal models in the brain together with their implications for the interpretation of neural activity recorded with various neuroimaging techniques. However, it is argued here that the strength of predictive coding frameworks reside in the specificity of the generative internal models not in their generality; specifically, internal models come with a set of rules applied on particular representational formats themselves depending on the levels and the network structure at which predictive operations occur. As such, predictive coding in AV speech owes to specify the level(s) and the kinds of internal predictions that are necessary to account for the perceptual benefits or illusions observed in the field. Among those specifications, the actual content of a prediction comes first and foremost, followed by the representational granularity of that prediction in time. This review specifically presents a focused discussion on these issues.},
}
@techreport{van_BenthemTR-ILLC-08,
       author = {van Benthem, Johan},
        title = {A brief history of natural logic},
  institution = {Institute for Logic, Language \& Computation},
         year = 2008,
        issue = {2008-05},
          url = {http://www.illc.uva.nl/Research/Publications/Reports/PP-2008-05.text.pdf}
}
@inproceedings{van_EijckSLLC-07,
       author = {Jan Van Eijck},
        title = {Natural Logic for Natural Language},
       editor = {Balder D. ten Cate and Henk W. Zeevat},
    booktitle = {Proceedings 6th International Tbilisi Symposium on Logic, Language, and Computation},
         year = 2007,
        pages = {216-230},
     abstract = {For a cognitive account of reasoning it is useful to factor out the syntactic aspect -- the aspect that has to do with pattern matching and simple substitution -- from the rest. The calculus of monotonicity, alias the calculus of natural logic, does precisely this, for it is a calculus of appropriate substitutions at marked positions in syntactic structures. We first introduce the semantic and the syntactic sides of monotonicity reasoning or 'natural logic', and propose an improvement to the syntactic monotonicity calculus, in the form of an improved algorithm for monotonicity marking. Next, we focus on the role of monotonicity in syllogistic reasoning. In particular, we show how the syllogistic inference rules (for traditional syllogistics, but also for a broader class of quantifiers) can be decomposed in a monotonicity component, an argument swap component, and an existential import component. Finally, we connect the decomposition of syllogistics to the doctrine of distribution.}
}
@inproceedings{vandeSandeetalICCV-11,
        title = {Segmentation As Selective Search for Object Recognition},
       author = {van de Sande, K. E. A. and Uijlings, J. R. R. and Gevers, T. and Smeulders, A. W. M.},
    booktitle = {{IEEE} International Conference on Computer Vision},
    publisher = {{IEEE} Computer Society},
         year = 2011,
        pages = {1879-1886},
}
@article{vandenOordCoRR-16,
       author = {A{\"{a}}ron van den Oord and Sander Dieleman and Heiga Zen and Karen Simonyan and Oriol Vinyals and Alex Graves and Nal Kalchbrenner and Andrew W. Senior and Koray Kavukcuoglu},
        title = {WaveNet: {A} Generative Model for Raw Audio},
      journal = {CoRR},
       volume = {arXiv:1609.03499},
         year = {2016},
     abstract = {This paper introduces WaveNet, a deep neural network for generating raw audio waveforms. The model is fully probabilistic and autoregressive, with the predictive distribution for each audio sample conditioned on all previous ones; nonetheless we show that it can be efficiently trained on data with tens of thousands of samples per second of audio. When applied to text-to-speech, it yields state-of-the-art performance, with human listeners rating it as significantly more natural sounding than the best parametric and concatenative systems for both English and Mandarin. A single WaveNet can capture the characteristics of many different speakers with equal fidelity, and can switch between them by conditioning on the speaker identity. When trained to model music, we find that it generates novel and often highly realistic musical fragments. We also show that it can be employed as a discriminative model, returning promising results for phoneme recognition.}
}
@article{vandenOordCoRR-17,
       author = {A{\"{a}}ron van den Oord and Oriol Vinyals and Koray Kavukcuoglu},
        title = {Neural Discrete Representation Learning},
      journal = {CoRR},
       volume = {arXiv:1711.00937},
         year = {2017},
     abstract = {Learning useful representations without supervision remains a key challenge in machine learning. In this paper, we propose a simple yet powerful generative model that learns such discrete representations. Our model, the Vector Quantised-Variational AutoEncoder (VQ-VAE), differs from VAEs in two key ways: the encoder network outputs discrete, rather than continuous, codes; and the prior is learnt rather than static. In order to learn a discrete latent representation, we incorporate ideas from vector quantisation (VQ). Using the VQ method allows the model to circumvent issues of "posterior collapse" -- where the latents are ignored when they are paired with a powerful autoregressive decoder -- typically observed in the VAE framework. Pairing these representations with an autoregressive prior, the model can generate high quality images, videos, and speech as well as doing high quality speaker conversion and unsupervised learning of phonemes, providing further evidence of the utility of the learnt representations.}
}
@article{vanderMaatenandHintonJMLR-08,
        title = {Visualizing Data using t-{SNE}},
       author = {Laurens van der Maaten and Geoffrey Hinton},
      journal = {Journal of Machine Learning Research},
       volume = 9,
         year = 2008,
        pages = {2579-2605},
     abstract = {We present a new technique called "t-SNE" that visualizes high-dimensional data by giving each datapoint a location in a two or three-dimensional map. The technique is a variation of Stochastic Neighbor Embedding (Hinton and Roweis, 2002) that is much easier to optimize, and produces significantly better visualizations by reducing the tendency to crowd points together in the center of the map. t-SNE is better than existing techniques at creating a single map that reveals structure at many different scales. This is particularly important for high-dimensional data that lie on several different, but related, low-dimensional manifolds, such as images of objects from multiple classes seen from multiple viewpoints. For visualizing the structure of very large data sets, we show how t-SNE can use random walks on neighborhood graphs to allow the implicit structure of all of the data to influence the way in which a subset of the data is displayed. We illustrate the performance of t-SNE on a wide variety of data sets and compare it with many other non-parametric visualization techniques, including Sammon mapping, Isomap, and Locally Linear Embedding. The visualizations produced by t-SNE are significantly better than those produced by the other techniques on almost all of the data sets.}
}
@book{vanderMade2013,
       author = {van der Made, Peter AJ},
        title = {Higher Intelligence: How to Create a Functional Artificial Brain},
         year = {2013},
    publisher = {Vivid Publishing},
      address = {Western Australia},
}
@techreport{vanderVeldePhD-03,
       author = {Frank van der Velde},
        title = {Neural Blackboard Architectures of Combinatorial Structures in Cognition},
        issue = {{PhD Dissertation}},
  institution = {Unit of Cognitive Psychology, Leiden University},
         year = 2003,
     abstract = {Human cognition is unique in the way in which it relies on combinatorial (or compositional) structures. Language provides ample evidence for the existence of combinatorial structures, but they can also be found in visual cognition. To understand the neural basis of human cognition, it is therefore essential to understand how combinatorial structures can be instantiated in neural terms. In his recent book on the foundations of language, Jackendoff formulated four fundamental problems for a neural instantiation of combinatorial structures: the massiveness of the binding problem, the problem of 2, the problem of variables and the transformation of combinatorial structures from working memory to long-term memory. This paper aims to show that these problems can be solved by means of neural blackboard architectures. For this purpose, a neural blackboard architecture for sentence structure is presented. In this architecture, neural structures that encode for words are temporarily bound in a manner that preserves the structure of the sentence. It is shown that the architecture solves the four problems presented by Jackendoff. The ability of the architecture to instantiate sentence structures is illustrated with examples of sentence complexity observed in human language performance. Similarities exist between the architecture for sentence structure and blackboard architectures for combinatorial structures in visual cognition, derived from the structure of the visual cortex. These architectures are briefly discussed, together with an example of a combinatorial structure in which the blackboard architectures for language and vision are combined. In this way, the architecture for language is grounded in perception.}
}
@article{vanderVeldeanddeKampsBBS-06,
       author = {van der Velde, F. and de Kamps, M.},
        title = {Neural blackboard architectures of combinatorial structures in cognition},
      journal = {Behavioral Brain Science},
         year = 2006,
       volume = 29,
        issue = 1,
        pages = {37-70},
     abstract = {Human cognition is unique in the way in which it relies on combinatorial (or compositional) structures. Language provides ample evidence for the existence of combinatorial structures, but they can also be found in visual cognition. To understand the neural basis of human cognition, it is therefore essential to understand how combinatorial structures can be instantiated in neural terms. In his recent book on the foundations of language, Jackendoff described four fundamental problems for a neural instantiation of combinatorial structures: the massiveness of the binding problem, the problem of 2, the problem of variables, and the transformation of combinatorial structures from working memory to long-term memory. This paper aims to show that these problems can be solved by means of neural "blackboard" architectures. For this purpose, a neural blackboard architecture for sentence structure is presented. In this architecture, neural structures that encode for words are temporarily bound in a manner that preserves the structure of the sentence. It is shown that the architecture solves the four problems presented by Jackendoff. The ability of the architecture to instantiate sentence structures is illustrated with examples of sentence complexity observed in human language performance. Similarities exist between the architecture for sentence structure and blackboard architectures for combinatorial structures in visual cognition, derived from the structure of the visual cortex. These architectures are briefly discussed, together with an example of a combinatorial structure in which the blackboard architectures for language and vision are combined. In this way, the architecture for language is grounded in perception. Perspectives and potential developments of the architectures are discussed.}
}
@article{vanderVeldeanddeKampsCN-15,
       author = {van der Velde, Frank and de Kamps, Marc},
        title = {The necessity of connection structures in neural models of variable binding},
      journal = {Cognitive Neurodynamics},
         year = {2015},
       number = {4},
       volume = {9},
        pages = {359-370},
     abstract = {In his review of neural binding problems, Feldman (Cogn Neurodyn 7:1-11, 2013) addressed two types of models as solutions of (novel) variable binding. The one type uses labels such as phase synchrony of activation. The other (`connectivity based') type uses dedicated connections structures to achieve novel variable binding. Feldman argued that label (synchrony) based models are the only possible candidates to handle novel variable binding, whereas connectivity based models lack the flexibility required for that. We argue and illustrate that Feldman's analysis is incorrect. Contrary to his conclusion, connectivity based models are the only viable candidates for models of novel variable binding because they are the only type of models that can produce behavior. We will show that the label (synchrony) based models analyzed by Feldman are in fact examples of connectivity based models. Feldman's analysis that novel variable binding can be achieved without existing connection structures seems to result from analyzing the binding problem in a wrong frame of reference, in particular in an outside instead of the required inside frame of reference. Connectivity based models can be models of novel variable binding when they possess a connection structure that resembles a small-world network, as found in the brain. We will illustrate binding with this type of model with episode binding and the binding of words, including novel words, in sentence structures.},
}
@inproceedings{vonAhnetalCHI-06,
       author = {Luis von Ahn and Ruoran Liu and Manuel Blum},
        title = {Peekaboom: a game for locating objects in images},
    booktitle = {Proceedings of the {SIGCHI} conference on Human Factors in computing systems},
     location = {Montr\'{e}al, Qu\'{e}bec, Canada},
    publisher = {ACM},
        pages = {55-64},
         year = {2006},
}
@article{vonBartheldetalJCN-16,
       author = {von Bartheld, Christopher S. and Bahney, Jami and Herculano-Houzel, Suzana},
        title = {The search for true numbers of neurons and glial cells in the human brain: A review of 150 years of cell counting},
      journal = {J Comp Neurol},
       volume = {524},
        issue = {18},
         year = {2016},
        pages = {3865--3895},
     abstract = {For half a century, the human brain was believed to contain about 100 billion neurons and one trillion glial cells, with a glia:neuron ratio of 10:1. A new counting method, the isotropic fractionator, has challenged the notion that glia outnumber neurons and revived a question that was widely thought to have been resolved. The recently validated isotropic fractionator demonstrates a glia:neuron ratio of less than 1:1 and a total number of less than 100 billion glial cells in the human brain. A survey of original evidence shows that histological data always supported a 1:1 ratio of glia to neurons in the entire human brain, and a range of 40-130 billion glial cells. We review how the claim of one trillion glial cells originated, was perpetuated, and eventually refuted. We compile how numbers of neurons and glial cells in the adult human brain were reported and we examine the reasons for an erroneous consensus about the relative abundance of glial cells in human brains that persisted for half a century. Our review includes a brief history of cell counting in human brains, types of counting methods that were and are employed, ranges of previous estimates, and the current status of knowledge about the number of cells. We also discuss implications and consequences of the new insights into true numbers of glial cells in the human brain, and the promise and potential impact of the newly validated isotropic fractionator for reliable quantification of glia and neurons in neurological and psychiatric diseases.},
} 
@book{vonEconomoandKoskinas1929,
        title = {The cytoarchitectonics of the human cerebral cortex},
       author = {von Economo, C. and Koskinas, G. N.},
    publisher = {Oxford University Press},
      comment = {Translated by Steven Parker},
       series = {Oxford Medical Publications},
      address = {London},
         year = 1929,
}
@article{vonMelchneretalNATURE-00,
       author = {von Melchner, Laurie and Pallas, Sarah L. and Sur, Mriganka},
        title = {Visual Behaviour Mediated by Retinal Projections Directed to the Auditory Pathway},
      journal = {Nature},
        issue = 6780,
        pages = {871-876},
       volume = 404,
         year = 2000,
}
@book{vonNeumann1958thebrain,
        title = {The computer and the brain},
       author = {John von Neumann},
       series = {The Silliman Memorial Lectures Series},
         year = {1958},
    publisher = {Yale University Press}
}
@book{vonNeumann53,
       author = {John von Neumann and Oskar Morgenstern},
        title = {Theory of Games and Economic Behavior, Third Edition},
    publisher = {Princeton University Press},
      address = {Princeton},
         year = 1953,
}
@book{vonUexk1926theoretical,
        title = {Theoretical Biology},
       author = {von Uexk\"{u}ll, J. and Mackinnon, D.L.},
       series = {{International Library of Psychology, Philosophy and Scientific Method}},
         year = {1926},
    publisher = {K. Paul, Trench, Trubner \& Company Limited}
}
@book{vonWinterfeldtandEdwards86,
       author = {{von Winterfeldt}, Detlof and Edwards, Ward},
        title = {Decision Analysis and Behavioral Research},
    publisher = {Cambridge University Press},
      address = {New York},
         year = 1986,
}
@techreport{vonderMalsburgMPI-81,
        title = {The Correlation Theory of Brain Function},
       author = {Christoph von der Malsburg},
        issue = {{Internal Report 81-2}},
  institution = {Max Planck Institute for Biophysical Chemistry},
         year = 1981
}
@incollection{vonderMalsburgPNN-94,
        title = {The Correlation Theory of Brain Function},
       author = {Christoph von der Malsburg},
    booktitle = {Models of Neural Networks: Physics of Neural Networks},
       editor = {Domany, E. and van Hemmen, J.L. and Schulten, K.},
    publisher = {Springer},
         year = {1994},
     abstract = {A summary of brain theory is given so far as it is contained within the framework of Localization Theory. Diffculties of this "conventional theory" are traced back to a specific deficiency: there is no way to express relations between active cells (as for instance their representing parts of the same object). A new theory is proposed to cure this deficiency. It introduces a new kind of dynamical control, termed synaptic modulation, according to which synapses switch between a conducting and a non- conducting state. The dynamics of this variable is controlled on a fast time scale by correlations in the temporal fine structure of cellular signals. Furthermore, conventional synaptic plasticity is replaced by a refined version. Synaptic modulation and plasticity form the basis for short-term and long-term memory, respectively. Signal correlations, shaped by the variable network, express structure and relationships within objects. In particular, the figure-ground problem may be solved in this way. Synaptic modulation introduces flexibility into cerebral networks which is necessary to solve the invariance problem. Since momentarily useless connections are deactivated, interference between different memory traces can be reduced, and memory capacity increased, in comparison with conventional associative memory.}
}
