CS379C: Computational Models of the Neocortex

Spring 2018

Description:


In CS379C this quarter, we consider the following three challenge problems:
  1. How would you design a personal assistant capable of maintaining relationships with each member of a family, managing a comprehensive episodic memory to enrich those relationships, adopting a different intentional stance appropriate for each household member and essentially behaving as another member of the family?
  2. What if you had all of the C++ — or Java or Python — code currently checked into GitHub including all the versions and all the diffs plus sample I/O, unit tests and documentation. How would you go about developing a neural network architecture that learns to write programs from sample I/O and natural language descriptions?
  3. Suppose you had the complete wiring diagram (connectome) of a fly and petabytes of recordings from each neuron in the fly's brain aligned with high-speed images recording every aspect of the fly's behavior and the environment in which those behaviors were carried out. How would you construct a model of the fly's brain?

  • Hypothesis: Each of these problems can be solved using a recurrent neural network architecture constructed from published component networks each of which is relatively well understood and has been applied successfully to solving simpler problems.
  • Participation: In class, we examine this hypthesis by designing networks for key parts of each problem, borrowing ideas from both systems and cognitive neuroscience. Students propose and complete related programming projects for their final grade.



Location and Time:


TTh 4:30 - 5:50pm in the Hewlett building, room 101




Staff:


Instructor: Thomas Dean

Email: tld [at] google [dot] com

Office hours: by appointment

 

Course Assistant: Daniel Fernandes

Email: dfern [at] stanford [dot] edu




Textbooks:


There are no required textbooks for this course but you are expected to do a lot of reading on your own and these three texts are good to have around for reference. I’ve yet to meet anyone who has read all three cover to cover but over the years I’ve probably read most of the chapters in one edition or the other and often found them relevant.

  • - Neuroscience: Exploring the Brain (Third Edition), Bear, Connors and Paradiso.

  • - The Cognitive Neurosciences (Third Edition), Gazzaniga.

  • - Principles of Neural Science (Fourth Edition), Kandel, Schwartz and Jessell.



Grading:

- Class participation including presentation (30%)

- Project proposal due around midterm (20%)

- Project report due around finals week (50%)