Research Projects

I am broadly interested in how people understand the emotions and other mental states of those around them (Affective and Social Cognition). I study this by building computational cognitive models of reasoning. That is, I investigate how people intuitively reason about those around them, and try to codify such reasoning using computational models (usually, via probabilistic approaches). Computational cognitive modeling (i) allows researchers to specify and test precise, quantitative hypotheses about cognition and affect, and (ii) opens the doors to many applications, such as enabling computers to "reason" about emotions and mental states in a human-like manner.

In my work, I take an interdisciplinary approach, theoretically grounded in cognitive science and affective science, along with tools from computer science (probabilistic modeling and probabilistic programming; machine learning; natural language processing; social network analyses) and behavioral economics (economic games and analyses). I also apply modeling to study empathy, Theory of Mind, and prosocial behavior, among others.

In addition to cognitive science research, I also explore many applications of the modeling work that I do. I am interested in two main areas of applications: applications to technology (e.g., Affective Computing and affective robotics), and applications to clinical contexts (e.g., building models of affective and social dysfunction).

Understanding how humans reason about emotions and mental states lets us build computational models and artificial intelligence that can similarly reason about emotions and mental states (right arrow). Furthermore, having better computational models will allow us to ask more precise questions about the nature of human cognition (left arrow). This results in a virtuous cycle where scientific progress in psychology fuels progress in artificial intelligence which in turn fuels more progress in psychology.

Below, I've provided a non-exhaustive list of projects that I've been working on for a while. My more recent lines of research, which do not yet have their own little writeups, include:
  • multimodal time-series emotion recognition (e.g., the Stanford Emotional Narratives Dataset)

    (Ong, Wu, Zhi-Xuan, Reddan, Kahhale, Mattek, & Zaki, IEEE Transactions on Affective Computing, In Press; Wu, Zhang, Zhi-Xuan, Zaki, & Ong, ACII 2019).

  • probabilistic programming and deep probabilistic modelling

    (Ong, Soh, Zaki, & Goodman, IEEE Transactions on Affective Computing, 2021; Zhi-Xuan, Soh, & Ong, AAAI 2020)

  • better emotion recognition models, usually via inspiration from human reasoning, or methodological innovations, or both (e.g., adding knowledge, adding structure over emotion classes...)

    (Suresh & Ong, EMNLP 2021; Suresh & Ong, ACII 2021; Wu & Ong, AAAI 2021)

  • Ethical Affective Computing

    (Ong, ACII 2021)

  • interpretability of NLP models of emotion recognition

    (Nguyen, Wu, & Ong, Knowledge-Based Systems, 2021; Wu, Nguyen, & Ong, BlackboxNLP @ EMNLP, 2020).

  • reasoning about other mental states like intentions and social norms

    (grant from National Robotics Programme; Tan & Ong, CogSci 2019; Teo & Ong, CogSci 2021)

I also have done some work in education.

(Chen*, Ong*, Ng, & Coppola, AERA Open, 2021; Chen, Chavez, Ong, & Gunderson, Psych Science, 2017)

Affective Cognition

When we see someone miss a bus, receive a present, or walk with a skip in their step, we have no difficulty inferring their emotions, their thoughts, and even what they might do next. The ability to effortlessly perform such reasoning, Affective Cognition (reasoning about affect), is crucial to our everyday lives.

Underpinning affective cognition is a set of rich knowledge about emotions, encapsulated in laypeople's intuitive theories of emotion. That is, in people's minds, what are emotions? What causes emotions? What do emotions cause people to do?

I take a probabilistic approach by modeling people's intuitive theories as generative, causal models (i.e., using Bayesian Networks and other similar models). Under this framework, laypeople's intuitive reasoning about emotions can be described by (Bayesian) inference within these causal models. (See here for a more detailed writeup.)

I use a variety of different approaches to address these questions (and much more!) in a variety of different contexts:

  • I study the cues (facial expressions, prosody, as well as contextual cues) that people use when reasoning about others.
  • I study reasoning across many different contexts: reasoning about "near-misses", reasoning about others across psychological distance, reasoning about emotion-driven actions, etc.
  • I study how children acquire such rich knowledge about emotions.
  • I study how such reasoning may be impacted by affective disorders.
My projects on affective cognition are also closely related to the next set of projects on empathic accuracy; they are distinct research questions but my work in one often informs the other.

Helping me along this exciting journey of understanding affective cognition are a bunch of amazing collaborators, including: Noah Goodman, Jamil Zaki, Mika Asaba, Hyowon Gweon.

Representative Publications:
Ong, Zaki, & Goodman, Cognition, 2015
Ong, Zaki, & Goodman, Topics in Cognitive Science, 2019
Ong, Goodman, & Zaki, Emotion, 2018
Asaba*, Ong*, & Gweon, Developmental Psychology, 2019
Ong, Doctoral Dissertation, 2017

Empathic Accuracy

When people reason about the emotions of those around them, are they actually accurate? Or perhaps, are there systematic biases in the way that people interpret different types of information when inferring the emotions of those around them? What are the computational, psychological, and neural bases underlying accurate emotion judgments?

In this set of projects, we are interested in people's accuracy in evaluating the emotions of those around them. My collaborators and I investigate empathic accuracy using challenging, naturalistic contexts. To this end, I use tools from computer science to extract features from videos (e.g., acoustic features, linguistic features) and examine how people use these features in making their judgments. This involves acoustic processing, natural language processing, as well as time-series probabilistic modeling.

In addition to computational modeling, we also study empathic accuracy (i) with neuroscientific approaches (fMRI, EEG), (ii) across demographic groups and across cultures, and (iii) in populations with mood disorders.

This work is done with many awesome collaborators, including: Jamil Zaki, Erika Weisz, June Gruber, Anat Perry.

Representative Publications:
Devlin, Zaki, Ong, & Gruber, Cognitive Therapy and Research, 2016
Jospe, Genzer, Klein-Selle, Ong, Zaki, & Perry, Cortex 2020

Social Support in a real-life dorm (PNAS)

A social support network in a real-life dormitory. Nodes (circles) represent students, and edges (arrows between nodes) represent support seeking.
[Figure from Morelli et al, 2017]

Analyzing Social Support in an online network

An ego-centric network examining how social support spreads outwards from the focal node (in the center).

Empathy and Social Support in real-life and online social networks

As they say, "No man is an island". People naturally turn to others in their social network for companionship and social support, in both good times and bad. When important life events happen, who do people turn to, and when is the provided social support effective?

I use social network analyses to look at how people navigate their social networks, and how emotions, empathy, and social support propagate within both in-person and online social networks. I also combine network analysis with other approaches (like analyzing natural language text) to examine how social interactions contribute to social support within the network.

Main Collaborators: Sylvia Morelli, Matthew Jackson, Jamil Zaki.

Representative Publications:
Morelli, Ong, Makati, Jackson, & Zaki, PNAS, 2017

Prosocial Behavior via Empathy norms (PSPB)

Average donation behavior after participants were exposed to either an empathic or a non-empathic social norm.
[Figure from Nook et al, 2016]

Prosocial Behavior

The emotions that we feel towards those around us influence our behavior towards them, and I'm interested in (social, emotional, empathic) factors that motivate people to be nice to others.

Main Collaborators: Jamil Zaki, June Gruber, Xuan Zhao, Erik Nook

Representative Publications:
Ong, Zaki, & Gruber, ‎J. Abnorm. Psych., 2017
Nook, Ong, Morelli, Mitchell, & Zaki, PSPB, 2016
Ong, undergraduate honors thesis, 2011

Past projects in physics and optics (from 2012 and earlier)

In the past, prior to my starting graduate school, in roughly reverse chronological order, I have done research in behavioral economics, vision, condensed matter physics, optics, and material science. I've written up short summaries of some of my previous, completed projects in optics and physics, below.

Laser Speckle Characterization
Laser Speckle Characterization

I developed a method of analyzing anisotropy in speckled images, by analyzing the power spectral distribution in a polar-coordinate representation.
[with Sanjeev Solanki, Xinan Liang, Xuewu Xu]

Enhanced Brownian diffusion of colloidal dimers in an oscillatory shear flow
Brownian motion of collodial dimers in a shear flow.

We investigated how applying an oscillatory shear flow affects the Brownian motion of collodial dimer particles (based on the geometry of the dimer particle).
[with Brian Leahy, Xiang Cheng, Itai Cohen]

Vacancy and dislocation dynamics in colloidal dimer crystals
Vacancy and dislocation dynamics in colloidal dimer crystals

We examined how the structure of a collodial crystal (having symmetric dimers in a crystal of spheres) affects the dislocation dynamics of the crystal.
[with Sharon Gerbode, Itai Cohen, and others]