Humans readily make nuanced and abstract inferences about the unobserved contents of others’ minds, including their beliefs, values, and emotions. Professor Saxe’ s Lab has begun to develop probabilistic models of people’ s intuitive reasoning about others’ emotions. The goal of my project is to extend these Bayesian models to capture observers’ cognition about others’ emotions in a range of more complex social games and ethical dilemmas. This will involve designing behavioral experiments, collecting human data, and building generative models of observers’ emotion attributions. This investigation of people’ s intuitive theory of emotion aims to advance both our understanding of human social cognition and our ability to recapitulate these functions in artificially intelligent systems.

I am participating in SuperUROP because I’ m interested in understanding more about our social cognition and applying the skills I’ ve acquired through my classes and past research projects to model such processes and potentially implement them in artificial agents.