Abstract: At any given moment, how should an intelligent agent decide what to think about, how to think about it, and how long to think for? My research attempts to answer these questions by focusing on the best examples of intelligent agents that we have: humans. In particular, I study how people use their "mental simulations", which can be thought of as samples from a rich generative model of the world. I show how people adaptively use their mental simulations to learn new things about the world; that they choose which simulations to run based on which they think will be more informative; and that they allocate their cognitive resources to spend less time on easy problems and more time on hard problems. Based on these results, I will illustrate how machine learning and cognitive science can complement one another by showing how ideas from cognitive science can inform and inspire new approaches to building artificially intelligent agents.
Bio: Jessica Hamrick is a Research Scientist at DeepMind in London, having recently completed her Ph.D. in Psychology at the University of California, Berkeley working with Tom Griffiths. Previously, she received her M.Eng. in Computer Science from MIT working with Josh Tenenbaum. Jessica's research focuses on model-based reasoning and planning, situated at the intersection of cognitive science, machine learning, and AI. In addition to research, Jessica is involved in several open source projects including Project Jupyter, and is the lead maintainer of nbgrader, a tool for grading Jupyter notebook assignments.
This talk is part of the Informatics: Institute for Language, Cognition and Computation/HCRC Seminar Series