Speaker: Jessica Hamrick, University of California, Berkeley
Title: Metareasoning and mental simulation
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. Moreover, I will discuss how the mathematical models of human cognition that I develop in my research can be incorporated into AI in order to build agents that are better able to reason about and communicate with human collaborators.
Jessica Hamrick is a Ph.D. candidate in the Psychology department 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, and did research for a summer at Google DeepMind. Jessica is a recipient of the NSF Graduate Fellowship, the Berkeley Fellowship, and the Outstanding Graduate Student Instructor Award. In addition to research, Jessica is a core contributor to the IPython/Jupyter notebook, and is a member of the Project Jupyter Steering Council.
For additional information contact Steve Easterbrook