I’m proud to announce that I'll be joining the MIT Media Lab this fall as a postdoc, working with Cynthia Breazeal’s Personal Robots group. My research will focus on a project called Socially Assistive Robots.
I research how robots and other agents can learn interactively from human teachers. Much of my work on this topic concerns how to design agents that learn from human-delivered reward, a problem we call interactive shaping. I also examine how other forms of learning can be improved by a model of a trainer’s reward. More broadly, I'm interested in reinforcement learning and other agent learning, human behavior, human-robot interaction, and general machine learning.
I am currently organizing the AAAI 2012 Fall Symposium on Robots Learning Interactively from Human Teachers (RLIHT), and I recently co-chaired the IJCAI 2011 Workshop on Agents Learning Interactively from Human Teachers (ALIHT). Since the spring of 2005, when I started taking undergraduate math and computer science classes as a psychology-degree-holding “undergrad” at UT-Austin, my research has been advised by Peter Stone within the LARG research group. Additionally, I led a reading group at my university called Agents that Learn from Humans.
Listed in my publications, the K-CAP 2009 paper on the TAMER framework describes our progress towards a general system for interactive shaping, and the AAMAS 2010 paper — winner of the Best Student Paper Award — investigates how to learn both from human reward and MDP reward signals.
