Before joining the faculty, I was a postdoctoral associate at MIT working with Julie Shah and the Interactive Robotics group in Computer Science and Artificial Intelligence Lab (CSAIL) on studying robotics with applications on manufacturing. I received my Ph.D. in Computer Science under the supervision of Victor Lesser from University of Massachusetts at Amherst in 2011.
To develop a general-purpose learning framework, I am interested in studying topics including, but not limited to, representation learning, world-model learning and reasoning, policy learning and planning, multi-agent learning.
Specifically, my goal is to build autonomous agents equipped with the capability of extracting information from perceptions, learning and reasoning world models that are generalized over tasks, predicting consequences of actions, envisioning and setting achievable goals, planning actions to achieve these goals, collaborating with other agents or humans to perform complex tasks, as well as generalizing learned knowledge and sharing learning experiences with other agents.
To approach my goal, my current research is building novel efficient models and methods of deep learning, reinforcement learning, and multi-agent systems, with applications in robotics and video games.