Loading Events

« All Events

  • This event has passed.

Colloquium Talk: Towards Knowledge Transfer and Generalization in Deep Reinforcement Learning

August 8 @ 10:30 am - 11:30 am

Title: Towards Knowledge Transfer and Generalization in Deep Reinforcement LearningSpeaker: Chongjie Zhang, Tsinghua UniversityLocation: Northeastern University, 805 Columbus Avenue, Interdisciplinary Science & Engineering Complex, 6th Floor, Room #655, Boston, Massachusetts 02118AbstractDeep reinforcement learning has recently shown considerable successes of achieving human-level control in some games. However, learning each task independently and from scratch requires vast experiences and is not tractable for most real-world problems. To address this challenge, in this talk, I will first present model-free policy reuse approaches that transfer past source policies to new tasks with theoretical optimality guarantees. I will then discuss an object-oriented dynamics learning approach for enabling generalizable model-based deep reinforcement learning. This approach enables a dynamics model learned from few environments to generalize over novel environments while providing semantically and visually interpretable representations.CCIS Faculty Host: Chris Amato