Tutorial1

Sample Efficient Reinforcement Learningļ¼šfrom AlphaGo to EfficientZero

9:00 - 10:00 & 10:15 - 12:15, Dec. 15, 2022, UTC+8.

Yang Gao

I am an Assistant Professor at Institute for Interdisciplinary Information Science (IIIS) at Tsinghua University. I am interested in the intersection between computer vision and robotics. Specifically, I want to explore how to utilize the prior knowledge we have from computer vision to do robot manipulation tasks both more efficiently and effectively. This not only involves understanding how to use the previous visual experiences but also potentially needs re-designing robotic learning algorithms to better handles the visual states. Thus, it is a co-design problem between vision and robotics. As testing benchmarks, I work on robot manipulation and autonomous driving applications, both in simulation and in the real world. Previously, I obtained my Ph.D. degree from UC Berkeley, advised by Prof. Trevor Darrell and B.E. from the Computer Science Department at Tsinghua University.