Yifan Zhu
About MeI am a first year PhD student in Electrical Engineering at Stanford University, advised by Professor Benjamin Van Roy. I received the BS in Mathematics, and MS in Electrical Engineering, all from Stanford University. I have been award the J. E. Wallace Sterling Award for Scholastic Achievement, and elected to Phi Beta Kappa. Research InterestCurrent reinforcement learning algorithms have achieved incredible success in computer games, showcasing superhuman performance. However, this achievement often comes at the cost of requiring an exceptionally large amount of data. For example, training the state-of-the-art A3C agent on Atari games requires hundreds of millions of frames. This inefficiency is a major obstacle for applying reinforcement learning to real-world problems, where data is often scarce. I am interested in developing methods to improve the sample efficiency of reinforcement learning algorithms, by learning from more than just the reward signal, and by efficient exploration. Other InterestsI make occasional contributions to the open source software I use (e.g., KDE, and Firefox Bug 148624). |