Abstract

Thompson sampling (TS) and itsvariants are popular algorithms for reinforcement learning and multi-armedbandits. In this tutorial, we will briefly review the basic concepts ofreinforcement learning, bandits, and TS. We will also discuss several practicalconsiderations when applying TS to real-world problems, as well as thehigh-level insights on how to analyze TS. Preliminary experiment results willalso be discussed.

Time

12月25日（周二）14:00-15:00

Speaker

Zheng Wen (温晸) iscurrently a senior research scientist at Adobe Research, his current researchfocuses on reinforcement learning, multi-armed bandit, and dynamic programming.Before joining Adobe Research, he worked as a research scientist at Yahoo!Labs. Prior to that, he received a PhD in Electrical Engineering from StanfordUniversity.

Venue

信息管理与工程学院102室

上海财经大学（第三教学楼西侧）

上海市杨浦区武东路100号