学术前沿讲座2022年第12期:Differentially private algorithms in reinforcement learning

268 views

报告摘要

Reinforcement learning could involve sensitive information in states, rewards, and transitions. In this talk, we discuss how differentially private algorithms could protect the information from being inferred by an attacker. The talk will focus on continuous RL settings and provide analyses on privacy and utility, while several recent works follow.

主讲人信息

王趵翔

香港中文大学(深圳)数据科学学院助理教授
Baoxiang Wang is an assistant professor at The Chinese University of Hong Kong, Shenzhen. He works on reinforcement learning and RL theory. He solved the gambler's problem in 2020 which was open since 1989. He obtained his Ph.D. in Computer Science at The Chinese University of Hong Kong in 2020, under Siu On Chan and Andrej Bogdanov. Before that, he obtained his B.E. in Information Security at Shanghai Jiao Tong University. Homepage at bxiangwang@github.io.