机器学习与数据科学博士生系列论坛（第八十四期）—— Beyond Value Functions, Optimizing General Functionals of MDP with Distributional RL
报告人： 彭洋（北京大学）
时间：2025-03-06 16:00-17:00
地点：腾讯会议 531-8098-3912
Abstract:
Traditional reinforcement learning (RL) focuses on maximizing the expectation of the return (i.e. value function) in Markov decision processes. However, in some fields such as healthcare, finance, and robotics, a poor return may lead to catastrophic outcomes. Therefore, simply considering the expectation of the return is insufficient. We need to incorporate risk into the decision-making process, which motivates the problem of optimizing general functionals of the return, such as the Conditional Value at Risk (CVaR). Distributional RL provides a natural framework for solving this problem by modeling the full distribution of the return.
In this talk, we will introduce the recent work in this field. We will discuss what functionals can be optimized and present the distributional dynamic programming algorithms with theoretical guarantees. Extensive experiments demonstrate that the algorithms can achieve risk-aware decision-making and outperform existing methods.
About the Speaker:
论坛每次邀请一位博士生就某个前沿课题做较为系统深入的介绍，主题包括但不限于机器学习、高维统计学、运筹优化和理论计算机科学。
Your participation is warmly welcomed!
欢迎扫码关注北大统计科学中心公众号，了解更多讲座信息!
