International Joint Conference On Theoretical Computer Science – Frontier of Algorithmic Wisdom

August 15-19, 2022, City University of Hong Kong, Hong Kong

 

Invited Speakers

Track C


A brief introduction to convergence analysis of reinforcement learning algorithms

Xingguo Chen

Nanjing University of Posts and Telecommunications

Abstract:
The success of reinforcement learning is inseparable from the convergence analysis of various algorithms over the past 40 years. This report will briefly introduce the convergence analysis of reinforcement learning algorithms based on tabular value functions, linear value functions and nonlinear value functions. We will analyze the target solutions of these algorithms from a fixed point perspective, introduce the main techniques based on the proofs of the contraction mappings and positive definite matrix, and look forward to future improvement directions.

Bio:
Xingguo Chen received his Ph.D. degree from Nanjing University in Dec. 2013, and has been teaching at Nanjing University of Posts and Telecommunications since Jan. 2014. His research interests mainly include Reinforcement Learning and Game AI. He has published more than 40 academic papers in IEEE TNNLS, TCIAIG, AAAI, GECCO, JPDC, WR, EP, etc.