New edition of Christopher Amato's intro book on cooperative multi-agent RL now available

dhadfieldmenell · x · 2026-09-17

Christopher Amato released a new version of his introductory book on cooperative multi-agent reinforcement learning (arXiv:2405.06161, v6), with a longer edition to be co-published with Frans Oliehoek. The text covers the setting where all agents share a joint reward and explains the three main paradigms:

Topics include applying single-agent methods to CTE, scalable multi-agent methods, independent Q-learning, and policy gradient approaches.

Original post →

More from Research

Research channel →