This tutorial provides a comprehensive overview of the field of Large Language Model based Multi-Agent Systems (LLM-MAS). As LLMs evolve into capable autonomous agents, the next frontier lies in orchestrating their collective intelligence to solve complex problems beyond the scope of any single agent. This tutorial bridges the gap between the foundational principles of multi-agent systems and the practical design of modern LLM-MAS.

This tutorial provides an overview of embodied intelligence, focusing on how robots can autonomously learn, plan, and interact with humans and their environment. We cover fundamental techniques in robot learning and planning, as well as approaches for enabling human-robot interaction in real-world scenarios. Through a combination of theoretical insights and practical examples, participants can gain a comprehensive understanding of how to build robotic frameworks that are capable of adaptive behavior and meaningful social connections.

Matching Markets is a cornerstone of economics and game theory. Bandit learning in matching markets address the stable matching equilibria under agents’ uncertain preferences with key metrics like regrets and incentive compatibility. This tutorial synthesizes recent advancements in these areas and points out valuable open problems and future directions.