Deviate or Not: Learning Coalition Structures with Multiple-bit Observations in Games

TL;DR

This paper introduces a method to efficiently learn hidden coalition structures in multi-agent systems by designing specific games and analyzing agents' deviations, achieving near-optimal learning rounds.

Contribution

It presents the first logarithmic-round algorithm for coalition structure learning in multi-agent systems, matching the theoretical lower bound, and extends it to practical game formats.

Findings

Learning can be achieved in O(log n) rounds with unrestricted game design.

Algorithms are developed for specific game formats with near-optimal complexity.

Theoretical bounds are established for the number of rounds needed to learn coalition structures.

Abstract

We consider the Coalition Structure Learning (CSL) problem in multi-agent systems, motivated by the existence of coalitions in many real-world systems, e.g., trading platforms and auction systems. In this problem, there is a hidden coalition structure within a set of $n$ agents, which affects the behavior of the agents in games. Our goal is to actively design a sequence of games for the agents to play, such that observations in these games can be used to learn the hidden coalition structure. In particular, we consider the setting where in each round, we design and present a game together with a strategy profile to the agents, and receive a multiple-bit observation -- for each agent, we observe whether or not they would like to deviate from the specified strategy. We show that we can learn the coalition structure in $O(\log n)$ rounds if we are allowed to design any normal-form game, matching the information-theoretical lower bound. For practicality, we extend the result to settings where we can only choose games of a specific format, and design algorithms to learn the coalition structure in these settings. For most settings, our complexity matches the theoretical lower bound up to a constant factor.