Learning Meta Representations for Agents in Multi-Agent Reinforcement Learning

TL;DR

This paper introduces Meta Representations for Agents (MRA), enabling multi-agent reinforcement learning agents to generalize across different population sizes by learning a set of strategies that capture both common and game-specific knowledge.

Contribution

The paper proposes a novel method, MRA, that models multi-modal policy sets with explicit game-common and game-specific knowledge, improving generalization in multi-agent RL.

Findings

MRA reaches Nash Equilibrium in various Markov Games.

MRA improves training performance and generalization.

Fast adaptation is possible with limited latent space.

Abstract

In multi-agent reinforcement learning, the behaviors that agents learn in a single Markov Game (MG) are typically confined to the given agent number. Every single MG induced by varying the population may possess distinct optimal joint strategies and game-specific knowledge, which are modeled independently in modern multi-agent reinforcement learning algorithms. In this work, our focus is on creating agents that can generalize across population-varying MGs. Instead of learning a unimodal policy, each agent learns a policy set comprising effective strategies across a variety of games. To achieve this, we propose Meta Representations for Agents (MRA) that explicitly models the game-common and game-specific strategic knowledge. By representing the policy sets with multi-modal latent policies, the game-common strategic knowledge and diverse strategic modes are discovered through an iterative optimization procedure. We prove that by approximately maximizing the resulting constrained mutual information objective, the policies can reach Nash Equilibrium in every evaluation MG when the latent space is sufficiently large. When deploying MRA in practical settings with limited latent space sizes, fast adaptation can be achieved by leveraging the first-order gradient information. Extensive experiments demonstrate the effectiveness of MRA in improving training performance and generalization ability in challenging evaluation games.