Attention-Guided Contrastive Role Representations for Multi-Agent Reinforcement Learning

TL;DR

This paper introduces ACORM, a novel attention-guided contrastive learning framework for multi-agent reinforcement learning that enhances role representation, promotes behavior diversity, and improves coordination, achieving state-of-the-art results in complex tasks.

Contribution

The paper proposes a new contrastive role representation learning method using attention mechanisms to improve multi-agent cooperation and performance.

Findings

Achieves state-of-the-art results on StarCraft II and Google research football tasks.

Demonstrates improved behavior heterogeneity and coordination among agents.

Outperforms existing methods in multi-agent reinforcement learning benchmarks.

Abstract

Real-world multi-agent tasks usually involve dynamic team composition with the emergence of roles, which should also be a key to efficient cooperation in multi-agent reinforcement learning (MARL). Drawing inspiration from the correlation between roles and agent's behavior patterns, we propose a novel framework of **A**ttention-guided **CO**ntrastive **R**ole representation learning for **M**ARL (**ACORM**) to promote behavior heterogeneity, knowledge transfer, and skillful coordination across agents. First, we introduce mutual information maximization to formalize role representation learning, derive a contrastive learning objective, and concisely approximate the distribution of negative pairs. Second, we leverage an attention mechanism to prompt the global state to attend to learned role representations in value decomposition, implicitly guiding agent coordination in a skillful role space to yield more expressive credit assignment. Experiments on challenging StarCraft II micromanagement and Google research football tasks demonstrate the state-of-the-art performance of our method and its advantages over existing approaches. Our code is available at [https://github.com/NJU-RL/ACORM](https://github.com/NJU-RL/ACORM).