Revisiting Multi-Agent World Modeling from a Diffusion-Inspired Perspective

TL;DR

This paper introduces DIMA, a diffusion-inspired world model for multi-agent reinforcement learning that simplifies environment modeling and achieves state-of-the-art results on multiple benchmarks.

Contribution

It proposes a novel diffusion-inspired approach to multi-agent world modeling, improving accuracy and efficiency over prior methods.

Findings

DIMA outperforms previous models in final return and sample efficiency.

The diffusion-inspired approach effectively captures agent dependencies.

State-of-the-art results on MAMuJoCo and Bi-DexHands benchmarks.

Abstract

World models have recently attracted growing interest in Multi-Agent Reinforcement Learning (MARL) due to their ability to improve sample efficiency for policy learning. However, accurately modeling environments in MARL is challenging due to the exponentially large joint action space and highly uncertain dynamics inherent in multi-agent systems. To address this, we reduce modeling complexity by shifting from jointly modeling the entire state-action transition dynamics to focusing on the state space alone at each timestep through sequential agent modeling. Specifically, our approach enables the model to progressively resolve uncertainty while capturing the structured dependencies among agents, providing a more accurate representation of how agents influence the state. Interestingly, this sequential revelation of agents' actions in a multi-agent system aligns with the reverse process in diffusion models--a class of powerful generative models known for their expressiveness and training stability compared to autoregressive or latent variable models. Leveraging this insight, we develop a flexible and robust world model for MARL using diffusion models. Our method, Diffusion-Inspired Multi-Agent world model (DIMA), achieves state-of-the-art performance across multiple multi-agent control benchmarks, significantly outperforming prior world models in terms of final return and sample efficiency, including MAMuJoCo and Bi-DexHands. DIMA establishes a new paradigm for constructing multi-agent world models, advancing the frontier of MARL research. Codes are open-sourced at https://github.com/breez3young/DIMA.