Coordinated Diffusion: Generating Multi-Agent Behavior Without Multi-Agent Demonstrations

TL;DR

CoDi introduces a diffusion-based framework that learns multi-agent coordination from single-agent demonstrations and a user-defined cost function, avoiding the need for multi-agent demonstration data.

Contribution

This work presents a novel diffusion-based method to generate multi-agent behaviors using only single-agent data and a cost function, bypassing the data bottleneck.

Findings

CoDi effectively learns coordinated multi-agent behaviors from single-agent data.

The method is more data-efficient than existing multi-agent baselines.

CoDi performs well in simulation and hardware experiments, demonstrating robustness and coordination.

Abstract

Imitation learning powered by generative models has proven effective for modeling complex single-agent behaviors. However, teaching multi-agent systems, like multiple arms or vehicles, to coordinate through imitation learning is hindered by a fundamental data bottleneck: as the joint state-action space grows exponentially with the number of agents, collecting a sufficient amount of coordinated multi-agent demonstrations becomes extremely costly. In this work, we ask: how can we leverage single-agent demonstration data to learn multi-agent policies? We present Coordinated Diffusion (CoDi), a framework that couples independently trained single-agent diffusion policies through a user-defined multi-agent cost function, without requiring any coordinated demonstrations. We derive a new diffusion-based sampling scheme wherein the diffusion score function decomposes into independent, single-agent pre-trained base policies plus a cost-driven guidance term that coordinates these base policies into cohesive multi-agent behavior. We show that this guidance term can be estimated in a gradient-free manner, making CoDi applicable to black-box, non-differentiable cost functions without additional training. Theoretically and empirically, we analyze the conditions under which this composition can faithfully approximate a target multi-agent behavior. We find a complementary role for demonstration data versus the cost function: single-agent demonstrations must cover the support of the desired multi-agent behavior, while the cost function must promote desired behavior from this product of single-agent policies. Our results in simulation and hardware experiments of a two-arm manipulation task show that CoDi discovers robust coordinated behavior from single-agent data, is more data-efficient than multi-agent baselines, and highlights the importance of joint guidance, base policy support, and cost design.