MetaMind: General and Cognitive World Models in Multi-Agent Systems by Meta-Theory of Mind
This addresses the problem of enabling decentralized coordination in multi-agent systems for applications like robotics or autonomous systems, though it appears incremental as it builds on existing theory of mind concepts.
The paper tackles the challenge of modeling interdependent agent dynamics and predicting interactive trajectories in multi-agent systems without centralized supervision, proposing MetaMind, a world model that uses meta-theory of mind to enable agents to reason about others' goals and beliefs from limited observations, achieving superior task performance and outperforming baselines in few-shot generalization.
A major challenge for world models in multi-agent systems is to understand interdependent agent dynamics, predict interactive multi-agent trajectories, and plan over long horizons with collective awareness, without centralized supervision or explicit communication. In this paper, MetaMind, a general and cognitive world model for multi-agent systems that leverages a novel meta-theory of mind (Meta-ToM) framework, is proposed. Through MetaMind, each agent learns not only to predict and plan over its own beliefs, but also to inversely reason goals and beliefs from its own behavior trajectories. This self-reflective, bidirectional inference loop enables each agent to learn a metacognitive ability in a self-supervised manner. Then, MetaMind is shown to generalize the metacognitive ability from first-person to third-person through analogical reasoning. Thus, in multi-agent systems, each agent with MetaMind can actively reason about goals and beliefs of other agents from limited, observable behavior trajectories in a zero-shot manner, and then adapt to emergent collective intention without an explicit communication mechanism. Extended simulation results on diverse multi-agent tasks demonstrate that MetaMind can achieve superior task performance and outperform baselines in few-shot multi-agent generalization.