MuMA-ToM: Multi-modal Multi-Agent Theory of Mind

TL;DR

MuMA-ToM introduces a pioneering multi-modal benchmark for multi-agent Theory of Mind reasoning, enabling AI to better understand complex social interactions through multi-modal data and advanced modeling.

Contribution

The paper presents MuMA-ToM, the first multi-modal ToM benchmark for embodied multi-agent interactions, and proposes LIMP, a novel multi-modal ToM model that outperforms existing methods.

Findings

LIMP significantly outperforms state-of-the-art models.

MuMA-ToM is validated with human experiments and baseline comparisons.

The benchmark enables evaluation of mental reasoning in realistic multi-modal scenarios.

Abstract

Understanding people's social interactions in complex real-world scenarios often relies on intricate mental reasoning. To truly understand how and why people interact with one another, we must infer the underlying mental states that give rise to the social interactions, i.e., Theory of Mind reasoning in multi-agent interactions. Additionally, social interactions are often multi-modal -- we can watch people's actions, hear their conversations, and/or read about their past behaviors. For AI systems to successfully and safely interact with people in real-world environments, they also need to understand people's mental states as well as their inferences about each other's mental states based on multi-modal information about their interactions. For this, we introduce MuMA-ToM, a Multi-modal Multi-Agent Theory of Mind benchmark. MuMA-ToM is the first multi-modal Theory of Mind benchmark that evaluates mental reasoning in embodied multi-agent interactions. In MuMA-ToM, we provide video and text descriptions of people's multi-modal behavior in realistic household environments. Based on the context, we then ask questions about people's goals, beliefs, and beliefs about others' goals. We validated MuMA-ToM in a human experiment and provided a human baseline. We also proposed a novel multi-modal, multi-agent ToM model, LIMP (Language model-based Inverse Multi-agent Planning). Our experimental results show that LIMP significantly outperforms state-of-the-art methods, including large multi-modal models (e.g., GPT-4o, Gemini-1.5 Pro) and a recent multi-modal ToM model, BIP-ALM.