AutoToM: Scaling Model-based Mental Inference via Automated Agent Modeling

TL;DR

AutoToM introduces an automated, scalable, and robust method for Theory of Mind reasoning that leverages Bayesian inverse planning with LLMs, outperforming existing approaches across multiple benchmarks.

Contribution

It presents AutoToM, a novel automated agent modeling technique that improves mental inference by iterative refinement guided by uncertainty, combining robustness with interpretability.

Findings

Outperforms existing ToM methods and large reasoning models on five benchmarks.

Produces human-like confidence estimates for mental inference.

Enables online mental inference for embodied decision-making.

Abstract

Theory of Mind (ToM), the ability to understand people's minds based on their behavior, is key to developing socially intelligent agents. Current approaches to ToM reasoning either rely on prompting Large Language Models (LLMs), which are prone to systematic errors, or use handcrafted, rigid agent models for model-based inference, which are more robust but fail to generalize across domains. In this work, we introduce AutoToM, an automated agent modeling method for scalable, robust, and interpretable mental inference. Given a ToM problem, AutoToM first proposes an initial agent model and then performs automated Bayesian inverse planning based on this model, leveraging an LLM backend. Guided by inference uncertainty, it iteratively refines the model by introducing additional mental variables and/or incorporating more timesteps in the context. Across five diverse benchmarks, AutoToM outperforms existing ToM methods and even large reasoning models. Additionally, we show that AutoToM can produce human-like confidence estimates and enable online mental inference for embodied decision-making.