Learning Dynamic Belief Graphs for Theory-of-mind Reasoning

2026-03-20

AI Summary

This work addresses the challenge that large language models struggle to accurately model humans’ dynamic and implicit beliefs under high uncertainty, hindering precise inference of their information-seeking and decision-making behaviors. To overcome this, the authors propose a Dynamic Belief Graph model that represents mental states as evolving probabilistic graphical structures over time, jointly inferring latent beliefs, their time-varying dependencies, and their connections to actions and information-seeking processes. Key innovations include a projection mechanism translating textual probabilistic statements into consistent graph updates, an energy-based factor graph representation of beliefs, and an ELBO objective integrating belief accumulation with delayed decision-making. Experiments demonstrate that the method significantly improves action prediction accuracy on multiple real-world disaster evacuation datasets and recovers interpretable belief trajectories aligned with human cognitive patterns.

Technology Category

• Knowledge Representation and Reasoning: Reasoning with Beliefs
• Reasoning under Uncertainty: Graphical Models
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Theory of Mind (ToM) reasoning with Large Language Models (LLMs) requires inferring how people's implicit, evolving beliefs shape what they seek and how they act under uncertainty -- especially in high-stakes settings such as disaster response, emergency medicine, and human-in-the-loop autonomy. Prior approaches either prompt LLMs directly or use latent-state models that treat beliefs as static and independent, often producing incoherent mental models over time and weak reasoning in dynamic contexts. We introduce a structured cognitive trajectory model for LLM-based ToM that represents mental state as a dynamic belief graph, jointly inferring latent beliefs, learning their time-varying dependencies, and linking belief evolution to information seeking and decisions. Our model contributes (i) a novel projection from textualized probabilistic statements to consistent probabilistic graphical model updates, (ii) an energy-based factor graph representation of belief interdependencies, and (iii) an ELBO-based objective that captures belief accumulation and delayed decisions. Across multiple real-world disaster evacuation datasets, our model significantly improves action prediction and recovers interpretable belief trajectories consistent with human reasoning, providing a principled module for augmenting LLMs with ToM in high-uncertainty environment. https://anonymous.4open.science/r/ICML_submission-6373/

Problem

• Theory of Mind
• Dynamic Belief
• Large Language Models
• Uncertainty Reasoning
• Mental State Inference

Innovation

• Dynamic Belief Graph
• Theory-of-Mind Reasoning
• Probabilistic Graphical Model
• Energy-Based Factor Graph
• ELBO Optimization