Neuro-Symbolic Synergy for Interactive World Modeling

2026-02-11

AI Summary

This work addresses the limitations of large language models (LLMs) as world models—namely, their tendency to generate hallucinations and difficulty adhering strictly to deterministic rules—while also overcoming the restricted semantic expressivity of purely symbolic systems. To bridge this gap, the authors propose NeSyS, a novel neuro-symbolic coordination framework that directly embeds symbolic rules into the output distribution of an LLM to constrain its behavior, complemented by targeted fine-tuning on trajectories not covered by the rules. Through an alternating training mechanism, NeSyS effectively integrates the probabilistic semantic priors of LLMs with executable symbolic constraints, achieving a unified balance between expressive semantics and logical robustness. Evaluated on ScienceWorld, WebShop, and Plancraft, NeSyS matches or exceeds baseline performance in prediction accuracy while reducing training data requirements by 50%.

Technology Category

• Machine Learning: Neuro-Symbolic Learning
• Cognitive Modeling & Cognitive Systems: Symbolic Representations
• Natural Language Processing: (Large) Language Models

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

Large language models (LLMs) exhibit strong general-purpose reasoning capabilities, yet they frequently hallucinate when used as world models (WMs), where strict compliance with deterministic transition rules--particularly in corner cases--is essential. In contrast, Symbolic WMs provide logical consistency but lack semantic expressivity. To bridge this gap, we propose Neuro-Symbolic Synergy (NeSyS), a framework that integrates the probabilistic semantic priors of LLMs with executable symbolic rules to achieve both expressivity and robustness. NeSyS alternates training between the two models using trajectories inadequately explained by the other. Unlike rule-based prompting, the symbolic WM directly constrains the LLM by modifying its output probability distribution. The neural WM is fine-tuned only on trajectories not covered by symbolic rules, reducing training data by 50% without loss of accuracy. Extensive experiments on three distinct interactive environments, i.e., ScienceWorld, Webshop, and Plancraft, demonstrate NeSyS's consistent advantages over baselines in both WM prediction accuracy and data efficiency.

Problem

• world modeling
• hallucination
• symbolic reasoning
• large language models
• deterministic transitions

Innovation

• Neuro-Symbolic Integration
• World Modeling
• Symbolic Constraints
• Data Efficiency
• Interactive Environments