RIMRULE: Improving Tool-Using Language Agents via MDL-Guided Rule Learning

2025-12-31

arXiv.org

AI Summary

This work addresses the limited tool-use capabilities of large language models (LLMs) when interacting with non-standard, poorly documented, or private APIs. To overcome this challenge, the authors propose a dynamic injection mechanism that automatically distills compact, interpretable bimodal rules—combining natural language and symbolic representations—from failed execution trajectories, guided by the Minimum Description Length (MDL) principle. These rules are efficiently retrieved and applied during inference without modifying the model’s weights. The approach significantly improves accuracy across multiple tool-use benchmarks, outperforming existing prompting strategies. Moreover, the distilled rules exhibit cross-architecture transferability, enabling reuse across different LLMs, and complement fine-tuning methods by offering a lightweight, modular enhancement to tool-augmented reasoning.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Large language models (LLMs) often struggle to use tools reliably in domain-specific settings, where APIs may be idiosyncratic, under-documented, or tailored to private workflows. This highlights the need for effective adaptation to task-specific tools. We propose RIMRULE, a neuro-symbolic approach for LLM adaptation based on dynamic rule injection. Compact, interpretable rules are distilled from failure traces and injected into the prompt during inference to improve task performance. These rules are proposed by the LLM itself and consolidated using a Minimum Description Length (MDL) objective that favors generality and conciseness. Each rule is stored in both natural language and a structured symbolic form, supporting efficient retrieval at inference time. Experiments on tool-use benchmarks show that this approach improves accuracy on both seen and unseen tools without modifying LLM weights. It outperforms prompting-based adaptation methods and complements finetuning. Moreover, rules learned from one LLM can be reused to improve others, including long reasoning LLMs, highlighting the portability of symbolic knowledge across architectures.

Problem

• tool use
• large language models
• domain-specific adaptation
• API idiosyncrasy
• reliable tool integration

Innovation

• neuro-symbolic
• dynamic rule injection
• Minimum Description Length (MDL)
• tool-use adaptation
• symbolic knowledge portability