LoSemB: Logic-Guided Semantic Bridging for Inductive Tool Retrieval

2025-08-11

AI Summary

Large language models (LLMs) struggle to generalize to unseen tools under dynamic tool library expansion—existing transductive approaches assume all tools are observed during training, leading to severe distribution shift and brittle similarity-based retrieval. Method: We propose the Logic-guided Semantic Bridging (LSB) framework, the first fine-tuning-free inductive tool retrieval method. LSB mitigates distribution shift via a logic embedding alignment module that transfers latent reasoning structures across tool distributions; it enhances retrieval robustness through a relation-augmented mechanism that reduces reliance on superficial lexical or syntactic similarity; and it jointly integrates semantic representation learning with graph-structured modeling of tool dependencies. Results: LSB achieves significant improvements over state-of-the-art methods across multiple benchmarks, demonstrating strong robustness and efficiency in both inductive and transductive settings.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Knowledge Representation and Reasoning: Logic Programming

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Tool learning has emerged as a promising paradigm for large language models (LLMs) to solve many real-world tasks. Nonetheless, with the tool repository rapidly expanding, it is impractical to contain all tools within the limited input length of LLMs. To alleviate these issues, researchers have explored incorporating a tool retrieval module to select the most relevant tools or represent tools as unique tokens within LLM parameters. However, most state-of-the-art methods are under transductive settings, assuming all tools have been observed during training. Such a setting deviates from reality as the real-world tool repository is evolving and incorporates new tools frequently. When dealing with these unseen tools, which refer to tools not encountered during the training phase, these methods are limited by two key issues, including the large distribution shift and the vulnerability of similarity-based retrieval. To this end, inspired by human cognitive processes of mastering unseen tools through discovering and applying the logical information from prior experience, we introduce a novel Logic-Guided Semantic Bridging framework for inductive tool retrieval, namely, LoSemB, which aims to mine and transfer latent logical information for inductive tool retrieval without costly retraining. Specifically, LoSemB contains a logic-based embedding alignment module to mitigate distribution shifts and implements a relational augmented retrieval mechanism to reduce the vulnerability of similarity-based retrieval. Extensive experiments demonstrate that LoSemB achieves advanced performance in inductive settings while maintaining desirable effectiveness in the transductive setting.

Problem

• Addresses inductive tool retrieval for unseen tools
• Mitigates distribution shifts in tool retrieval
• Enhances similarity-based retrieval vulnerability

Innovation

• Logic-Guided Semantic Bridging for inductive retrieval
• Logic-based embedding alignment reduces distribution shifts
• Relational augmented retrieval enhances similarity-based methods