Enhancing the Patent Matching Capability of Large Language Models via the Memory Graph

2025-04-21

AI Summary

Current large language models (LLMs) for patent matching overly rely on lexical keyword matching, neglecting the hierarchical classification structure and ontological semantic relationships inherent in patents—resulting in poor generalizability and limited interpretability. To address this, we propose MemGraph: a framework that constructs a dynamic memory graph grounded in parametric memory of LLMs. Through prompt-guided entity extraction, ontology alignment, and semantic-enhanced reasoning, MemGraph enables hierarchy-aware and relation-driven patent matching. Our approach is model-agnostic—compatible with diverse open- and closed-source LLMs—without requiring fine-tuning. On the PatentMatch benchmark, MemGraph achieves a 17.68% relative improvement over strong baselines, demonstrating significantly enhanced cross-domain generalization and intrinsic reasoning interpretability. This work establishes a novel paradigm for intelligent intellectual property management.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Linked Open Data, Knowledge Graphs & KB Completion
• Natural Language Processing: Information Extraction

Application Category

• 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
• Search and Retrieval-Augmented AI: Large language models for search

Abstract

Intellectual Property (IP) management involves strategically protecting and utilizing intellectual assets to enhance organizational innovation, competitiveness, and value creation. Patent matching is a crucial task in intellectual property management, which facilitates the organization and utilization of patents. Existing models often rely on the emergent capabilities of Large Language Models (LLMs) and leverage them to identify related patents directly. However, these methods usually depend on matching keywords and overlook the hierarchical classification and categorical relationships of patents. In this paper, we propose MemGraph, a method that augments the patent matching capabilities of LLMs by incorporating a memory graph derived from their parametric memory. Specifically, MemGraph prompts LLMs to traverse their memory to identify relevant entities within patents, followed by attributing these entities to corresponding ontologies. After traversing the memory graph, we utilize extracted entities and ontologies to improve the capability of LLM in comprehending the semantics of patents. Experimental results on the PatentMatch dataset demonstrate the effectiveness of MemGraph, achieving a 17.68% performance improvement over baseline LLMs. The further analysis highlights the generalization ability of MemGraph across various LLMs, both in-domain and out-of-domain, and its capacity to enhance the internal reasoning processes of LLMs during patent matching. All data and codes are available at https://github.com/NEUIR/MemGraph.

Problem

• Improving patent matching by leveraging hierarchical classification
• Enhancing LLM semantics understanding via memory graph traversal
• Addressing keyword reliance in existing patent matching models

Innovation

• Memory graph enhances LLM patent matching
• Traverses memory to identify patent entities
• Attributes entities to ontologies for semantics