An automatic patent literature retrieval system based on LLM-RAG

With the acceleration of technological innovation, efficient retrieval and classification of patent literature have become essential for intellectual property management and enterprise R&D. Traditional keyword- and rule-based retrieval methods often fail to address complex query intents or capture semantic associations across technical domains, resulting in incomplete and low-relevance results. This study presents an automated patent retrieval framework integrating Large Language Models (LLMs) with Retrieval-Augmented Generation (RAG) technology. The system comprises three components: (1) a preprocessing module for patent data standardization, (2) a high-efficiency vector retrieval engine leveraging LLM-generated embeddings, and (3) a RAG-enhanced query module that combines external document retrieval with context-aware response generation. Evaluations were conducted on the Google Patents dataset (2006–2024), containing millions of global patent records with metadata such as filing date, domain, and status. The proposed gpt-3.5-turbo-0125+RAG configuration achieved 80.5% semantic matching accuracy and 92.1% recall, surpassing baseline LLM methods by 28 percentage points. The framework also demonstrated strong generalization in cross-domain classification and semantic clustering tasks. These results validate the effectiveness of LLM–RAG integration for intelligent patent retrieval, providing a foundation for next-generation AI-driven intellectual property analysis platforms.