Towards Automated Patent Workflows: AI-Orchestrated Multi-Agent Framework for Intellectual Property Management and Analysis

TL;DR

PatExpert is presented, an autonomous multi-agent conversational framework designed to streamline and optimize patent-related tasks and offers a robust solution for automating and optimizing patent analysis.

Abstract

Patents are the currency of innovation, and like any currency, they need to be managed and protected (Gavin Potenza). Patents, as legal documents that secure intellectual property rights, play a critical role in technological innovation. The growing complexity of patent documents and the surge in patent applications have created a need for automated solutions in patent analysis. In this work, we present PatExpert, an autonomous multi-agent conversational framework designed to streamline and optimize patent-related tasks. The framework consists of a metaagent that coordinates task-specific expert agents for various patent-related tasks and a critique agent for error handling and feedback provision. The meta-agent orchestrates specialized expert agents, each fine-tuned for specific tasks such as patent classification, acceptance, claim generation, abstractive summarization, multi-patent analysis, and scientific hypothesis generation. For multi-patent analysis, the framework incorporates advanced methods like Graph Retrieval-Augmented Generation (GRAG) to enhance response accuracy and relevance by combining semantic similarity with knowledge graphs. Error handling is managed by critique agents (Gold-LLM-as-a-Judge and Reward-LLM-as-a-Judge), which evaluate output responses for accuracy and provide iterative feedback. The framework also prioritizes explainability, ensuring transparent justifications for decisions made during patent analysis. Its comprehensive capabilities make it a valuable tool for automating complex patent workflows, enhancing efficiency, accuracy, and compliance in patent-related tasks. Empirical evidence demonstrates significant improvements in patent processing tasks, concluding that the framework offers a robust solution for automating and optimizing patent analysis.

Figures (1)

• Figure 1: The figure shows the architecture of the multi-agent conversational framework, PatExpert. The meta-agent oversees various sub-agents, each responsible for specialized tasks such as patent classification, acceptance prediction, claim generation, abstractive summarization, hypothesis generation, and multi-patent analysis. User input flows through the meta-agent (which utilizes Meta-Llama-3.1-405B to interpret user queries), delegating tasks to the relevant sub-agents (using fine-tuned GPT-4o mini) to provide accurate responses. Critique agents, including the Reward-LLM-as-a-Judge (Nvidia Nemotron-4-340B-Reward) and LLM-as-a-Judge (OpenAI GPT-4o), evaluate the outputs for accuracy. These agents provide critical feedback to refine responses, ensuring that the framework adheres to high standards of precision, compliance, and quality in patent-related tasks. The knowledge graph stores structured information extracted from patents, while the conversational database holds historical interactions, helping the framework maintain context and continuity in multi-turn queries. The knowledge graph enables efficient semantic retrieval across multiple patents for ODQA tasks, improving response accuracy. In summary, PatExpert handles complex patent-related tasks and generates accurate, coherent responses through its structured, multi-agent framework.