ToC: Tree-of-Claims Search with Multi-Agent Language Models

2025-11-21

AI Summary

Patent claim optimization requires balancing novelty against breadth of protection, yet manual drafting suffers from high cost and low consistency, while conventional large language models lack structured, iterative reasoning capabilities. This paper proposes Tree-of-Claims, the first framework integrating Monte Carlo Tree Search (MCTS) with multi-agent language model reasoning: an editing agent performs fine-grained semantic rewriting; a reviewing agent evaluates claim quality via chain-of-thought reasoning and a multi-objective reward function; and MCTS orchestrates an interpretable, controllable tree-based search process. The method significantly improves claim novelty, coverage breadth, and semantic coherence. On a benchmark of 1,145 patent claims, it achieves an average 8% improvement (up to 9%) over zero-shot and few-shot baselines, demonstrating effectiveness, stability, and legal robustness.

Technology Category

• Search and Optimization: Learning to Search
• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: (Large) Language Models

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work
• Search and Retrieval-Augmented AI: Agentic search

Abstract

Optimizing patent claims is a critical yet challenging task, demanding careful balance between maximizing novelty and preserving legal scope. Manual claim drafting is labor-intensive, costly, and inherently inconsistent, while conventional Large Language Models (LLMs) often lack the structured, iterative reasoning essential for precise claim refinement. To address these challenges, we introduce Tree of Claims (ToC), an innovative framework that redefines claim editing as a guided search problem. ToC synergistically integrates Monte Carlo Tree Search (MCTS) with a collaborative multi-agent system, comprising an LLM-based EditorAgent that proposes contextually grounded edits, and an ExaminerAgent that mimics patent examiner critiques through structured, chain-of-thought analyses of novelty and prior art disclosure. Driven by a carefully designed multi-objective reward function, ToC jointly optimizes novelty, scope retention, and semantic coherence. Experimental evaluation on a benchmark of 1145 claims demonstrates that ToC significantly outperforms standard LLMs in zero-shot and few-shot scenarios, achieving an average composite score improvement of 8%, and up to 9% in certain cases. Extensive experiments, including detailed ablation studies, validate ToC's efficacy in generating superior, legally robust claim revisions. Overall, ToC establishes a transparent, controllable, and interpretable methodology that effectively bridges advanced LLM reasoning capabilities with strategic MCTS planning for structured patent claim optimization.The source code is available at https://github.com/ysy2003/ToC.

Problem

• Optimizing patent claims by balancing novelty and legal scope
• Addressing labor-intensive manual drafting and inconsistent LLM outputs
• Improving structured iterative reasoning for precise claim refinement

Innovation

• Integrates Monte Carlo Tree Search with multi-agent system
• Uses EditorAgent for edits and ExaminerAgent for critiques
• Optimizes novelty, scope retention, and semantic coherence