🤖 AI Summary
Patent similarity assessment requires simultaneous consideration of technical features, application domains, and claim scope—yet existing methods largely ignore the composite semantic structure of patent documents. To address this, we propose the Multi-dimensional Reasoning Graph (MARG), a novel framework that employs a four-stage dynamic weighted graph reasoning mechanism to enable decoupled modeling and context-aware fusion across these three dimensions. We introduce the first patent-specific multi-dimensional disentangled representation paradigm and construct PatentSimBench—the first expert-annotated, fine-grained benchmark for patent similarity evaluation. Experiments demonstrate that MARG achieves a high correlation (r = 0.938) with expert judgments on PatentSimBench, significantly outperforming state-of-the-art embedding models and prompt-engineering approaches. This work establishes a new paradigm for interpretable, high-fidelity patent similarity analysis.
📝 Abstract
Patent similarity evaluation plays a critical role in intellectual property analysis. However, existing methods often overlook the intricate structure of patent documents, which integrate technical specifications, legal boundaries, and application contexts. We introduce PatentMind, a novel framework for patent similarity assessment based on a Multi-Aspect Reasoning Graph (MARG). PatentMind decomposes patents into three core dimensions: technical feature, application domain, and claim scope, to compute dimension-specific similarity scores. These scores are dynamically weighted through a four-stage reasoning process which integrates contextual signals to emulate expert-level judgment. To support evaluation, we construct PatentSimBench, a human-annotated benchmark comprising 500 patent pairs. Experimental results demonstrate that PatentMind achieves a strong correlation ($r=0.938$) with expert annotations, significantly outperforming embedding-based models and advanced prompt engineering methods.These results highlight the effectiveness of modular reasoning frameworks in overcoming key limitations of embedding-based methods for analyzing patent similarity.