Adaptive Multi-Stage Patent Claim Generation with Unified Quality Assessment

TL;DR

This paper presents a three-stage patent claim generation framework that improves cross-jurisdictional generalization, semantic modeling, and quality assessment, significantly outperforming existing models on multiple benchmarks.

Contribution

It introduces a novel multi-stage approach with relationship-aware similarity, domain-adaptive generation, and unified quality assessment, advancing automated patent claim generation.

Findings

7.6-point ROUGE-L improvement over GPT-4o

8.3% BERTScore enhancement over Llama-3.1-8B

0.847 correlation with human experts

Abstract

Current patent claim generation systems face three fundamental limitations: poor cross-jurisdictional generalization, inadequate semantic relationship modeling between claims and prior art, and unreliable quality assessment. We introduce a novel three-stage framework that addresses these challenges through relationship-aware similarity analysis, domain-adaptive claim generation, and unified quality assessment. Our approach employs multi-head attention with eight specialized heads for explicit relationship modeling, integrates curriculum learning with dynamic LoRA adapter selection across five patent domains, and implements cross-attention mechanisms between evaluation aspects for comprehensive quality assessment. Extensive experiments on USPTO HUPD dataset, EPO patent collections, and Patent-CE benchmark demonstrate substantial improvements: 7.6-point ROUGE-L gain over GPT-4o, 8.3\% BERTScore enhancement over Llama-3.1-8B, and 0.847 correlation with human experts compared to 0.623 for separate evaluation models. Our method maintains 89.4\% cross-jurisdictional performance retention versus 76.2\% for baselines, establishing a comprehensive solution for automated patent prosecution workflows.