Adaptive Multi-Stage Patent Claim Generation with Unified Quality Assessment

TL;DR

This work tackles the core challenges of patent claim generation by proposing a three-stage framework that combines relationship-aware similarity analysis, domain-adaptive claim generation via dynamic LoRA adapters, and unified quality assessment with cross-attention across evaluation aspects. It leverages eight specialized attention heads, curriculum learning across five patent domains, and a Longformer-based evaluation backbone to achieve strong cross-jurisdictional generalization and better alignment with human judgments. Empirical results on USPTO HUPD, EPO, and Patent-CE benchmarks show substantial improvements in ROUGE-L, BERTScore, and expert correlation, with high cross-jurisdictional retention and faster convergence. The approach offers a practical pathway to robust, automated patent prosecution workflows across diverse jurisdictions and technical domains.

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.

Figures (7)

• Figure 1: Motivation of the adaptive multi-stage framework for patent claim generation and evaluation.
• Figure 2: Architecture of the proposed three-stage patent claim generation and unified quality assessment system.
• Figure 3: Unified evaluation metric correlation.
• Figure 4: Main benchmark performance on patent claim generation.
• Figure 5: Cross-jurisdictional claim adaptation (USPTO, EPO, CNIPA).