A Hybrid Architecture with Efficient Fine Tuning for Abstractive Patent Document Summarization

2025-03-13

AI Summary

Patent documents pose significant challenges for automatic summarization due to their exceptional length and dense integration of technical and legal information. To address this, we propose a hybrid extractive-abstractive summarization method. First, LexRank—a graph-based sentence-ranking algorithm—is employed to extract salient sentences, providing structured guidance for subsequent abstraction. Second, a BART model is efficiently fine-tuned via Low-Rank Adaptation (LoRA) to generate high-quality abstractive summaries. Third, a meta-learning mechanism is incorporated to enhance cross-domain generalization across diverse technological fields. This work introduces the first three-tier architecture—“extraction-guided abstraction with domain adaptation”—specifically designed for patent summarization. Evaluated on a multi-domain patent dataset, our approach achieves substantial improvements in ROUGE-L (+4.2) and information density, while demonstrating high summary coherence and strong domain transferability. The framework balances theoretical novelty with practical deployability, offering a scalable solution for industrial patent analytics.

Technology Category

• Natural Language Processing: Summarization
• Data Mining & Knowledge Management: Data Visualization & Summarization
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Automatic patent summarization approaches that help in the patent analysis and comprehension procedure are in high demand due to the colossal growth of innovations. The development of natural language processing (NLP), text mining, and deep learning has notably amplified the efficacy of text summarization models for abundant types of documents. Summarizing patent text remains a pertinent challenge due to the labyrinthine writing style of these documents, which includes technical and legal intricacies. Additionally, these patent document contents are considerably lengthier than archetypal documents, which intricates the process of extracting pertinent information for summarization. Embodying extractive and abstractive text summarization methodologies into a hybrid framework, this study proposes a system for efficiently creating abstractive summaries of patent records. The procedure involves leveraging the LexRank graph-based algorithm to retrieve the important sentences from input parent texts, then utilizing a Bidirectional Auto-Regressive Transformer (BART) model that has been fine-tuned using Low-Ranking Adaptation (LoRA) for producing text summaries. This is accompanied by methodical testing and evaluation strategies. Furthermore, the author employed certain meta-learning techniques to achieve Domain Generalization (DG) of the abstractive component across multiple patent fields.

Problem

• Develops a hybrid system for patent document summarization.
• Addresses challenges in summarizing lengthy, technical patent texts.
• Utilizes LexRank and fine-tuned BART for efficient summarization.

Innovation

• Hybrid framework combining extractive and abstractive summarization
• LexRank algorithm for sentence extraction
• Fine-tuned BART model with LoRA for summarization