RELexED: Retrieval-Enhanced Legal Summarization with Exemplar Diversity

2025-01-23

AI Summary

Existing legal document summarization methods rely solely on source documents, leading to topic drift and stylistic inconsistency. To address this, we propose a retrieval-augmented, exemplar-guided generation framework featuring a two-stage exemplar selection strategy: first, influence functions quantify the causal contribution of each exemplar to the target summary; second, a Determinantal Point Process (DPP) optimizes the trade-off between exemplar relevance and diversity. The model jointly leverages both the source document and high-quality reference summaries to guide generation, is fine-tuned on legal-domain data, and integrates a Retrieval-Augmented Generation (RAG) architecture. Experiments on two legal summarization benchmarks demonstrate significant improvements over exemplar-free baselines and conventional similarity-based exemplar selection methods, yielding substantial gains in topic consistency and linguistic coherence. Our approach establishes a novel, interpretable, and controllable paradigm for legal text summarization.

Technology Category

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

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

This paper addresses the task of legal summarization, which involves distilling complex legal documents into concise, coherent summaries. Current approaches often struggle with content theme deviation and inconsistent writing styles due to their reliance solely on source documents. We propose RELexED, a retrieval-augmented framework that utilizes exemplar summaries along with the source document to guide the model. RELexED employs a two-stage exemplar selection strategy, leveraging a determinantal point process to balance the trade-off between similarity of exemplars to the query and diversity among exemplars, with scores computed via influence functions. Experimental results on two legal summarization datasets demonstrate that RELexED significantly outperforms models that do not utilize exemplars and those that rely solely on similarity-based exemplar selection.

Problem

• Legal Document Summarization
• Theme Deviation
• Style Inconsistency

Innovation

• RELexED
• example abstracts
• two-step strategy