RELexED: Retrieval-Enhanced Legal Summarization with Exemplar Diversity

TL;DR

Legal summarization often suffers from content drift and stylistic mismatch when models rely solely on source documents. RELexED mitigates this by a retrieval-augmented approach that uses exemplar summaries and a two-stage exemplar selection with a DPP to balance exemplar quality and diversity, with quality and similarity derived from influence functions. Empirical results on SuperSCOTUS and CivilSum show that RELexED consistently outperforms baselines, particularly in coherence and faithfulness, by leveraging diverse and informative exemplars. This work demonstrates that carefully curated exemplars can boost supervised legal summarization without resorting to substantial model scaling, with implications for scalable, domain-specific text generation in legal contexts.

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.