Considering Length Diversity in Retrieval-Augmented Summarization

TL;DR

This work addresses controlling summary length within retrieval-augmented summarization by examining exemplar length effects and introducing Diverse Length-aware Maximal Marginal Relevance ($DL\text{-}MMR$). By storing only target-length information and using a length-diversity criterion, $DL\text{-}MMR$ reduces the need for exhaustive exemplar-exemplar comparisons, achieving massive memory (approximately $781{,}513\times$) and computation (approximately $500{,}092\times$) savings while maintaining informativeness. Experiments on Google, Broad, and BNC with multiple backbones demonstrate that $DL\text{-}MMR$ outperforms NN, is competitive with MMR, and is validated by human evaluations showing improved conciseness and informativeness. The findings highlight the practical impact of length diversity in retrieval-augmented summarization and point toward multilingual extensions and open-source release for broader adoption.

Abstract

This study investigates retrieval-augmented summarization by specifically examining the impact of exemplar summary lengths under length constraints, not covered by previous work. We propose a Diverse Length-aware Maximal Marginal Relevance (DL-MMR) algorithm to better control summary lengths. This algorithm combines the query relevance with diverse target lengths in retrieval-augmented summarization. Unlike previous methods that necessitate exhaustive exemplar exemplar relevance comparisons using MMR, DL-MMR considers the exemplar target length as well and avoids comparing exemplars to each other, thereby reducing computational cost and conserving memory during the construction of an exemplar pool. Experimental results showed the effectiveness of DL-MMR, which considers length diversity, compared to the original MMR algorithm. DL-MMR additionally showed the effectiveness in memory saving of 781,513 times and computational cost reduction of 500,092 times, while maintaining the same level of informativeness.