Improving Update Summarization by Revisiting the MMR Criterion

TL;DR

This paper introduces a novel update summarization method using a modified Maximal Marginal Relevance criterion, Smmr, to select diverse, topic-relevant sentences for coherent multi-document summaries, validated through TAC 2008.

Contribution

It proposes a new sentence selection criterion, Smmr, for update summarization that balances relevance and novelty, improving summary quality.

Findings

Achieved promising results in TAC 2008 evaluation

Demonstrated effectiveness of Smmr in selecting diverse sentences

Enhanced summary coherence through linguistic post-processing

Abstract

This paper describes a method for multi-document update summarization that relies on a double maximization criterion. A Maximal Marginal Relevance like criterion, modified and so called Smmr, is used to select sentences that are close to the topic and at the same time, distant from sentences used in already read documents. Summaries are then generated by assembling the high ranked material and applying some ruled-based linguistic post-processing in order to obtain length reduction and maintain coherency. Through a participation to the Text Analysis Conference (TAC) 2008 evaluation campaign, we have shown that our method achieves promising results.