A New Approach to Overgenerating and Scoring Abstractive Summaries

TL;DR

This paper introduces a two-stage method for generating and selecting diverse, length-controlled abstractive summaries that improve faithfulness and relevance, achieving state-of-the-art results on benchmark datasets.

Contribution

It presents a novel two-stage framework for generating multiple diverse summaries with precise length control and improved faithfulness, outperforming existing methods.

Findings

Achieves state-of-the-art performance on benchmark datasets.

Effectively controls summary length according to user needs.

Enhances faithfulness and diversity in generated summaries.

Abstract

We propose a new approach to generate multiple variants of the target summary with diverse content and varying lengths, then score and select admissible ones according to users' needs. Abstractive summarizers trained on single reference summaries may struggle to produce outputs that achieve multiple desirable properties, i.e., capturing the most important information, being faithful to the original, grammatical and fluent. In this paper, we propose a two-staged strategy to generate a diverse set of candidate summaries from the source text in stage one, then score and select admissible ones in stage two. Importantly, our generator gives a precise control over the length of the summary, which is especially well-suited when space is limited. Our selectors are designed to predict the optimal summary length and put special emphasis on faithfulness to the original text. Both stages can be effectively trained, optimized and evaluated. Our experiments on benchmark summarization datasets suggest that this paradigm can achieve state-of-the-art performance.