Supervising the Centroid Baseline for Extractive Multi-Document Summarization

TL;DR

This paper enhances the centroid-based extractive multi-document summarization method by integrating beam search and a centroid estimation attention model, achieving improved results across multiple datasets and languages.

Contribution

It introduces a novel combination of beam search and centroid estimation attention to refine the centroid summarization approach.

Findings

Improved summarization performance on multiple datasets

Effective in multilingual scenarios

Enhanced sentence selection accuracy

Abstract

The centroid method is a simple approach for extractive multi-document summarization and many improvements to its pipeline have been proposed. We further refine it by adding a beam search process to the sentence selection and also a centroid estimation attention model that leads to improved results. We demonstrate this in several multi-document summarization datasets, including in a multilingual scenario.