Topic-Guided Abstractive Multi-Document Summarization

TL;DR

This paper introduces a novel multi-document summarization model that uses a heterogeneous graph and neural topic modeling to improve summary quality by capturing cross-document semantics and global topics.

Contribution

It proposes a graph-to-sequence framework combined with joint topic modeling and multi-task learning for enhanced abstractive multi-document summarization.

Findings

Outperforms previous state-of-the-art models on Multi-News dataset

Achieves higher Rouge scores and better human evaluation results

Learns high-quality, meaningful topics that aid summary generation

Abstract

A critical point of multi-document summarization (MDS) is to learn the relations among various documents. In this paper, we propose a novel abstractive MDS model, in which we represent multiple documents as a heterogeneous graph, taking semantic nodes of different granularities into account, and then apply a graph-to-sequence framework to generate summaries. Moreover, we employ a neural topic model to jointly discover latent topics that can act as cross-document semantic units to bridge different documents and provide global information to guide the summary generation. Since topic extraction can be viewed as a special type of summarization that "summarizes" texts into a more abstract format, i.e., a topic distribution, we adopt a multi-task learning strategy to jointly train the topic and summarization module, allowing the promotion of each other. Experimental results on the Multi-News dataset demonstrate that our model outperforms previous state-of-the-art MDS models on both Rouge metrics and human evaluation, meanwhile learns high-quality topics.