Graph-based Neural Multi-Document Summarization

TL;DR

This paper introduces a neural multi-document summarization system that leverages sentence relation graphs and Graph Convolutional Networks to improve extractive summarization performance, demonstrating advantages over traditional methods.

Contribution

The paper presents a novel neural MDS approach combining GCNs with sentence relation graphs, enhancing salience estimation and redundancy avoidance.

Findings

Outperforms traditional graph-based extractive methods

Achieves competitive results against state-of-the-art systems

Demonstrates the benefit of integrating sentence relations with neural networks

Abstract

We propose a neural multi-document summarization (MDS) system that incorporates sentence relation graphs. We employ a Graph Convolutional Network (GCN) on the relation graphs, with sentence embeddings obtained from Recurrent Neural Networks as input node features. Through multiple layer-wise propagation, the GCN generates high-level hidden sentence features for salience estimation. We then use a greedy heuristic to extract salient sentences while avoiding redundancy. In our experiments on DUC 2004, we consider three types of sentence relation graphs and demonstrate the advantage of combining sentence relations in graphs with the representation power of deep neural networks. Our model improves upon traditional graph-based extractive approaches and the vanilla GRU sequence model with no graph, and it achieves competitive results against other state-of-the-art multi-document summarization systems.