Graph-based Neural Multi-Document Summarization
This work addresses the problem of generating concise summaries from multiple documents for applications like news aggregation, though it is incremental in combining existing graph and neural methods.
The paper tackles multi-document summarization by incorporating sentence relation graphs with Graph Convolutional Networks to estimate sentence salience, achieving competitive results on the DUC 2004 benchmark.
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.