SummPip: Unsupervised Multi-Document Summarization with Sentence Graph Compression
This work addresses the challenge of resource-intensive training data for multi-document summarization, offering an unsupervised solution that is incremental in improving domain applicability.
The paper tackles the problem of multi-document summarization without requiring training data by proposing SummPip, an unsupervised method that uses sentence graph compression and spectral clustering, achieving competitive results on Multi-News and DUC-2004 datasets compared to both unsupervised and supervised neural approaches.
Obtaining training data for multi-document summarization (MDS) is time consuming and resource-intensive, so recent neural models can only be trained for limited domains. In this paper, we propose SummPip: an unsupervised method for multi-document summarization, in which we convert the original documents to a sentence graph, taking both linguistic and deep representation into account, then apply spectral clustering to obtain multiple clusters of sentences, and finally compress each cluster to generate the final summary. Experiments on Multi-News and DUC-2004 datasets show that our method is competitive to previous unsupervised methods and is even comparable to the neural supervised approaches. In addition, human evaluation shows our system produces consistent and complete summaries compared to human written ones.