Sparse Optimization for Unsupervised Extractive Summarization of Long Documents with the Frank-Wolfe Algorithm

TL;DR

This paper introduces an efficient unsupervised extractive summarization method for long documents using a sparse auto-regression model solved by the Frank-Wolfe algorithm, improving results especially with paraphrased summaries.

Contribution

It proposes a novel sparse auto-regression framework for unsupervised extractive summarization and an efficient Frank-Wolfe algorithm tailored for long documents.

Findings

Achieves better ROUGE scores than existing methods.

Efficiently generates summaries with approximately k iterations.

Performs well with embedding-based semantic evaluation.

Abstract

We address the problem of unsupervised extractive document summarization, especially for long documents. We model the unsupervised problem as a sparse auto-regression one and approximate the resulting combinatorial problem via a convex, norm-constrained problem. We solve it using a dedicated Frank-Wolfe algorithm. To generate a summary with $k$ sentences, the algorithm only needs to execute $\approx k$ iterations, making it very efficient. We explain how to avoid explicit calculation of the full gradient and how to include sentence embedding information. We evaluate our approach against two other unsupervised methods using both lexical (standard) ROUGE scores, as well as semantic (embedding-based) ones. Our method achieves better results with both datasets and works especially well when combined with embeddings for highly paraphrased summaries.