GUSUM: Graph-Based Unsupervised Summarization using Sentence-BERT and Sentence Features

Abstract

Unsupervised extractive document summarization aims to extract salient sentences from a document without requiring a labelled corpus. In existing graph-based methods, vertex and edge weights are mostly created by calculating sentence similarities. In this paper, we develop a Graph-Based Unsupervised Summarization method for extractive text summarization. We revive traditional graph ranking algorithms with recent sentence embedding models and sentence features and modify how sentence centrality is computed. We first use Sentence-BERT, a state-of-the-art method for obtaining sentence embeddings to better capture the sentence meaning. In this way, we define the edges of a graph where semantic similarities are represented. Then, we create an undirected graph in which the calculated sentence feature scores of each sentence are represented in the vertices. In the last stage, we determine the most important sentences in the document with the ranking method we suggested on the graph created. Experiments on CNN/Daily Mail and New York Times datasets show our approach achieves high performance on unsupervised graph-based summarization when evaluated both automatically and by humans.

Tasks

Reproductions