Sentence Centrality Revisited for Unsupervised Summarization

TL;DR

This paper proposes an unsupervised summarization method that enhances sentence centrality computation using BERT embeddings and directed graphs to better capture sentence importance across diverse languages and styles.

Contribution

It introduces a novel graph-based ranking algorithm that incorporates BERT and directed edges for improved unsupervised summarization.

Findings

Outperforms strong baselines on multiple datasets

Effective across different languages and writing styles

Demonstrates the benefit of directed graphs and neural embeddings

Abstract

Single document summarization has enjoyed renewed interests in recent years thanks to the popularity of neural network models and the availability of large-scale datasets. In this paper we develop an unsupervised approach arguing that it is unrealistic to expect large-scale and high-quality training data to be available or created for different types of summaries, domains, or languages. We revisit a popular graph-based ranking algorithm and modify how node (aka sentence) centrality is computed in two ways: (a)~we employ BERT, a state-of-the-art neural representation learning model to better capture sentential meaning and (b)~we build graphs with directed edges arguing that the contribution of any two nodes to their respective centrality is influenced by their relative position in a document. Experimental results on three news summarization datasets representative of different languages and writing styles show that our approach outperforms strong baselines by a wide margin.