SumREN: Summarizing Reported Speech about Events in News

TL;DR

This paper introduces SumREN, a new benchmark and method for summarizing reported speech about events in news articles, emphasizing reactions of speakers rather than just event details.

Contribution

It presents a novel task, creates a benchmark dataset, and develops a pipeline-based summarization framework that improves factuality and abstraction in reported speech summaries.

Findings

Smaller models like BART can match GPT-3 performance with silver data training.

The proposed pipeline generates more abstractive and factual summaries.

SUMREN benchmark includes 745 summaries from 633 articles about 132 events.

Abstract

A primary objective of news articles is to establish the factual record for an event, frequently achieved by conveying both the details of the specified event (i.e., the 5 Ws; Who, What, Where, When and Why regarding the event) and how people reacted to it (i.e., reported statements). However, existing work on news summarization almost exclusively focuses on the event details. In this work, we propose the novel task of summarizing the reactions of different speakers, as expressed by their reported statements, to a given event. To this end, we create a new multi-document summarization benchmark, SUMREN, comprising 745 summaries of reported statements from various public figures obtained from 633 news articles discussing 132 events. We propose an automatic silver training data generation approach for our task, which helps smaller models like BART achieve GPT-3 level performance on this task. Finally, we introduce a pipeline-based framework for summarizing reported speech, which we empirically show to generate summaries that are more abstractive and factual than baseline query-focused summarization approaches.