Template-based Abstractive Microblog Opinion Summarization

• Published in: Transactions of the Association for Computational Linguistics Vol. 10; pp. 1229 - 1248
• Main Authors: Bilal, Iman Munire, Wang, Bo, Tsakalidis, Adam, Nguyen, Dong, Procter, Rob, Liakata, Maria
• Format: Journal Article
• Language: English
• Published: One Broadway, 12th Floor, Cambridge, Massachusetts 02142, USA MIT Press 22.11.2022; MIT Press Journals, The; The MIT Press
• Subjects: Attitudes; Automatic summarization; Computer science; Coronaviruses; COVID-19; Datasets; Homeless people; Microblogs; Performance enhancement; Performance evaluation; Product reviews; Social media; Social networks; Summaries; User generated content
• ISSN: 2307-387X; 2307-387X
• DOI: 10.1162/tacl_a_00516

We introduce the task of microblog opinion summarization (MOS) and share a dataset of 3100 gold-standard opinion summaries to facilitate research in this domain. The dataset contains summaries of tweets spanning a 2-year period and covers more topics than any other public Twitter summarization dataset. Summaries are abstractive in nature and have been created by journalists skilled in summarizing news articles following a template separating factual information (main story) from author opinions. Our method differs from previous work on generating gold-standard summaries from social media, which usually involves selecting representative posts and thus favors extractive summarization models. To showcase the dataset’s utility and challenges, we benchmark a range of abstractive and extractive state-of-the-art summarization models and achieve good performance, with the former outperforming the latter. We also show that fine-tuning is necessary to improve performance and investigate the benefits of using different sample sizes.