Domain Adaptation with Pre-trained Transformers for Query-Focused Abstractive Text Summarization

• Published in: Computational linguistics - Association for Computational Linguistics Vol. 48; no. 2; pp. 279 - 320
• Main Authors: Laskar, Md Tahmid Rahman, Hoque, Enamul, Huang, Jimmy Xiangji
• Format: Journal Article
• Language: English
• Published: One Broadway, 12th Floor, Cambridge, Massachusetts 02142, USA MIT Press 09.06.2022; MIT Press Journals, The; The MIT Press
• Subjects: Abstracts; Adaptation; Automatic summarization; Datasets; Documents; Natural language processing; Queries; Summaries; Supervised learning
• ISSN: 0891-2017; 1530-9312
• DOI: 10.1162/coli_a_00434

The Query-Focused Text Summarization (QFTS) task aims at building systems that generate the summary of the text document(s) based on the given query. A key challenge in addressing this task is the lack of large labeled data for training the summarization model. In this article, we address this challenge by exploring a series of domain adaptation techniques. Given the recent success of pre-trained transformer models in a wide range of natural language processing tasks, we utilize such models to generate abstractive summaries for the QFTS task for both single-document and multi-document scenarios. For domain adaptation, we apply a variety of techniques using pre-trained transformer-based summarization models including transfer learning, weakly supervised learning, and distant supervision. Extensive experiments on six datasets show that our proposed approach is very effective in generating abstractive summaries for the QFTS task while setting a new state-of-the-art result in several datasets across a set of automatic and human evaluation metrics.