Improve Query Focused Abstractive Summarization by Incorporating Answer Relevance

Findings of the Association for Computational Linguistics: ACL-IJCNLP 2021, pages 3124–3131 August 1–6, 2021. ©2021 Association for Computational Linguistics 3124 Improve Query Focused Abstractive Summarization by Incorporating Answer Relevance Dan Su∗, Tiezheng Yu∗, Pascale Fung Center for Artificial Intelligence Research (CAiRE) Department of Electronic and Computer Engineering The Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong {dsu,tyuah}@connect.ust.hk, pascale@ece.ust.hk Abstract Query focused summarization (QFS) models aim to generate summaries from source docu- ments that can answer the given query. Most previous work on QFS only considers the query relevance criterion when producing the summary. However, studying the effect of an- swer relevance in the summary generating pro- cess is also important. In this paper, we pro- pose QFS-BART, a model that incorporates the explicit answer relevance of the source doc- uments given the query via a question answer- ing model, to generate coherent and answer- related summaries. Furthermore, our model can take advantage of large pre-trained mod- els which improve the summarization perfor- mance significantly. Empirical results on the Debatepedia dataset show that the proposed model achieves the new state-of-the-art perfor- mance.1 1 Introduction Query focused summarization (QFS) models aim to extract essential information from a source doc- ument(s) and organize it into a summary that can answer a query (Dang, 2005). The input can be either a single document that has multiple views or multiple documents that contain multiple top- ics, and the output summary should be focused on the given query. QFS has various applications (e.g., a personalized search engine that provides the user with an overview summary based on their query (Su et al., 2020b)). Early work on the QFS task mainly focused on generating extractive summaries (Davis et al., 2012; Daum´e III and Marcu, 2006; Feigenblat et al., 2017; Xu and Lapata, 2020b), which may contain un- readable sentence ordering and lack cohesiveness. ∗ The two authors contribute equally. 1The code is released at: https://github.com/ HLTCHKUST/QFS Document: Interrogator Ali Soufan said in an April op-ed article in the New York Times: “It is inaccurate to say that Abu Zubaydah had been un- cooperative [and that enhanced interrogation tech- niques supplies interrogators with previously un- obtainable information]. Along with another f.b.i. agent and with several c.i.a. officers present I ques- tioned him from March to June before the harsh techniques were introduced later in August. Under traditional interrogation methods he provided us with important actionable intelligence.” Query: Are traditional interrogation methods in- sufficient? Summary: The same info can be obtained by tra- ditional interrogations. Table 1: An example of QFS. The input is a document and a corresponding query, and the highlight sentence is the answer from our QA module. We observe that the summary and the answers are very correlated. Other work on abstractive QFS incorporated the query relevance into existing neural summarization models (Nema et al., 2017; Baumel et al., 2018). The closest work to ours was done by (Su et al., 2020a) and (Xu and Lapata, 2020a,b), who lever- aged an …

Content truncated to 500 words