Multi-hop Inference for Question-driven Summarization

Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing, pages 6734–6744, November 16–20, 2020. c⃝2020 Association for Computational Linguistics 6734 Multi-hop Inference for Question-driven Summarization∗ Yang Deng, Wenxuan Zhang, Wai Lam The Chinese University of Hong Kong {ydeng,wxzhang,wlam}@se.cuhk.edu.hk Abstract Question-driven summarization has been re- cently studied as an effective approach to summarizing the source document to pro- duce concise but informative answers for non- factoid questions. In this work, we propose a novel question-driven abstractive summa- rization method, Multi-hop Selective Genera- tor (MSG), to incorporate multi-hop reason- ing into question-driven summarization and, meanwhile, provide justifications for the gen- erated summaries. Specifically, we jointly model the relevance to the question and the interrelation among different sentences via a human-like multi-hop inference module, which captures important sentences for justify- ing the summarized answer. A gated selective pointer generator network with a multi-view coverage mechanism is designed to integrate diverse information from different perspec- tives. Experimental results show that the pro- posed method consistently outperforms state- of-the-art methods on two non-factoid QA datasets, namely WikiHow and PubMedQA. 1 Introduction Recent years have witnessed several attempts on exploring question-driven summarization, which aims at summarizing the source document with re- spect to a specific question, to produce a concise but informative answer in non-factoid question an- swering (QA) (Tomasoni and Huang, 2010; Chan et al., 2012; Song et al., 2017). Unlike factoid QA (Rajpurkar et al., 2016), e.g., “Who is the au- thor of Harry Potter?”, whose answer is generally a single phrase or a short sentence with limited infor- mation, the answers for non-factoid questions are supposed to be more informative, involving some ∗ The work described in this paper is substantially sup- ported by a grant from the Research Grant Council of the Hong Kong Special Administrative Region, China (Project Codes: 14200719). detailed analysis to explain or justify the final an- swers, such as questions in community QA (Ishida et al., 2018; Deng et al., 2020a) or explainable QA (Fan et al., 2019; Nakatsuji and Okui, 2020). As the example from PubMedQA (Jin et al., 2019) presented in Figure 1, the answer can be regarded as the summary over the document driven by the reasoning process of the given question. Most of related studies focus on query-based summarization approaches for summarizing the query-related content from the source docu- ment (Shen and Li, 2011; Wang et al., 2013; Cao et al., 2016; Nema et al., 2017). However, these approaches fall short of tackling question-driven summarization problem in QA scenario, since the query-based summarization process is typically based on semantic relevance measurement with- out a careful reasoning or inference process, which is essential to question-driven summarization. Cur- rently, question-driven summarization is mainly explored by traditional information retrieval meth- ods to select sentences from the source document to construct the final answer (Wang et al., 2014; Song et al., 2017; Yulianti et al., 2018), which heavily rely on hand-crafted features or tedious multi-stage pipelines. Besides, compared to extrac- tive summarization (Cao et al., …

Content truncated to 500 words