Open-Domain Why-Question Answering with Adversarial Learning to Encode Answer Texts

39549

Authors

nan

Abstract

AbstractIn this paper, we propose a method for why-question answering (why-QA) that uses an adversarial learning framework. Existing why-QA methods retrieve “answer passages” that usually consist of several sentences. These multi-sentence passages contain not only the reason sought by a why-question and its connection to the why-question, but also redundant and/or unrelated parts. We use our proposed “Adversarial networks for Generating compact-answer Representation” (AGR) to generate from a passage a vector representation of the non-redundant reason sought by a why-question and exploit the representation for judging whether the passage actually answers the why-question. Through a series of experiments using Japanese why-QA datasets, we show that these representations improve the performance of our why-QA neural model as well as that of a BERT-based why-QA model. We show that they also improve a state-of-the-art distantly supervised open-domain QA (DS-QA) method on publicly available English datasets, even though the target task is not a why-QA.

Keywords

nan

Full Content

Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4227–4237 Florence, Italy, July 28 - August 2, 2019. c⃝2019 Association for Computational Linguistics 4227 Open-Domain Why-Question Answering with Adversarial Learning to Encode Answer Texts Jong-Hoon Oh§ Kazuma Kadowaki§‡ Julien Kloetzer§ Ryu Iida§¶ Kentaro Torisawa§¶ Data-driven Intelligent System Research Center (DIRECT), National Institute of Information and Communications Technology (NICT)§ Advanced Technology Laboratory, The Japan Research Institute, Limited (JRI)‡ Graduate School of Science and Technology, NAIST¶ {rovellia, kadowaki, julien, ryu.iida, torisawa}@nict.go.jp Abstract In this paper, we propose a method for why- question answering (why-QA) that uses an ad- versarial learning framework. Existing why- QA methods retrieve answer passages that usually consist of several sentences. These multi-sentence passages contain not only the reason sought by a why-question and its con- nection to the why-question, but also redun- dant and/or unrelated parts. We use our proposed Adversarial networks for Generat- ing compact-answer Representation (AGR) to generate from a passage a vector representa- tion of the non-redundant reason sought by a why-question and exploit the representa- tion for judging whether the passage actu- ally answers the why-question. Through a se- ries of experiments using Japanese why-QA datasets, we show that these representations improve the performance of our why-QA neu- ral model as well as that of a BERT-based why-QA model. We show that they also im- prove a state-of-the-art distantly supervised open-domain QA (DS-QA) method on pub- licly available English datasets, even though the target task is not a why-QA. 1 Introduction Why-question answering (why-QA) tasks retrieve from a text archive answers to such why-questions as “Why does honey last such a long time?” Previ- ous why-QA methods retrieve from a text archive answer passages, each of which consists of sev- eral sentences, like A in Table 1 (Girju, 2003; Higashinaka and Isozaki, 2008; Oh et al., 2012, 2013, 2016, 2017; dos Santos et al., 2016; Sharp et al., 2016; Tan et al., 2016; Verberne et al., 2011), and then determine whether the passages answer the question. A proper answer passage must con- tain (1) a paraphrase of the why-question (e.g., the underlined texts in Table 1) and (2) the reasons or the causes (e.g., the bold texts in Table 1) of Q Why does honey last a long time? A While excavating Egypt’s pyramids, archaeologists have found pots of honey in an ancient tomb: thou- sands of years old and still preserved. Honey can last a long time due to three special properties. Its average pH is 3.9, which is quite acidic. Such high level of acidity is certainly hostile and hin- ders the growth of many microbes. Though honey contains around 17–18% water, its water activity is too low to support the growth of microbes. More- over honey contains hydrogen peroxide, which is thought to help prevent the growth of microbes in honey. Despite these properties, honey can be con- taminated under certain circumstances. C Because its acidity, low water activity, and hydrogen peroxide together hinder the growth of …

Content truncated to 500 words

Labeled Text

AbstractIn this paper, we propose a method for why-question answering (why-QA) that uses an adversarial learning framework. Existing why-QA methods retrieve “answer passages” that usually consist of several sentences. These multi-sentence passages contain not only the reason sought by a why-question and its connection to the why-question, but also redundant and/or unrelated parts. We use our proposed “Adversarial networks for Generating compact-answer Representation” (AGR) to generate from a passage a vector representation of the non-redundant reason sought by a why-question and exploit the representation for judging whether the passage actually answers the why-question. Through a series of experiments using Japanese why-QA datasets, we show that these representations improve the performance of our why-QA neural model as well as that of a BERT-based why-QA model. We show that they also improve a state-of-the-art distantly supervised open-domain QA (DS-QA) method on publicly available English datasets, even though the target task is not a why-QA.

Publication Details

Paper ID:
39549
Year:
2019
Publication:
TIPSTER
Publication Type:
conf
DOI:
nan