QASA: Advanced Document Retriever for Open-Domain Question Answering by Learning to Rank Question-Aware Self-Attentive Document Representations

Nguyen, Minh Trang and Tran, Van Lien and Can, Duy Cat and Ha, Quang Thuy and Vu, Thi Ly and Chng, Eng-Siong (2019) QASA: Advanced Document Retriever for Open-Domain Question Answering by Learning to Rank Question-Aware Self-Attentive Document Representations. In: The 3rd International Conference on Machine Learning and Soft Computing, 25-28 January 2019, Da Lat, Vietnam.

PDF - Published Version Restricted to Repository staff only Available under License Creative Commons Attribution Non-commercial No Derivatives. Download (463kB) | Request a copy

Abstract

For information consumers, being able to obtain a short and accurate answer for a query is one of the most desirable features. This motivation, along with the rise of deep learning, has led to a boom in open-domain Question Answering (QA) research. While the problem of machine comprehension has received multiple success with the help of large training corpora and the emergence of attention mechanism, the development of document retrieval in open-domain QA is lagged behind. In this work, we propose a novel encoding method for learning question-aware self-attentive document representations. By applying pair-wise ranking approach to these encodings, we build a Document Retriever, called QASA, which is then integrated with a machine reader to form a complete open-domain QA system. Our system is thoroughly evaluated using QUASAR-T dataset and shows surpassing results compared to other state-of-the-art methods.

Item Type:	Conference or Workshop Item (Paper)
Subjects:	Information Technology (IT)
Divisions:	Faculty of Information Technology (FIT)
Depositing User:	Trang Minh Nguyễn
Date Deposited:	20 Jun 2019 22:49
Last Modified:	20 Jun 2019 22:49
URI:	http://eprints.uet.vnu.edu.vn/eprints/id/eprint/3511

Actions (login required)

View Item