Please use this identifier to cite or link to this item:
https://hdl.handle.net/2440/131346
Citations | ||
Scopus | Web of Science® | Altmetric |
---|---|---|
?
|
?
|
Type: | Conference paper |
Title: | BERT-CoQAC: BERT-based conversational question answering in context |
Author: | Zaib, M. Tran, D.H. Sagar, S. Mahmood, A. Zhang, W.E. Sheng, Q.Z. |
Citation: | Communications in Computer and Information Science, 2021, vol.1362, pp.47-57 |
Publisher: | Springer |
Publisher Place: | Singapore |
Issue Date: | 2021 |
Series/Report no.: | Communications in Computer and Information Science; 1362 |
ISBN: | 9789811600098 |
ISSN: | 1865-0929 1865-0937 |
Conference Name: | International Symposium on Parallel Architectures, Algorithms and Programming (PAAP) (28 Dec 2020 - 30 Dec 2020 : Shenzhen, China) |
Statement of Responsibility: | Munazza Zaib, Dai Hoang Tran, Subhash Sagar, Adnan Mahmood, Wei E. Zhang, Quan Z. Sheng |
Abstract: | As one promising way to inquire about any particular information through a dialog with the bot, question answering dialog systems have gained increasing research interests recently. Designing interactive QA systems has always been a challenging task in natural language processing and used as a benchmark to evaluate machine’s ability of natural language understanding. However, such systems often struggle when the question answering is carried out in multiple turns by the users to seek more information based on what they have already learned, thus, giving rise to another complicated form called Conversational Question Answering (CQA). CQA systems are often criticized for not understanding or utilizing the previous context of the conversation when answering the questions. To address the research gap, in this paper, we explore how to integrate the conversational history into the neural machine comprehension system. On one hand, we introduce a framework based on publicly available pre-trained language model called BERT for incorporating history turns into the system. On the other hand, we propose a history selection mechanism that selects the turns that are relevant and contributes the most to answer the current question. Experimentation results revealed that our framework is comparable in performance with the state-of-the-art models on the QuAC (http://quac.ai/) leader board. We also conduct a number of experiments to show the side effects of using entire context information which brings unnecessary information and noise signals resulting in a decline in the model’s performance. |
Keywords: | Machine comprehension; Information retrieval; Deep learning; Deep learning applications |
Description: | The Joint International Conference PDCAT-PAAP 2020, the 21st International Conference on Parallel and Distributed Computing, Applications and Technologies (PDCAT’20) and the 11th International Symposium on Parallel Architectures, Algorithms and Programming (PAAP’20) |
Rights: | © Springer Nature Singapore Pte Ltd. 2021 |
DOI: | 10.1007/978-981-16-0010-4_5 |
Published version: | https://link.springer.com/book/10.1007/978-981-16-0010-4 |
Appears in Collections: | Aurora harvest 4 Computer Science publications |
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.