Jafari, R orcid.org/0000-0001-7298-2363, Razvarz, S and Gegov, A (2020) End-to-End Memory Networks: A Survey. In: Arai, K, Kapoor, S and Bhatia, R, (eds.) Advances in Intelligent Systems and Computing. SAI 2020 Computing Conference, 16-17 Jul 2020, Online. Springer Verlag , pp. 291-300. ISBN 978-3-030-52245-2
Abstract
Constructing a dialog system which can speak naturally with a human is considered as a major challenge of artificial intelligence. End-to-end dialog system is taken to be a primary research topic in the area of conversational systems. Since an end-to-end dialog system is structured based on learning a dialog policy from transactional dialogs in a defined extent, therefore, useful datasets are required for evaluating the learning procedures. In this paper, different deep learning techniques are applied to the Dialog bAbI datasets. On this dataset, the performance of the proposed techniques is analyzed. The performance results demonstrate that all the proposed techniques attain decent precisions on the Dialog bAbI datasets. The best performance is obtained utilizing end-to-end memory network with a unified weight tying scheme (UN2N).
Metadata
Item Type: | Proceedings Paper |
---|---|
Authors/Creators: |
|
Editors: |
|
Copyright, Publisher and Additional Information: | © Springer Nature Switzerland AG 2020. This is an author produced version of an article published in Advances in Intelligent Systems and Computing. Uploaded in accordance with the publisher's self-archiving policy. |
Keywords: | Memory networks; Deep learning; Dialog bAbI dataset |
Dates: |
|
Institution: | The University of Leeds |
Academic Units: | The University of Leeds > Faculty of Arts, Humanities and Cultures (Leeds) > School of Design (Leeds) |
Depositing User: | Symplectic Publications |
Date Deposited: | 27 Jan 2020 10:57 |
Last Modified: | 04 Jul 2021 00:38 |
Status: | Published |
Publisher: | Springer Verlag |
Identification Number: | 10.1007/978-3-030-52246-9_20 |
Open Archives Initiative ID (OAI ID): | oai:eprints.whiterose.ac.uk:156092 |