End-to-end task-oriented dialogue system with distantly supervised knowledge base retriever

Libo Qin; Yijia Liu; Wanxiang Che; Haoyang Wen; Ting Liu

doi:10.1007/978-3-030-01716-3_20

End-to-end task-oriented dialogue system with distantly supervised knowledge base retriever

Libo Qin
, Yijia Liu
, Wanxiang Che^*
, Haoyang Wen
, Ting Liu

^*Corresponding author for this work

Faculty of Computing

Harbin Institute of Technology

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

2 Scopus citations

Abstract

Task-oriented dialog systems usually face the challenge of querying knowledge base. However, it usually cannot be explicitly modeled due to the lack of annotation. In this paper, we introduce an explicit KB retrieval component (KB retriever) into the seq2seq dialogue system. We first use the KB retriever to get the most relevant entry according to the dialogue history and KB, and then apply the copying mechanism to retrieve entities from the retrieved KB in decoding time. Moreover, the KB retriever is trained with distant supervision, which does not need any annotation efforts. Experiments on Stanford Multi-turn Task-oriented Dialogue Dataset shows that our framework significantly outperforms other sequence-to-sequence based baseline models on both automatic and human evaluation.

Original language	English
Title of host publication	Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data - 17th China National Conference, CCL 2018, and 6th International Symposium, NLP-NABD 2018, Proceedings
Editors	Xiaojie Wang, Ting Liu, Maosong Sun, Zhiyuan Liu, Yang Liu
Publisher	Springer Verlag
Pages	238-249
Number of pages	12
ISBN (Print)	9783030017156
DOIs	https://doi.org/10.1007/978-3-030-01716-3_20
State	Published - 2018
Event	17th China National Conference on Computational Linguistics, CCL 2018 and 6th International Symposium on Natural Language Processing Based on Naturally Annotated Big Data, NLP-NABD 2018 - Changsha, China Duration: 19 Oct 2018 → 21 Oct 2018

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	11221 LNAI
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	17th China National Conference on Computational Linguistics, CCL 2018 and 6th International Symposium on Natural Language Processing Based on Naturally Annotated Big Data, NLP-NABD 2018
Country/Territory	China
City	Changsha
Period	19/10/18 → 21/10/18

Keywords

Knowledge base
Sequence-to-sequence
Task-oriented dialog systems

Access to Document

10.1007/978-3-030-01716-3_20

Cite this

Qin, L., Liu, Y., Che, W., Wen, H., & Liu, T. (2018). End-to-end task-oriented dialogue system with distantly supervised knowledge base retriever. In X. Wang, T. Liu, M. Sun, Z. Liu, & Y. Liu (Eds.), Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data - 17th China National Conference, CCL 2018, and 6th International Symposium, NLP-NABD 2018, Proceedings (pp. 238-249). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11221 LNAI). Springer Verlag. https://doi.org/10.1007/978-3-030-01716-3_20

Qin, Libo ; Liu, Yijia ; Che, Wanxiang et al. / End-to-end task-oriented dialogue system with distantly supervised knowledge base retriever. Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data - 17th China National Conference, CCL 2018, and 6th International Symposium, NLP-NABD 2018, Proceedings. editor / Xiaojie Wang ; Ting Liu ; Maosong Sun ; Zhiyuan Liu ; Yang Liu. Springer Verlag, 2018. pp. 238-249 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{cb955760c93844d38d9eafec43cef28f,

title = "End-to-end task-oriented dialogue system with distantly supervised knowledge base retriever",

abstract = "Task-oriented dialog systems usually face the challenge of querying knowledge base. However, it usually cannot be explicitly modeled due to the lack of annotation. In this paper, we introduce an explicit KB retrieval component (KB retriever) into the seq2seq dialogue system. We first use the KB retriever to get the most relevant entry according to the dialogue history and KB, and then apply the copying mechanism to retrieve entities from the retrieved KB in decoding time. Moreover, the KB retriever is trained with distant supervision, which does not need any annotation efforts. Experiments on Stanford Multi-turn Task-oriented Dialogue Dataset shows that our framework significantly outperforms other sequence-to-sequence based baseline models on both automatic and human evaluation.",

keywords = "Knowledge base, Sequence-to-sequence, Task-oriented dialog systems",

author = "Libo Qin and Yijia Liu and Wanxiang Che and Haoyang Wen and Ting Liu",

note = "Publisher Copyright: {\textcopyright} Springer Nature Switzerland AG 2018.; 17th China National Conference on Computational Linguistics, CCL 2018 and 6th International Symposium on Natural Language Processing Based on Naturally Annotated Big Data, NLP-NABD 2018 ; Conference date: 19-10-2018 Through 21-10-2018",

year = "2018",

doi = "10.1007/978-3-030-01716-3\_20",

language = "英语",

isbn = "9783030017156",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "238--249",

editor = "Xiaojie Wang and Ting Liu and Maosong Sun and Zhiyuan Liu and Yang Liu",

booktitle = "Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data - 17th China National Conference, CCL 2018, and 6th International Symposium, NLP-NABD 2018, Proceedings",

address = "德国",

}

Qin, L, Liu, Y, Che, W, Wen, H & Liu, T 2018, End-to-end task-oriented dialogue system with distantly supervised knowledge base retriever. in X Wang, T Liu, M Sun, Z Liu & Y Liu (eds), Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data - 17th China National Conference, CCL 2018, and 6th International Symposium, NLP-NABD 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11221 LNAI, Springer Verlag, pp. 238-249, 17th China National Conference on Computational Linguistics, CCL 2018 and 6th International Symposium on Natural Language Processing Based on Naturally Annotated Big Data, NLP-NABD 2018, Changsha, China, 19/10/18. https://doi.org/10.1007/978-3-030-01716-3_20

End-to-end task-oriented dialogue system with distantly supervised knowledge base retriever. / Qin, Libo; Liu, Yijia; Che, Wanxiang et al.
Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data - 17th China National Conference, CCL 2018, and 6th International Symposium, NLP-NABD 2018, Proceedings. ed. / Xiaojie Wang; Ting Liu; Maosong Sun; Zhiyuan Liu; Yang Liu. Springer Verlag, 2018. p. 238-249 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11221 LNAI).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - End-to-end task-oriented dialogue system with distantly supervised knowledge base retriever

AU - Qin, Libo

AU - Liu, Yijia

AU - Che, Wanxiang

AU - Wen, Haoyang

AU - Liu, Ting

N1 - Publisher Copyright: © Springer Nature Switzerland AG 2018.

PY - 2018

Y1 - 2018

N2 - Task-oriented dialog systems usually face the challenge of querying knowledge base. However, it usually cannot be explicitly modeled due to the lack of annotation. In this paper, we introduce an explicit KB retrieval component (KB retriever) into the seq2seq dialogue system. We first use the KB retriever to get the most relevant entry according to the dialogue history and KB, and then apply the copying mechanism to retrieve entities from the retrieved KB in decoding time. Moreover, the KB retriever is trained with distant supervision, which does not need any annotation efforts. Experiments on Stanford Multi-turn Task-oriented Dialogue Dataset shows that our framework significantly outperforms other sequence-to-sequence based baseline models on both automatic and human evaluation.

AB - Task-oriented dialog systems usually face the challenge of querying knowledge base. However, it usually cannot be explicitly modeled due to the lack of annotation. In this paper, we introduce an explicit KB retrieval component (KB retriever) into the seq2seq dialogue system. We first use the KB retriever to get the most relevant entry according to the dialogue history and KB, and then apply the copying mechanism to retrieve entities from the retrieved KB in decoding time. Moreover, the KB retriever is trained with distant supervision, which does not need any annotation efforts. Experiments on Stanford Multi-turn Task-oriented Dialogue Dataset shows that our framework significantly outperforms other sequence-to-sequence based baseline models on both automatic and human evaluation.

KW - Knowledge base

KW - Sequence-to-sequence

KW - Task-oriented dialog systems

UR - https://www.scopus.com/pages/publications/85055420457

U2 - 10.1007/978-3-030-01716-3_20

DO - 10.1007/978-3-030-01716-3_20

M3 - 会议稿件

AN - SCOPUS:85055420457

SN - 9783030017156

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 238

EP - 249

BT - Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data - 17th China National Conference, CCL 2018, and 6th International Symposium, NLP-NABD 2018, Proceedings

A2 - Wang, Xiaojie

A2 - Liu, Ting

A2 - Sun, Maosong

A2 - Liu, Zhiyuan

A2 - Liu, Yang

PB - Springer Verlag

T2 - 17th China National Conference on Computational Linguistics, CCL 2018 and 6th International Symposium on Natural Language Processing Based on Naturally Annotated Big Data, NLP-NABD 2018

Y2 - 19 October 2018 through 21 October 2018

ER -

Qin L, Liu Y, Che W, Wen H, Liu T. End-to-end task-oriented dialogue system with distantly supervised knowledge base retriever. In Wang X, Liu T, Sun M, Liu Z, Liu Y, editors, Chinese Computational Linguistics and Natural Language Processing Based on Naturally Annotated Big Data - 17th China National Conference, CCL 2018, and 6th International Symposium, NLP-NABD 2018, Proceedings. Springer Verlag. 2018. p. 238-249. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-030-01716-3_20

End-to-end task-oriented dialogue system with distantly supervised knowledge base retriever

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this