Active learning for deep semantic parsing

Long Duong; Hadi Afshar; Dominique Estival; Glen Pink; Philip Cohen; Mark Johnson

doi:10.18653/v1/p18-2008

Active learning for deep semantic parsing

Long Duong, Hadi Afshar, Dominique Estival, Glen Pink, Philip Cohen, Mark Johnson

Western Sydney University

Research output: Chapter in Book / Conference Paper › Conference Paper › peer-review

22 Citations (Scopus)

Abstract

Semantic parsing requires training data that is expensive and slow to collect. We apply active learning to both traditional and "overnight" data collection approaches. We show that it is possible to obtain good training hyperparameters from seed data which is only a small fraction of the full dataset. We show that uncertainty sampling based on least confidence score is competitive in traditional data collection but not applicable for overnight collection. We evaluate several active learning strategies for overnight data collection and show that different example selection strategies per domain perform best.

Original language	English
Title of host publication	The 56th Annual Meeting of the Association for Computational Linguistics: Proceedings of the Conference, Vol. 2 (Short Papers), July 15 - 20, 2018, Melbourne, Australia
Publisher	Association for Computational Linguistics
Pages	43-48
Number of pages	6
ISBN (Print)	9781948087346
DOIs	https://doi.org/10.18653/v1/p18-2008
Publication status	Published - 2018
Event	Association for Computational Linguistics. Meeting - Duration: 15 Jul 2018 → …

Conference

Conference	Association for Computational Linguistics. Meeting
Period	15/07/18 → …

Access to Document

10.18653/v1/p18-2008

Cite this

Duong, L., Afshar, H., Estival, D., Pink, G., Cohen, P., & Johnson, M. (2018). Active learning for deep semantic parsing. In The 56th Annual Meeting of the Association for Computational Linguistics: Proceedings of the Conference, Vol. 2 (Short Papers), July 15 - 20, 2018, Melbourne, Australia (pp. 43-48). Association for Computational Linguistics. https://doi.org/10.18653/v1/p18-2008

@inproceedings{845944e0462a4e59b1a6d26a746892b8,

title = "Active learning for deep semantic parsing",

abstract = "Semantic parsing requires training data that is expensive and slow to collect. We apply active learning to both traditional and {"}overnight{"} data collection approaches. We show that it is possible to obtain good training hyperparameters from seed data which is only a small fraction of the full dataset. We show that uncertainty sampling based on least confidence score is competitive in traditional data collection but not applicable for overnight collection. We evaluate several active learning strategies for overnight data collection and show that different example selection strategies per domain perform best.",

author = "Long Duong and Hadi Afshar and Dominique Estival and Glen Pink and Philip Cohen and Mark Johnson",

year = "2018",

doi = "10.18653/v1/p18-2008",

language = "English",

isbn = "9781948087346",

pages = "43--48",

booktitle = "The 56th Annual Meeting of the Association for Computational Linguistics: Proceedings of the Conference, Vol. 2 (Short Papers), July 15 - 20, 2018, Melbourne, Australia",

publisher = "Association for Computational Linguistics",

note = "Association for Computational Linguistics. Meeting ; Conference date: 15-07-2018",

}

Duong, L, Afshar, H, Estival, D, Pink, G, Cohen, P & Johnson, M 2018, Active learning for deep semantic parsing. in The 56th Annual Meeting of the Association for Computational Linguistics: Proceedings of the Conference, Vol. 2 (Short Papers), July 15 - 20, 2018, Melbourne, Australia. Association for Computational Linguistics, pp. 43-48, Association for Computational Linguistics. Meeting, 15/07/18. https://doi.org/10.18653/v1/p18-2008

Active learning for deep semantic parsing. / Duong, Long; Afshar, Hadi; Estival, Dominique et al.
The 56th Annual Meeting of the Association for Computational Linguistics: Proceedings of the Conference, Vol. 2 (Short Papers), July 15 - 20, 2018, Melbourne, Australia. Association for Computational Linguistics, 2018. p. 43-48.

Research output: Chapter in Book / Conference Paper › Conference Paper › peer-review

TY - GEN

T1 - Active learning for deep semantic parsing

AU - Duong, Long

AU - Afshar, Hadi

AU - Estival, Dominique

AU - Pink, Glen

AU - Cohen, Philip

AU - Johnson, Mark

PY - 2018

Y1 - 2018

N2 - Semantic parsing requires training data that is expensive and slow to collect. We apply active learning to both traditional and "overnight" data collection approaches. We show that it is possible to obtain good training hyperparameters from seed data which is only a small fraction of the full dataset. We show that uncertainty sampling based on least confidence score is competitive in traditional data collection but not applicable for overnight collection. We evaluate several active learning strategies for overnight data collection and show that different example selection strategies per domain perform best.

AB - Semantic parsing requires training data that is expensive and slow to collect. We apply active learning to both traditional and "overnight" data collection approaches. We show that it is possible to obtain good training hyperparameters from seed data which is only a small fraction of the full dataset. We show that uncertainty sampling based on least confidence score is competitive in traditional data collection but not applicable for overnight collection. We evaluate several active learning strategies for overnight data collection and show that different example selection strategies per domain perform best.

UR - https://hdl.handle.net/1959.7/uws:66931

U2 - 10.18653/v1/p18-2008

DO - 10.18653/v1/p18-2008

M3 - Conference Paper

SN - 9781948087346

SP - 43

EP - 48

BT - The 56th Annual Meeting of the Association for Computational Linguistics: Proceedings of the Conference, Vol. 2 (Short Papers), July 15 - 20, 2018, Melbourne, Australia

PB - Association for Computational Linguistics

T2 - Association for Computational Linguistics. Meeting

Y2 - 15 July 2018

ER -

Active learning for deep semantic parsing

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this