Multi-task Learning of Pairwise Sequence Classification Tasks Over Disparate Label Spaces

Isabelle Augenstein; Sebastian  Ruder; Anders Søgaard

doi:10.18653/v1/N18-1172

Multi-task Learning of Pairwise Sequence Classification Tasks Over Disparate Label Spaces

Isabelle Augenstein, Sebastian Ruder, Anders Søgaard

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

381 Downloads (Pure)

Abstract

We combine multi-task learning and semi-supervised learning by inducing a joint embedding space between disparate label spaces and learning transfer functions between label embeddings, enabling us to jointly leverage unlabelled data and auxiliary, annotated datasets. We evaluate our approach on a variety of tasks with disparate label spaces. We outperform strong single and multi-task baselines and achieve a new state of the art for aspect-based and topic-based sentiment analysis.

Original language	English
Title of host publication	Proceedings, 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies : (Long Papers)
Number of pages	11
Volume	1
Publisher	Association for Computational Linguistics
Publication date	2018
Pages	1896–1906
DOIs	https://doi.org/10.18653/v1/N18-1172
Publication status	Published - 2018
Event	16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies - New Orleans, United States Duration: 1 Jun 2018 → 6 Jun 2018

Conference

Conference	16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies
Country/Territory	United States
City	New Orleans
Period	01/06/2018 → 06/06/2018

Access to Document

10.18653/v1/N18-1172

AugensteinSøgaardSubmitted manuscript, 742 KB

https://arxiv.org/abs/1802.09913

Cite this

Augenstein, I., Ruder, S., & Søgaard, A. (2018). Multi-task Learning of Pairwise Sequence Classification Tasks Over Disparate Label Spaces. In Proceedings, 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies : (Long Papers) (Vol. 1, pp. 1896–1906). Association for Computational Linguistics. https://doi.org/10.18653/v1/N18-1172

Multi-task Learning of Pairwise Sequence Classification Tasks Over Disparate Label Spaces. / Augenstein, Isabelle; Ruder, Sebastian ; Søgaard, Anders.

Proceedings, 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies : (Long Papers). Vol. 1 Association for Computational Linguistics, 2018. p. 1896–1906.

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

Augenstein, I, Ruder, S & Søgaard, A 2018, Multi-task Learning of Pairwise Sequence Classification Tasks Over Disparate Label Spaces. in Proceedings, 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies : (Long Papers). vol. 1, Association for Computational Linguistics, pp. 1896–1906, 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, New Orleans, United States, 01/06/2018. https://doi.org/10.18653/v1/N18-1172

Augenstein I, Ruder S, Søgaard A. Multi-task Learning of Pairwise Sequence Classification Tasks Over Disparate Label Spaces. In Proceedings, 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies : (Long Papers). Vol. 1. Association for Computational Linguistics. 2018. p. 1896–1906 https://doi.org/10.18653/v1/N18-1172

Augenstein, Isabelle ; Ruder, Sebastian ; Søgaard, Anders. / Multi-task Learning of Pairwise Sequence Classification Tasks Over Disparate Label Spaces. Proceedings, 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies : (Long Papers). Vol. 1 Association for Computational Linguistics, 2018. pp. 1896–1906

@inproceedings{549949285482479cb402e683b9b39c43,

title = "Multi-task Learning of Pairwise Sequence Classification Tasks Over Disparate Label Spaces",

abstract = "We combine multi-task learning and semi-supervised learning by inducing a joint embedding space between disparate label spaces and learning transfer functions between label embeddings, enabling us to jointly leverage unlabelled data and auxiliary, annotated datasets. We evaluate our approach on a variety of tasks with disparate label spaces. We outperform strong single and multi-task baselines and achieve a new state of the art for aspect-based and topic-based sentiment analysis.",

author = "Isabelle Augenstein and Sebastian Ruder and Anders S{\o}gaard",

year = "2018",

doi = "10.18653/v1/N18-1172",

language = "English",

volume = "1",

pages = "1896–1906",

booktitle = "Proceedings, 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies",

publisher = "Association for Computational Linguistics",

note = "16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL 2018 ; Conference date: 01-06-2018 Through 06-06-2018",

}

TY - GEN

T1 - Multi-task Learning of Pairwise Sequence Classification Tasks Over Disparate Label Spaces

AU - Augenstein, Isabelle

AU - Ruder, Sebastian

AU - Søgaard, Anders

PY - 2018

Y1 - 2018

N2 - We combine multi-task learning and semi-supervised learning by inducing a joint embedding space between disparate label spaces and learning transfer functions between label embeddings, enabling us to jointly leverage unlabelled data and auxiliary, annotated datasets. We evaluate our approach on a variety of tasks with disparate label spaces. We outperform strong single and multi-task baselines and achieve a new state of the art for aspect-based and topic-based sentiment analysis.

AB - We combine multi-task learning and semi-supervised learning by inducing a joint embedding space between disparate label spaces and learning transfer functions between label embeddings, enabling us to jointly leverage unlabelled data and auxiliary, annotated datasets. We evaluate our approach on a variety of tasks with disparate label spaces. We outperform strong single and multi-task baselines and achieve a new state of the art for aspect-based and topic-based sentiment analysis.

U2 - 10.18653/v1/N18-1172

DO - 10.18653/v1/N18-1172

M3 - Article in proceedings

VL - 1

SP - 1896

EP - 1906

BT - Proceedings, 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies

PB - Association for Computational Linguistics

T2 - 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies

Y2 - 1 June 2018 through 6 June 2018

ER -