Sluice Resolution without Hand-Crafted Features over Brittle Syntax Trees

Ola Rønning, Daniel Hardt, Anders Søgaard

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

Abstract

Sluice resolution in English is the problem of finding antecedents of wh-fronted ellipses. Previous work has relied on handcrafted features over syntax trees that scale poorly to other languages and domains; in particular, to dialogue, which is one of the most interesting applications of sluice resolution. Syntactic information is arguably important for sluice resolution, but we show that multi-task learning with partial parsing as auxiliary tasks effectively closes the gap and buys us an additional 9% error reduction over previous work. Since we are not directly relying on features from partial parsers, our system is more robust to domain shifts, giving a 26% error reduction on embedded sluices in dialogue.
Sluice resolution in English is the problem of finding antecedents of wh-fronted ellipses. Previous work has relied on handcrafted features over syntax trees that scale poorly to other languages and domains; in particular, to dialogue, which is one of the most interesting applications of sluice resolution. Syntactic information is arguably important for sluice resolution, but we show that multi-task learning with partial parsing as auxiliary tasks effectively closes the gap and buys us an additional 9% error reduction over previous work. Since we are not directly relying on features from partial parsers, our system is more robust to domain shifts, giving a 26% error reduction on embedded sluices in dialogue.
LanguageEnglish
Title of host publicationProceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics : Human Language Technologies
EditorsStephanie Lukin, Margaret Mitchell
Volume2
Place of PublicationNew Orleans, LA
PublisherAssociation for Computational Linguistics
Date2018
Pages236–241
ISBN (Print)9781948087292
StatePublished - 2018
EventThe 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies - New Orleans, United States
Duration: 1 Jun 20186 Jun 2018
Conference number: 16
http://naacl2018.org/

Conference

ConferenceThe 16th Annual Conference of the North American Chapter of the Association for Computational Linguistics
Number16
CountryUnited States
CityNew Orleans
Period01/06/201806/06/2018
Internet address

Cite this

Rønning, O., Hardt, D., & Søgaard, A. (2018). Sluice Resolution without Hand-Crafted Features over Brittle Syntax Trees. In S. Lukin, & M. Mitchell (Eds.), Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Vol. 2, pp. 236–241). New Orleans, LA: Association for Computational Linguistics.
Rønning, Ola ; Hardt, Daniel ; Søgaard, Anders. / Sluice Resolution without Hand-Crafted Features over Brittle Syntax Trees. Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. editor / Stephanie Lukin ; Margaret Mitchell. Vol. 2 New Orleans, LA : Association for Computational Linguistics, 2018. pp. 236–241
@inproceedings{d43ae962a2264b32a45a71caecc5e178,
title = "Sluice Resolution without Hand-Crafted Features over Brittle Syntax Trees",
abstract = "Sluice resolution in English is the problem of finding antecedents of wh-fronted ellipses. Previous work has relied on handcrafted features over syntax trees that scale poorly to other languages and domains; in particular, to dialogue, which is one of the most interesting applications of sluice resolution. Syntactic information is arguably important for sluice resolution, but we show that multi-task learning with partial parsing as auxiliary tasks effectively closes the gap and buys us an additional 9{\%} error reduction over previous work. Since we are not directly relying on features from partial parsers, our system is more robust to domain shifts, giving a 26{\%} error reduction on embedded sluices in dialogue.",
author = "Ola R{\o}nning and Daniel Hardt and Anders S{\o}gaard",
year = "2018",
language = "English",
isbn = "9781948087292",
volume = "2",
pages = "236–241",
editor = "Stephanie Lukin and Margaret Mitchell",
booktitle = "Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics",
publisher = "Association for Computational Linguistics",
address = "United States",

}

Rønning, O, Hardt, D & Søgaard, A 2018, Sluice Resolution without Hand-Crafted Features over Brittle Syntax Trees. in S Lukin & M Mitchell (eds), Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. vol. 2, Association for Computational Linguistics, New Orleans, LA, pp. 236–241, New Orleans, United States, 01/06/2018.

Sluice Resolution without Hand-Crafted Features over Brittle Syntax Trees. / Rønning, Ola; Hardt, Daniel; Søgaard, Anders.

Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. ed. / Stephanie Lukin; Margaret Mitchell. Vol. 2 New Orleans, LA : Association for Computational Linguistics, 2018. p. 236–241.

Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

TY - GEN

T1 - Sluice Resolution without Hand-Crafted Features over Brittle Syntax Trees

AU - Rønning,Ola

AU - Hardt,Daniel

AU - Søgaard,Anders

PY - 2018

Y1 - 2018

N2 - Sluice resolution in English is the problem of finding antecedents of wh-fronted ellipses. Previous work has relied on handcrafted features over syntax trees that scale poorly to other languages and domains; in particular, to dialogue, which is one of the most interesting applications of sluice resolution. Syntactic information is arguably important for sluice resolution, but we show that multi-task learning with partial parsing as auxiliary tasks effectively closes the gap and buys us an additional 9% error reduction over previous work. Since we are not directly relying on features from partial parsers, our system is more robust to domain shifts, giving a 26% error reduction on embedded sluices in dialogue.

AB - Sluice resolution in English is the problem of finding antecedents of wh-fronted ellipses. Previous work has relied on handcrafted features over syntax trees that scale poorly to other languages and domains; in particular, to dialogue, which is one of the most interesting applications of sluice resolution. Syntactic information is arguably important for sluice resolution, but we show that multi-task learning with partial parsing as auxiliary tasks effectively closes the gap and buys us an additional 9% error reduction over previous work. Since we are not directly relying on features from partial parsers, our system is more robust to domain shifts, giving a 26% error reduction on embedded sluices in dialogue.

M3 - Article in proceedings

SN - 9781948087292

VL - 2

SP - 236

EP - 241

BT - Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics

PB - Association for Computational Linguistics

CY - New Orleans, LA

ER -

Rønning O, Hardt D, Søgaard A. Sluice Resolution without Hand-Crafted Features over Brittle Syntax Trees. In Lukin S, Mitchell M, editors, Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies. Vol. 2. New Orleans, LA: Association for Computational Linguistics. 2018. p. 236–241.