A neural network model for low-resource universal dependency parsing

Long Duong; Trevor Cohn; Steven Bird; Paul Cook

A neural network model for low-resource universal dependency parsing

Long Duong, Trevor Cohn, Steven Bird, Paul Cook

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper published in Proceedings › peer-review

Abstract

Accurate dependency parsing requires large treebanks, which are only available for a few languages. We propose a method that takes advantage of shared structure across languages to build a mature parser using less training data. We propose a model for learning a shared "universal" parser that operates over an interlingual continuous representation of language, along with language-specific mapping components. Compared with supervised learning, our methods give a consistent 8-10% improvement across several treebanks in low-resource simulations.

Original language	English
Title of host publication	Conference Proceedings - EMNLP 2015
Subtitle of host publication	Conference on Empirical Methods in Natural Language Processing
Place of Publication	Lisbon, Portugal
Publisher	Association for Computational Linguistics (ACL)
Pages	339-348
Number of pages	10
ISBN (Electronic)	9781941643327
Publication status	Published - 2015
Externally published	Yes
Event	Conference on Empirical Methods in Natural Language Processing, EMNLP 2015 - Lisbon, Portugal Duration: 17 Sep 2015 → 21 Sep 2015

Conference

Conference	Conference on Empirical Methods in Natural Language Processing, EMNLP 2015
Country/Territory	Portugal
City	Lisbon
Period	17/09/15 → 21/09/15

Access to Document

http://www.emnlp2015.org/proceedings/EMNLP/pdf/EMNLP040.pdf

Cite this

@inproceedings{f3cc60e8135448c7b4ce3c7144168468,

title = "A neural network model for low-resource universal dependency parsing",

abstract = "Accurate dependency parsing requires large treebanks, which are only available for a few languages. We propose a method that takes advantage of shared structure across languages to build a mature parser using less training data. We propose a model for learning a shared {"}universal{"} parser that operates over an interlingual continuous representation of language, along with language-specific mapping components. Compared with supervised learning, our methods give a consistent 8-10% improvement across several treebanks in low-resource simulations.",

author = "Long Duong and Trevor Cohn and Steven Bird and Paul Cook",

year = "2015",

language = "English",

pages = "339--348",

booktitle = "Conference Proceedings - EMNLP 2015",

publisher = "Association for Computational Linguistics (ACL)",

note = "Conference on Empirical Methods in Natural Language Processing, EMNLP 2015 ; Conference date: 17-09-2015 Through 21-09-2015",

}

Duong, L, Cohn, T, Bird, S & Cook, P 2015, A neural network model for low-resource universal dependency parsing. in Conference Proceedings - EMNLP 2015: Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics (ACL), Lisbon, Portugal, pp. 339-348, Conference on Empirical Methods in Natural Language Processing, EMNLP 2015, Lisbon, Portugal, 17/09/15. <http://www.emnlp2015.org/proceedings/EMNLP/pdf/EMNLP040.pdf>

A neural network model for low-resource universal dependency parsing. / Duong, Long; Cohn, Trevor; Bird, Steven et al.

Conference Proceedings - EMNLP 2015: Conference on Empirical Methods in Natural Language Processing. Lisbon, Portugal : Association for Computational Linguistics (ACL), 2015. p. 339-348.

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper published in Proceedings › peer-review

TY - GEN

T1 - A neural network model for low-resource universal dependency parsing

AU - Duong, Long

AU - Cohn, Trevor

AU - Bird, Steven

AU - Cook, Paul

PY - 2015

Y1 - 2015

N2 - Accurate dependency parsing requires large treebanks, which are only available for a few languages. We propose a method that takes advantage of shared structure across languages to build a mature parser using less training data. We propose a model for learning a shared "universal" parser that operates over an interlingual continuous representation of language, along with language-specific mapping components. Compared with supervised learning, our methods give a consistent 8-10% improvement across several treebanks in low-resource simulations.

AB - Accurate dependency parsing requires large treebanks, which are only available for a few languages. We propose a method that takes advantage of shared structure across languages to build a mature parser using less training data. We propose a model for learning a shared "universal" parser that operates over an interlingual continuous representation of language, along with language-specific mapping components. Compared with supervised learning, our methods give a consistent 8-10% improvement across several treebanks in low-resource simulations.

UR - http://www.scopus.com/inward/record.url?scp=84959925904&partnerID=8YFLogxK

UR - http://www.emnlp2015.org/proceedings/EMNLP/

M3 - Conference Paper published in Proceedings

AN - SCOPUS:84959925904

SP - 339

EP - 348

BT - Conference Proceedings - EMNLP 2015

PB - Association for Computational Linguistics (ACL)

CY - Lisbon, Portugal

T2 - Conference on Empirical Methods in Natural Language Processing, EMNLP 2015

Y2 - 17 September 2015 through 21 September 2015

ER -

A neural network model for low-resource universal dependency parsing

Abstract

Conference

Access to Document

Other files and links

Fingerprint

Cite this