A Joint Dependency Model of Morphological and Syntactic Structure for Statistical Machine Translation

Rico Sennrich; Barry Haddow

doi:10.18653/v1/D15-1248

A Joint Dependency Model of Morphological and Syntactic Structure for Statistical Machine Translation

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

Abstract

When translating between two languages that differ in their degree of morphological synthesis, syntactic structures in one language may be realized as morphological structures in the other, and SMT models need a mechanism to learn such translations. Prior work has used morpheme splitting with flat representations that do not encode the hierarchical structure between morphemes, but this structure is relevant for learning morphosyntactic constraints and selectional preferences. We propose to model syntactic and morphological structure jointly in a dependency translation model, allowing the system to generalize to the level of morphemes. We present a dependency representation of German compounds and particle verbs that results in improvements in translation quality of 1.4–1.8 BLEU in the WMT English–German translation task.

Original language	English
Title of host publication	Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing
Place of Publication	Lisbon, Portugal
Publisher	Association for Computational Linguistics
Pages	2081-2087
Number of pages	7
DOIs	https://doi.org/10.18653/v1/D15-1248
Publication status	Published - 1 Sept 2015
Event	2015 Conference on Empirical Methods in Natural Language Processing - Lisbon, Portugal Duration: 17 Sept 2015 → 21 Sept 2015 http://www.emnlp2015.org/

Conference

Conference	2015 Conference on Empirical Methods in Natural Language Processing
Abbreviated title	EMNLP 2015
Country/Territory	Portugal
City	Lisbon
Period	17/09/15 → 21/09/15
Internet address	http://www.emnlp2015.org/

Access to Document

10.18653/v1/D15-1248Licence: Creative Commons: Attribution (CC-BY)

D15-1248Final published version, 123 KBLicence: Creative Commons: Attribution (CC-BY)

http://aclweb.org/anthology/D15-1248.pdfLicence: Creative Commons: Attribution (CC-BY)

Translation for Massive Open Online Courses- TraMooc
Koehn, P. (Principal Investigator) & Birch-Mayne, A. (Co-investigator)
EU government bodies
1/02/15 → 31/01/18
Project: Research
QT21: Quality Translation 21
Koehn, P. (Principal Investigator), Haddow, B. (Co-investigator) & Huck, M. (Co-investigator)
EU government bodies
1/02/15 → 31/01/18
Project: Research
HimL: Health in my Language
Haddow, B. (Principal Investigator), Birch-Mayne, A. (Co-investigator) & Webber, B. (Co-investigator)
EU government bodies
1/02/15 → 31/01/18
Project: Research

Barry Haddow
- bhaddowinf.ed.acuk
- School of Informatics - Senior Research Fellow
- Institute of Language, Cognition and Computation
- Language, Interaction, and Robotics
Person: Academic: Research Active (Research Assistant)

Cite this

@inproceedings{b5530c5a5f474c8a82778b9e855d6cec,

title = "A Joint Dependency Model of Morphological and Syntactic Structure for Statistical Machine Translation",

abstract = "When translating between two languages that differ in their degree of morphological synthesis, syntactic structures in one language may be realized as morphological structures in the other, and SMT models need a mechanism to learn such translations. Prior work has used morpheme splitting with flat representations that do not encode the hierarchical structure between morphemes, but this structure is relevant for learning morphosyntactic constraints and selectional preferences. We propose to model syntactic and morphological structure jointly in a dependency translation model, allowing the system to generalize to the level of morphemes. We present a dependency representation of German compounds and particle verbs that results in improvements in translation quality of 1.4–1.8 BLEU in the WMT English–German translation task.",

author = "Rico Sennrich and Barry Haddow",

year = "2015",

month = sep,

day = "1",

doi = "10.18653/v1/D15-1248",

language = "English",

pages = "2081--2087",

booktitle = "Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing",

publisher = "Association for Computational Linguistics",

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

url = "http://www.emnlp2015.org/",

}

Sennrich, R & Haddow, B 2015, A Joint Dependency Model of Morphological and Syntactic Structure for Statistical Machine Translation. in Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, Lisbon, Portugal, pp. 2081-2087, 2015 Conference on Empirical Methods in Natural Language Processing , Lisbon, Portugal, 17/09/15. https://doi.org/10.18653/v1/D15-1248

A Joint Dependency Model of Morphological and Syntactic Structure for Statistical Machine Translation. / Sennrich, Rico; Haddow, Barry.
Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing. Lisbon, Portugal: Association for Computational Linguistics, 2015. p. 2081-2087.

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

TY - GEN

T1 - A Joint Dependency Model of Morphological and Syntactic Structure for Statistical Machine Translation

AU - Sennrich, Rico

AU - Haddow, Barry

PY - 2015/9/1

Y1 - 2015/9/1

N2 - When translating between two languages that differ in their degree of morphological synthesis, syntactic structures in one language may be realized as morphological structures in the other, and SMT models need a mechanism to learn such translations. Prior work has used morpheme splitting with flat representations that do not encode the hierarchical structure between morphemes, but this structure is relevant for learning morphosyntactic constraints and selectional preferences. We propose to model syntactic and morphological structure jointly in a dependency translation model, allowing the system to generalize to the level of morphemes. We present a dependency representation of German compounds and particle verbs that results in improvements in translation quality of 1.4–1.8 BLEU in the WMT English–German translation task.

AB - When translating between two languages that differ in their degree of morphological synthesis, syntactic structures in one language may be realized as morphological structures in the other, and SMT models need a mechanism to learn such translations. Prior work has used morpheme splitting with flat representations that do not encode the hierarchical structure between morphemes, but this structure is relevant for learning morphosyntactic constraints and selectional preferences. We propose to model syntactic and morphological structure jointly in a dependency translation model, allowing the system to generalize to the level of morphemes. We present a dependency representation of German compounds and particle verbs that results in improvements in translation quality of 1.4–1.8 BLEU in the WMT English–German translation task.

U2 - 10.18653/v1/D15-1248

DO - 10.18653/v1/D15-1248

M3 - Conference contribution

SP - 2081

EP - 2087

BT - Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing

PB - Association for Computational Linguistics

CY - Lisbon, Portugal

T2 - 2015 Conference on Empirical Methods in Natural Language Processing

Y2 - 17 September 2015 through 21 September 2015

ER -

A Joint Dependency Model of Morphological and Syntactic Structure for Statistical Machine Translation

Abstract

Conference

Access to Document

Fingerprint

Projects

Translation for Massive Open Online Courses- TraMooc

QT21: Quality Translation 21

HimL: Health in my Language

Profiles

Barry Haddow

Cite this