A Verb Lexicon Model with Source-side Syntactic Context for String-to-Tree Machine Translation

Maria Nadejde; Alexandra Birch; Philipp Koehn

A Verb Lexicon Model with Source-side Syntactic Context for String-to-Tree Machine Translation

Maria Nadejde, Alexandra Birch, Philipp Koehn

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

Abstract

String-to-tree MT systems translate verbs without lexical or syntactic context on the source side and with limited targetside context. The lack of context is one reason why verb translation recall is as low as 45.5%.
We propose a verb lexicon model trained with a feedforward neural network that predicts the target verb conditioned on a wide source-side context. We show that a syntactic context extracted from the dependency parse of the source sentence improves the model’s accuracy by 1.5% over a baseline trained on a window context.
When used as an extra feature for re-ranking the n-best list produced by the string-to-tree MT system, the verb lexicon model improves verb translation recall by more than 7%.

Original language	English
Title of host publication	Proceedings of the International Workshop on Spoken Language Translation (IWSLT)
Number of pages	9
Publication status	Published - 1 Dec 2016
Event	13th International Workshop on Spoken Language Translation 2016 - Seattle, United States Duration: 8 Dec 2016 → 9 Dec 2016 https://workshop2016.iwslt.org/

Conference

Conference	13th International Workshop on Spoken Language Translation 2016
Abbreviated title	IWSLT 2016
Country/Territory	United States
City	Seattle
Period	8/12/16 → 9/12/16
Internet address	https://workshop2016.iwslt.org/

Access to Document

IWSLT_2016_paper_10Final published version, 178 KB

http://workshop2016.iwslt.org/downloads/IWSLT_2016_paper_10.pdf

2 Finished

HimL: Health in my Language
Haddow, B., Birch-Mayne, A. & Webber, B.
EU government bodies
1/02/15 → 31/01/18
Project: Research
QT21: Quality Translation 21
Koehn, P., Haddow, B. & Huck, M.
EU government bodies
1/02/15 → 31/01/18
Project: Research

Cite this

@inproceedings{cfb167d9c2354cc380a9eecc7bd83be8,

title = "A Verb Lexicon Model with Source-side Syntactic Context for String-to-Tree Machine Translation",

abstract = "String-to-tree MT systems translate verbs without lexical or syntactic context on the source side and with limited targetside context. The lack of context is one reason why verb translation recall is as low as 45.5%.We propose a verb lexicon model trained with a feedforward neural network that predicts the target verb conditioned on a wide source-side context. We show that a syntactic context extracted from the dependency parse of the source sentence improves the model{\textquoteright}s accuracy by 1.5% over a baseline trained on a window context.When used as an extra feature for re-ranking the n-best list produced by the string-to-tree MT system, the verb lexicon model improves verb translation recall by more than 7%.",

author = "Maria Nadejde and Alexandra Birch and Philipp Koehn",

year = "2016",

month = dec,

day = "1",

language = "English",

booktitle = "Proceedings of the International Workshop on Spoken Language Translation (IWSLT)",

note = "13th International Workshop on Spoken Language Translation 2016, IWSLT 2016 ; Conference date: 08-12-2016 Through 09-12-2016",

url = "https://workshop2016.iwslt.org/",

}

Nadejde, M, Birch, A & Koehn, P 2016, A Verb Lexicon Model with Source-side Syntactic Context for String-to-Tree Machine Translation. in Proceedings of the International Workshop on Spoken Language Translation (IWSLT). 13th International Workshop on Spoken Language Translation 2016, Seattle, Washington, United States, 8/12/16. <http://workshop2016.iwslt.org/downloads/IWSLT_2016_paper_10.pdf>

TY - GEN

T1 - A Verb Lexicon Model with Source-side Syntactic Context for String-to-Tree Machine Translation

AU - Nadejde, Maria

AU - Birch, Alexandra

AU - Koehn, Philipp

PY - 2016/12/1

Y1 - 2016/12/1

N2 - String-to-tree MT systems translate verbs without lexical or syntactic context on the source side and with limited targetside context. The lack of context is one reason why verb translation recall is as low as 45.5%.We propose a verb lexicon model trained with a feedforward neural network that predicts the target verb conditioned on a wide source-side context. We show that a syntactic context extracted from the dependency parse of the source sentence improves the model’s accuracy by 1.5% over a baseline trained on a window context.When used as an extra feature for re-ranking the n-best list produced by the string-to-tree MT system, the verb lexicon model improves verb translation recall by more than 7%.

AB - String-to-tree MT systems translate verbs without lexical or syntactic context on the source side and with limited targetside context. The lack of context is one reason why verb translation recall is as low as 45.5%.We propose a verb lexicon model trained with a feedforward neural network that predicts the target verb conditioned on a wide source-side context. We show that a syntactic context extracted from the dependency parse of the source sentence improves the model’s accuracy by 1.5% over a baseline trained on a window context.When used as an extra feature for re-ranking the n-best list produced by the string-to-tree MT system, the verb lexicon model improves verb translation recall by more than 7%.

M3 - Conference contribution

BT - Proceedings of the International Workshop on Spoken Language Translation (IWSLT)

T2 - 13th International Workshop on Spoken Language Translation 2016

Y2 - 8 December 2016 through 9 December 2016

ER -

A Verb Lexicon Model with Source-side Syntactic Context for String-to-Tree Machine Translation

Abstract

Conference

Access to Document

Fingerprint

Projects

HimL: Health in my Language

QT21: Quality Translation 21

Cite this