HSSA tree structures for BTG-based preordering in machine translation

Yujia Zhang; Hao Wang; Yves Lepage

HSSA tree structures for BTG-based preordering in machine translation

Yujia Zhang, Hao Wang, Yves Lepage

Graduate School of Information, Production and Systems

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

Abstract

The Hierarchical Sub-Sentential Alignment (HSSA) method is a method to obtain aligned binary tree structures for two aligned sentences in translation correspondence. We propose to use the binary aligned tree structures delivered by this method as training data for preordering prior to machine translation. For that, we learn a Bracketing Transduction Grammar (BTG) from these binary aligned tree structures. In two oracle experiments in English to Japanese and Japanese to English translation, we show that it is theoretically possible to outperform a baseline system with a default distortion limit of 6, by about 2.5 and 5 BLEU points and, 7 and 10 RIBES points respectively, when preordering the source sentences using the learnt preordering model and using a distortion limit of 0. An attempt at learning a preordering model and its results are also reported.

Original language	English
Title of host publication	Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016
Editors	Jong C. Park, Jin-Woo Chung
Publisher	Institute for the Study of Language and Information
Pages	123-132
Number of pages	10
ISBN (Electronic)	9788968174285
Publication status	Published - 2016
Event	30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016 - Seoul, Korea, Republic of Duration: 2016 Oct 28 → 2016 Oct 30

Publication series

Name	Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016

Other

Other	30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016
Country	Korea, Republic of
City	Seoul
Period	16/10/28 → 16/10/30

ASJC Scopus subject areas

Language and Linguistics
Computer Science (miscellaneous)
Information Systems

Fingerprint Dive into the research topics of 'HSSA tree structures for BTG-based preordering in machine translation'. Together they form a unique fingerprint.

Cite this

Zhang, Y., Wang, H., & Lepage, Y. (2016). HSSA tree structures for BTG-based preordering in machine translation. In J. C. Park, & J-W. Chung (Eds.), Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016 (pp. 123-132). (Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016). Institute for the Study of Language and Information.

HSSA tree structures for BTG-based preordering in machine translation. / Zhang, Yujia; Wang, Hao; Lepage, Yves.

Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016. ed. / Jong C. Park; Jin-Woo Chung. Institute for the Study of Language and Information, 2016. p. 123-132 (Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016).

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

Zhang, Y, Wang, H & Lepage, Y 2016, HSSA tree structures for BTG-based preordering in machine translation. in JC Park & J-W Chung (eds), Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016. Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016, Institute for the Study of Language and Information, pp. 123-132, 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016, Seoul, Korea, Republic of, 16/10/28.

Zhang Y, Wang H, Lepage Y. HSSA tree structures for BTG-based preordering in machine translation. In Park JC, Chung J-W, editors, Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016. Institute for the Study of Language and Information. 2016. p. 123-132. (Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016).

Zhang, Yujia ; Wang, Hao ; Lepage, Yves. / HSSA tree structures for BTG-based preordering in machine translation. Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016. editor / Jong C. Park ; Jin-Woo Chung. Institute for the Study of Language and Information, 2016. pp. 123-132 (Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016).

@inproceedings{7c4711f73ae64e81921672856114cf1f,

title = "HSSA tree structures for BTG-based preordering in machine translation",

abstract = "The Hierarchical Sub-Sentential Alignment (HSSA) method is a method to obtain aligned binary tree structures for two aligned sentences in translation correspondence. We propose to use the binary aligned tree structures delivered by this method as training data for preordering prior to machine translation. For that, we learn a Bracketing Transduction Grammar (BTG) from these binary aligned tree structures. In two oracle experiments in English to Japanese and Japanese to English translation, we show that it is theoretically possible to outperform a baseline system with a default distortion limit of 6, by about 2.5 and 5 BLEU points and, 7 and 10 RIBES points respectively, when preordering the source sentences using the learnt preordering model and using a distortion limit of 0. An attempt at learning a preordering model and its results are also reported.",

author = "Yujia Zhang and Hao Wang and Yves Lepage",

note = "Funding Information: The second author is supported in part by China Scholarship Council (CSC) under CSC Grant No. 201406890026. We would like to thank Tetsuji Nakagawa for his most helpful comments on the experiment setting details.; 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016 ; Conference date: 28-10-2016 Through 30-10-2016",

year = "2016",

language = "English",

series = "Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016",

publisher = "Institute for the Study of Language and Information",

pages = "123--132",

editor = "Park, {Jong C.} and Jin-Woo Chung",

booktitle = "Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016",

}

TY - GEN

T1 - HSSA tree structures for BTG-based preordering in machine translation

AU - Zhang, Yujia

AU - Wang, Hao

AU - Lepage, Yves

N1 - Funding Information: The second author is supported in part by China Scholarship Council (CSC) under CSC Grant No. 201406890026. We would like to thank Tetsuji Nakagawa for his most helpful comments on the experiment setting details.

PY - 2016

Y1 - 2016

N2 - The Hierarchical Sub-Sentential Alignment (HSSA) method is a method to obtain aligned binary tree structures for two aligned sentences in translation correspondence. We propose to use the binary aligned tree structures delivered by this method as training data for preordering prior to machine translation. For that, we learn a Bracketing Transduction Grammar (BTG) from these binary aligned tree structures. In two oracle experiments in English to Japanese and Japanese to English translation, we show that it is theoretically possible to outperform a baseline system with a default distortion limit of 6, by about 2.5 and 5 BLEU points and, 7 and 10 RIBES points respectively, when preordering the source sentences using the learnt preordering model and using a distortion limit of 0. An attempt at learning a preordering model and its results are also reported.

AB - The Hierarchical Sub-Sentential Alignment (HSSA) method is a method to obtain aligned binary tree structures for two aligned sentences in translation correspondence. We propose to use the binary aligned tree structures delivered by this method as training data for preordering prior to machine translation. For that, we learn a Bracketing Transduction Grammar (BTG) from these binary aligned tree structures. In two oracle experiments in English to Japanese and Japanese to English translation, we show that it is theoretically possible to outperform a baseline system with a default distortion limit of 6, by about 2.5 and 5 BLEU points and, 7 and 10 RIBES points respectively, when preordering the source sentences using the learnt preordering model and using a distortion limit of 0. An attempt at learning a preordering model and its results are also reported.

UR - http://www.scopus.com/inward/record.url?scp=85015885084&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015885084&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:85015885084

T3 - Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016

SP - 123

EP - 132

BT - Proceedings of the 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016

A2 - Park, Jong C.

A2 - Chung, Jin-Woo

PB - Institute for the Study of Language and Information

T2 - 30th Pacific Asia Conference on Language, Information and Computation, PACLIC 2016

Y2 - 28 October 2016 through 30 October 2016

ER -