NTT’s Neural Machine Translation Systems for WMT 2018

Makoto Morishita; Jun Suzuki; Masaaki Nagata

doi:10.18653/v1/W18-64048

NTT’s Neural Machine Translation Systems for WMT 2018

Makoto Morishita, Jun Suzuki, Masaaki Nagata

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

8 Citations (Scopus)

Abstract

This paper describes NTT’s neural machine translation systems submitted to the WMT 2018 English-German and German-English news translation tasks. Our submission has three main components: the Transformer model, corpus cleaning, and right-to-left nbest re-ranking techniques. Through our experiments, we identified two keys for improving accuracy: filtering noisy training sentences and right-to-left re-ranking. We also found that the Transformer model requires more training data than the RNN-based model, and the RNN-based model sometimes achieves better accuracy than the Transformer model when the corpus is small.

Original language	English
Title of host publication	Shared Task Papers
Publisher	Association for Computational Linguistics (ACL)
Pages	461-466
Number of pages	6
ISBN (Electronic)	9781948087810
DOIs	https://doi.org/10.18653/v1/W18-64048
Publication status	Published - 2018
Externally published	Yes
Event	3rd Conference on Machine Translation, WMT 2018 at the Conference on Empirical Methods in Natural Language Processing, EMNLP 2018 - Brussels, Belgium Duration: 2018 Oct 31 → 2018 Nov 1

Publication series

Name	WMT 2018 - 3rd Conference on Machine Translation, Proceedings of the Conference
Volume	2

Conference

Conference	3rd Conference on Machine Translation, WMT 2018 at the Conference on Empirical Methods in Natural Language Processing, EMNLP 2018
Country/Territory	Belgium
City	Brussels
Period	18/10/31 → 18/11/1

ASJC Scopus subject areas

Computer Science Applications
Information Systems
Computational Theory and Mathematics

Access to Document

10.18653/v1/W18-64048

Cite this

NTT’s Neural Machine Translation Systems for WMT 2018. / Morishita, Makoto; Suzuki, Jun; Nagata, Masaaki.

Shared Task Papers. Association for Computational Linguistics (ACL), 2018. p. 461-466 (WMT 2018 - 3rd Conference on Machine Translation, Proceedings of the Conference; Vol. 2).

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

Morishita, M, Suzuki, J & Nagata, M 2018, NTT’s Neural Machine Translation Systems for WMT 2018. in Shared Task Papers. WMT 2018 - 3rd Conference on Machine Translation, Proceedings of the Conference, vol. 2, Association for Computational Linguistics (ACL), pp. 461-466, 3rd Conference on Machine Translation, WMT 2018 at the Conference on Empirical Methods in Natural Language Processing, EMNLP 2018, Brussels, Belgium, 18/10/31. https://doi.org/10.18653/v1/W18-64048

@inproceedings{f921dd0b1525446c92c869e409d404ad,

title = "NTT{\textquoteright}s Neural Machine Translation Systems for WMT 2018",

abstract = "This paper describes NTT{\textquoteright}s neural machine translation systems submitted to the WMT 2018 English-German and German-English news translation tasks. Our submission has three main components: the Transformer model, corpus cleaning, and right-to-left nbest re-ranking techniques. Through our experiments, we identified two keys for improving accuracy: filtering noisy training sentences and right-to-left re-ranking. We also found that the Transformer model requires more training data than the RNN-based model, and the RNN-based model sometimes achieves better accuracy than the Transformer model when the corpus is small.",

author = "Makoto Morishita and Jun Suzuki and Masaaki Nagata",

note = "Publisher Copyright: {\textcopyright}2018 Association for Computational Linguistics; 3rd Conference on Machine Translation, WMT 2018 at the Conference on Empirical Methods in Natural Language Processing, EMNLP 2018 ; Conference date: 31-10-2018 Through 01-11-2018",

year = "2018",

doi = "10.18653/v1/W18-64048",

language = "English",

series = "WMT 2018 - 3rd Conference on Machine Translation, Proceedings of the Conference",

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

pages = "461--466",

booktitle = "Shared Task Papers",

}

TY - GEN

T1 - NTT’s Neural Machine Translation Systems for WMT 2018

AU - Morishita, Makoto

AU - Suzuki, Jun

AU - Nagata, Masaaki

PY - 2018

Y1 - 2018

N2 - This paper describes NTT’s neural machine translation systems submitted to the WMT 2018 English-German and German-English news translation tasks. Our submission has three main components: the Transformer model, corpus cleaning, and right-to-left nbest re-ranking techniques. Through our experiments, we identified two keys for improving accuracy: filtering noisy training sentences and right-to-left re-ranking. We also found that the Transformer model requires more training data than the RNN-based model, and the RNN-based model sometimes achieves better accuracy than the Transformer model when the corpus is small.

AB - This paper describes NTT’s neural machine translation systems submitted to the WMT 2018 English-German and German-English news translation tasks. Our submission has three main components: the Transformer model, corpus cleaning, and right-to-left nbest re-ranking techniques. Through our experiments, we identified two keys for improving accuracy: filtering noisy training sentences and right-to-left re-ranking. We also found that the Transformer model requires more training data than the RNN-based model, and the RNN-based model sometimes achieves better accuracy than the Transformer model when the corpus is small.

UR - http://www.scopus.com/inward/record.url?scp=85094766967&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85094766967&partnerID=8YFLogxK

U2 - 10.18653/v1/W18-64048

DO - 10.18653/v1/W18-64048

M3 - Conference contribution

AN - SCOPUS:85094766967

T3 - WMT 2018 - 3rd Conference on Machine Translation, Proceedings of the Conference

SP - 461

EP - 466

BT - Shared Task Papers

PB - Association for Computational Linguistics (ACL)

T2 - 3rd Conference on Machine Translation, WMT 2018 at the Conference on Empirical Methods in Natural Language Processing, EMNLP 2018

Y2 - 31 October 2018 through 1 November 2018

ER -

NTT’s Neural Machine Translation Systems for WMT 2018

Abstract

Publication series

Conference

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this