Letting a neural network decide which machine translation system to use for black-box fuzzy-match repair

John E. Ortega; Weiyi Lu; Adam Meyers; Kyunghyun Cho

Letting a neural network decide which machine translation system to use for black-box fuzzy-match repair

John E. Ortega, Weiyi Lu, Adam Meyers, Kyunghyun Cho

Computer Science

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

Abstract

While systems using the Neural Network-based Machine Translation (NMT) paradigm achieve the highest scores on recent shared tasks, phrase-based (PBMT) systems, rule-based (RBMT) systems and other systems may get better results for individual examples. Therefore, combined systems should achieve the best results for MT, particularly if the system combination method can take advantage of the strengths of each paradigm. In this paper, we describe a system that predicts whether a NMT, PBMT or RBMT will get the best Spanish translation result for a particular English sentence in DGT-TM 2016¹. Then we use fuzzy-match repair (FMR) as a mechanism to show that the combined system outperforms individual systems in a black-box machine translation setting.

Original language	English (US)
Title of host publication	EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation
Editors	Juan Antonio Perez-Ortiz, Felipe Sanchez-Martinez, Miquel Espla-Gomis, Maja Popovic, Celia Rico, Andre Martins, Joachim Van den Bogaert, Mikel L. Forcada
Publisher	European Association for Machine Translation
Pages	209-218
Number of pages	10
ISBN (Electronic)	9788409019014
State	Published - Jan 1 2018
Event	21st Annual Conference of the European Association for Machine Translation, EAMT 2018 - Alacant, Spain Duration: May 28 2018 → May 30 2018

Publication series

Name	EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation

Conference

Conference	21st Annual Conference of the European Association for Machine Translation, EAMT 2018
Country/Territory	Spain
City	Alacant
Period	5/28/18 → 5/30/18

ASJC Scopus subject areas

Human-Computer Interaction
Software
Language and Linguistics

Cite this

Ortega, J. E., Lu, W., Meyers, A., & Cho, K. (2018). Letting a neural network decide which machine translation system to use for black-box fuzzy-match repair. In J. A. Perez-Ortiz, F. Sanchez-Martinez, M. Espla-Gomis, M. Popovic, C. Rico, A. Martins, J. Van den Bogaert, & M. L. Forcada (Eds.), EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation (pp. 209-218). (EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation). European Association for Machine Translation.

Letting a neural network decide which machine translation system to use for black-box fuzzy-match repair. / Ortega, John E.; Lu, Weiyi; Meyers, Adam et al.
EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation. ed. / Juan Antonio Perez-Ortiz; Felipe Sanchez-Martinez; Miquel Espla-Gomis; Maja Popovic; Celia Rico; Andre Martins; Joachim Van den Bogaert; Mikel L. Forcada. European Association for Machine Translation, 2018. p. 209-218 (EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation).

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

Ortega, JE, Lu, W, Meyers, A & Cho, K 2018, Letting a neural network decide which machine translation system to use for black-box fuzzy-match repair. in JA Perez-Ortiz, F Sanchez-Martinez, M Espla-Gomis, M Popovic, C Rico, A Martins, J Van den Bogaert & ML Forcada (eds), EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation. EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation, European Association for Machine Translation, pp. 209-218, 21st Annual Conference of the European Association for Machine Translation, EAMT 2018, Alacant, Spain, 5/28/18.

Ortega JE, Lu W, Meyers A, Cho K. Letting a neural network decide which machine translation system to use for black-box fuzzy-match repair. In Perez-Ortiz JA, Sanchez-Martinez F, Espla-Gomis M, Popovic M, Rico C, Martins A, Van den Bogaert J, Forcada ML, editors, EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation. European Association for Machine Translation. 2018. p. 209-218. (EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation).

Ortega, John E. ; Lu, Weiyi ; Meyers, Adam et al. / Letting a neural network decide which machine translation system to use for black-box fuzzy-match repair. EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation. editor / Juan Antonio Perez-Ortiz ; Felipe Sanchez-Martinez ; Miquel Espla-Gomis ; Maja Popovic ; Celia Rico ; Andre Martins ; Joachim Van den Bogaert ; Mikel L. Forcada. European Association for Machine Translation, 2018. pp. 209-218 (EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation).

@inproceedings{fc9c680e93ff4b848db985e68eaab9e2,

title = "Letting a neural network decide which machine translation system to use for black-box fuzzy-match repair",

abstract = "While systems using the Neural Network-based Machine Translation (NMT) paradigm achieve the highest scores on recent shared tasks, phrase-based (PBMT) systems, rule-based (RBMT) systems and other systems may get better results for individual examples. Therefore, combined systems should achieve the best results for MT, particularly if the system combination method can take advantage of the strengths of each paradigm. In this paper, we describe a system that predicts whether a NMT, PBMT or RBMT will get the best Spanish translation result for a particular English sentence in DGT-TM 20161. Then we use fuzzy-match repair (FMR) as a mechanism to show that the combined system outperforms individual systems in a black-box machine translation setting.",

author = "Ortega, {John E.} and Weiyi Lu and Adam Meyers and Kyunghyun Cho",

note = "Publisher Copyright: {\textcopyright} 2018 The authors. This article is licensed under a Creative Commons 3.0 licence, no derivative works, attribution, CC-BY-ND.; 21st Annual Conference of the European Association for Machine Translation, EAMT 2018 ; Conference date: 28-05-2018 Through 30-05-2018",

year = "2018",

month = jan,

day = "1",

language = "English (US)",

series = "EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation",

publisher = "European Association for Machine Translation",

pages = "209--218",

editor = "Perez-Ortiz, {Juan Antonio} and Felipe Sanchez-Martinez and Miquel Espla-Gomis and Maja Popovic and Celia Rico and Andre Martins and {Van den Bogaert}, Joachim and Forcada, {Mikel L.}",

booktitle = "EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation",

}

TY - GEN

T1 - Letting a neural network decide which machine translation system to use for black-box fuzzy-match repair

AU - Ortega, John E.

AU - Lu, Weiyi

AU - Meyers, Adam

AU - Cho, Kyunghyun

PY - 2018/1/1

Y1 - 2018/1/1

N2 - While systems using the Neural Network-based Machine Translation (NMT) paradigm achieve the highest scores on recent shared tasks, phrase-based (PBMT) systems, rule-based (RBMT) systems and other systems may get better results for individual examples. Therefore, combined systems should achieve the best results for MT, particularly if the system combination method can take advantage of the strengths of each paradigm. In this paper, we describe a system that predicts whether a NMT, PBMT or RBMT will get the best Spanish translation result for a particular English sentence in DGT-TM 20161. Then we use fuzzy-match repair (FMR) as a mechanism to show that the combined system outperforms individual systems in a black-box machine translation setting.

AB - While systems using the Neural Network-based Machine Translation (NMT) paradigm achieve the highest scores on recent shared tasks, phrase-based (PBMT) systems, rule-based (RBMT) systems and other systems may get better results for individual examples. Therefore, combined systems should achieve the best results for MT, particularly if the system combination method can take advantage of the strengths of each paradigm. In this paper, we describe a system that predicts whether a NMT, PBMT or RBMT will get the best Spanish translation result for a particular English sentence in DGT-TM 20161. Then we use fuzzy-match repair (FMR) as a mechanism to show that the combined system outperforms individual systems in a black-box machine translation setting.

UR - http://www.scopus.com/inward/record.url?scp=85074905660&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85074905660&partnerID=8YFLogxK

M3 - Conference contribution

T3 - EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation

SP - 209

EP - 218

BT - EAMT 2018 - Proceedings of the 21st Annual Conference of the European Association for Machine Translation

A2 - Perez-Ortiz, Juan Antonio

A2 - Sanchez-Martinez, Felipe

A2 - Espla-Gomis, Miquel

A2 - Popovic, Maja

A2 - Rico, Celia

A2 - Martins, Andre

A2 - Van den Bogaert, Joachim

A2 - Forcada, Mikel L.

PB - European Association for Machine Translation

T2 - 21st Annual Conference of the European Association for Machine Translation, EAMT 2018

Y2 - 28 May 2018 through 30 May 2018

ER -

Letting a neural network decide which machine translation system to use for black-box fuzzy-match repair

Abstract

Publication series

Conference

ASJC Scopus subject areas

Other files and links

Fingerprint

Cite this