Automatic Post-Editing of machine translation: a neural programmer-interpreter approach

Thuy-Trang Vu, Reza Haffari

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

Abstract

Automated Post-Editing (PE) is the task of automatically correcting common and repetitive errors found in machine translation (MT)
output. In this paper, we present a neural programmer-interpreter approach to this task, resembling the way that humans perform postediting using discrete edit operations, which we refer to as programs. Our model outperforms previous neural models for inducing
PE programs on the WMT17 APE task for German-English up to +1 BLEU score and - 0.7 TER scores.
Original languageEnglish
Title of host publicationEMNLP 2018
Subtitle of host publicationBrussels, Belgium Oct. 31-Nov. 4
EditorsDavid Chiang
Place of PublicationStroudsburg PA 18360
PublisherThe Association for Computational Linguistics
Pages3048-3053
Number of pages6
ISBN (Electronic)9781948087841
Publication statusPublished - 2018
EventEmpirical Methods in Natural Language Processing 2018 - Brussels, Belgium
Duration: 31 Oct 20184 Nov 2018
https://emnlp2018.org/

Conference

ConferenceEmpirical Methods in Natural Language Processing 2018
Abbreviated titleEMNLP 2018
CountryBelgium
CityBrussels
Period31/10/184/11/18
Internet address

Cite this

Vu, T-T., & Haffari, R. (2018). Automatic Post-Editing of machine translation: a neural programmer-interpreter approach. In D. Chiang (Ed.), EMNLP 2018: Brussels, Belgium Oct. 31-Nov. 4 (pp. 3048-3053). Stroudsburg PA 18360: The Association for Computational Linguistics.
Vu, Thuy-Trang ; Haffari, Reza. / Automatic Post-Editing of machine translation : a neural programmer-interpreter approach. EMNLP 2018: Brussels, Belgium Oct. 31-Nov. 4. editor / David Chiang. Stroudsburg PA 18360 : The Association for Computational Linguistics, 2018. pp. 3048-3053
@inproceedings{5df5318a26824846b47f7b04249bb476,
title = "Automatic Post-Editing of machine translation: a neural programmer-interpreter approach",
abstract = "Automated Post-Editing (PE) is the task of automatically correcting common and repetitive errors found in machine translation (MT)output. In this paper, we present a neural programmer-interpreter approach to this task, resembling the way that humans perform postediting using discrete edit operations, which we refer to as programs. Our model outperforms previous neural models for inducingPE programs on the WMT17 APE task for German-English up to +1 BLEU score and - 0.7 TER scores.",
author = "Thuy-Trang Vu and Reza Haffari",
year = "2018",
language = "English",
pages = "3048--3053",
editor = "Chiang, {David }",
booktitle = "EMNLP 2018",
publisher = "The Association for Computational Linguistics",

}

Vu, T-T & Haffari, R 2018, Automatic Post-Editing of machine translation: a neural programmer-interpreter approach. in D Chiang (ed.), EMNLP 2018: Brussels, Belgium Oct. 31-Nov. 4. The Association for Computational Linguistics, Stroudsburg PA 18360, pp. 3048-3053, Empirical Methods in Natural Language Processing 2018, Brussels, Belgium, 31/10/18.

Automatic Post-Editing of machine translation : a neural programmer-interpreter approach. / Vu, Thuy-Trang; Haffari, Reza.

EMNLP 2018: Brussels, Belgium Oct. 31-Nov. 4. ed. / David Chiang. Stroudsburg PA 18360 : The Association for Computational Linguistics, 2018. p. 3048-3053.

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

TY - GEN

T1 - Automatic Post-Editing of machine translation

T2 - a neural programmer-interpreter approach

AU - Vu, Thuy-Trang

AU - Haffari, Reza

PY - 2018

Y1 - 2018

N2 - Automated Post-Editing (PE) is the task of automatically correcting common and repetitive errors found in machine translation (MT)output. In this paper, we present a neural programmer-interpreter approach to this task, resembling the way that humans perform postediting using discrete edit operations, which we refer to as programs. Our model outperforms previous neural models for inducingPE programs on the WMT17 APE task for German-English up to +1 BLEU score and - 0.7 TER scores.

AB - Automated Post-Editing (PE) is the task of automatically correcting common and repetitive errors found in machine translation (MT)output. In this paper, we present a neural programmer-interpreter approach to this task, resembling the way that humans perform postediting using discrete edit operations, which we refer to as programs. Our model outperforms previous neural models for inducingPE programs on the WMT17 APE task for German-English up to +1 BLEU score and - 0.7 TER scores.

M3 - Conference Paper

SP - 3048

EP - 3053

BT - EMNLP 2018

A2 - Chiang, David

PB - The Association for Computational Linguistics

CY - Stroudsburg PA 18360

ER -

Vu T-T, Haffari R. Automatic Post-Editing of machine translation: a neural programmer-interpreter approach. In Chiang D, editor, EMNLP 2018: Brussels, Belgium Oct. 31-Nov. 4. Stroudsburg PA 18360: The Association for Computational Linguistics. 2018. p. 3048-3053