Neural Machine Translation Techniques for Named Entity Transliteration

Roman Grundkiewicz; Kenneth Heafield

Neural Machine Translation Techniques for Named Entity Transliteration

Roman Grundkiewicz
, Kenneth Heafield

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

Abstract

Transliterating named entities from one language into another can be approached as neural machine translation (NMT) problem, for which we use deep attentional RNN encoder-decoder models. To build a strong transliteration system, we apply well-established techniques from NMT, such as dropout regularization, model ensembling, rescoring with right-to-left models, and back-translation. Our submission to the NEWS 2018 Shared Task on Named Entity Transliteration ranked first in several tracks

Original language	English
Title of host publication	Proceedings of the Seventh Named Entities Workshop
Publisher	Association for Computational Linguistics (ACL)
Pages	89–94
Number of pages	6
Publication status	Published - 2018
Event	NEWS 2018: The Seventh Named Entities Workshop - Melbourne, Australia Duration: 20 Jul 2018 → … http://workshop.colips.org/news2018/index.html

Workshop

Workshop	NEWS 2018: The Seventh Named Entities Workshop
Abbreviated title	NEWS 2018
Country/Territory	Australia
City	Melbourne
Period	20/07/18 → …
Internet address	http://workshop.colips.org/news2018/index.html

Access to Document

news-2018-sharedAccepted author manuscript, 146 KB

Cite this

@inproceedings{1b070b5ef95f4136b065195a41ad07dd,

title = "Neural Machine Translation Techniques for Named Entity Transliteration",

abstract = "Transliterating named entities from one language into another can be approached as neural machine translation (NMT) problem, for which we use deep attentional RNN encoder-decoder models. To build a strong transliteration system, we apply well-established techniques from NMT, such as dropout regularization, model ensembling, rescoring with right-to-left models, and back-translation. Our submission to the NEWS 2018 Shared Task on Named Entity Transliteration ranked first in several tracks",

author = "Roman Grundkiewicz and Kenneth Heafield",

year = "2018",

language = "English",

pages = "89–94 ",

booktitle = "Proceedings of the Seventh Named Entities Workshop",

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

note = "NEWS 2018: The Seventh Named Entities Workshop, NEWS 2018 ; Conference date: 20-07-2018",

url = "http://workshop.colips.org/news2018/index.html",

}

TY - GEN

T1 - Neural Machine Translation Techniques for Named Entity Transliteration

AU - Grundkiewicz, Roman

AU - Heafield, Kenneth

PY - 2018

Y1 - 2018

N2 - Transliterating named entities from one language into another can be approached as neural machine translation (NMT) problem, for which we use deep attentional RNN encoder-decoder models. To build a strong transliteration system, we apply well-established techniques from NMT, such as dropout regularization, model ensembling, rescoring with right-to-left models, and back-translation. Our submission to the NEWS 2018 Shared Task on Named Entity Transliteration ranked first in several tracks

AB - Transliterating named entities from one language into another can be approached as neural machine translation (NMT) problem, for which we use deep attentional RNN encoder-decoder models. To build a strong transliteration system, we apply well-established techniques from NMT, such as dropout regularization, model ensembling, rescoring with right-to-left models, and back-translation. Our submission to the NEWS 2018 Shared Task on Named Entity Transliteration ranked first in several tracks

M3 - Conference contribution

SP - 89

EP - 94

BT - Proceedings of the Seventh Named Entities Workshop

PB - Association for Computational Linguistics (ACL)

T2 - NEWS 2018: The Seventh Named Entities Workshop

Y2 - 20 July 2018

ER -

Neural Machine Translation Techniques for Named Entity Transliteration

Abstract

Workshop

Access to Document

Fingerprint

Cite this