Integrating an Unsupervised Transliteration Model into Statistical Machine Translation

Nadir Durrani; Hassan Sajjad; Hieu Hoang; Philipp Koehn

Integrating an Unsupervised Transliteration Model into Statistical Machine Translation

Nadir Durrani, Hassan Sajjad, Hieu Hoang, Philipp Koehn

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

Abstract

We investigate three methods for integrating an unsupervised transliteration model into an end-to-end SMT system. We induce a transliteration model from parallel data and use it to translate OOV words. Our approach is fully unsupervised and language independent. In the methods to integrate transliterations, we observed
improvements from 0.23-0.75 (∆ 0.41) BLEU points across 7 language pairs. We
also show that our mined transliteration corpora provide better rule coverage and
translation quality compared to the gold standard transliteration corpora.

Original language	English
Title of host publication	Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2014, April 26-30, 2014, Gothenburg, Sweden
Pages	148-153
Number of pages	6
Publication status	Published - 2014

Access to Document

Durrani et al 2014 Integrating an Unsupervised Transliteration ModelAccepted author manuscript, 201 KB

http://aclweb.org/anthology/E/E14/E14-4029.pdf

Fingerprint Dive into the research topics of 'Integrating an Unsupervised Transliteration Model into Statistical Machine Translation'. Together they form a unique fingerprint.

1 Finished

EU-Bridge
Renals, S., King, S., Koehn, P. & Osborne, M.
EU government bodies
1/02/12 → 31/01/15
Project: Research

Cite this

@inproceedings{183240bada9f41f5b4dded0442981f9f,

title = "Integrating an Unsupervised Transliteration Model into Statistical Machine Translation",

abstract = "We investigate three methods for integrating an unsupervised transliteration model into an end-to-end SMT system. We induce a transliteration model from parallel data and use it to translate OOV words. Our approach is fully unsupervised and language independent. In the methods to integrate transliterations, we observedimprovements from 0.23-0.75 (∆ 0.41) BLEU points across 7 language pairs. Wealso show that our mined transliteration corpora provide better rule coverage andtranslation quality compared to the gold standard transliteration corpora.",

author = "Nadir Durrani and Hassan Sajjad and Hieu Hoang and Philipp Koehn",

year = "2014",

language = "English",

pages = "148--153",

booktitle = "Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2014, April 26-30, 2014, Gothenburg, Sweden",

}

Integrating an Unsupervised Transliteration Model into Statistical Machine Translation. / Durrani, Nadir; Sajjad, Hassan; Hoang, Hieu; Koehn, Philipp.

Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2014, April 26-30, 2014, Gothenburg, Sweden. 2014. p. 148-153.

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

TY - GEN

T1 - Integrating an Unsupervised Transliteration Model into Statistical Machine Translation

AU - Durrani, Nadir

AU - Sajjad, Hassan

AU - Hoang, Hieu

AU - Koehn, Philipp

PY - 2014

Y1 - 2014

N2 - We investigate three methods for integrating an unsupervised transliteration model into an end-to-end SMT system. We induce a transliteration model from parallel data and use it to translate OOV words. Our approach is fully unsupervised and language independent. In the methods to integrate transliterations, we observedimprovements from 0.23-0.75 (∆ 0.41) BLEU points across 7 language pairs. Wealso show that our mined transliteration corpora provide better rule coverage andtranslation quality compared to the gold standard transliteration corpora.

AB - We investigate three methods for integrating an unsupervised transliteration model into an end-to-end SMT system. We induce a transliteration model from parallel data and use it to translate OOV words. Our approach is fully unsupervised and language independent. In the methods to integrate transliterations, we observedimprovements from 0.23-0.75 (∆ 0.41) BLEU points across 7 language pairs. Wealso show that our mined transliteration corpora provide better rule coverage andtranslation quality compared to the gold standard transliteration corpora.

M3 - Conference contribution

SP - 148

EP - 153

BT - Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, EACL 2014, April 26-30, 2014, Gothenburg, Sweden

ER -

Integrating an Unsupervised Transliteration Model into Statistical Machine Translation

Abstract

Access to Document

EU-Bridge

Cite this