Unsupervised constraint driven learning for transliteration discovery

Ming Wei Chang; Dan Goldwasser; Dan Roth; Yuancheng Tu

doi:10.3115/1620754.1620798

Unsupervised constraint driven learning for transliteration discovery

Ming Wei Chang, Dan Goldwasser, Dan Roth, Yuancheng Tu

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

Abstract

This paper introduces a novel unsupervised constraint-driven learning algorithm for identifying named-entity (NE) transliterations in bilingual corpora. The proposed method does not require any annotated data or aligned corpora. Instead, it is bootstrapped using a simple resource - a romanization table. We show that this resource, when used in conjunction with constraints, can efficiently identify transliteration pairs. We evaluate the proposed method on transliterating English NEs to three different languages - Chinese, Russian and Hebrew. Our experiments show that constraint driven learning can significantly outperform existing unsupervised models and achieve competitive results to existing supervised models.

Original language	English (US)
Title of host publication	NAACL HLT 2009 - Human Language Technologies
Subtitle of host publication	The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference
Publisher	Association for Computational Linguistics (ACL)
Pages	299-307
Number of pages	9
ISBN (Print)	9781932432411
DOIs	https://doi.org/10.3115/1620754.1620798
State	Published - 2009
Externally published	Yes
Event	Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, NAACL HLT 2009 - Boulder, CO, United States Duration: May 31 2009 → Jun 5 2009

Publication series

Name	NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference

Other

Other	Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, NAACL HLT 2009
Country/Territory	United States
City	Boulder, CO
Period	5/31/09 → 6/5/09

ASJC Scopus subject areas

Language and Linguistics
Social Sciences (miscellaneous)

Online availability

10.3115/1620754.1620798

Library availability

Discover UIUC Full Text

Cite this

Chang, M. W., Goldwasser, D., Roth, D., & Tu, Y. (2009). Unsupervised constraint driven learning for transliteration discovery. In NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference (pp. 299-307). (NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference). Association for Computational Linguistics (ACL). https://doi.org/10.3115/1620754.1620798

Unsupervised constraint driven learning for transliteration discovery. / Chang, Ming Wei; Goldwasser, Dan; Roth, Dan et al.
NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference. Association for Computational Linguistics (ACL), 2009. p. 299-307 (NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference).

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

Chang, MW, Goldwasser, D, Roth, D & Tu, Y 2009, Unsupervised constraint driven learning for transliteration discovery. in NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference. NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference, Association for Computational Linguistics (ACL), pp. 299-307, Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, NAACL HLT 2009, Boulder, CO, United States, 5/31/09. https://doi.org/10.3115/1620754.1620798

Chang MW, Goldwasser D, Roth D, Tu Y. Unsupervised constraint driven learning for transliteration discovery. In NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference. Association for Computational Linguistics (ACL). 2009. p. 299-307. (NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference). doi: 10.3115/1620754.1620798

Chang, Ming Wei ; Goldwasser, Dan ; Roth, Dan et al. / Unsupervised constraint driven learning for transliteration discovery. NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference. Association for Computational Linguistics (ACL), 2009. pp. 299-307 (NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference).

@inproceedings{d42a2bff308e4254848457e2760a2a40,

title = "Unsupervised constraint driven learning for transliteration discovery",

abstract = "This paper introduces a novel unsupervised constraint-driven learning algorithm for identifying named-entity (NE) transliterations in bilingual corpora. The proposed method does not require any annotated data or aligned corpora. Instead, it is bootstrapped using a simple resource - a romanization table. We show that this resource, when used in conjunction with constraints, can efficiently identify transliteration pairs. We evaluate the proposed method on transliterating English NEs to three different languages - Chinese, Russian and Hebrew. Our experiments show that constraint driven learning can significantly outperform existing unsupervised models and achieve competitive results to existing supervised models.",

author = "Chang, {Ming Wei} and Dan Goldwasser and Dan Roth and Yuancheng Tu",

year = "2009",

doi = "10.3115/1620754.1620798",

language = "English (US)",

isbn = "9781932432411",

series = "NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference",

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

pages = "299--307",

booktitle = "NAACL HLT 2009 - Human Language Technologies",

note = "Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, NAACL HLT 2009 ; Conference date: 31-05-2009 Through 05-06-2009",

}

TY - GEN

T1 - Unsupervised constraint driven learning for transliteration discovery

AU - Chang, Ming Wei

AU - Goldwasser, Dan

AU - Roth, Dan

AU - Tu, Yuancheng

PY - 2009

Y1 - 2009

N2 - This paper introduces a novel unsupervised constraint-driven learning algorithm for identifying named-entity (NE) transliterations in bilingual corpora. The proposed method does not require any annotated data or aligned corpora. Instead, it is bootstrapped using a simple resource - a romanization table. We show that this resource, when used in conjunction with constraints, can efficiently identify transliteration pairs. We evaluate the proposed method on transliterating English NEs to three different languages - Chinese, Russian and Hebrew. Our experiments show that constraint driven learning can significantly outperform existing unsupervised models and achieve competitive results to existing supervised models.

AB - This paper introduces a novel unsupervised constraint-driven learning algorithm for identifying named-entity (NE) transliterations in bilingual corpora. The proposed method does not require any annotated data or aligned corpora. Instead, it is bootstrapped using a simple resource - a romanization table. We show that this resource, when used in conjunction with constraints, can efficiently identify transliteration pairs. We evaluate the proposed method on transliterating English NEs to three different languages - Chinese, Russian and Hebrew. Our experiments show that constraint driven learning can significantly outperform existing unsupervised models and achieve competitive results to existing supervised models.

UR - http://www.scopus.com/inward/record.url?scp=77958023635&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=77958023635&partnerID=8YFLogxK

U2 - 10.3115/1620754.1620798

DO - 10.3115/1620754.1620798

M3 - Conference contribution

AN - SCOPUS:77958023635

SN - 9781932432411

T3 - NAACL HLT 2009 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Proceedings of the Conference

SP - 299

EP - 307

BT - NAACL HLT 2009 - Human Language Technologies

PB - Association for Computational Linguistics (ACL)

T2 - Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, NAACL HLT 2009

Y2 - 31 May 2009 through 5 June 2009

ER -

Unsupervised constraint driven learning for transliteration discovery

Abstract

Publication series

Other

ASJC Scopus subject areas

Online availability

Library availability

Related links

Fingerprint

Cite this