A localized prediction model for statistical machine translation

Christoph Tillmann; Tong Zhang

doi:10.3115/1219840.1219909

A localized prediction model for statistical machine translation

Christoph Tillmann^*, Tong Zhang

^*Corresponding author for this work

Research output: Chapter in Book/Conference Proceeding/Report › Conference Paper published in a book › peer-review

46 Citations (Scopus)

Abstract

In this paper, we present a novel training method for a localized phrase-based prediction model for statistical machine translation (SMT). The model predicts blocks with orientation to handle local phrase re-ordering. We use a maximum likelihood criterion to train a log-linear block bigram model which uses realvalued features (e.g. a language model score) as well as binary features based on the block identities themselves, e.g. block bigram features. Our training algorithm can easily handle millions of features. The best system obtains a 18.6% improvement over the baseline on a standard Arabic-English translation task.

Original language	English
Title of host publication	ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference
Publisher	Association for Computational Linguistics (ACL)
Pages	557-564
Number of pages	8
ISBN (Print)	1932432515, 9781932432510
DOIs	https://doi.org/10.3115/1219840.1219909
Publication status	Published - 2005
Externally published	Yes
Event	43rd Annual Meeting of the Association for Computational Linguistics, ACL-05 - Ann Arbor, MI, United States Duration: 25 Jun 2005 → 30 Jun 2005

Publication series

Name	ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference

Conference

Conference	43rd Annual Meeting of the Association for Computational Linguistics, ACL-05
Country/Territory	United States
City	Ann Arbor, MI
Period	25/06/05 → 30/06/05

Access to Document

10.3115/1219840.1219909

Cite this

Tillmann, C., & Zhang, T. (2005). A localized prediction model for statistical machine translation. In ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference (pp. 557-564). (ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference). Association for Computational Linguistics (ACL). https://doi.org/10.3115/1219840.1219909

Tillmann, Christoph ; Zhang, Tong. / A localized prediction model for statistical machine translation. ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference. Association for Computational Linguistics (ACL), 2005. pp. 557-564 (ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference).

@inproceedings{9ec9b79970bb42bdadb12620c4089071,

title = "A localized prediction model for statistical machine translation",

abstract = "In this paper, we present a novel training method for a localized phrase-based prediction model for statistical machine translation (SMT). The model predicts blocks with orientation to handle local phrase re-ordering. We use a maximum likelihood criterion to train a log-linear block bigram model which uses realvalued features (e.g. a language model score) as well as binary features based on the block identities themselves, e.g. block bigram features. Our training algorithm can easily handle millions of features. The best system obtains a 18.6\% improvement over the baseline on a standard Arabic-English translation task.",

author = "Christoph Tillmann and Tong Zhang",

year = "2005",

doi = "10.3115/1219840.1219909",

language = "English",

isbn = "1932432515",

series = "ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference",

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

pages = "557--564",

booktitle = "ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference",

note = "43rd Annual Meeting of the Association for Computational Linguistics, ACL-05 ; Conference date: 25-06-2005 Through 30-06-2005",

}

Tillmann, C & Zhang, T 2005, A localized prediction model for statistical machine translation. in ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference. ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference, Association for Computational Linguistics (ACL), pp. 557-564, 43rd Annual Meeting of the Association for Computational Linguistics, ACL-05, Ann Arbor, MI, United States, 25/06/05. https://doi.org/10.3115/1219840.1219909

A localized prediction model for statistical machine translation. / Tillmann, Christoph; Zhang, Tong.
ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference. Association for Computational Linguistics (ACL), 2005. p. 557-564 (ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference).

Research output: Chapter in Book/Conference Proceeding/Report › Conference Paper published in a book › peer-review

TY - GEN

T1 - A localized prediction model for statistical machine translation

AU - Tillmann, Christoph

AU - Zhang, Tong

PY - 2005

Y1 - 2005

N2 - In this paper, we present a novel training method for a localized phrase-based prediction model for statistical machine translation (SMT). The model predicts blocks with orientation to handle local phrase re-ordering. We use a maximum likelihood criterion to train a log-linear block bigram model which uses realvalued features (e.g. a language model score) as well as binary features based on the block identities themselves, e.g. block bigram features. Our training algorithm can easily handle millions of features. The best system obtains a 18.6% improvement over the baseline on a standard Arabic-English translation task.

AB - In this paper, we present a novel training method for a localized phrase-based prediction model for statistical machine translation (SMT). The model predicts blocks with orientation to handle local phrase re-ordering. We use a maximum likelihood criterion to train a log-linear block bigram model which uses realvalued features (e.g. a language model score) as well as binary features based on the block identities themselves, e.g. block bigram features. Our training algorithm can easily handle millions of features. The best system obtains a 18.6% improvement over the baseline on a standard Arabic-English translation task.

UR - http://www.scopus.com/inward/record.url?scp=84859908602&partnerID=8YFLogxK

U2 - 10.3115/1219840.1219909

DO - 10.3115/1219840.1219909

M3 - Conference Paper published in a book

AN - SCOPUS:84859908602

SN - 1932432515

SN - 9781932432510

T3 - ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference

SP - 557

EP - 564

BT - ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference

PB - Association for Computational Linguistics (ACL)

T2 - 43rd Annual Meeting of the Association for Computational Linguistics, ACL-05

Y2 - 25 June 2005 through 30 June 2005

ER -

Tillmann C, Zhang T. A localized prediction model for statistical machine translation. In ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference. Association for Computational Linguistics (ACL). 2005. p. 557-564. (ACL-05 - 43rd Annual Meeting of the Association for Computational Linguistics, Proceedings of the Conference). doi: 10.3115/1219840.1219909

A localized prediction model for statistical machine translation

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this