Fast, Scalable Phrase-Based SMT Decoding

Hieu Hoang; Nikolay Bogoychev; Lane Schwartz; Marcin Junczys-Dowmunt

Fast, Scalable Phrase-Based SMT Decoding

Hieu Hoang, Nikolay Bogoychev, Lane Schwartz, Marcin Junczys-Dowmunt

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

Abstract / Description of output

The utilization of statistical machine translation (SMT) has grown enormously over the last decade, many using open-source software developed by the NLP community. As commercial use has increased, there is need for software that is optimized for commercial requirements, in particular, fast phrase-based decoding and more efficient utilization of modern multicore servers.
In this paper we re-examine the major components of phrase-based decoding and decoder implementation with particular emphasis on speed and scalability on multicore machines. The result is a drop-in replacement for the Moses decoder which is up to fifteen times faster and scales monotonically with the number of cores.

Original language	English
Title of host publication	The Twelfth Conference of The Association for Machine Translation in the Americas 2016
Place of Publication	Austin, Texas, USA
Pages	40-52
Number of pages	13
Volume	1
Publication status	Published - 1 Nov 2016
Event	Twelfth Conference of The Association for Machine Translation in the Americas - Austin, United States Duration: 28 Oct 2016 → 1 Nov 2016 http://www.amta2016.org/ https://amtaweb.org/amta-2016-proceedings-are-available/

Conference

Conference	Twelfth Conference of The Association for Machine Translation in the Americas
Abbreviated title	AMTA 2016
Country/Territory	United States
City	Austin
Period	28/10/16 → 1/11/16
Internet address	http://www.amta2016.org/ https://amtaweb.org/amta-2016-proceedings-are-available/

Access to Document

https://amtaweb.org/wp-content/uploads/2016/10/AMTA2016_Research_Proceedings_v7.pdf#page=46

Cite this

@inproceedings{bc588e9475984178b3449bbac447f386,

title = "Fast, Scalable Phrase-Based SMT Decoding",

abstract = "The utilization of statistical machine translation (SMT) has grown enormously over the last decade, many using open-source software developed by the NLP community. As commercial use has increased, there is need for software that is optimized for commercial requirements, in particular, fast phrase-based decoding and more efficient utilization of modern multicore servers.In this paper we re-examine the major components of phrase-based decoding and decoder implementation with particular emphasis on speed and scalability on multicore machines. The result is a drop-in replacement for the Moses decoder which is up to fifteen times faster and scales monotonically with the number of cores.",

author = "Hieu Hoang and Nikolay Bogoychev and Lane Schwartz and Marcin Junczys-Dowmunt",

year = "2016",

month = nov,

day = "1",

language = "English",

volume = "1",

pages = "40--52",

booktitle = "The Twelfth Conference of The Association for Machine Translation in the Americas 2016",

note = "Twelfth Conference of The Association for Machine Translation in the Americas, AMTA 2016 ; Conference date: 28-10-2016 Through 01-11-2016",

url = "http://www.amta2016.org/, https://amtaweb.org/amta-2016-proceedings-are-available/",

}

Hoang, H, Bogoychev, N, Schwartz, L & Junczys-Dowmunt, M 2016, Fast, Scalable Phrase-Based SMT Decoding. in The Twelfth Conference of The Association for Machine Translation in the Americas 2016. vol. 1, Austin, Texas, USA, pp. 40-52, Twelfth Conference of The Association for Machine Translation in the Americas, Austin, Texas, United States, 28/10/16. <https://amtaweb.org/wp-content/uploads/2016/10/AMTA2016_Research_Proceedings_v7.pdf#page=46>

TY - GEN

T1 - Fast, Scalable Phrase-Based SMT Decoding

AU - Hoang, Hieu

AU - Bogoychev, Nikolay

AU - Schwartz, Lane

AU - Junczys-Dowmunt, Marcin

PY - 2016/11/1

Y1 - 2016/11/1

N2 - The utilization of statistical machine translation (SMT) has grown enormously over the last decade, many using open-source software developed by the NLP community. As commercial use has increased, there is need for software that is optimized for commercial requirements, in particular, fast phrase-based decoding and more efficient utilization of modern multicore servers.In this paper we re-examine the major components of phrase-based decoding and decoder implementation with particular emphasis on speed and scalability on multicore machines. The result is a drop-in replacement for the Moses decoder which is up to fifteen times faster and scales monotonically with the number of cores.

AB - The utilization of statistical machine translation (SMT) has grown enormously over the last decade, many using open-source software developed by the NLP community. As commercial use has increased, there is need for software that is optimized for commercial requirements, in particular, fast phrase-based decoding and more efficient utilization of modern multicore servers.In this paper we re-examine the major components of phrase-based decoding and decoder implementation with particular emphasis on speed and scalability on multicore machines. The result is a drop-in replacement for the Moses decoder which is up to fifteen times faster and scales monotonically with the number of cores.

M3 - Conference contribution

VL - 1

SP - 40

EP - 52

BT - The Twelfth Conference of The Association for Machine Translation in the Americas 2016

CY - Austin, Texas, USA

T2 - Twelfth Conference of The Association for Machine Translation in the Americas

Y2 - 28 October 2016 through 1 November 2016

ER -

Fast, Scalable Phrase-Based SMT Decoding

Abstract / Description of output

Conference

Access to Document

Fingerprint

Cite this