The MIT-LL/AFRL IWSLT 2012 MT system

Jennifer Drexler, Wade Shen, Tim Anderson, Raymond Slyh, Brian Ore, Eric Hansen, Terry Gleason


Abstract
This paper describes the MIT-LL/AFRL statistical MT system and the improvements that were developed during the IWSLT 2012 evaluation campaign. As part of these efforts, we experimented with a number of extensions to the standard phrase-based model that improve performance on the Arabic to English and English to French TED-talk translation task. We also applied our existing ASR system to the TED-talk lecture ASR task, and combined our ASR and MT systems for the TED-talk SLT task. We discuss the architecture of the MIT-LL/AFRL MT system, improvements over our 2011 system, and experiments we ran during the IWSLT-2012 evaluation. Specifically, we focus on 1) cross-domain translation using MAP adaptation, 2) cross-entropy filtering of MT training data, and 3) improved Arabic morphology for MT preprocessing.
Anthology ID:
2012.iwslt-evaluation.14
Volume:
Proceedings of the 9th International Workshop on Spoken Language Translation: Evaluation Campaign
Month:
December 6-7
Year:
2012
Address:
Hong Kong, Table of contents
Venue:
IWSLT
SIG:
SIGSLT
Publisher:
Note:
Pages:
109–116
Language:
URL:
https://aclanthology.org/2012.iwslt-evaluation.14
DOI:
Bibkey:
Cite (ACL):
Jennifer Drexler, Wade Shen, Tim Anderson, Raymond Slyh, Brian Ore, Eric Hansen, and Terry Gleason. 2012. The MIT-LL/AFRL IWSLT 2012 MT system. In Proceedings of the 9th International Workshop on Spoken Language Translation: Evaluation Campaign, pages 109–116, Hong Kong, Table of contents.
Cite (Informal):
The MIT-LL/AFRL IWSLT 2012 MT system (Drexler et al., IWSLT 2012)
Copy Citation:
PDF:
https://aclanthology.org/2012.iwslt-evaluation.14.pdf