Kriya - The SFU system for translation task at WMT-12

Authors:
Majid Razmara;Baskaran Sankaran;Ann Clifton;Anoop Sarkar
Affiliations:
Simon Fraser University, Burnaby BC, Canada;Simon Fraser University, Burnaby BC, Canada;Simon Fraser University, Burnaby BC, Canada;Simon Fraser University, Burnaby BC, Canada
Venue:
WMT '12 Proceedings of the Seventh Workshop on Statistical Machine Translation
Year:
2012

Citing 8
Cited 1

Minimum error rate training in statistical machine translation

ACL '03 Proceedings of the 41st Annual Meeting on Association for Computational Linguistics - Volume 1
Unsupervised models for morpheme segmentation and morphology learning

ACM Transactions on Speech and Language Processing (TSLP)
Hierarchical Phrase-Based Translation

Computational Linguistics
Moses: open source toolkit for statistical machine translation

ACL '07 Proceedings of the 45th Annual Meeting of the ACL on Interactive Poster and Demonstration Sessions
CONDOR, a new parallel, constrained extension of Powell's UOBYQA algorithm: Experimental results and comparison with the DFO algorithm

Journal of Computational and Applied Mathematics
Combining morpheme-based machine translation with post-processing morpheme prediction

HLT '11 Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics: Human Language Technologies - Volume 1
KenLM: faster and smaller language model queries

WMT '11 Proceedings of the Sixth Workshop on Statistical Machine Translation
Mixing multiple translation models in statistical machine translation

ACL '12 Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics: Long Papers - Volume 1

Findings of the 2012 workshop on statistical machine translation

WMT '12 Proceedings of the Seventh Workshop on Statistical Machine Translation

Quantified Score

Hi-index	0.00

Visualization

Abstract

This paper describes our submissions for the WMT-12 translation task using Kriya - our hierarchical phrase-based system. We submitted systems in French-English and English-Czech language pairs. In addition to the baseline system following the standard MT pipeline, we tried ensemble decoding for French-English. The ensemble decoding method improved the BLEU score by 0.4 points over the baseline in newstest-2011. For English-Czech, we segmented the Czech side of the corpora and trained two different segmented models in addition to our baseline system.