Minimum Error-Rate Training in Statistical Machine Translation Using Structural SVMs

Authors:
Jesús González-Rubio;Daniel Ortiz-Martinez;Francisco Casacuberta
Affiliations:
Instituto Tecnológico de Informática, Universidad Politécnica de Valencia, Spain;Instituto Tecnológico de Informática, Universidad Politécnica de Valencia, Spain;Departamento de Sistemas Informáticos y Computación, Universidad Politécnica de Valencia, Spain
Venue:
IbPRIA '09 Proceedings of the 4th Iberian Conference on Pattern Recognition and Image Analysis
Year:
2009

Citing 7
Cited 0

Discriminative training and maximum entropy models for statistical machine translation

ACL '02 Proceedings of the 40th Annual Meeting on Association for Computational Linguistics
Statistical phrase-based translation

NAACL '03 Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology - Volume 1
Minimum error rate training in statistical machine translation

ACL '03 Proceedings of the 41st Annual Meeting on Association for Computational Linguistics - Volume 1
Large Margin Methods for Structured and Interdependent Output Variables

The Journal of Machine Learning Research
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
Cutting-plane training of structural SVMs

Machine Learning
Manual and automatic evaluation of machine translation between European languages

StatMT '06 Proceedings of the Workshop on Statistical Machine Translation

Quantified Score

Hi-index	0.00

Visualization

Abstract

Different works on training of log-linear interpolation models for statistical machine translation reported performance improvements by optimizing parameters with respect to translation quality, rather than to likelihood oriented criteria. This work presents an alternative minimum error-rate training procedure based on structural support vector machines (SSVMs) for log-linear interpolation models which is not limited to the model scaling factors and needs only few iterations to converge. Experimental results are reported on the Spanish---English Europarl corpus.