Sciweavers

ACL
2012

Improving the IBM Alignment Models Using Variational Bayes

11 years 6 months ago
Improving the IBM Alignment Models Using Variational Bayes
Bayesian approaches have been shown to reduce the amount of overfitting that occurs when running the EM algorithm, by placing prior probabilities on the model parameters. We apply one such Bayesian technique, variational Bayes, to the IBM models of word alignment for statistical machine translation. We show that using variational Bayes improves the performance of the widely used GIZA++ software, as well as improving the overall performance of the Moses machine translation system in terms of BLEU score.
Darcey Riley, Daniel Gildea
Added 29 Sep 2012
Updated 29 Sep 2012
Type Journal
Year 2012
Where ACL
Authors Darcey Riley, Daniel Gildea
Comments (0)