Paper: Improving the IBM Alignment Models Using Variational Bayes

ACL ID P12-2060
Title Improving the IBM Alignment Models Using Variational Bayes
Venue Annual Meeting of the Association of Computational Linguistics
Session Short Paper
Year 2012
Authors

Bayesian approaches have been shown to re- duce 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, vari- ational Bayes, to the IBM models of word alignment for statistical machine translation. We show that using variational Bayes im- proves 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.