Paper: Discriminative Feature-Tied Mixture Modeling for Statistical Machine Translation

ACL ID P11-2074
Title Discriminative Feature-Tied Mixture Modeling for Statistical Machine Translation
Venue Annual Meeting of the Association of Computational Linguistics
Session Main Conference
Year 2011
Authors

In this paper we present a novel discrimi- native mixture model for statistical machine translation (SMT). We model the feature space with a log-linear combination of multiple mix- ture components. Each component contains a large set of features trained in a maximum- entropy framework. All features within the same mixture component are tied and share the same mixture weights, where the mixture weights are trained discriminatively to max- imize the translation performance. This ap- proach aims at bridging the gap between the maximum-likelihood training and the discrim- inative training for SMT. It is shown that the feature space can be partitioned in a vari- ety of ways, such as based on feature types, word alignments, or domains, for various ap- plications. The proposed approach improves t...