We address the problem of training the free parameters of a statistical machine translation system. We show significant improvements over a state-of-the-art minimum error rate tr...
Often, the training procedure for statistical machine translation models is based on maximum likelihood or related criteria. A general problem of this approach is that there is on...
We compare two pivot strategies for phrase-based statistical machine translation (SMT), namely phrase translation and sentence translation. The phrase translation strategy means t...
Och's (2003) minimum error rate training (MERT) procedure is the most commonly used method for training feature weights in statistical machine translation (SMT) models. The u...
This paper presents a novel online relevant set algorithm for a linearly-scored block sequence translation model. The key component is a new procedure to directly optimize the glob...