Beyond Log-Linear Models: Boosted Minimum Error Rate Training for N-best Re-ranking

Current re-ranking algorithms for machine translation rely on log-linear models, which have the potential problem of underfitting the training data. We present BoostedMERT, a novel boosting algorithm that uses Minimum Error Rate Training (MERT) as a weak learner and builds a re-ranker far more expressive than log-linear models. BoostedMERT is easy to implement, inherits the efficient optimization properties of MERT, and can quickly boost the BLEU score on N-best re-ranking tasks. In this paper, we describe the general algorithm and present preliminary results on the IWSLT 2007 Arabic-English task.

• Type: Conference
• Year: 2008
• Where: ACL
• Authors: Kevin Duh, Katrin Kirchhoff
• Keywords: ACL 2008, Computational Linguistics, Current Re-ranking Algorithms, Log-linear Models, Minimum Error Rate