... 37–40,Columbus, Ohio, USA, June 2008.c2008 Association for Computational LinguisticsBeyond Log -Linear Models:Boosted Minimum Error Rate Training for N-best Re-rankingKevin Duh∗Dept. of Electrical ... betweenboosting for classification and boosting for rank-ing. For instance, the distribution update scheme5There are other ways to define a ranking output that areworth exploring. For example, ... problem is more serious for log -linear modelsof around 10 features and focus on it in this work.To truly achieve the benefits of re-ranking in MT,one must go beyond the log -linear model. The re-ranker...