... phrase- based system usually include LM, reordering model, word and phrase counts, and phrase and lexicon translation models. Given the focus of this paper, we review only the phrase and lexicon ... with the baseline, training phrase or lexicon models alone gives a gain of 0.7 and 0.5 BLEU points, respectively, on the test set. For a full training of both phrase and lexicon models, we ... Republic of Korea, 8-14 July 2012.c2012 Association for Computational LinguisticsMaximum Expected BLEU Training of Phrase and Lexicon Translation Models Xiaodong He Li Deng Microsoft Research...