Considerations in maximum mutual information and minimum classification error training for statistical machine translation

3462

Knowledge Triples