This paper proposes a novel lexicalized approach for rule selection for syntax-based statistical machine translation (SMT). We build maximum entropy (MaxEnt) models which combine ...
This paper presents a method for improving phrase-based Statistical Machine Translation systems by enriching the original translation model with information derived from a multilin...
This paper proposes a novel maximum entropy based rule selection (MERS) model for syntax-based statistical machine translation (SMT). The MERS model combines local contextual info...
We proposed a method of machine translation using inductive learning with genetic algorithms, and confirmed the effectiveness of applying genetic algorithms. However, the system b...
We show for the first time that incorporating the predictions of a word sense disambiguation system within a typical phrase-based statistical machine translation (SMT) model cons...