Expanding the Language model in a low-resource hybrid MT system

The present article investigates the fusion of different language models to improve transla-tion accuracy. A hybrid MT system, recently-developed in the European Commission-funded PRESEMT project that combines ex-ample-based MT and Statistical MT princi-ples is used as a starting point. In this article, the syntactically-defined phrasal language models (NPs, VPs etc.) used by this MT sys-tem are supplemented by n-gram language models to improve translation accuracy. For specific structural patterns, n-gram statistics are consulted to determine whether the pat-tern instantiations are corroborated. Experi-ments indicate improvements in translation accuracy.