Institute

Studying

Research


 

Models of Morphosyntax for Statistical Machine Translation

Statistical approaches to machine translation (MT) have shown themselves to be effective in the last few years. However, when translating into a morphologically rich language this is not true, particularly when there is also significant syntactic divergence between the two languages. The quality of statistical machine translation is poor in this case because of independence assumptions made between the models of morphology, syntax and translation that do not reflect linguistic reality.

The proposed project would use advances in automatic linguistic analysis of syntax and morphology to advance statistical MT. The dependencies between morphology, syntax and translation should be directly modeled. This will lead to the creation of translation models and search algorithms that will dramatically improve translation quality for morphologically rich languages.

The proposed project would use advances in automatic linguistic analysis of syntax and morphology to advance statistical MT. The dependencies between morphology, syntax and translation should be directly modeled. This will lead to the creation of translation models and search algorithms that will dramatically improve translation quality for morphologically rich languages.