Dr. Nizar Habash receives a Google Research Award

Preserving Morphological Richness in Poor-Pivot-based Statistical Machine Translation

The proposed effort addresses the weakness of using a morphologically poor language (English) as a pivot/bridge in statistical machine translation between morphologically rich languages (Hebrew and Arabic).

 

Our approach considers multiple mechanisms to pass, preserve, and utilize the morphological features of the source (input) language in the generation of the target (output) language.  The award will fund a student for one year.