Arabizi transliteration

Our Arabizi-to-Arabic transliteration software is available here.
The corresponding Arabizi-English bitext crawled from the web can be downloaded here.

This software component and data set were used for OpenMT 2015, and were presented in the COLING 2016 Workshop on Noisy User-generated Text (WNUT2016) as “A Simple but Effective Approach to Improve Arabizi-to-English Statistical Machine Translation” by Marlies van der Wees, Arianna Bisazza and Christof Monz.