Translation requires a deep understanding of language yet this ability lies far away on the horizon of machine translation. Today’s statistical methods are however moving in the right direction, providing imperfect yet useful translations. Short-term progress is likely to come from stronger statistical learning combined with linguistic knowledge and better modelling of the target language.   

This is the 5th article in the XRCE 20th anniversary series.  

Follow @xrce for notification of the publication of these articles.