Recently I came across a statement from Google Translate where they claim their TM corpus used in their MT engine is improved over time. Does Microsoft do anything in this regard to improve the quality of the output over time?
↧