Machine translation works well for sentences but turns out to falter at the
   claim that neural machine translation now matches the performance of
   [machine-translation.png?sw=600&cx=0&cy=0&cw=1420&ch=53
   At issue is how machine translation should be evaluated. This is
   translation is determined by professional human translators who read
   both the original text and the translation to see how well it expresses
   see only the translation and determine how well it is expressed in
   translations at the sentence level, whereas humans also evaluate text
   examiners rated each translation for adequacy and fluency at the
   translations are rated as more adequate and more fluent than machine
   translations. “Human raters assessing adequacy and fluency show a
   stronger preference for human over machine translation when evaluating
   document-level evaluation unveils errors such as mistranslation of an
   This suggests that the way machine translation is evaluated needs to
   “As machine translation quality improves, translations will become
   understand the original text and its translation, and also exposes
   translation errors related to discourse phenomena which remain
   That change should help machine translation improve. Which means it is
   still set to surpass human translation—just not yet.
   Ref: arxiv.org/abs/1808.07048 : Has Machine Translation Achieved Human