May 2018
Beginner
490 pages
13h 16m
English
The final error (translation) probability produced is interesting, as shown in the following output:
>>------Summary------>>Neural Machine Translation model:nmt>>Google Translate: True: 2 False: 8 ETP 0.67>>Customized Google Translate: True: 2 False: 7 ETP 0.58>>Google Translate acceptable: True: 4.5 False: 5.5 ETP 0.46>>Customized Google Translate acceptable: True: 4.5 False: 4.5 ETP 0.38
Even with its NMT model, Google Translate is still struggling.
This provides great opportunities for AI linguists, as shown with some of the methods presented to improve Google Translate at a local level that could even go further.
This experiment with Google Translate shows that Google has just ...
Read now
Unlock full access