Google Translate now utilizes a form of machine learning that allow the program to actively learn languages without being specifically coded to do so.
With GNMT, Google translate is able to provide more natural and accurate language translations due to it translating whole sentences rather than individual words. Thanks to the Google Neural Machine Translation (GNMT), Google is able to translate whole sentences instead of translating each word individually. This provides users a more natural and accurate translation with better grammar and syntax.
“Let’s say we train a multilingual system with Japanese to English and Korean to English examples… GNMT shares its parameters to translate between these four different language pairs… It can generate reasonable Korean to Japanese translations, even though it has never been taught to do so… To the best of our knowledge, this is the first time this type of transfer learning has worked in Machine Translation,” written in Google Research blog.
The new system is currently being used now; Google hopes this to be one way for the company to make head way in its development of artificial intelligence.
Source: Cnet
