Do you prefer a traditional dictionary or the new Google translate system?
(Mirror Daily, United States) Officials from Google have recently announced that Google Translate is about to get significant improvements. It is all due to a new system, called Google Neural Machine Translation (GNMT), and it was designed by an artificial intelligence specialized team. The system is now able to provide more accurate translations than it used to, and here is how.
According to officially released information, the GNMT system has a great advantage which prevails over the features of the old Google translate system. What the new system does it to process a phrase or a sentence as a whole, and then translates it function of the whole meaning of the sentences. This procedure assures a much more accurate translation, as well as a proper exchange of information.
We are all familiar with the old Google translate system. It is based on individual words rather than word structures. However, translating mot-à-mot is not always the best option, as both users and developers could already acknowledge. This is why officials from Google account for the efficiency of their most recent product.
The new Google Neural Machine Translation is said to translate as good as a human translator. One of the first tasks for the system was translating from Chinese into English. Chinese is known as one of the most difficult languages to learn, and the improvement of the translate system with this particular language can turn out to be very useful for learners.
Developers also revealed information on the efficiency of the GNMT system. According to them, there is a lower rate of mistakes. The system can prevent errors for more than fifty percent, and up to eighty-five percent.
Specialists have also released a graphic which shows the achievements of human translators, old systems used for translation, and the GNMT. An ideal translation is rarely achieved by humans, but they are most likely to get the high score. However, traditional systems fall behind, with a gap of two points between their results and the ideal ones. Out of these options, GNMT is the artificial intelligence system which can replace a human translator successfully.
Nevertheless, developers know that their latest product is far from being perfect. Not only Chinese, but several other languages will be added. Producers intend to reduce the rate of errors as much as possible. For the time being, translating phrases function of the context is still regarded as a great achievement.
Google endeavors to provide users with the best tools for translation. Do you like the idea, or are you a fan of traditional and on-line dictionaries?
Image courtesy of: Flickr