📜 ⬆️ ⬇️

Google translate

All greetings, I just talked to KO here, he asked me to convey:
I would like to once again clarify a little about the Google translator.
The basis of the translation algorithm is a statistical analysis of the “templates” of the text, based on their vast base of indexed text. Large chunks of text are analyzed and the most frequent corresponding repetition in the pair language is chosen. (Of course, this is the most simplified model).
It’s the same with the words, so it turns out that what we have here is Linkedin, we have My Circle, since the content of the text, which features linkedin in Russian, is most relevant to the content of the text where My circle appears.
Hence all the nuances and fun.
The main goal of this approach is a much faster process of translating text than classical algorithms. Just do not forget to add self-learning algorithm. Therefore, theoretically, in the long run, this approach has a chance to even become a leader.
PS Naturally, everything written above can be just an erroneous guess of many bloggers , specialists in this field, or my personal delusion, but in fact the truth is much worse, and more than 999,999.9 Chinese work on translation

')

Source: https://habr.com/ru/post/69977/


All Articles