Regarding the architecture of the translation program
- Vous devez vous identifier ou créer un compte pour écrire des commentaires
So, anyway I am thinking that making the translation program which is superior google's one is the urgent challenge and the resonse measure.
What do you think about it?
I have heard that google translation has the different architecture from other translation programs. I don't know well, though.
I think that even we newbies might be able to conceive some great ideas of architecture even if we don't know how to write code if we could construct the basic or dialectical architecture of translation programs in our brains.
Would you be able to give us the excellent summary of the architecture which enough to be able to grasp its fundamental framework like Fleud or Deleuze's ones? for finding their talent and letting them bloom.
For example, even if you try to take time to pieces, it remains as space image because as you know, it is merely all 1 and 0 as Bergson says.
Thinking the new architecture as time image might bring the new ideas by using e.g. my and our own abilities you know well.
A random number generator would not extract its potencial enough. You might have to leave from the idea of 1 and 0. Impoossible? Have you tried then say so? Thank you.
I do not know of any easy way to present LSTM: https://en.wikipedia.org/wiki/Long_short-term_memory
I don't understand at all the explanation of the page, even in Japanese. More easy one, please. You can do it.
edit/p.s: I think I have written this on one or two emails or the text boxes of the browser or elsewhere, it does not make only those figures. It also make letters. But it is difficult to recognize those letters if it is written on a lump of ash of tobacco.
- Vous devez vous identifier ou créer un compte pour écrire des commentaires