Re:Why Deep Learning Cannot Be Applied to Natural Languages ...
by nl on Monday January 02, 2017 @06:00PM
(#13305255)
Why do you say German is notoriously bad? I don't speak German, but that linked page reads fine to me, and their neural translation system has state-of-the-art performance for both English->German and German->English at least[1]. I would have thought that the well-structured nature of German makes it reasonably easy to translate, assuming you have a NN architecture with sufficient memory range.
Re:Why Deep Learning Cannot Be Applied to Natural Languages ...
by fnl on Tuesday January 03, 2017 @03:30AM
(#13307680)
The translation isn't "only nonsense", and in many cases even grammatically correct, but mostly semantically wrong, or rather weird. In any case, a long shot from a human translation. But I wouldn't agree that it's "quite good". And, this is the easy case. EN->DE works even worse.
Re:Why Deep Learning Cannot Be Applied to Natural Languages ...
by Houshalter on Monday January 02, 2017 @06:06PM
(#13305284)
Doesn't German have a ton more compound words though? That could be an issue because their system is word based.
[1] Google Translate on the German Wikipedia entry for Weihnachten (X-mas): https://translate.googleusercontent.com/translate_c?depth=1&...