A human does not use just the text. A human uses the full extent of shared consciousness, culture, current events, imagination, etc. Translation is a creative endeavour, not a mechanical one.
-
-
Replying to @marcan42 @TigerGold59 and
If you really think translation is just turning meaning from one language into another, you should try actually doing it some time. It's not that easy. People who have never translated anything serious often have no idea about just how hard it is.
1 reply 0 retweets 2 likes -
Replying to @marcan42 @TigerGold59 and
Translating means understanding jokes and coming up with new ones in the target language. Translating means keeping in your head the full context of whatever you're translating and reasoning about it. Translating means knowing what is going to happen in the story in the *future*.
1 reply 0 retweets 2 likes -
Replying to @marcan42 @TigerGold59 and
Have you ever read a professionally translated book? Ever encounter a pun? Do you think that came straight from the exact same pun in the original book? It didn't. The translator put that in, either as a replacement for a different pun that is untranslatable, or to fit the mood.
2 replies 0 retweets 1 like -
This Tweet is unavailable.
-
An AI that could do that might as well be human itself and be given human rights. Machine learning cannot do that, no matter how much you "train" it. You're taking about true Artificial Intelligence, which does not exist yet.
1 reply 0 retweets 0 likes -
This Tweet is unavailable.
-
That was my point from the very beginning. And machine learning is not capable of that. Just ask any machine learning expert. That's not how it works.
1 reply 0 retweets 0 likes -
This Tweet is unavailable.
-
Yes, there is. Because a true AI is not something you "train" and then "use" just like a human is not something you "train" and then "use". A true AI is something that has to *live, experience, and learn on its own*.
1 reply 0 retweets 0 likes
It doesn't matter how many layers you use. Machine learning networks are not capable of the moral equivalent of consciousness. They do not have the architecture required for that, no matter how big you make them.
-
-
This Tweet is unavailable.
-
No. Not with the architectures and approaches we currently use for machine learning, no matter how large you make the network nor the training set.
1 reply 0 retweets 1 like
End of conversation
-
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.