”Perhaps GPT-2 is a bizarre blind alley passing through an extremely virtuosic kind of simulated competence that will, despite appearances, never quite lead into real competence.” .. yes, preciselyhttps://twitter.com/nostalgebraist/status/1211801012945637376 …
-
-
that would seem to require a vastly different neural architecture .... In deep computer vision something analogous was attempted in the form of InfoGan variants with Bayes nets of latent variables... but nobody got these to work after much trying...
-
it could be that such complex semantics friendly deep NN architectures will work if one upgrades backdrop to some better learning algorithm... but I'm not so sure ...may be that recognizing huge weighted libraries of simple patterns is what deep NN perception nets are good at...
- 3 more replies
New 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.