Agree with limitations of various methodologies (deep learning, svms, etc), but Hinton not critiquing neural architectures, rather back prop
Deep Learning is not enough, and we need to start over -- Hinton confirms what I have been saying for two decades.https://www.axios.com/ai-pioneer-advocates-starting-over-2485537027.html …
-
-
-
Hinton was also critiquing the reliance on massive data.
- 1 more reply
New conversation -
-
-
Gary, what Hinton said here is that neural nets need more unsupervised learning. Didn't think that's what you've been saying for 2 decades?
-
mostly predicted need to start over & not rely on backprop & massive labeled data, but re unsupervised see story of my daughter in nyt oped
- 1 more reply
New conversation -
-
-
Has any reasonable person ever said that gradient-based function approximation, by itself, solves all problems relating to intelligence?
-
Do you consider Elon Musk reasonable? :)
- 1 more reply
New conversation -
-
-
I think he's saying backprop may not be enough not that deep learning in general is not enough...?
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
I'm deeply suspicious of "start-over" narratives. Science is more incremental than our romantic intuitions (or Kuhn) would have it.
-
However, it’s never too late to question the fundamentals, the truth - those that we “believe” we fully understand.
End of conversation
New conversation -
-
-
Also, this quote... The future depends on some graduate student who is deeply suspicious of everything I have said.
- 2 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.