Also I think "the brain minimizes an objective function" is a vacuous statement. All dynamical systems can be reframed as solving a variational problem.
These last two are concrete and actionable. They may mean that you can only really test more specific and refined implementation-level models, as opposed to general classes. But maybe not.
-
-
Well getting back to the tweet that set off this thread, I do hope in 100 years we have a much clearer view of things.
-
I don't see why it has to be backprop or belief prop. It could be something that takes on facets of both and more depending. I think this is where the evidence is already pointing. So while backprop is physically unlikely, it can work as long as subpopulations of neurons can
- 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.