Just watched this AI DEBATE with Yoshua Bengio and @GaryMarcus https://youtu.be/EeqwFjqFvJA and was inspired to write a blogpost about innate priors and how much information is encoded in the genome (coming soon...)
-
-
i dont understand this distinction. Hopi's expts show that genes determine whether the burrow is short or long. One mech to encode this in the genome involves a reward function on the length of the burrow (and a bit more). probably need more than "stop now" for birdspic.twitter.com/WspWDX3iAM
-
It doesn’t contradict the genomic encoding at all. But it puts a different emphasis as far as the need to *additionally* have a quite powerful RL system operating, to get the right behavior. It suggests a different AI emphasis: meta learning or evolving very specific reward fxns.
- 10 more replies
New conversation -
-
-
Well, that's the question! We know many of those behaviors can be pre-wired - we don't know how!
-
I don't know what
@GaryMarcus means by "pre-wired:" early in development? Species-typical? Universal? - 2 more replies
New conversation -
-
-
It has to be the former, yes? Otherwise the demand on the genome is impossible. And of course an allelic difference is rarely specific to one behaviour/trait, or has the same relationship to behavioural differences in all genetic contexts.
-
My hand wavy guess would be burrowing is a hardwired synaptic motor program and state-dependent modulatory systems that influence a range of covarying behaviours tell you when to start and stop.
End of conversation
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.