The most remarkable thing about the brain is that it wasn't designed. As a result, how the brain works is largely an irrelevant question.
-
-
Replying to @fchollet
Brain architecture is the result of an optimization process. So what really matters for reproducibility is the optimization problem itself.
10 replies 41 retweets 100 likes -
Replying to @fchollet
This is basically what
@KordingLab and@NKriegeskorte were arguing for at COSYNE this year. I think you all have a hammer so only see nails
2 replies 0 retweets 6 likes -
Replying to @the_moliver @fchollet and
Given a fully observable trained net, can we recover its cost function & learning rule? Else this call to arms has no ansatz for the brain.
1 reply 1 retweet 0 likes -
Replying to @desipoika @the_moliver and
Representations are almost impossible to interpret or invert to get at cost/learning as they reflect the world. Need to measure learning.
1 reply 2 retweets 2 likes -
Replying to @KordingLab @the_moliver and
hard! how successful has that approach (IOC) been in the motor learning field? how do we distinguish learning from developmental / innate?
1 reply 0 retweets 0 likes -
Replying to @desipoika @the_moliver and
Patch clamp cells and stimulate pre-synaptic and neuromodulatory cells. Also, motor control field could quantify RF changes with perturbtns.
1 reply 0 retweets 1 like -
Replying to @KordingLab @desipoika and
E.g.
@mattlark style experiments are way cool for asking these questions.1 reply 0 retweets 1 like
Look for the error signals! Is it doing video frame prediction, GAN-like comparisons, or what? Should be possible to track evolutionarily...
-
-
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
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.