Kosta Derpanis (sabbatical in Zurich)#932
Jack Gallant (UC Berkeley), in his talk “Representation of the World in the Human Brain,” made a really interesting observation about world models: evidence that the mammalian brain implements a Generative World Model is "shockingly poor".
Many users appreciate the neuroscientist's claim of shockingly poor evidence for mammalian generative world models and praise him as ahead of his time, while some dismiss the post as self-promotion.
"Generative World Models are highly correlated with other theories."
Isn’t that expected, given that the laws of nature apply to humans as well?
Cool!
@roeiherzig So was his "finding" (or lack thereof) mean that we shouldn't look for it either? 😅 And therefore all of our brain is a discriminative machine?
I should tell him to read out paper, as a slightly stronger piece of evidence :) https://www.nature.com/articles/s41562-023-01759-7
@roeiherzig is there a video of this
@stochasticchasm @roeiherzig here: https://www.youtube.com/live/cFwHAMf0kEc?si=63v85Cp_tnNt1aNh
@roeiherzig Relevant: https://aeon.co/essays/your-brain-does-not-process-information-and-it-is-not-a-computer
@roeiherzig Jack Gallant, always one step ahead of his time!
@Amir_Arsalan_ @roeiherzig cant go on for 1 minute on science twitter without someone plugging their shit
@roeiherzig Thanks for sharing!
Kosta Derpanis (sabbatical in Zurich)#932Jack Gallant (UC Berkeley), in his talk “Representation of the World in the Human Brain,” made a really interesting observation about world models: evidence that the mammalian brain implements a Generative World Model is "shockingly poor".