astonished that anyone knowledgeable could claim that neural nets are (obviously) an “abstraction of neural processing” when we don’t yet know how brains work. if you don’t know how Y works you can’t really speak with certainty about whether X is an abstraction of Y. Period.https://twitter.com/tyrell_turing/status/1200072223299657728 …
-
I'd say the initial inspiration for artificial neural nets (especially conv nets) comes from our understanding of how certain aspects of neural processing might work.
@ylecun how much did neuroscience inspire CNNs? how seriously do you take CNNs as a model of neuroscience? have you written about these two questions? i recall you being very critical of current neuromorphic work.
