Tomaso Poggio: Brains, Minds, and Machines
Lex Fridman Podcast
00:00
How Many Neurons Do You Need to Approximate?
The universality, universal approximation theorem says you can approximate any computable function with just a finite number of nurons in a single hidden layer. But it's possible that there's alast those solutions that aren't very good. How many monomials i need to get a good approximation depends on the dimensionality of your function and how many variables you have.
Transcript
Play full episode
