aright, i am trying to figure how neural networks works so i can create an OCR and take over the universe. i've come to a point where i don't really understand hidden layers.
how do i know how many hidden nodes do i need in a layer? or how do i know how many layers i need anyway? what's the purpose of hidden layers? can't i just connect the input nodes to the output nodes and let it do the processing?
[link][12 comments]