Jump to content

Talk:Cybenko theorem

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

What are the parameters w1, w2, ..., wn? What are alpha and theta? I suppose w's are weights. I think this part needs more clarification. 77.240.251.193 (talk) 16:49, 9 December 2009 (UTC)[reply]

Nevermind, I took a look at the theorem and found out that w's are weights from input to hidden layer, thetas are biases to the hidden layer and alphas are weights from the hidden layer to the output neuron. 213.164.108.65 (talk) 19:54, 9 December 2009 (UTC)[reply]