3 ms·
Another interesting thing is that perceptron can approximate any non-linear function with just a single hidden layer. http://cstheory.stackexchange.com/questio
by supercarrot 10y ago
Another interesting thing is that perceptron can approximate any non-linear function with just a single hidden layer.
http://cstheory.stackexchange.com/questions/17545/universal-approximation-theorem-neural-networks http://cstheory.stackexchange.com/questions/17545/universal-...