Comment by taeric

My concern with a lot of signal processing techniques used in ml is that sometimes they presuppose things that may not be true.

That is, signal processing had Nyquist's rates. And typically knows there is an underlying signal. Does ml have either?