Dear all,
I am looking for a measure of complexity of a mapping between two high-dimensional spaces.
Imagine I have a mapping:
(x1,...,xn) = F(y1,...,ym)
Imagine, I can represent this mapping as a composition of some other functions, say
(x1,..,xn) = f1(y1,...,yk)+f2(yk+1, ..., ym)
I want some measure according to which the complexity of the function F that does not allow for such representation would be higher than the summed complexity of the three functions: f1, f2, "+".
Best,
Alex