For practical purposes, functionals are just infinite-dimensional functions. In practice, a function is always represented numerically using a finite basis (gri

For practical purposes, functionals are just infinite-dimensional functions. In practice, a function is always represented numerically using a finite basis (grid, splines, Fourier), so functionals become just high-dimensional functions. At which point we are back to ordinary machine learning. So I’m not sure what’s the point here.