Approximation beats concentration? An approximation view on inference with smooth radial kernels

Open in new window