Groups
Category
An f-divergence measures how different two probability distributions P and Q are by averaging a convex function f of the density ratio p(x)/q(x) under Q.
Rademacher complexity is a data-dependent measure of how well a function class can fit random noise on a given sample.