Nonparametric Density Estimation under Adversarial Losses

In a general framework, we study how the choice of loss and the assumed smoothness of the underlying density together determine the minimax rate.