Adaptive Empirical Distributions in the Framework of Inverse Problems