In mathematical statistics and statistical learning theory, the problem I study concerns the precise characterization of the population excess risk in supervised learning — arguably the most fundamental problem in the field. Specifically, we focus on the following mathematical model.