Oracle inequalities for computationally budgeted model selection

[edit]

Alekh Agarwal, John C. Duchi, Peter L. Bartlett, Clement Levrard ;
Proceedings of the 24th Annual Conference on Learning Theory, PMLR 19:69-86, 2011.

Abstract

We analyze general model selection procedures using penalized empirical loss minimization under computational constraints. While classical model selection approaches do not consider computational aspects of performing model selection, we argue that any practical model selection procedure must not only trade off estimation and approximation error, but also the effects of the computational effort required to compute empirical minimizers for different function classes. We provide a framework for analyzing such problems, and we give algorithms for model selection under a computational budget. These algorithms satisfy oracle inequalities that show that the risk of the selected model is not much worse than if we had devoted all of our computational budget to the best function class.

Related Material