Worst-Case Bounds for Gaussian Process Models

Part of Advances in Neural Information Processing Systems 18 (NIPS 2005)

Bibtex Metadata Paper


Sham M. Kakade, Matthias W. Seeger, Dean P. Foster


We present a competitive analysis of some non-parametric Bayesian al- gorithms in a worst-case online learning setting, where no probabilistic assumptions about the generation of the data are made. We consider models which use a Gaussian process prior (over the space of all func- tions) and provide bounds on the regret (under the log loss) for com- monly used non-parametric Bayesian algorithms — including Gaussian regression and logistic regression — which show how these algorithms can perform favorably under rather general conditions. These bounds ex- plicitly handle the infinite dimensionality of these non-parametric classes in a natural way. We also make formal connections to the minimax and minimum description length (MDL) framework. Here, we show precisely how Bayesian Gaussian regression is a minimax strategy.