A Unifying View of Sparse Approximate Gaussian Process Regression

Abstract

We provide a new unifying view, including all existing proper probabilistic\nsparse approximations for Gaussian process regression. Our approach relies on\nexpressing the effective prior which the methods are using. This\nallows new insights to be gained, and highlights the relationship between\nexisting methods. It also allows for a clear theoretically justified ranking\nof the closeness of the known approximations to the corresponding full GPs.\nFinally we point directly to designs of new better sparse approximations,\ncombining the best of the existing strategies, within attractive\ncomputational constraints.