Finite-Dimensional Approximation of Gaussian Processes

Part of Advances in Neural Information Processing Systems 11 (NIPS 1998)

Bibtex Metadata Paper


Giancarlo Ferrari-Trecate, Christopher Williams, Manfred Opper


Gaussian process (GP) prediction suffers from O(n3) scaling with the data set size n. By using a finite-dimensional basis to approximate the GP predictor, the computational complexity can be reduced. We de(cid:173) rive optimal finite-dimensional predictors under a number of assump(cid:173) tions, and show the superiority of these predictors over the Projected Bayes Regression method (which is asymptotically optimal). We also show how to calculate the minimal model size for a given n. The calculations are backed up by numerical experiments.