Gaussian process variance reduction by location selection

Open Webpage

Lorenzo Bottarelli, Marco Loog

Published: 2019, Last Modified: 15 May 2025Pattern Recognit. Lett. 2019EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•Gradient Descent is proposed to minimize the posterior variance of a Gaussian process.•An additional fast heuristic is offered as an alternative to Gradient Descent.•Empirical evaluations show the benefits when compared to standard submodular approach.