Locally Smoothed Gaussian Process Regression

TL;DR

This paper introduces a localized Gaussian process regression framework that accelerates computation by down-weighting distant data points, achieving competitive accuracy with significant speedups over traditional GPR methods.

Contribution

The paper presents a novel localization kernel approach for GPR that enhances computational efficiency while maintaining predictive performance.

Findings

Achieves comparable accuracy to full GPR and deep Gaussian processes.

Provides significant computational speedups due to sparsification.

Demonstrates effectiveness through extensive experiments.

Abstract

We develop a novel framework to accelerate Gaussian process regression (GPR). In particular, we consider localization kernels at each data point to down-weigh the contributions from other data points that are far away, and we derive the GPR model stemming from the application of such localization operation. Through a set of experiments, we demonstrate the competitive performance of the proposed approach compared to full GPR, other localized models, and deep Gaussian processes. Crucially, these performances are obtained with considerable speedups compared to standard global GPR due to the sparsification effect of the Gram matrix induced by the localization operation.