A Distance-based Framework for Gaussian Processes over Probability Distributions

TL;DR

This paper introduces a framework for Gaussian process regression where inputs are probability distributions, utilizing distance measures between distributions to handle noisy or distributional inputs effectively.

Contribution

It proposes a novel distance-based approach for Gaussian processes that extends traditional methods to distributional inputs, addressing practical scenarios with noisy or uncertain data.

Findings

Framework effectively models distributional inputs

Demonstrated with a numerical example

Addresses practical noisy data scenarios

Abstract

Gaussian processes constitute a very powerful and well-understood method for non-parametric regression and classification. In the classical framework, the training data consists of deterministic vector-valued inputs and the corresponding (noisy) measurements whose joint distribution is assumed to be Gaussian. In many practical applications, however, the inputs are either noisy, i.e., each input is a vector-valued sample from an unknown probability distribution, or the probability distributions are the inputs. In this paper, we address Gaussian process regression with inputs given in form of probability distributions and propose a framework that is based on distances between such inputs. To this end, we review different admissible distance measures and provide a numerical example that demonstrates our framework.