English

Know Your Boundaries: Constraining Gaussian Processes by Variational Harmonic Features

Machine Learning 2019-04-11 v1 Machine Learning

Abstract

Gaussian processes (GPs) provide a powerful framework for extrapolation, interpolation, and noise removal in regression and classification. This paper considers constraining GPs to arbitrarily-shaped domains with boundary conditions. We solve a Fourier-like generalised harmonic feature representation of the GP prior in the domain of interest, which both constrains the GP and attains a low-rank representation that is used for speeding up inference. The method scales as O(nm2)\mathcal{O}(nm^2) in prediction and O(m3)\mathcal{O}(m^3) in hyperparameter learning for regression, where nn is the number of data points and mm the number of features. Furthermore, we make use of the variational approach to allow the method to deal with non-Gaussian likelihoods. The experiments cover both simulated and empirical data in which the boundary conditions allow for inclusion of additional physical information.

Keywords

Cite

@article{arxiv.1904.05207,
  title  = {Know Your Boundaries: Constraining Gaussian Processes by Variational Harmonic Features},
  author = {Arno Solin and Manon Kok},
  journal= {arXiv preprint arXiv:1904.05207},
  year   = {2019}
}

Comments

Appearing in Proceedings of AISTATS 2019

R2 v1 2026-06-23T08:35:27.990Z