Abstract:Gaussian process (GP) regression is a fundamental tool in Bayesian statistics. It is also known as kriging and is the Bayesian counterpart to the frequentist kernel ridge regression. Most of the theoretical work on GP regression has focused on a large-$n$ asymptotics, i.e. as the amount of data increases. Fixed-sample analysis is much more difficult outside of simple cases, such as locations on a regular grid. In this work we perform a fixed-sample analysis that was first studied in the context of approximation theory by Driscoll & Fornberg (2002), called the ``flat limit''. In flat-limit asymptotics, the goal is to characterise kernel methods as the length-scale of the kernel function tends to infinity, so that kernels appear flat over the range of the data. Surprisingly, this limit is well-defined, and displays interesting behaviour: Driscoll & Fornberg showed that radial basis interpolation converges in the flat limit to polynomial interpolation, if the kernel is Gaussian. Subsequent work showed that this holds true in the multivariate setting as well, but that kernels other than the Gaussian may have (polyharmonic) splines as the limit interpolant. Leveraging recent results on the spectral behaviour of kernel matrices in the flat limit, we study the flat limit of Gaussian process regression. Results show that Gaussian process regression tends in the flat limit to (multivariate) polynomial regression, or (polyharmonic) spline regression, depending on the kernel. Importantly, this holds for both the predictive mean and the predictive variance, so that the posterior predictive distributions become equivalent. Our results have practical consequences: for instance, they show that optimal GP predictions in the sense of leave-one-out loss may occur at very large length-scales, which would be invisible to current implementations because of numerical difficulties.

Images of Gaussian and other stochastic processes under closed, densely-defined, unbounded linear operators

Images of Gaussian and other stochastic processes under closed, densely-defined, unbounded linear operators

Kernel Generated Two-Time Parameter Gaussian Processes and Some of Their Path Properties

Path Properties of Kernel Generated Two-Time Parameter Gaussian Processes

On Moduli of Continuity for Gaussian and L2-Norm Squared Processes Generated by Ornstein-Uhlenbeck Processes

ON LARGE INCREMENTS OF L P VALUED GAUSSIAN PROCESSES

Gaussian processes in Non-commutative probability

Quasi-invariant Theorem on the Gaussian Path Space

Hilbert space valued Gaussian processes, their kernels, factorizations, and covariance structure

Uniform approximation of common Gaussian process kernels using equispaced Fourier grids

Stochastic Calculus with respect to Gaussian Processes: Part I

Gaussian Process Regression in the Flat Limit

Variational Gaussian Processes: A Functional Analysis View

Sample Path Regularity of Gaussian Processes from the Covariance Kernel

Stationary Kernels and Gaussian Processes on Lie Groups and their Homogeneous Spaces I: the compact case

Compactly-supported nonstationary kernels for computing exact Gaussian processes on big data

Wiener Chaos in Kernel Regression: Towards Untangling Aleatoric and Epistemic Uncertainty

Stationary Kernels and Gaussian Processes on Lie Groups and their Homogeneous Spaces II: non-compact symmetric spaces

Integration in reproducing kernel Hilbert spaces of Gaussian kernels

A Unifying Perspective on Non-Stationary Kernels for Deeper Gaussian Processes