Scalable Bayesian Quadrature via Operator-Adapted Features

DOI10.48550/ARXIV.2610.38212MaRDI QIDQ7409136arXivFDO

Authors Huanyan Zhu, Cheng Li

Publication date 28 October 2026

Published in arXiv

Tldr We introduce scalable Bayesian quadrature via an operator-adapted finite-feature approximation to kernel methods. Our method constructs a finite-dimensional working kernel from eigenvalues and eigenfunctions of a second-order differential operator, achieving optimal statistical rates with $m \lesssim n$ basis functions.

Full work available at URL https://doi.org/10.48550/ARXIV.2610.38212

arXiv classification

math.NAmath.STstat.ML

Cites work

• Adaptive Bayesian estimation using a Gaussian random field with inverse gamma bandwidth
• An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach
• Bayes-Hermite quadrature
• Contraction rates for sparse variational approximations in Gaussian process regression
• Convergence Analysis of Deterministic Kernel-Based Quadrature Rules in Misspecified Settings
• Convergence Guarantees for Gaussian Process Means With Misspecified Likelihoods and Smoothness
• Convergence of sparse variational inference in Gaussian processes regression
• Convergences of regularized algorithms and stochastic gradient methods with random projections
• Efficient Numerical Integration in Reproducing Kernel Hilbert Spaces via Leverage Scores Sampling
• Equispaced Fourier representations for efficient Gaussian process regression from a billion data points
• Estimating the parameters of some common Gaussian random fields with nugget under fixed-domain asymptotics
• Fast Bayesian Inference with Batch Bayesian Quadrature via Kernel Recombination
• Functional analysis, Sobolev spaces and partial differential equations
• Gaussian Processes for Machine Learning
• Hilbert space methods for reduced-rank Gaussian process regression
• Kernel Thinning
• On some nonlinear fractional equations involving the Bessel operator
• On the distribution of points in a cube and the approximate evaluation of integrals
• On the equivalence between kernel quadrature rules and random feature expansions
• Optimal estimation of variance in nonparametric regression with random design
• Optimal Hardy-weights for the (p,A)-Laplacian with a potential term
• Pointwise uncertainty quantification for sparse variational Gaussian process regression with a Brownian motion prior
• Positively Weighted Kernel Quadrature via Subsampling
• Posterior concentration for Gaussian process priors under rescaled and hierarchical Matérn and confluent hypergeometric covariance functions
• Probabilistic integration: a role in statistical computation?
• Random Fourier Features for Kernel Ridge Regression: Approximation Bounds and Statistical Guarantees
• Randomized sketches for kernels: fast and optimal nonparametric regression
• Residual variance estimation using a nearest neighbor statistic
• Sampling-based Nyström Approximation and Kernel Quadrature
• The covering radius of randomly distributed points on a manifold
• The Legendre equation and its self-adjoint operators
• Towards a unified analysis of random Fourier features
• Uncertainty quantification for sparse spectral variational approximations in Gaussian process regression
• Uniform approximation of common Gaussian process kernels using equispaced Fourier grids

This page was built for publication: Scalable Bayesian Quadrature via Operator-Adapted Features

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7409136)