Hierarchical Bayesian Quadrature

DOI10.48550/arXiv.2607.10793MaRDI QIDQ7283984arXivFDO

Authors Philipp Hennig, Tim Weiland, Toni Karvonen

Publication date 12 July 2026

Published in arXiv

Tldr We propose a hierarchical approach to adaptively partition the domain of numerical integration into local regions with varying covariance functions, which allows Bayesian quadrature to handle non-stationarity in the integrand. We achieve improved accuracy and efficiency over standard Gaussian process-based methods on benchmark problems and real-world applications.

Full work available at URL https://doi.org/10.48550/arXiv.2607.10793

arXiv classification

cs.LGmath.NA

Cites work

• A simple method for comparing complex models: Bayesian model comparison for hierarchical multinomial processing tree models using Warp-III bridge sampling
• Bayes-Hermite quadrature
• Bayesian Probabilistic Numerical Integration with Tree-Based Models
• Bayesian treed Gaussian process models with an application to computer modeling
• Constructing Priors that Penalize the Complexity of Gaussian Random Fields
• Convergence Analysis of Deterministic Kernel-Based Quadrature Rules in Misspecified Settings
• Convergence Guarantees for Adaptive Bayesian Quadrature Methods
• Easily Computed Marginal Likelihoods from Posterior Simulation Using the THAMES Estimator
• Estimating the dimension of a model
• Fast Bayesian Inference with Batch Bayesian Quadrature via Kernel Recombination
• Gaussian Processes for Machine Learning
• Nested kriging predictions for datasets with a large number of observations
• Patchwork kriging for large-scale Gaussian process regression
• Physics-Informed Gaussian Process Regression Generalizes Linear PDE Solvers
• Probabilistic integration: a role in statistical computation?
• Probabilistic numerics and uncertainty in computations
• Remarks on algorithm 006: An adaptive algorithm for numerical integration over an N-dimensional rectangular region

This page was built for publication: Hierarchical Bayesian Quadrature

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7283984)