PDE-Residual-Informed Kriging
Abstract
Physics-informed Gaussian process (GP) methods provide a probabilistic framework for PDE solving. Existing approaches incorporate PDE information in different ways: pointwise conditioning uses the governing equation only at finitely many selected locations, globally constrained GPs restrict the admissible functions to satisfy the PDE exactly, and residual-penalized methods control the PDE mismatch over the whole domain through an additional penalty weight. What is still missing is a GP formulation that directly incorporates global PDE-residual information of the entire domain into its covariance structure, while retaining model flexibility and standard GP conditioning without a separate residual-weight parameter. To solve this issue, this study proposes PDE-residual-informed Kriging (PRIK) based on a PDE-residual kernel. The PDE-residual kernel is constructed by equipping a kernel-section space with an inner product defined by the L2 inner product of PDE residuals. This construction makes PRIK yield the posterior mean with minimum integrated squared PDE residual. It also provides direct statistical interpretations: the variance-scale estimate equals the normalized PDE-residual energy (i.e., the integrated squared PDE residual of the posterior mean divided by the number of labeled observations), and the profile likelihood balances this residual energy against model complexity. Under the stated regularity and stability conditions, we establish H2-quasi-optimality and convergence rates for the posterior mean, together with a nonasymptotic bound for its PDE-residual norm. We further show that the normalized posterior kernel converges to a variational Green-type kernel and derive a uniform nonasymptotic contraction bound for the posterior variance. Numerical studies provide empirical evidence consistent with the theory and demonstrate the PDE-solving capability of PRIK.
Related
- A Graph-informed Kriging model for sample-efficient Bayesian optimization of wind farm layouts
- No-Harm Physics-Informed Inverse Learning with Residual-Calibrated Uncertainty
- ARISE: An adaptive residual-informed stability ensemble for feature selection in small-sample biomedical omics
- Distance-Residual Physics-Informed Neural Networks: A Deep Learning Framework for Differential and Partial Differential Inclusions
- Rational Quadratic Kriging
BibTeX
@misc{tan2026pderesidualinformed,
title = {{PDE-Residual-Informed Kriging}},
author = {Yu Tan and Shichang Du and Yongxiang Li},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.41590},
eprint = {2610.41590},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.41590},
}
RIS
TY - UNPB
TI - PDE-Residual-Informed Kriging
AU - Yu Tan
AU - Shichang Du
AU - Yongxiang Li
PY - 2026
DA - 2026/10/30
PB - arXiv
DO - 10.48550/arxiv.2610.41590
UR - https://arxiv.org/abs/2610.41590
AB - Physics-informed Gaussian process (GP) methods provide a probabilistic framework for PDE solving. Existing approaches incorporate PDE information in different ways: pointwise conditioning uses the governing equation only at finitely many selected locations, globally constrained GPs restrict the admissible functions to satisfy the PDE exactly, and residual-penalized methods control the PDE mismatch over the whole domain through an additional penalty weight. What is still missing is a GP formulation that directly incorporates global PDE-residual information of the entire domain into its covariance structure, while retaining model flexibility and standard GP conditioning without a separate residual-weight parameter. To solve this issue, this study proposes PDE-residual-informed Kriging (PRIK) based on a PDE-residual kernel. The PDE-residual kernel is constructed by equipping a kernel-section space with an inner product defined by the L^2 inner product of PDE residuals. This construction makes PRIK yield the posterior mean with minimum integrated squared PDE residual. It also provides direct statistical interpretations: the variance-scale estimate equals the normalized PDE-residual energy (i.e., the integrated squared PDE residual of the posterior mean divided by the number of labeled observations), and the profile likelihood balances this residual energy against model complexity. Under the stated regularity and stability conditions, we establish H^2-quasi-optimality and convergence rates for the posterior mean, together with a nonasymptotic bound for its PDE-residual norm. We further show that the normalized posterior kernel converges to a variational Green-type kernel and derive a uniform nonasymptotic contraction bound for the posterior variance. Numerical studies provide empirical evidence consistent with the theory and demonstrate the PDE-solving capability of PRIK.
ER -