Solving and learning nonlinear PDEs with Gaussian processes

DOI10.1016/j.jcp.2021.110668OpenAlexW3198653666WikidataQ114163419 ScholiaQ114163419MaRDI QIDQ2133484

Bamdad Hosseini, Andrew M. Stuart, Houman Owhadi, Yi-Fan Chen

Publication date: 29 April 2022

Published in: Journal of Computational Physics (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/2103.12959

zbMATH Keywords

Gaussian processesinverse problemsnonlinear partial differential equationskernel methodsoptimal recovery

Mathematics Subject Classification ID

Artificial intelligence (68Txx) Nonparametric inference (62Gxx) Probabilistic methods, stochastic differential equations (65Cxx)

Related Items

Numerical methods for mean field games based on Gaussian processes and Fourier features, Physics-informed Karhunen-Loéve and neural network approximations for solving inverse differential equation problems, Learning dynamical systems from data: a simple cross-validation perspective. III: Irregularly-sampled time series, A framework for machine learning of model error in dynamical systems, Covariance models and Gaussian process regression for the wave equation. Application to related inverse problems, Manifold-constrained Gaussian process inference for time-varying parameters in dynamic systems, A framework based on symbolic regression coupled with eXtended physics-informed neural networks for gray-box learning of equations of motion from data, Physics-informed radial basis network (PIRBN): a local approximating neural network for solving nonlinear partial differential equations, Learning dynamical systems from data: a simple cross-validation perspective. IV: Case with partial observations, Localized Model Reduction for Nonlinear Elliptic Partial Differential Equations: Localized Training, Partition of Unity, and Adaptive Enrichment, Learning black- and gray-box chemotactic PDEs/closures from agent based Monte Carlo simulation data, Long-time integration of parametric evolution equations with physics-informed DeepONets, Gaussian process hydrodynamics, Tutorial on Amortized Optimization, Sobolev regularity of Gaussian random fields, Physics-informed information field theory for modeling physical systems with uncertainty quantification, Data-driven control of agent-based models: an equation/variable-free machine learning approach, Sparse Gaussian processes for solving nonlinear PDEs, Stochastic PDE representation of random fields for large-scale Gaussian process regression and statistical finite element analysis, Kernel methods are competitive for operator learning, A note on microlocal kernel design for some slow-fast stochastic differential equations with critical transitions and application to EEG signals, Learning ``best kernels from data in Gaussian process regression. With application to aerodynamics, One-shot learning of stochastic differential equations with data adapted kernels, Do ideas have shape? Idea registration as the continuous limit of artificial neural networks

Uses Software

• PDE-Net
• DGM
• PINNsNTK

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Unnamed Item
• Compression, inversion, and approximate PCA of dense kernel matrices at near-linear computational complexity
• Gamblets for opening the complexity-bottleneck of implicit schemes for hyperbolic and parabolic ODEs/PDEs with rough coefficients
• Average case optimality for linear problems
• Applied functional analysis. Functional analysis, Sobolev spaces and elliptic differential equations
• Meshless methods based on collocation with radial basis functions
• Bayesian deep convolutional encoder-decoder networks for surrogate modeling and uncertainty quantification
• The Deep Ritz Method: a deep learning-based numerical algorithm for solving variational problems
• The screening effect in kriging
• DGM: a deep learning algorithm for solving partial differential equations
• Model reduction and neural networks for parametric PDEs
• Kernel flows: from learning kernels from data into the abyss
• On the eigenvector bias of Fourier feature networks: from regression to solving multi-scale PDEs with physics-informed neural networks
• Physics-informed neural networks: a deep learning framework for solving forward and inverse problems involving nonlinear partial differential equations
• Some results on Tchebycheffian spline functions and stochastic processes
• Gaussian measure in Hilbert space and applications in numerical analysis
• Inverse problems: A Bayesian perspective
• Multigrid with Rough Coefficients and Multiresolution Operator Decomposition from Hierarchical Information Games
• Approximation of Bayesian Inverse Problems for PDEs
• Numerical Gaussian Processes for Time-Dependent and Nonlinear Partial Differential Equations
• Kernel techniques: From machine learning to meshless methods
• Optimization with PDE Constraints
• Pushing the Limits of Contemporary Statistics: Contributions in Honor of Jayanta K. Ghosh
• Shorted Operators. II
• On the small balls problem for equivalent Gaussian measures
• A review of numerical methods for nonlinear partial differential equations
• Can a finite element method perform arbitrarily badly?
• Consistency of empirical Bayes and kernel flow for hierarchical parameter estimation
• Operator-Adapted Wavelets, Fast Solvers, and Numerical Homogenization
• Solving high-dimensional partial differential equations using deep learning
• Sparse Cholesky Factorization by Kullback--Leibler Minimization
• Hidden fluid mechanics: Learning velocity and pressure fields from flow visualizations
• Parameter Estimation for Differential Equations: a Generalized Smoothing Approach
• Statistical Numerical Approximation
• Learning data-driven discretizations for partial differential equations
• Bayesian Probabilistic Numerical Methods
• Bayesian Numerical Homogenization
• Probabilistic numerics and uncertainty in computations
• Parameter Estimation for Differential Equation Models Using a Framework of Measurement Error in Regression Models
• A Correspondence Between Bayesian Estimation on Stochastic Processes and Smoothing by Splines
• MAP estimators and their consistency in Bayesian nonparametric inverse problems
• The elements of statistical learning. Data mining, inference, and prediction
• MCMC methods for functions: modifying old algorithms to make them faster