Probabilistic Numerical Methods for PDE-constrained Bayesian Inverse Problems

  title={Probabilistic Numerical Methods for PDE-constrained Bayesian Inverse Problems},
  author={Jon Cockayne and Chris. J. Oates and Timothy John Sullivan and Mark A. Girolami},
This paper develops meshless methods for probabilistically describing discretisation error in the numerical solution of partial differential equations. This construction enables the solution of Bayesian inverse problems while accounting for the impact of the discretisation of the forward problem. In particular, this drives statistical inferences to be more conservative in the presence of significant solver error. Theoretical results are presented describing rates of convergence for the… 

Figures from this paper

Sampling Methods for Bayesian Inference Involving Convergent Noisy Approximations of Forward Maps

This paper considers mean-square convergent random approximations, and quantifies how Monte Carlo errors propagate from the forward map to the solution of the inverse problems.

Differentiable Likelihoods for Fast Inversion of 'Likelihood-Free' Dynamical Systems

Gaussian ODE filtering is employed (a probabilistic numerical method for ODEs) to construct a local Gaussian approximation to the likelihood, which yields tractable estimators for the gradient and Hessian of the (log-)likelihood.

A Bayesian Conjugate Gradient Method (with Discussion)

This paper proposes a novel statistical model for this error, set in a Bayesian framework, and is a strict generalisation of the conjugate gradient method, which is recovered as the posterior mean for a particular choice of prior.

Physics-Informed Gaussian Process Regression Generalizes Linear PDE Solvers

The results enable the seamless integration of mechanistic models as modular building blocks into probabilistic models by blurring the boundaries between numerical analysis and Bayesian inference.

Strong convergence rates of probabilistic integrators for ordinary differential equations

This work extends the convergence analysis of probabilistic integrators for deterministic ordinary differential equations, as proposed by Conrad et al, to establish mean-square convergence in the uniform norm on discrete- or continuous-time solutions under relaxed regularity assumptions on the driving vector fields and their induced flows.

Inference of Nonlinear Partial Differential Equations via Constrained Gaussian Processes

This paper proposes a novel method for estimating unknown parameters in PDEs, called PDE-Informed Gaussian Process Inference (PIGPI), which completely bypasses the numerical solver for PDE, thus achieving drastic savings in computation time, especially for nonlinear P DEs.

Randomised one-step time integration methods for deterministic operator differential equations

This work proves pathwise and expected error bounds on the random trajectories, given an assumption on the local truncation error of the underlying deterministic time integration and an assumption that the absolute moments of the random variables decay with the time step.

Randomised one-step time integration methods for deterministic operator differential equations

Uncertainty quantification plays an important role in problems that involve inferring a parameter of an initial value problem from observations of the solution. Conrad et al. (Stat Comput

Random time step probabilistic methods for uncertainty quantification in chaotic and geometric numerical integration

A novel probabilistic numerical method for quantifying the uncertainty induced by the time integration of ordinary differential equations (ODEs) is introduced. Departing from the classical strategy



Probabilistic Meshless Methods for Partial Differential Equations and Bayesian Inverse Problems

A probabilistic numerical method for solution of partial differential equations (PDEs) and application of that method to PDE-constrained inverse problems, which enables the solution of challenging inverse problems whilst accounting for the impact of discretisation error due to numerical solution of the PDE.

Inverse problems: A Bayesian perspective

The Bayesian approach to regularization is reviewed, developing a function space viewpoint on the subject, which allows for a full characterization of all possible solutions, and their relative probabilities, whilst simultaneously forcing significant modelling issues to be addressed in a clear and precise fashion.

Approximation of stochastic partial differential equations by a kernel-based collocation method

In this paper we present the theoretical framework needed to justify the use of a kernel-based collocation method (meshfree approximation method) to estimate the solution of high-dimensional

Bayesian Numerical Homogenization | Multiscale Modeling & Simulation | Vol. 13, No. 3 | Society for Industrial and Applied Mathematics

Numerical homogenization, i.e., the finite-dimensional approximation of solution spaces of PDEs with arbitrary rough coefficients, requires the identification of accurate basis elements. These basis

Statistical analysis of differential equations: introducing probability measures on numerical solutions

It is shown that a wide variety of existing solvers can be randomised, inducing a probability measure over the solutions of ordinary and partial differential equation models, and the formal means to incorporate this uncertainty in a statistical model and its subsequent analysis are provided.

Solving differential equations with radial basis functions: multilevel methods and smoothing

Some of the meshless radial basis function methods used for the numerical solution of partial differential equations are reviewed and the important role of smoothing within a multilevel framework is demonstrated.

Bayesian Numerical Homogenization

  • H. Owhadi
  • Computer Science
    Multiscale Model. Simul.
  • 2015
It is suggested that the answer to the above questions could be positive based on the reformulation of numerical homogenization as a Bayesian inference problem in which a given PDE with rough coefficients is excited with noise and one tries to estimate the value of the solution at a given point based on a finite number of observations.

Deflation Techniques for Finding Distinct Solutions of Nonlinear Partial Differential Equations

This paper presents an infinite-dimensional deflation algorithm for systematically modifying the residual of a nonlinear PDE problem to eliminate known solutions from consideration, and enables the Newton--Kantorovitch iteration to converge to several different solutions, even starting from the same initial guess.

Probabilistic numerics and uncertainty in computations

It is shown that the probabilistic view suggests new algorithms that can flexibly be adapted to suit application specifics, while delivering improved empirical performance.

Prior distributions for variance parameters in hierarchical models (comment on article by Browne and Draper)

Various noninformative prior distributions have been suggested for scale parameters in hierarchical models. We construct a new folded-noncentral-t family of conditionally conjugate priors for