Projected Wasserstein Gradient Descent for High-Dimensional Bayesian Inference
From MaRDI portal
Abstract: We propose a projected Wasserstein gradient descent method (pWGD) for high-dimensional Bayesian inference problems. The underlying density function of a particle system of WGD is approximated by kernel density estimation (KDE), which faces the long-standing curse of dimensionality. We overcome this challenge by exploiting the intrinsic low-rank structure in the difference between the posterior and prior distributions. The parameters are projected into a low-dimensional subspace to alleviate the approximation error of KDE in high dimensions. We formulate a projected Wasserstein gradient flow and analyze its convergence property under mild assumptions. Several numerical experiments illustrate the accuracy, convergence, and complexity scalability of pWGD with respect to parameter dimension, sample size, and processor cores.
Recommendations
- Bayesian learning with Wasserstein barycenters
- Scalable Bayes via barycenter in Wasserstein space
- Approximate Bayesian Computation with the Wasserstein Distance
- Wasserstein gradients for the temporal evolution of probability distributions
- High-dimensional Bayesian optimization with projections using quantile Gaussian processes
- Robust Wasserstein profile inference and applications to machine learning
- Posterior contraction in Gaussian process regression using Wasserstein approximations
- High-dimensional Bayesian inference via the unadjusted Langevin algorithm
- Projected stochastic gradients for convex constrained problems in Hilbert spaces
- Wasserstein-based projections with applications to inverse problems
Cites work
- A computational framework for infinite-dimensional Bayesian inverse problems. I: The linearized case, with application to global seismic inversion
- A Fast and Scalable Computational Framework for Large-Scale High-Dimensional Bayesian Optimal Experimental Design
- A Fast and Scalable Method for A-Optimal Design of Experiments for Infinite-dimensional Bayesian Nonlinear Inverse Problems
- A-optimal encoding weights for nonlinear inverse problems, with application to the Helmholtz inverse problem
- Accelerating diffusions
- An explicit link between Gaussian fields and Gaussian Markov random fields: the stochastic partial differential equation approach
- Bayesian inference of heterogeneous epidemic models: application to COVID-19 spread accounting for long-term care facilities
- Certified dimension reduction in nonlinear Bayesian inverse problems
- Derivative-informed projected neural networks for high-dimensional parametric maps governed by PDEs
- Dimension-independent likelihood-informed MCMC
- Discrete gradients for computational Bayesian inference
- Fokker-Planck particle systems for Bayesian inference: computational approaches
- Geometric MCMC for infinite-dimensional inverse problems
- Gradient flows in metric spaces and in the space of probability measures
- Hessian-based adaptive sparse quadrature for infinite-dimensional Bayesian inverse problems
- Hessian-based model reduction for large-scale systems with initial-condition inputs
- HESSIAN-BASED SAMPLING FOR HIGH-DIMENSIONAL MODEL REDUCTION
- scientific article; zbMATH DE number 1909499 (Why is no real title available?)
- scientific article; zbMATH DE number 6159604 (Why is no real title available?)
- Interacting Langevin diffusions: gradient structure and ensemble Kalman sampler
- Inverse problems: a Bayesian perspective
- Is there an analog of Nesterov acceleration for gradient-based MCMC?
- MCMC using Hamiltonian dynamics
- Mean-variance risk-averse optimal control of systems governed by PDEs with random parameter fields using quadratic approximations
- Nonparametric kernel density estimation and its computational aspects
- Optimal design of acoustic metamaterial cloaks under uncertainty
- Optimal Transport
- Randomized algorithms for generalized Hermitian eigenvalue problems with application to computing Karhunen-Loève expansion.
- Rank-one modification of the symmetric eigenproblem
- Sparse approximation of triangular transports. II: The infinite-dimensional case
- Sparse deterministic approximation of Bayesian inverse problems
- Sparse, adaptive Smolyak quadratures for Bayesian inverse problems
- Sparse-grid, reduced-basis Bayesian inversion
- Sparse-grid, reduced-basis Bayesian inversion: nonaffine-parametric nonlinear equations
- Stochastic spectral methods for efficient Bayesian solution of inverse problems
- Taylor approximation and variance reduction for PDE-constrained optimal control under uncertainty
- Tensor train construction from tensor actions, with application to compression of large high order derivative tensors
- The Variational Formulation of the Fokker--Planck Equation
Cited in
(13)- Accelerated information gradient flow
- A note on parametric Bayesian inference via gradient flows
- Approximate Bayesian Computation with the Wasserstein Distance
- Optimal neural network approximation of Wasserstein gradient direction via convex optimization
- Lipschitz-regularized gradient flows and generative particle algorithms for high-dimensional scarce data
- First- and second-order adjoint methods for stochastic identification problems.
- Tensor train based sampling algorithms for approximating regularized Wasserstein proximal operators
- Sequential infinite-dimensional Bayesian optimal experimental design with derivative-informed latent attention neural operator
- Sliced-Wasserstein distances and flows on Cartan-Hadamard manifolds
- Accurate, scalable, and efficient Bayesian optimal experimental design with derivative-informed neural operators
- Sequential Bayesian design for efficient surrogate construction in the inversion of Darcy flows
- Nesterov acceleration in optimizing over probability measures
- A library of mirrors: deep neural nets in low dimensions are convex Lasso models with reflection features
This page was built for publication: Projected Wasserstein Gradient Descent for High-Dimensional Bayesian Inference
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5880609)
