Bayesian learning with Wasserstein barycenters

DOI10.1051/PS/2022015MaRDI QIDQ6175889zbMATH OpenOpenAlexWikidataFDO

Authors Julio D. Backhoff Veraguas, Joaquin Fontbona, Gonzalo Rios, Felipe Tobar

Publication date 21 August 2023

Published in ESAIM: Probability and Statistics (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/1805.10833

zbMATH Keywords

consistencyMCMCBayesian learningnon-parametric estimationstochastic gradient descent in Wasserstein spaceWasserstein distance and barycenter

Mathematics Subject Classification ID

Bayesian inference (62F15) Nonparametric estimation (62G05) Asymptotic properties of nonparametric inference (62G20) Multivariate distribution of statistics (62H10)

Abstract: We introduce and study a novel model-selection strategy for Bayesian learning, based on optimal transport, along with its associated predictive posterior law: the Wasserstein population barycenter of the posterior law over models. We first show how this estimator, termed Bayesian Wasserstein barycenter (BWB), arises naturally in a general, parameter-free Bayesian model-selection framework, when the considered Bayesian risk is the Wasserstein distance. Examples are given, illustrating how the BWB extends some classic parametric and non-parametric selection strategies. Furthermore, we also provide explicit conditions granting the existence and statistical consistency of the BWB, and discuss some of its general and specific properties, providing insights into its advantages compared to usual choices, such as the model average estimator. Finally, we illustrate how this estimator can be computed using the stochastic gradient descent (SGD) algorithm in Wasserstein space introduced in a companion paper arXiv:2201.04232v2 [math.OC], and provide a numerical example for experimental validation of the proposed method.

Cites work

• scientific article; zbMATH DE number 1909499 (Why is no real title available?)
• scientific article; zbMATH DE number 3053873 (Why is no real title available?)
• A Stochastic Approximation Method
• A class of Wasserstein metrics for probability distributions
• A fixed-point approach to barycenters in Wasserstein space
• A smoothed dual approach for variational Wasserstein problems
• An introduction to MCMC for machine learning
• An invitation to statistics in Wasserstein space
• Asymptotic equivalence of empirical likelihood and Bayesian MAP
• Barycenters in the Wasserstein space
• Bayesian inference with optimal maps
• Bayesian learning with Wasserstein barycenters
• Characterization of barycenters in the Wasserstein space by averaging optimal transport maps
• Computational optimal transport. With applications to data sciences
• Concentration inequalities and model selection. Ecole d'Eté de Probabilités de Saint-Flour XXXIII -- 2003.
• Ensemble samplers with affine invariance
• Entropy-regularized 2-Wasserstein distance between Gaussian measures
• Existence and consistency of Wasserstein barycenters
• Fréchet means and Procrustes analysis in Wasserstein space
• Fundamentals of nonparametric Bayesian inference
• Gradient flows in metric spaces and in the space of probability measures
• Handbook of Markov Chain Monte Carlo
• Learning to generate Wasserstein barycenters
• Limiting Behavior of Posterior Distributions when the Model is Incorrect
• Machine learning. A probabilistic perspective
• On Bayes procedures
• On the consistency of Bayes estimates
• Optimal Transport
• Optimal coupling of multivariate distributions and stochastic processes
• Optimal transport for applied mathematicians. Calculus of variations, PDEs, and modeling
• Optimal transportation with infinitely many marginals
• Sanov's theorem in the Wasserstein distance: a necessary and sufficient condition
• The Bernstein-von Mises theorem under misspecification
• The Frechet distance between multivariate normal distributions
• Wasserstein Barycenters Are NP-Hard to Compute
• Wasserstein barycenters over Riemannian manifolds
• Wide consensus aggregation in the Wasserstein space. Application to location-scatter families

Cited in

(8)

• Bayesian learning with Wasserstein barycenters
• Projected Wasserstein Gradient Descent for High-Dimensional Bayesian Inference
• Stochastic gradient descent for barycenters in Wasserstein space
• The Wasserstein space of stochastic processes
• Generalized Bayes Quantification Learning under Dataset Shift
• Wasserstein medians: robustness, PDE characterization, and numerics
• Stochastic approximation versus sample average approximation for Wasserstein barycenters
• Approximate Bayesian Computation with the Wasserstein Distance

This page was built for publication: Bayesian learning with Wasserstein barycenters

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6175889)