Entropic optimal transport is maximum-likelihood deconvolution
From MaRDI portal
Abstract: We give a statistical interpretation of entropic optimal transport by showing that performing maximum-likelihood estimation for Gaussian deconvolution corresponds to calculating a projection with respect to the entropic optimal transport distance. This structural result gives theoretical support for the wide adoption of these tools in the machine learning community.
Recommendations
- Entropic optimal transport: geometry and large deviations
- Computational optimal transport. With applications to data sciences
- Convergence of entropic schemes for optimal transport and gradient flows
- Empirical regularized optimal transport: statistical theory and applications
- Entropic regularization of continuous optimal transport problems
Cites work
- A survey of the Schrödinger problem and some of its connections with optimal transport
- Aggregation by exponential weighting, sharp PAC-Bayesian bounds and sparsity
- Convolutional Wasserstein distances: efficient optimal transportation on geometric domains
- Deconvolution for the Wasserstein metric and geometric inference
- Exponential screening and optimal rates of sparse estimation
- Kullback-Leibler aggregation and misspecified generalized linear models
- Learning by mirror averaging
- Mathematical foundations of infinite-dimensional statistical models
- Measurement Error in Nonlinear Models
- Mirror averaging with sparsity priors
- Mixture models: theory, geometry and applications
- Numerical resolution of an “unbalanced” mass transport problem
- On minimum Kantorovich distance estimators
- On the estimation of quadratic functionals
- Optimal Rates of Convergence for Deconvolving a Density
- Optimal entropy-transport problems and a new Hellinger-Kantorovich distance between positive measures
- Optimal transport for applied mathematicians. Calculus of variations, PDEs, and modeling
- Sparse estimation by exponential weighting
- Sparse regression learning by aggregation and Langevin Monte-Carlo
- The earth mover's distance as a metric for image retrieval
Cited in
(15)- Toward a mathematical theory of trajectory inference
- Fast convergence of empirical barycenters in Alexandrov spaces and the Wasserstein space
- Cortically based optimal transport
- Asymptotics for semidiscrete entropic optimal transport
- Asymptotics of discrete Schrödinger bridges via chaos decomposition
- Semi-discrete optimal transport: hardness, regularization and numerical solution
- An Improved Central Limit Theorem and Fast Convergence Rates for Entropic Transportation Costs
- A registration method for reduced basis problems using linear optimal transport
- Minimax estimation of smooth optimal transport maps
- Minimax estimation of smooth densities in Wasserstein distance
- Recent theoretical advances in decentralized distributed convex optimization
- Domain decomposition for entropy regularized optimal transport
- Estimation of Wasserstein distances in the spiked transport model
- Recovering Latent Variables by Matching
- Multivariate ranks and quantiles using optimal transport: consistency, rates and nonparametric testing
This page was built for publication: Entropic optimal transport is maximum-likelihood deconvolution
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1632830)
