Optimal Transport Bounds for Latent Density Smoothing on MTW(K>0) Manifolds

DOI10.48550/arXiv.2609.33629MaRDI QIDQ7362013zbMATH OpenarXivFDO

Authors Wonjun Lee, Wenyan Luo

Publication date 27 September 2026

Published in arXiv

Tldr This paper derives optimal transport bounds for three types of smoothing on manifolds: intrinsic, ambient, and latent; and demonstrates that sufficiently small intrinsic smoothing improves upon unsmoothed empirical measures under the Ma--Trudinger--Wang condition with positive cross-curvature.

Full work available at URL https://doi.org/10.48550/arXiv.2609.33629

arXiv classification

math.NA

Cites work

• A CENTRAL LIMIT THEOREM AND A STRONG MIXING CONDITION
• A Connection Between Score Matching and Denoising Autoencoders
• A Riemannian interpolation inequality à la Borell, Brascamp and Lieb
• Convergence of denoising diffusion models under the manifold hypothesis
• Convergence of Smoothed Empirical Measures With Applications to Entropy Estimation
• Curvature Measures
• Diffusion maps
• Embedding Riemannian manifolds by their heat kernel
• Estimation of non-normalized statistical models by score matching
• Geometry-Preserving Encoder/Decoder in Latent Generative Models
• Interpolating between Optimal Transport and MMD using Sinkhorn Divergences
• Laplacian Eigenmaps for Dimensionality Reduction and Data Representation
• Measure estimation on manifolds: an optimal transport approach
• Minimax estimation of smooth densities in Wasserstein distance
• On Estimation of a Probability Density Function and Mode
• On the mean speed of convergence of empirical and occupation measures in Wasserstein distance
• On the rate of convergence in Wasserstein distance of the empirical measure
• On the regularity of solutions of optimal transportation problems
• Partial regularity for optimal transport maps
• Regularity of optimal maps on the sphere: the quadratic cost and the reflector antenna
• Regularity of potential functions of the optimal transportation problem
• Riemannian geometry and geometric analysis
• Score-Based Generative Models Detect Manifolds
• Sharp asymptotic and finite-sample rates of convergence of empirical measures in Wasserstein distance
• Smooth p-Wasserstein Distance: Structure, Empirical Approximation, and Statistical Applications
• Statistical inference with regularized optimal transport
• The Speed of Mean Glivenko-Cantelli Convergence
• What regularized auto-encoders learn from the data-generating distribution

This page was built for publication: Optimal Transport Bounds for Latent Density Smoothing on MTW($K>0$) Manifolds

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7362013)