Sharp bounds for max-sliced Wasserstein distances

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autore principale: Boedihardjo, March T.
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866916495501557760
author Boedihardjo, March T.
author_facet Boedihardjo, March T.
contents We obtain essentially matching upper and lower bounds for the expected max-sliced 1-Wasserstein distance between a probability measure on a separable Hilbert space and its empirical distribution from $n$ samples. By proving a Banach space version of this result, we also obtain an upper bound, that is sharp up to a log factor, for the expected max-sliced 2-Wasserstein distance between a symmetric probability measure $μ$ on a Euclidean space and its symmetrized empirical distribution in terms of the operator norm of the covariance matrix of $μ$ and the diameter of the support of $μ$.
format Preprint
id arxiv_https___arxiv_org_abs_2403_00666
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Sharp bounds for max-sliced Wasserstein distances
Boedihardjo, March T.
Probability
Machine Learning
We obtain essentially matching upper and lower bounds for the expected max-sliced 1-Wasserstein distance between a probability measure on a separable Hilbert space and its empirical distribution from $n$ samples. By proving a Banach space version of this result, we also obtain an upper bound, that is sharp up to a log factor, for the expected max-sliced 2-Wasserstein distance between a symmetric probability measure $μ$ on a Euclidean space and its symmetrized empirical distribution in terms of the operator norm of the covariance matrix of $μ$ and the diameter of the support of $μ$.
title Sharp bounds for max-sliced Wasserstein distances
topic Probability
Machine Learning
url https://arxiv.org/abs/2403.00666