ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2209.07058
33
12

Structure preservation via the Wasserstein distance

15 September 2022
Daniel Bartl
S. Mendelson
ArXivPDFHTML
Abstract

We show that under minimal assumptions on a random vector X∈RdX\in\mathbb{R}^dX∈Rd and with high probability, given mmm independent copies of XXX, the coordinate distribution of each vector (⟨Xi,θ⟩)i=1m(\langle X_i,\theta \rangle)_{i=1}^m(⟨Xi​,θ⟩)i=1m​ is dictated by the distribution of the true marginal ⟨X,θ⟩\langle X,\theta \rangle⟨X,θ⟩. Specifically, we show that with high probability, \[\sup_{\theta \in S^{d-1}} \left( \frac{1}{m}\sum_{i=1}^m \left|\langle X_i,\theta \rangle^\sharp - \lambda^\theta_i \right|^2 \right)^{1/2} \leq c \left( \frac{d}{m} \right)^{1/4},\] where λiθ=m∫(i−1m,im]F⟨X,θ⟩−1(u) du\lambda^{\theta}_i = m\int_{(\frac{i-1}{m}, \frac{i}{m}]} F_{ \langle X,\theta \rangle }^{-1}(u)\,duλiθ​=m∫(mi−1​,mi​]​F⟨X,θ⟩−1​(u)du and a♯a^\sharpa♯ denotes the monotone non-decreasing rearrangement of aaa. Moreover, this estimate is optimal. The proof follows from a sharp estimate on the worst Wasserstein distance between a marginal of XXX and its empirical counterpart, 1m∑i=1mδ⟨Xi,θ⟩\frac{1}{m} \sum_{i=1}^m \delta_{\langle X_i, \theta \rangle}m1​∑i=1m​δ⟨Xi​,θ⟩​.

View on arXiv
Comments on this paper