English

Random $\epsilon$-Cover on Compact Symmetric Space

Probability 2025-09-03 v1 Computational Geometry

Abstract

A randomized scheme that succeeds with probability 1δ1-\delta (for any δ>0\delta>0) has been devised to construct (1) an equidistributed ϵ\epsilon-cover of a compact Riemannian symmetric space M\mathbb M of dimension dMd_{\mathbb M} and antipodal dimension dˉM\bar{d}_{\mathbb M}, and (2) an approximate (λr,2)(\lambda_r,2)-design, using n(ϵ,δ)n(\epsilon,\delta)-many Haar-random isometries of M\mathbb M, where \begin{equation}n(\epsilon,\delta):=O_{\mathbb M}\left(d_{\mathbb M}\ln \left(\frac 1\epsilon\right)+\log\left(\frac 1\delta\right)\right)\,,\end{equation} and λr\lambda_r is the rr-th smallest eigenvalue of the Laplace-Beltrami operator on M\mathbb M. The ϵ\epsilon-cover so-produced can be used to compute the integral of 1-Lipschitz functions within additive O~(ϵ)\tilde O(\epsilon)-error, as well as in comparing persistence homology computed from data cloud to that of a hypothetical data cloud sampled from the uniform measure.

Keywords

Cite

@article{arxiv.2304.07622,
  title  = {Random $\epsilon$-Cover on Compact Symmetric Space},
  author = {Somnath Chakraborty},
  journal= {arXiv preprint arXiv:2304.07622},
  year   = {2025}
}
R2 v1 2026-06-28T10:07:08.256Z