Approximating $L_p$ unit balls via random sampling

TL;DR

This paper demonstrates that a small, linear number of random samples from an isotropic vector can efficiently approximate the $L_p$ unit ball in high-dimensional space, improving previous bounds especially for log-concave vectors.

Contribution

It introduces a new sampling method that achieves $1 \u00b1 \u03b5$ approximation of $L_p$ balls with high probability using fewer samples than prior methods, especially for log-concave distributions.

Findings

Approximation with $N = c(p,q,\u03b5) d$ samples is sufficient.

Probability of success depends on the relationship between $q$ and $p$.

Improves previous bounds for log-concave vectors.

Abstract

Let $X$ be an isotropic random vector in $R^d$ that satisfies that for every $v \in S^{d-1}$, $\|<X,v>\|_{L_q} \leq L \|<X,v>\|_{L_p}$ for some $q \geq 2p$. We show that for $0<\varepsilon<1$, a set of $N = c(p,q,\varepsilon) d$ random points, selected independently according to $X$, can be used to construct a $1 \pm \varepsilon$ approximation of the $L_p$ unit ball endowed on $R^d$ by $X$. Moreover, $c(p,q,\varepsilon) \leq c^p \varepsilon^{-2}\log(2/\varepsilon)$; when $q=2p$ the approximation is achieved with probability at least $1-2\exp(-cN \varepsilon^2/\log^2(2/\varepsilon))$ and if $q$ is much larger than $p$---say, $q=4p$, the approximation is achieved with probability at least $1-2\exp(-cN \varepsilon^2)$. In particular, when $X$ is a log-concave random vector, this estimate improves the previous state-of-the-art---that $N=c^\prime(p,\varepsilon) d^{p/2}\log d$ random points are enough, and that the approximation is valid with constant probability.