Paperlane

no. 16 · Tuesday 17 November 2026

Weak and Strong Moments of Log-Concave Measures

Haozhou Zhang

math.PR

Abstract

Let $X\in\mathbb{R}^d$ be a log-concave random vector, and let $N$ be a seminorm on $\mathbb{R}^d$. For $p\ge 1$, we show that $\|N(X)\|_p$ is at most $\mathbb{E} N(X)+C s_p$, where $s_p$ is the supremum of $\|u(X)\|_p$ over all linear functionals $u$ satisfying $|u(x)|\le N(x)$ for every $x\in\mathbb{R}^d$. A choice for $C$ is $2+5e/2$. This bound implies Latała's conjecture on weak and strong moments.

fields
Probability
pdf ↗ page arXiv:2611.18867 ↗