English

Exponential integrability for log-concave measures

Probability 2023-08-23 v2 Analysis of PDEs Classical Analysis and ODEs

Abstract

Talagrand observed that finiteness of Ee12f(X)2\mathbb{E}\, e^{\frac{1}{2}|\nabla f(X)|^{2}} implies finiteness of Eef(X)\mathbb{E}\, e^{\, f(X)} where XX is the standard Gaussian vector in Rn\mathbb{R}^{n} and ff is a smooth function with zero average. However, in this paper we show that finiteness of Ee12f2(1+f)1 \mathbb{E}\, e^{\frac{1}{2}|\nabla f|^{2}} (1+|\nabla f|)^{-1} implies finiteness of Eef(X)\mathbb{E}\, e^{\, f(X)}, and we also obtain quantitative bounds \begin{align*} \log\, \mathbb{E}\, e^{\, f} \leq 10\, \mathbb{E}\, e^{\frac{1}{2}|\nabla f|^{2}} (1+|\nabla f|)^{-1}. \end{align*} Moreover, the extra factor (1+f)1(1+|\nabla f|)^{-1} is the best possible in the sense that there is smooth ff with Eef=\mathbb{E}\, e^{\,f} =\infty but Ee12f2(1+f)c<\mathbb{E}\, e^{\frac{1}{2}|\nabla f|^{2}} (1+|\nabla f|)^{-c}<\infty for all c>1c>1. As an application we show corresponding dual inequalities for the discrete time dyadic martingales and its quadratic variations.

Keywords

Cite

@article{arxiv.2004.09704,
  title  = {Exponential integrability for log-concave measures},
  author = {Paata Ivanisvili and Ryan Russell},
  journal= {arXiv preprint arXiv:2004.09704},
  year   = {2023}
}

Comments

We included an application showing what is the corresponding dual inequality for the dyadic square function

R2 v1 2026-06-23T14:59:05.855Z