2504.03114

#1 Entropy and functional forms of the dimensional Brunn--Minkowski inequality in Gauss space [PDF] [Copy] [Kimi] [REL]

Authors: Gautam Aishwarya, Dongbin Li

Given even strongly log-concave random vectors $X_{0}$ and $X_{1}$ in $\mathbb{R}^n$, we show that a natural joint distribution $(X_{0},X_{1})$ satisfies, $$\begin{equation} e^{ - \frac{1}{n}D ((1-t)X_{0} + t X_{1} \Vert Z)} \geq (1-t) e^{ - \frac{1}{n}D (X_{0} \Vert Z)} + t e^{ - \frac{1}{n}D ( X_{1} \Vert Z)}, \end{equation}$$ where $Z$ is distributed according to the standard Gaussian measure $\gamma$ on $\mathbb{R}^n$, $t \in [0,1]$, and $D(\cdot \Vert Z)$ is the Gaussian relative entropy. This extends and provides a different viewpoint on the corresponding geometric inequality proved by Eskenazis and Moschidis, namely that $$\begin{equation} \gamma \left( (1-t) K_{0} + t K_{1} \right)^{\frac{1}{n}} \geq (1-t) \gamma (K_{0})^{\frac{1}{n}} + t \gamma (K_{1})^{\frac{1}{n}}, \end{equation}$$ when $K_{0}, K_{1} \subseteq \mathbb{R}^n$ are origin-symmetric convex bodies. As an application, using Donsker--Varadhan duality, we obtain Gaussian Borell--Brascamp--Lieb inequalities applicable to even log-concave functions, which serve as functional forms of the Eskenazis--Moschidis inequality.

Subjects: Probability , Functional Analysis , Metric Geometry

Publish: 2025-04-04 01:55:02 UTC

---