Regularization under diffusion and anti-concentration of the information content

Under the Ornstein-Uhlenbeck semigroup $\{U_t\}$, any non-negative measurable $f : \mathbb R^n \to \mathbb R_+$ exhibits a uniform tail bound better than that implied by Markov's inequality and conservation of mass: For every $\alpha \geq e^3$, and $t > 0$, $$\gamma_n\left(\left\{x \in \mathbb R^n : U_t f(x) > \alpha \int f\,d\gamma_n\right\}\right) \leq C(t) \frac{1}{\alpha} \sqrt{\frac{\log \log \alpha}{\log \alpha}}$$ where $\gamma_n$ is the $n$-dimensional Gaussian measure and $C(t)$ is a constant depending only on $t$. This confirms positively the Gaussian limiting case of Talagrand's convolution conjecture (1989). This is shown to follow from a more general phenomenon. Suppose that $f : \mathbb{R}^n \to \mathbb{R}_+$ is {\em semi-log-convex} in the sense that for some $\beta > 0$, for all $x \in \mathbb{R}^n$, the eigenvalues of $\nabla^2 \log f(x)$ are at least $-\beta$. Then $f$ satisfies a tail bound asymptotically better than that implied by Markov's inequality.