Transport proofs of some functional inverse Santaló inequalities (Q6150871): Difference between revisions

The focus of this research paper revolves around the concepts of Entropy-Transport inequality and Inverse Santaló inequality. The authors introduced the symbol \(ET_n(c)\) to represent the Entropy-Transport inequality, which is defined as follows: For any constant \(c > 0\), the inequality \(ET_n(c)\) holds when, for all \(\log\)-concave probability measures \(\eta_1\) and \(\eta_2\) with essentially continuous densities on \(\mathbb{R}^n\), the following inequality is satisfied: \[ H(\eta_1)+H(\eta_2)\leq -n\log (ce^2)+\mathcal{T}(\nu_1,\nu_2),\tag{1} \] where \(\nu_1\) and \(\nu_2\) denote the moment measures of \(\eta_1\) and \(\eta_2\), respectively. Similarly, the authors introduced \(ET_{n,s}(c)\), which satisfies equation (1) for all \(\log\)-concave and symmetric measures \(\eta_1\) and \(\eta_2\). Here, symmetry means that \(\nu_i(A)=\nu_i(-A)\) for all measurable sets \(A\). Additionally, the Inverse Santaló inequality was represented by the symbol \(IS_n(c)\) and defined as follows: The inequality \(IS_n(c)\) is satisfied for some \(c\), if for all functions \(f \in \mathcal{F}(\mathbb{R}^n)\) such that both \(\int e^{-f(x)} dx\) and \(\int e^{-f^{\ast}(x)}dx\) are positive, it holds \[ \int e^{-f(x)} dx \int e^{-f^{\ast}(x)}dx \geq c^n.\tag{2} \] Similarly, \(IS_{n,s}(c)\) is defined as the satisfaction of equation (2) for all even functions in \(\mathcal{F}(\mathbb{R}^n)\). In this paper a new proof for the implication \[ ET_n(c)\Rightarrow IS_n(c) \] was presented. Furthermore, the authors established the satisfaction of \(ET_1(e)\) and \(ET_{1,s}(4)\) by employing transport arguments and correlation inequalities. Lastly, they proposed a short proof of \(IS_{n,s}(4).\) For the entire collection see [Zbl 07730209].