Triangular transformations of measures

A new identity for the entropy of a non-linear image of a measure on $\mathbb R^n$ is obtained, which yields the well-known Talagrand's inequality. Triangular mappings on $\mathbb R^n$ and $\mathbb R^\infty$ are studied, that is, mappings $T$ such that the $i$th coordinate function $T_i$ depends only on the variables $x_1,\dots,x_i$. With the help of such mappings the well-known open problem on the representability of each probability measure that is absolutely continuous with respect to a Gaussian measure $\gamma$ on an infinite dimensional space as the image of $\gamma$ under a map of the form $T(x)=x+F(x)$ where $F$ takes values in the Cameron–Martin space of the measure $\gamma$ is solved in the affirmative. As an application, a generalized logarithmic Sobolev inequality is also proved.