Gradient flows of the entropy for jump processes

From MaRDI portal



Abstract: We introduce a new transportation distance between probability measures that is built from a L'evy jump kernel. It is defined via a non-local variant of the Benamou-Brenier formula. We study geometric and topological properties of this distance, in particular we prove existence of geodesics. For translation invariant jump kernels we identify the semigroup generated by the associated non-local operator as the gradient flow of the relative entropy w.r.t. the new distance and show that the entropy is convex along geodesics.


``In the last two decades, the theory of optimal transportation has found applications to many areas of mathematics such as partial differential equations, geometry and probability theory. [\dots] In particular, optimal transport has proved very useful in the study of diffusion processes (from the introduction).NEWLINENEWLINE In the present article, the author introduces ``a new transport distance between probability measures on \(\mathbb{R}^d\) that is built from a Lévy jump kernel. It is defined via a non-local variant of the Benamou-Brenier formula. The author studies ``geometric and topological properties of this distance, in particular [\dots] existence of geodesics. For translation invariant jump kernels the author identifies ``the semigroup generated by the associated non-local operator as the gradient flow of the relative entropy w.r.t. the new distance and show[s] that the entropy is convex along geodesics (from the author's abstract). Applications of an approach via the gradient flow might include the study of rates of convergence to the equilibrium as well as logarithmic Sobolev inequalities.



Cites work


Cited in
(34)








This page was built for publication: Gradient flows of the entropy for jump processes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q405499)