On Sidorenko's conjecture for determinants and Gaussian Markov random fields

arXiv1701.03632MaRDI QIDQ6281891FDOQ6281891

Authors: Balázs Szegedy 

Publication date: 13 January 2017

Abstract: We study a class of determinant inequalities that are closely related to Sidorenko's famous conjecture (Also conjectured by ErdH os and Simonovits in a different form). Our results can also be interpreted as entropy inequalities for Gaussian Markov random fields (GMRF). We call a GMRF on a finite graph $G$ homogeneous if the marginal distributions on the edges are all identical. We show that if $G$ satisfies Sidorenko's conjecture then the differential entropy of any homogeneous GMRF on $G$ is at least $|E(G)|$ times the edge entropy plus $|V(G)|-2|E(G)|$ times the point entropy. We also prove this inequality in a large class of graphs for which Sidorenko's conjecture is not verified including the so-called M"obius ladder: $K_{5,5}setminus{C}_{10}$. The connection between Sidorenko's conjecture and GMRF's is established via a large deviation principle on high dimensional spheres combined with graph limit theory.

This page was built for publication: On Sidorenko's conjecture for determinants and Gaussian Markov random fields

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q6281891)