MODIFIED LOG-SOBOLEV INEQUALITIES FOR STRONG-RAYLEIGH MEASURES
成果类型：
Article
署名作者：
Hermon, Jonathan; Salez, Justin
署名单位：
University of British Columbia; Universite PSL; Universite Paris-Dauphine; Universite PSL
刊物名称：
ANNALS OF APPLIED PROBABILITY
ISSN/ISSBN：
1050-5164
DOI：
10.1214/22-AAP1847
发表日期：
2023
页码：
1301-1314
关键词：
摘要：
We establish universal modified log-Sobolev inequalities for reversible Markov chains on the boolean lattice {0, 1}n, when the invariant law pi sat-isfies a form of negative dependence known as the stochastic covering prop-erty. This condition is strictly weaker than the strong Rayleigh property, and is satisfied in particular by all determinantal measures, as well as the uniform distribution over the set of bases of any balanced matroid. In the special case where pi is k-homogeneous, our results imply the celebrated concentration in-equality for Lipschitz functions due to Pemantle and Peres (Combin. Probab. Comput. 23 (2014) 140-160). As another application, we deduce that the nat-ural Monte-Carlo Markov chain used to sample from pi has mixing time at most knlog log 1 pi(x) when initialized in state x. To the best of our knowledge, this is the first work relating negative dependence and modified log-Sobolev inequalities.