A probabilistic approach to convex ()-entropy decay for Markov chains

DOI10.1214/21-AAP1700MaRDI QIDQ2134287zbMATH OpenOpenAlexFDO

Authors Giovanni Conforti

Publication date 6 May 2022

Published in The Annals of Applied Probability (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/2004.10850

zbMATH Keywords

Bakry-Émery methodBeckner inequalitiesconvexity of the entropycouplingdiscrete convex Sobolev inequalitiesentropy functionalsexponential dissipationGlauber dynamicshardcore modelsMarkov chainsSobolev inequality

Mathematics Subject Classification ID

Ergodic theorems, spectral theory, Markov operators (37A30) Entropy and other invariants, isomorphism, classification in ergodic theory (37A35) Dynamical systems and their relations with probability theory and stochastic processes (37A50) Dynamical aspects of statistical mechanics (37A60) General theory of functional equations and inequalities (39B05) Functional inequalities, including subadditivity, convexity, etc. (39B62) Markov chains (discrete-time Markov processes on discrete state spaces) (60J10) Branching processes (Galton-Watson, birth-and-death, etc.) (60J80) Interacting random processes; statistical mechanics type models; percolation theory (60K35) Disordered systems (random Ising models, random Schrödinger operators, etc.) in equilibrium statistical mechanics (82B44)

Abstract: We study the exponential dissipation of entropic functionals for continuous time Markov chains and the associated convex Sobolev inequalities, including MLSI and Beckner inequalities. We propose a method that combines the Bakry 'Emery approach and coupling arguments, which we use as a probabilistic alternative to the discrete Bochner identities. The method is well suited to work in a non perturbative setting and we obtain new estimates for interacting random walks beyond the high temperature/weak interaction regime. In this framework, we also show that the exponential contraction of the Wasserstein distance along the semigroup implies MLSI. We also revisit classical examples often obtaining new inequalities and sometimes improving on the best known constants. In particular, we analyse the zero range dynamics, hardcore and Bernoulli-Laplace models and the Glauber dynamics for the Curie Weiss and Ising model.

Review text

Functional inequalities are a powerful tool to quantify the trend (hypercontractivity, and dissipation time-rates) to equilibrium of Markov semigroups. For diffusion processes on a Riemannan manifold the link with the celebrated logarithmic Sobolev inequality has been established by \textit{D. Bakry} and \textit{M. Émery} [Lect. Notes Math. 1123, 177--206 (1985; Zbl 0561.60080)]. A considerable effort has been put into transferring the ideas and concepts of the Bakry-Émery theory to the setting of continuous-time Markov chains. Specifically, the exponential decay of general entropic functionals, called \(\phi\)-entropies, has been recently investigated in terms of discrete Beckner inequalities [\textit{A. Jüngel} and \textit{W. Yue}, Ann. Appl. Probab. 27, No. 4, 2238--2269 (2017; Zbl 1374.60144)]. A problem about functional inequalities and convergence to equilibrium for nonlinear Markov chains has been recently addressed in [\textit{M. Erbar} et al., ALEA, Lat. Am. J. Probab. Math. Stat. 17, No. 1, 445--471 (2020; Zbl 1441.82017)]. The main purpose of the present paper is to develop a probabilistic approach to establish convex Sobolev inequalities and quantify the exponential decay of \(\phi\)-entropies for continuous-time Markov chains via the Bakry-Émery method. The method is applied to specific model examples and some explicit lower bounds are obtained. A probabilistic alternative was proposed to the discrete Bochner identities in [\textit{P. Caputo} et al., Ann. Inst. Henri Poincaré, Probab. Stat. 45, No. 3, 734--753 (2009; Zbl 1181.60142)], upon which a large part of the results about the modified logarithmic Sobolev inequality recently obtained in connection with the Bakry-Émery method relies. The proposed method provides a unified framework for the study of general convex Sobolev inequalities, modified logarithmic Sobolev inequalities and Beckner inequalities.

Recommendations

• Convex entropy decay via the Bochner-Bakry-Emery approach
• Discrete Beckner inequalities via the Bochner-Bakry-Emery approach for Markov chains
• Entropy decay for interacting systems via the Bochner-Bakry-Émery approach
• Geodesic convexity of the relative entropy in reversible Markov chains
• Trajectorial dissipation and gradient flow for the relative entropy in Markov chains

Cites work

• A discrete log-Sobolev inequality under a Bakry-Émery type condition
• A Generalized Poincare Inequality for Gaussian Measures
• A note on the Glauber dynamics for sampling independent sets
• Analysis and geometry of Markov diffusion operators
• Convex entropy decay via the Bochner-Bakry-Emery approach
• Discrete Beckner inequalities via the Bochner-Bakry-Emery approach for Markov chains
• Discrete Ricci curvature bounds for Bernoulli-Laplace and random transposition models
• Displacement convexity of entropy and related inequalities on graphs
• Entropic curvature and convergence to equilibrium for mean-field dynamics on discrete spaces
• Entropic Ricci curvature bounds for discrete interacting systems
• Entropy decay for interacting systems via the Bochner-Bakry-Émery approach
• Entropy dissipation estimates for inhomogeneous zero-range processes
• Exponential decay of entropy in the random transposition and Bernoulli-Laplace models
• Fast convergence of the Glauber dynamics for sampling independent sets
• Geodesic convexity of the relative entropy in reversible Markov chains
• Gradient flows of the entropy for finite Markov chains
• scientific article; zbMATH DE number 3894218 (Why is no real title available?)
• Inequalities in Fourier analysis
• Logarithmic Sobolev inequalities and stochastic Ising models
• Markov Chains
• Markov chains and mixing times. With a chapter on ``Coupling from the past by James G. Propp and David B. Wilson.
• Modified logarithmic Sobolev inequalities for some models of random walk
• Modified logarithmic Sobolev inequalities in discrete settings
• ON CONVEX SOBOLEV INEQUALITIES AND THE RATE OF CONVERGENCE TO EQUILIBRIUM FOR FOKKER-PLANCK TYPE EQUATIONS
• On estimation of the logarithmic Sobolev constant and gradient estimates of heat semigroups
• On Markov Chains for Independent Sets
• Optimal markovian couplings and applications
• Optimal Transport
• Ricci curvature bounds for weakly interacting Markov chains
• Ricci curvature of finite Markov chains via convexity of the entropy
• Ricci curvature of Markov chains on metric spaces
• Spectral gap estimates for interacting particle systems via a Bochner-type identity
• Time to Reach Stationarity in the Bernoulli–Laplace Diffusion Model
• Transport inequalities, gradient estimates, entropy and Ricci curvature

Cited in

(12)

• Discrete Beckner inequalities via the Bochner-Bakry-Emery approach for Markov chains
• Generalized \(\Gamma\) calculus and application to interacting particles on a graph
• Entropy-information inequalities under curvature-dimension conditions for continuous-time Markov chains
• On mixing of Markov chains: coupling, spectral independence, and entropy factorization
• \(\Phi\)-entropy inequality and application for SDEs with jumps
• Interpolation between modified logarithmic Sobolev and Poincaré inequalities for quantum Markovian dynamics
• Trajectorial dissipation of \({\Phi}\)-entropies for interacting particle systems
• Upgrading MLSI to LSI for reversible Markov chains
• Contractive coupling rates and curvature lower bounds for Markov chains
• Entropy and curvature: beyond the Peres-Tetali conjecture
• Long-time behavior for discretization schemes of Fokker-Planck equations via couplings
• Convex entropy decay via the Bochner-Bakry-Emery approach

This page was built for publication: A probabilistic approach to convex \((\phi)\)-entropy decay for Markov chains

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2134287)