The entropy per coordinate of a random vector is highly constrained under convexity conditions

Abstract

The entropy per coordinate in a log-concave random vector of any dimension with given density at the mode is shown to have a range of just 1. Uniform distributions on convex bodies are at the lower end of this range, the distribution with i.i.d. exponentially distributed coordinates is at the upper end, and the normal is exactly in the middle. Thus in terms of the amount of randomness as measured by entropy per coordinate, any log-concave random vector of any dimension contains randomness that differs from that in the normal random variable with the same maximal density value by at most 1/2. As applications, we obtain an information-theoretic formulation of the famous hyperplane conjecture in convex geometry, entropy bounds for certain infinitely divisible distributions, and quantitative estimates for the behavior of the density at the mode on convolution. More generally, one may consider so-called convex or hyperbolic probability measures on Euclidean spaces; we give new constraints on entropy per coordinate for this class of measures, which generalize our results under the log-concavity assumption, expose the extremal role of multivariate Pareto-type distributions, and give some applications.

15 pages, revised for IEEE Transactions on Information Theory

References in corpus (3)

• Reverse Brunn-Minkowski and reverse entropy power inequalities for convex measures
• Log-concavity, ultra-log-concavity, and a maximum entropy property of discrete compound Poisson measures
• On an Inequality of Karlin and Rinott Concerning Weighted Sums of i.i.d. Random Variables

Cited by in corpus (30)

• Reverse Brunn-Minkowski and reverse entropy power inequalities for convex measures
• Beyond the entropy power inequality, via rearrangements
• Forward and Reverse Entropy Power Inequalities in Convex Geometry
• Concentration of the information in data with log-concave distributions
• Optimal Concentration of Information Content For Log-Concave Densities
• Log-concavity, ultra-log-concavity, and a maximum entropy property of discrete compound Poisson measures
• A lower bound on the differential entropy of log-concave random vectors with applications
• Entropy bounds on abelian groups and the Ruzsa divergence
• On the problem of reversibility of the entropy power inequality
• Dimensional behaviour of entropy and information
• Analysis of Langevin Monte Carlo via convex optimization
• Bounds on Variance for Unimodal Distributions
• The Differential Entropy of Mixtures: New Bounds and Applications
• Sharp moment-entropy inequalities and capacity bounds for log-concave distributions
• Concentration of information content for convex measures
• Bernoulli sums and Rényi entropy inequalities
• A combinatorial approach to small ball inequalities for sums and differences
• Reversals of Rényi Entropy Inequalities under Log-Concavity
• Dimensional variance inequalities of Brascamp-Lieb type and a local approach to dimensional Prékopa's theorem
• Two-sided inequalities for the density function's maximum of weighted sum of chi-square variables
• Charge fluctuation entropy of Hawking radiation: a replica-free way to find large entropy
• Entropy jumps for isotropic log-concave random vectors and spectral gap
• Sumset estimates in convex geometry
• Dual Loomis-Whitney inequalities via information theory
• Entropy-variance inequalities for discrete log-concave random variables via degree of freedom
• Information-theoretic lower bounds for distributed function computation
• Rogers-Shephard inequality for log-concave functions
• A Lower Bound on the Entropy Rate for a Large Class of Stationary Processes and its Relation to the Hyperplane Conjecture
• Constrained Functional Value under General Convexity Conditions with Applications to Distributed Simulation
• Inequalities between $L^p$-norms for log-concave distributions