English

On Renyi Entropy Power Inequalities

Information Theory 2016-07-21 v3 math.IT Probability

Abstract

This paper gives improved R\'{e}nyi entropy power inequalities (R-EPIs). Consider a sum Sn=k=1nXkS_n = \sum_{k=1}^n X_k of nn independent continuous random vectors taking values on Rd\mathbb{R}^d, and let α[1,]\alpha \in [1, \infty]. An R-EPI provides a lower bound on the order-α\alpha R\'enyi entropy power of SnS_n that, up to a multiplicative constant (which may depend in general on n,α,dn, \alpha, d), is equal to the sum of the order-α\alpha R\'enyi entropy powers of the nn random vectors {Xk}k=1n\{X_k\}_{k=1}^n. For α=1\alpha=1, the R-EPI coincides with the well-known entropy power inequality by Shannon. The first improved R-EPI is obtained by tightening the recent R-EPI by Bobkov and Chistyakov which relies on the sharpened Young's inequality. A further improvement of the R-EPI also relies on convex optimization and results on rank-one modification of a real-valued diagonal matrix.

Keywords

Cite

@article{arxiv.1601.06555,
  title  = {On Renyi Entropy Power Inequalities},
  author = {Eshed Ram and Igal Sason},
  journal= {arXiv preprint arXiv:1601.06555},
  year   = {2016}
}

Comments

Revised version of a submission to the IEEE Trans. on Information Theory. Presented in part at ISIT 2016, Barcelona, July 2016

R2 v1 2026-06-22T12:35:56.979Z