Minimax Rényi Redundancy

Semih Yagli; Yucel Altug; Sergio Verdu

doi:10.1109/TIT.2018.2803070

Minimax Rényi Redundancy

Semih Yagli, Yucel Altug, Sergio Verdu

Research output: Contribution to journal › Article › peer-review

5 Scopus citations

Abstract

The redundancy for universal lossless compression of discrete memoryless sources in Campbell's setting is characterized as a minimax Rényi divergence, which is shown to be equal to the maximal \alpha -mutual information via a generalized redundancy-capacity theorem. Special attention is placed on the analysis of the asymptotics of minimax Rényi divergence, which is determined up to a term vanishing in blocklength.

Original language	English (US)
Pages (from-to)	3715-3733
Number of pages	19
Journal	IEEE Transactions on Information Theory
Volume	64
Issue number	5
DOIs	https://doi.org/10.1109/TIT.2018.2803070
State	Published - May 2018

All Science Journal Classification (ASJC) codes

Information Systems
Computer Science Applications
Library and Information Sciences

Keywords

Jeffreys' prior
Rényi divergence
Universal lossless compression
generalized redundancy-capacity theorem
minimax redundancy
minimax regret
risk aversion
α-mutual information

Access to Document

10.1109/TIT.2018.2803070

Cite this

@article{a900eeca178444caac31f1bec1473ccc,

title = "Minimax R{\'e}nyi Redundancy",

abstract = "The redundancy for universal lossless compression of discrete memoryless sources in Campbell's setting is characterized as a minimax R{\'e}nyi divergence, which is shown to be equal to the maximal \alpha -mutual information via a generalized redundancy-capacity theorem. Special attention is placed on the analysis of the asymptotics of minimax R{\'e}nyi divergence, which is determined up to a term vanishing in blocklength.",

keywords = "Jeffreys' prior, R{\'e}nyi divergence, Universal lossless compression, generalized redundancy-capacity theorem, minimax redundancy, minimax regret, risk aversion, α-mutual information",

author = "Semih Yagli and Yucel Altug and Sergio Verdu",

note = "Publisher Copyright: {\textcopyright} 1963-2012 IEEE.",

year = "2018",

month = may,

doi = "10.1109/TIT.2018.2803070",

language = "English (US)",

volume = "64",

pages = "3715--3733",

journal = "IEEE Transactions on Information Theory",

issn = "0018-9448",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "5",

}

TY - JOUR

T1 - Minimax Rényi Redundancy

AU - Yagli, Semih

AU - Altug, Yucel

AU - Verdu, Sergio

PY - 2018/5

Y1 - 2018/5

N2 - The redundancy for universal lossless compression of discrete memoryless sources in Campbell's setting is characterized as a minimax Rényi divergence, which is shown to be equal to the maximal \alpha -mutual information via a generalized redundancy-capacity theorem. Special attention is placed on the analysis of the asymptotics of minimax Rényi divergence, which is determined up to a term vanishing in blocklength.

AB - The redundancy for universal lossless compression of discrete memoryless sources in Campbell's setting is characterized as a minimax Rényi divergence, which is shown to be equal to the maximal \alpha -mutual information via a generalized redundancy-capacity theorem. Special attention is placed on the analysis of the asymptotics of minimax Rényi divergence, which is determined up to a term vanishing in blocklength.

KW - Jeffreys' prior

KW - Rényi divergence

KW - Universal lossless compression

KW - generalized redundancy-capacity theorem

KW - minimax redundancy

KW - minimax regret

KW - risk aversion

KW - α-mutual information

UR - http://www.scopus.com/inward/record.url?scp=85041535711&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85041535711&partnerID=8YFLogxK

U2 - 10.1109/TIT.2018.2803070

DO - 10.1109/TIT.2018.2803070

M3 - Article

AN - SCOPUS:85041535711

SN - 0018-9448

VL - 64

SP - 3715

EP - 3733

JO - IEEE Transactions on Information Theory

JF - IEEE Transactions on Information Theory

IS - 5

ER -

Minimax Rényi Redundancy

Abstract

All Science Journal Classification (ASJC) codes

Keywords

Access to Document

Other files and links

Fingerprint

Cite this