Skip to main navigation Skip to search Skip to main content

On mismatched list decoding

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

The setup of a general channel is considered in the mismatched case, i.e., when the decoder uses a general decoding metric. An expression for the average error probability in list decoding with block length n, metric qn, list size enΘn and rate R, denoted ϵ(n)qn(R, Θn), is established. Further, a general multi-letter formula for the mismatched capacity with list decoding is derived. It is shown that similarly to the matched capacity of the discrete memoryless channel, if the list size grows exponentially at a fixed rate Θ, then the increase in capacity is Θ bits per channel use. Additionally, a random coding lower bound on ϵ(n)qn(R, Θn) is presented. We conclude by presenting an inequality that can be regarded as an extension of Fano's inequality in the mismatched case with list decoding. As a special case, we derive a lower bound on the average probability of error at rates above the erasures only capacity of the discrete memoryless channel.

Original languageEnglish
Title of host publicationProceedings - 2015 IEEE International Symposium on Information Theory, ISIT 2015
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages526-530
Number of pages5
ISBN (Electronic)9781467377041
DOIs
StatePublished - 28 Sep 2015
Event2015 IEEE International Symposium on Information Theory, ISIT 2015 - Hong Kong, China
Duration: 14 Jun 201519 Jun 2015

Publication series

NameIEEE International Symposium on Information Theory - Proceedings
Volume2015-June

Conference

Conference2015 IEEE International Symposium on Information Theory, ISIT 2015
Country/TerritoryChina
CityHong Kong
Period14/06/1519/06/15

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Information Systems
  • Modelling and Simulation
  • Applied Mathematics

Fingerprint

Dive into the research topics of 'On mismatched list decoding'. Together they form a unique fingerprint.

Cite this