• DocumentCode
    2300921
  • Title

    Proof of replica formulas in the high noise regime for communication using LDGM codes

  • Author

    Kudekar, Shrinivas ; Macris, Nicolas

  • Author_Institution
    Sch. of Comput. & Commun. Sci., Ecole Polytech. Fed. de Lausanne, Lausanne
  • fYear
    2008
  • fDate
    5-9 May 2008
  • Firstpage
    416
  • Lastpage
    420
  • Abstract
    We consider communication over a binary input memoryless output symmetric channel with low density generator matrix codes and optimal maximum a posteriori decoding. It is known that the problem of computing the average conditional entropy, over such code ensembles in the asymptotic limit of large block length, is closely related to computing the free energy of a mean field spin glass in the thermodynamic limit. Tentative explicit formulas for these quantities have been derived thanks to the replica method (of spin glass theory) and are generally conjectured to be exact. In this contribution we show that the replica solution is indeed exact in the high noise regime, where it coincides with density evolution equations. Our method uses ideas coming from high temperature expansions in spin glass theory.
  • Keywords
    binary codes; entropy codes; maximum likelihood decoding; memoryless systems; parity check codes; LDGM codes; binary input memoryless output symmetric channel; conditional entropy; density evolution equations; high noise regime; low density generator matrix codes; mean field spin glass; optimal maximum a posteriori decoding; replica formulas; spin glass theory; temperature expansions; Decoding; Entropy; Equations; Glass; Interpolation; Noise generators; Noise level; Symmetric matrices; Temperature; Thermodynamics;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Information Theory Workshop, 2008. ITW '08. IEEE
  • Conference_Location
    Porto
  • Print_ISBN
    978-1-4244-2269-2
  • Electronic_ISBN
    978-1-4244-2271-5
  • Type

    conf

  • DOI
    10.1109/ITW.2008.4578698
  • Filename
    4578698