English

LP decoding of expander codes: a simpler proof

Information Theory 2012-06-13 v1 math.IT

Abstract

A code C\F2nC \subseteq \F_2^n is a (c,ϵ,δ)(c,\epsilon,\delta)-expander code if it has a Tanner graph, where every variable node has degree cc, and every subset of variable nodes L0L_0 such that L0δn|L_0|\leq \delta n has at least ϵcL0\epsilon c |L_0| neighbors. Feldman et al. (IEEE IT, 2007) proved that LP decoding corrects 3ϵ22ϵ1(δn1)\frac{3\epsilon-2}{2\epsilon-1} \cdot (\delta n-1) errors of (c,ϵ,δ)(c,\epsilon,\delta)-expander code, where ϵ>2/3+13c\epsilon > 2/3+\frac{1}{3c}. In this paper, we provide a simpler proof of their result and show that this result holds for every expansion parameter ϵ>2/3\epsilon > 2/3.

Cite

@article{arxiv.1206.2568,
  title  = {LP decoding of expander codes: a simpler proof},
  author = {Michael Viderman},
  journal= {arXiv preprint arXiv:1206.2568},
  year   = {2012}
}
R2 v1 2026-06-21T21:18:06.755Z