Computer Science, Information Theory, Information Theory (cs.IT)

journal:

--

date:

2023-12-26 00:00:00

Abstract

Tanner codes are graph-based linear codes whose parity-check matrices can be characterized by a bipartite graph $G$ together with an inner code $C_0$. Expander codes are Tanner codes whose defining bipartite graph $G$ has good expansion property. The landmark work of Sipser and Spielman showed that every bipartite expander $G$ with expansion ratio $\delta>3/4$ together with a parity-check code defines an expander code which corrects $\Omega(n)$ errors in $O(n)$ time, where $n$ is the code length. Viderman showed that $\delta>2/3-\Omega(1)$ is already sufficient. Our paper is motivated by the following natural and fundamental problem in decoding expander codes: \textbf{Question:} What are the sufficient and necessary conditions that $\delta$ and $d_0$ should satisfy so that {\it every} bipartite expander $G$ with expansion ratio $\delta$ and {\it every} inner code $C_0$ with minimum distance $d_0$ together define an expander code which corrects $\Omega(n)$ errors in $O(n)$ time? We give a near-optimal solution to the question above, showing that $\delta d_0>3$ is sufficient and $\delta d_0>1$ is necessary. Our result significantly improves the previously known result of Dowling and Gao, who showed that $d_0=\Omega(c\delta^{-2})$ is sufficient, where $c$ is the left-degree of $G$. We suspect that $\delta d_0>1$ is also sufficient to solve the question above.