Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Log inSign up
Challenges
Datasets
Workspaces
Papers
Discussions
Leaderboard
Blog
Job Board
Q3AS

© 2026 Aqora Quantum S.A.S.

TermsPrivacyLegal Notice

Code representation and performance of graph -based decoding

Junsheng Han

Published Jan 1, 2008• 0 citations
OpenAlex

Abstract

Key to the success of modern error correcting codes is the effectiveness of message-passing iterative decoding (MPID). Unlike maximum-likelihood (ML) decoding, the performance of MPID depends not only on the code, but on how the code is represented. In particular, the performance of MPID is potentially improved by using a redundant representation. We focus on Tanner graphs and study combinatorial structures therein that help explain the performance disparity among different representations of the same code. Emphasis is placed on the complexity- performance tradeoff, as more and more check nodes are allowed in the graph. Our discussion applies to MPID as well as linear programming decoding (LPD), which we collectively refer to as graph-based decoding. On an erasure channel, it is well-known that the performance of MPID or LPD is determined by stopping sets. Following Schwartz and Vardy, we define the stopping redundancy as the smallest number of check nodes in a Tanner graph such that smallest size of a non-empty stopping set is equal to the minimum Hamming distance of the code. Roughly speaking, stopping redundancy measures the complexity requirement (in number of check nodes) for MPID of a redundant graph representation to achieve performance comparable to ML decoding (up to a constant factor for small channel erasure probability). General upper bounds on stopping redundancy are obtained. One of our main contribution is a new upper bound based on probabilistic analysis, which is shown to be by far the strongest. From this bound, it can be shown, for example, that for a fixed minimum distance, the stopping redundancy grows just linearly with the redundancy (codimension). Specific results on the stopping redundancy of Golay and Reed-Muller codes are also obtained. We show that the stopping redundancy of maximum distance separable (MDS) codes is bounded in between a Turan number and a single-exclusion (SE) number --- a purely combinatorial quantity that we introduce. By studying upper bounds on the SE number, new results on the stopping redundancy of MDS codes are obtained. Schwartz and Vardy conjecture that the stopping redundancy of an MDS code should only depend on its length and minimum distance. Our results provide partial confirmation, both exact and asymptotic, to this conjecture. Stopping redundancy can be large for some codes. We observe that significantly fewer checks are needed if a small number of small stopping sets are allowed. These small stopping sets can then be dealt with by ̀g̀uessing'' during the iterative decoding process. Correspondingly, the guess-g stopping redundancy is defined and it is shown that the savings in number of required check nodes are potentially significant. Another theoretically interesting question is when MPID of a Tanner graph achieves the same word error rate an ML decoder. This prompts us to define and study ML redundancy. Applicability and possible extensions of the current work to a non-erasure channel are discussed. A framework based on pseudo-codewords is considered and shown to be relevant. However, it is also observed that the polytope characterization of pseudo-codewords is not complete enough to be an accurate indicator of MPID performance. Finally, in a separate piece of work, the probability of undetected error (PUE) for over-extended Reed-Solomon codes is studied through the weight distribution bounds of the code. The resulting PUE expressions are shown to be tight in a well-defined sense

Related papers

On the stopping distance and the stopping redundancy of codes

M. Schwartz, Alexander Vardy

It is now well known that the performance of a linear code Copf under iterative decoding on a binary erasure channel (and other channels) is determined by the size of the smallest stopping set in the Tanner graph for Copf. Several recent papers refer to this parameter as the stopping distance s of Copf. This is somewhat of a misnomer since the size of the smallest stopping set in the Tanner graph for Copf depends on the corresponding choice of a parity-check matrix. It is easy to see that s les d, where d is the minimum Hamming distance of Copf, and we show that it is always possible to choose a parity-check matrix for Copf (with sufficiently many dependent rows) such that s=d. We thus introduce a new parameter, the stopping redundancy of Copf, defined as the minimum number of rows in a parity- check matrix H for Copf such that the corresponding stopping distance s(H) attains its largest possible value, namely, s(H)=d. We then derive general bounds on the stopping redundancy of linear codes. We also examine several simple ways of constructing codes from other codes, and study the effect of these constructions on the stopping redundancy. Specifically, for the family of binary Reed-Muller codes (of all orders), we prove that their stopping redundancy is at most a constant times their conventional redundancy. We show that the stopping redundancies of the binary and ternary extended Golay codes are at most 34 and 22, respectively. Finally, we provide upper and lower bounds on the stopping redundancy of MDS codes

IEEE Transactions on Information Theory•Mar 1, 2006•178 citations

On the Stopping Distance and the Stopping Redundancy of Codes

Schwartz, Moshe, Alexander Vardy

It is now well known that the performance of a linear code $C$ under iterative decoding on a binary erasure channel (and other channels) is determined by the size of the smallest stopping set in the Tanner graph for $C$. Several recent papers refer to this parameter as the \emph{stopping distance} $s$ of $C$. This is somewhat of a misnomer since the size of the smallest stopping set in the Tanner graph for $C$ depends on the corresponding choice of a parity-check matrix. It is easy to see that $s \le d$, where $d$ is the minimum Hamming distance of $C$, and we show that it is always possible to choose a parity-check matrix for $C$ (with sufficiently many dependent rows) such that $s = d$. We thus introduce a new parameter, termed the \emph{stopping redundancy} of $C$, defined as the minimum number of rows in a parity-check matrix $H$ for $C$ such that the corresponding stopping distance $s(H)$ attains its largest possible value, namely $s(H) = d$. We then derive general bounds on the stopping redundancy of linear codes. We also examine several simple ways of constructing codes from other codes, and study the effect of these constructions on the stopping redundancy. Specifically, for the family of binary Reed-Muller codes (of all orders), we prove that their stopping redundancy is at most a constant times their conventional redundancy. We show that the stopping redundancies of the binary and ternary extended Golay codes are at most 35 and 22, respectively. Finally, we provide upper and lower bounds on the stopping redundancy of MDS codes.

arXiv:cs/0503058•Mar 23, 2005•0 citations

On the stopping distance and the stopping redundancy of codes

M. Schwartz, Alexander Vardy

It is now well known that the performance of a linear code Copf under iterative decoding on a binary erasure channel (and other channels) is determined by the size of the smallest stopping set in the Tanner graph for Copf. Several recent papers refer to this parameter as the stopping distance s of Copf. This is somewhat of a misnomer since the size of the smallest stopping set in the Tanner graph for Copf depends on the corresponding choice of a parity-check matrix. It is easy to see that s les d, where d is the minimum Hamming distance of Copf, and we show that it is always possible to choose a parity-check matrix for Copf (with sufficiently many dependent rows) such that s = d. We thus introduce a new parameter, termed the stopping redundancy of Copf, defined as the minimum number of rows in a parity-check matrix H for Copf such that the corresponding stopping distance s(H) attains its largest possible value, namely s(H) = d. We then derive general bounds on the stopping redundancy of linear codes. We also examine several simple ways of constructing codes from other codes, and study the effect of these constructions on the stopping redundancy. Specifically, for the family of binary Reed-Muller codes (of all orders), we prove that their stopping redundancy is at most a constant times their conventional redundancy. We show that the stopping redundancies of the binary and ternary extended Golay codes are at most 34 and 22, respectively. Finally, we provide upper and lower bounds on the stopping redundancy of MDS codes

IEEE International Symposium on Information Theory (ISIT)•Jan 1, 2005•27 citations

On the pseudocodeword weight and parity-check matrix redundancy of linear codes

Christine A. Kelley, Deepak Sridhara

The minimum pseudocodeword weightwminof a linear graph-based code is more influential in determining decoding performance when decoded via iterative and linear programming decoding algorithms than the classical minimum distancedminunder standard maximum-likelihood decoding Moreover, unlike the minimum distance which is unique to the code regardless of representation, the set of pseudocodewords, and therefore also the minimum pseudocodeword weight, depends on the graph representation used in decoding as well as on the communication channel. This means that a judicious choice of parity-check matrix is crucial for realizing the best potential of any graph-based code. In this paper, we introduce the notion ofpseudoweightredundancyfor the memoryless binary symmetric channel (BSC). Analogous to the stopping redundancy in the literature, this parameter gives the smallest number of rows needed for a parity-check matrix to havedmin=wmin. We provide some upper bounds on the BSC-pseudoweight redundancy and illustrate the concept with some results for Hamming codes, tree-based and finite geometry LDPC codes, Reed-Muller codes and Hadamard codes.

IEEE Information Theory Workshop (ITW)•Sep 1, 2007•12 citations

On ML redundancy of codes

Junsheng Han, Paul H. Siegel

The ML redundancy of a code is defined as the smallest number of rows in its parity-check matrix such that a message-passing decoder working in the corresponding Tanner graph achieves maximum-likelihood (ML) performance on an erasure channel. General upper bounds on ML redundancy are obtained. In particular, it is shown that the ML redundancy of a q-ary code is at most the number of minimal codewords in its dual code, divided by q−1. Special upper bounds are derived for codes whose dual code contains a covering design. For example, the ML redundancy of a Simplex code of length n is shown to be no greater than (n2− 4n + 9)/6.

IEEE International Symposium on Information Theory (ISIT)•Jul 1, 2008•5 citations

Bounds on the Expansion Properties of Tanner Graphs

Mingrui Zhu, K.M. Chugg

This work focuses on the expansion properties of a Tanner Graph because they are known to be related to the performance of associated iterative message-passing algorithms over various channels. By analyzing the eigenvalues and corresponding eigenvectors of the normalized incidence matrix representing a Tanner Graph, lower bounds on these expansion properties are derived. Specifically, for the binary erasure channel, these results lead to two lower bounds on stopping distance for any given binary linear code and an upper bound on stopping redundancy for the family of difference-set codes (type-I 2-D projective geometry low-density parity-check (LDPC) codes).

IEEE Transactions on Information Theory•Dec 1, 2007•3 citations
Key to the success of modern error correcting codes is the effectiveness of message-passing iterative decoding (MPID). Unlike maximum-likelihood (ML) decoding, the performance of MPID depends not only on the code, but on how the code is represented. In particular, the performance of MPID is potentially improved by using a redundant representation. We focus on Tanner graphs and study combinatorial structures therein that help explain the performance disparity among different representations of the same code. Emphasis is placed on the complexity- performance tradeoff, as more and more check nodes are allowed in the graph. Our discussion applies to MPID as well as linear programming decoding (LPD), which we collectively refer to as graph-based decoding. On an erasure channel, it is well-known that the performance of MPID or LPD is determined by stopping sets. Following Schwartz and Vardy, we define the stopping redundancy as the smallest number of check nodes in a Tanner graph such that smallest size of a non-empty stopping set is equal to the minimum Hamming distance of the code. Roughly speaking, stopping redundancy measures the complexity requirement (in number of check nodes) for MPID of a redundant graph representation to achieve performance comparable to ML decoding (up to a constant factor for small channel erasure probability). General upper bounds on stopping redundancy are obtained. One of our main contribution is a new upper bound based on probabilistic analysis, which is shown to be by far the strongest. From this bound, it can be shown, for example, that for a fixed minimum distance, the stopping redundancy grows just linearly with the redundancy (codimension). Specific results on the stopping redundancy of Golay and Reed-Muller codes are also obtained. We show that the stopping redundancy of maximum distance separable (MDS) codes is bounded in between a Turan number and a single-exclusion (SE) number --- a purely combinatorial quantity that we introduce. By studying upper bounds on the SE number, new results on the stopping redundancy of MDS codes are obtained. Schwartz and Vardy conjecture that the stopping redundancy of an MDS code should only depend on its length and minimum distance. Our results provide partial confirmation, both exact and asymptotic, to this conjecture. Stopping redundancy can be large for some codes. We observe that significantly fewer checks are needed if a small number of small stopping sets are allowed. These small stopping sets can then be dealt with by ̀g̀uessing'' during the iterative decoding process. Correspondingly, the guess-g stopping redundancy is defined and it is shown that the savings in number of required check nodes are potentially significant. Another theoretically interesting question is when MPID of a Tanner graph achieves the same word error rate an ML decoder. This prompts us to define and study ML redundancy. Applicability and possible extensions of the current work to a non-erasure channel are discussed. A framework based on pseudo-codewords is considered and shown to be relevant. However, it is also observed that the polytope characterization of pseudo-codewords is not complete enough to be an accurate indicator of MPID performance. Finally, in a separate piece of work, the probability of undetected error (PUE) for over-extended Reed-Solomon codes is studied through the weight distribution bounds of the code. The resulting PUE expressions are shown to be tight in a well-defined sense
It is now well known that the performance of a linear code Copf under iterative decoding on a binary erasure channel (and other channels) is determined by the size of the smallest stopping set in the Tanner graph for Copf. Several recent papers refer to this parameter as the stopping distance s of Copf. This is somewhat of a misnomer since the size of the smallest stopping set in the Tanner graph for Copf depends on the corresponding choice of a parity-check matrix. It is easy to see that s les d, where d is the minimum Hamming distance of Copf, and we show that it is always possible to choose a parity-check matrix for Copf (with sufficiently many dependent rows) such that s=d. We thus introduce a new parameter, the stopping redundancy of Copf, defined as the minimum number of rows in a parity- check matrix H for Copf such that the corresponding stopping distance s(H) attains its largest possible value, namely, s(H)=d. We then derive general bounds on the stopping redundancy of linear codes. We also examine several simple ways of constructing codes from other codes, and study the effect of these constructions on the stopping redundancy. Specifically, for the family of binary Reed-Muller codes (of all orders), we prove that their stopping redundancy is at most a constant times their conventional redundancy. We show that the stopping redundancies of the binary and ternary extended Golay codes are at most 34 and 22, respectively. Finally, we provide upper and lower bounds on the stopping redundancy of MDS codes
It is now well known that the performance of a linear code CCC under iterative decoding on a binary erasure channel (and other channels) is determined by the size of the smallest stopping set in the Tanner graph for CCC. Several recent papers refer to this parameter as the \emph{stopping distance} sss of CCC. This is somewhat of a misnomer since the size of the smallest stopping set in the Tanner graph for CCC depends on the corresponding choice of a parity-check matrix. It is easy to see that s≤ds \le ds≤d, where ddd is the minimum Hamming distance of CCC, and we show that it is always possible to choose a parity-check matrix for CCC (with sufficiently many dependent rows) such that s=ds = ds=d. We thus introduce a new parameter, termed the \emph{stopping redundancy} of CCC, defined as the minimum number of rows in a parity-check matrix HHH for CCC such that the corresponding stopping distance s(H)s(H)s(H) attains its largest possible value, namely s(H)=ds(H) = ds(H)=d. We then derive general bounds on the stopping redundancy of linear codes. We also examine several simple ways of constructing codes from other codes, and study the effect of these constructions on the stopping redundancy. Specifically, for the family of binary Reed-Muller codes (of all orders), we prove that their stopping redundancy is at most a constant times their conventional redundancy. We show that the stopping redundancies of the binary and ternary extended Golay codes are at most 35 and 22, respectively. Finally, we provide upper and lower bounds on the stopping redundancy of MDS codes.
It is now well known that the performance of a linear code Copf under iterative decoding on a binary erasure channel (and other channels) is determined by the size of the smallest stopping set in the Tanner graph for Copf. Several recent papers refer to this parameter as the stopping distance s of Copf. This is somewhat of a misnomer since the size of the smallest stopping set in the Tanner graph for Copf depends on the corresponding choice of a parity-check matrix. It is easy to see that s les d, where d is the minimum Hamming distance of Copf, and we show that it is always possible to choose a parity-check matrix for Copf (with sufficiently many dependent rows) such that s = d. We thus introduce a new parameter, termed the stopping redundancy of Copf, defined as the minimum number of rows in a parity-check matrix H for Copf such that the corresponding stopping distance s(H) attains its largest possible value, namely s(H) = d. We then derive general bounds on the stopping redundancy of linear codes. We also examine several simple ways of constructing codes from other codes, and study the effect of these constructions on the stopping redundancy. Specifically, for the family of binary Reed-Muller codes (of all orders), we prove that their stopping redundancy is at most a constant times their conventional redundancy. We show that the stopping redundancies of the binary and ternary extended Golay codes are at most 34 and 22, respectively. Finally, we provide upper and lower bounds on the stopping redundancy of MDS codes
The minimum pseudocodeword weightwminof a linear graph-based code is more influential in determining decoding performance when decoded via iterative and linear programming decoding algorithms than the classical minimum distancedminunder standard maximum-likelihood decoding Moreover, unlike the minimum distance which is unique to the code regardless of representation, the set of pseudocodewords, and therefore also the minimum pseudocodeword weight, depends on the graph representation used in decoding as well as on the communication channel. This means that a judicious choice of parity-check matrix is crucial for realizing the best potential of any graph-based code. In this paper, we introduce the notion ofpseudoweightredundancyfor the memoryless binary symmetric channel (BSC). Analogous to the stopping redundancy in the literature, this parameter gives the smallest number of rows needed for a parity-check matrix to havedmin=wmin. We provide some upper bounds on the BSC-pseudoweight redundancy and illustrate the concept with some results for Hamming codes, tree-based and finite geometry LDPC codes, Reed-Muller codes and Hadamard codes.
The ML redundancy of a code is defined as the smallest number of rows in its parity-check matrix such that a message-passing decoder working in the corresponding Tanner graph achieves maximum-likelihood (ML) performance on an erasure channel. General upper bounds on ML redundancy are obtained. In particular, it is shown that the ML redundancy of a q-ary code is at most the number of minimal codewords in its dual code, divided by q−1. Special upper bounds are derived for codes whose dual code contains a covering design. For example, the ML redundancy of a Simplex code of length n is shown to be no greater than (n2− 4n + 9)/6.
This work focuses on the expansion properties of a Tanner Graph because they are known to be related to the performance of associated iterative message-passing algorithms over various channels. By analyzing the eigenvalues and corresponding eigenvectors of the normalized incidence matrix representing a Tanner Graph, lower bounds on these expansion properties are derived. Specifically, for the binary erasure channel, these results lead to two lower bounds on stopping distance for any given binary linear code and an upper bound on stopping redundancy for the family of difference-set codes (type-I 2-D projective geometry low-density parity-check (LDPC) codes).