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

arXiv:2601.10376

A Hybrid Reliability--Weight Framework for Construction of Polar Codes

Mohammad Rowshan, Vlad Drăgoi

Published Jan 15, 2026• 0 citations
arXivHTMLPDF

Abstract

Polar codes are usually constructed by ranking synthetic bit-channels according to reliability, which guarantees capacity-achieving behavior but can yield poor low-weight spectra at short and moderate lengths. Recent algebraic results express the contribution of individual bit-channels to the multiplicities of minimum and near-minimum weight codewords in closed form. In this work we combine these insights into a mixed (reliability--weight) bit-channel ordering. We define a per-bit cost whose distance term is derived from orbit enumeration of minimum-weight codewords and scaled by a Bhattacharyya-type factor, and show that the resulting mixed construction minimises a truncated SC/ML union-bound surrogate within a class of decreasing monomial codes. We relate the mixed metric to error events in SCL decoding via a pruning/ML decomposition, and prove that mixed designs act as local perturbations of reliability-based constructions whose asymptotic impact vanishes as code-length approaches infinity. Numerical results for short and moderate lengths on BPSK-AWGN, implemented via Gaussian approximation and closed-form weight contributions, illustrate the trade-off between pure reliability-based and mixed constructions in terms of minimum distance, multiplicity, and union-bound approximations. All proofs are deferred to the appendices.

DOI
10.48550/arxiv.2601.10376

Related papers

On the Closed-Form Weight Enumeration of Polar Codes: 1.5 d -Weight Codewords

Vlad Drăgoi, Mohammad Rowshan, Jinhong Yuan

The weight distribution of an error correction code is a critical determinant of its error-correcting performance. In the case of polar codes, the minimum weight wmin(equal to the minimum distanced) is the only weight for which an explicit enumerator formula is currently available. Having closed-form weight enumerators for polar codewords with weights greater than the minimum weight not only simplifies the enumeration process but also provides valuable insights towards constructing better polar-like codes. In this paper, we contribute towards understanding the algebraic structure underlying higher weights by analyzing Minkowski sums of orbits. Our approach builds upon the lower triangular affine (LTA) group of decreasing monomial codes. Specifically, we propose a closed-form expression for the enumeration of codewords with weight 1.5wmin. The key insight for code design is that the enumeration of codewords with weight wmin and 1.5wminrelies on the set of maximum degree monomials. This set corresponds to the indices of minimum weight rows of the polar transformGNbelonging to the information setI. Consequently, reducing the cardinality of this set can lead to a reduction of the number of codewords in both weight categories.

IEEE Transactions on Communications•May 6, 2024•18 citations

Hybrid Constrcution Method for Multi-kernel Polar Codes

Haotian Fan, Bing Zhu, Yanlong Zhao +1 more

Polar codes are the first class of error correction codes that achieve the Shannon channel capacity when their code length approaches infinity, but the code lengths of polar codes are limited to powers of 2. Therefore, multi-kernel polar codes are proposed, and their generator matrices are constructed from a variety of polarized kernels with different sizes. In this paper, we study the effect of different kernel orders on the channel polarization in the Gaussian approximation construction for multi-kernel polar codes. Besides, we study the minimum-distance spectrum of polarized kernels and the generator matrices of multi-kernel polar codes, and we calculate the minimum-distance spectrum for some larger kernels. We experimentally compare the advantages and disadvantages of the Gaussian approximation and minimum-distance construction. Finally, we propose a hybrid construction method which considers both the reliability and distance, and experimental results show that the proposed construction method outperforms the construction method that only considers distance and reliability.

International Conference on Imaging, Signal Processing and Communications (ICISPC)•Jul 1, 2022•1 citation

Improving the Error Coefficient of Polar Codes

Mohammad Rowshan, Son Hoang Dau, Emanuele Viterbo

Polar codes are normally constructed based on the reliability of the sub-channels in the polarized vector channel. Code construction based on reliability is compatible with successive cancellation decoding. However, due to poor Hamming distance properties, the designed codes cannot perform well with near maximum likelihood decoders. In this work, we propose a new approach that modifies polar codes and PAC codes to significantly lower the number of codewords with minimum distance (a.k.a. error coefficient). This approach is based on the recognition of all the rows of polar transform involved in the formation of the minimum-weight codewords. The numerical results show that the designed codes outperform polar codes and PAC codes under list decoding.

IEEE Information Theory Workshop (ITW)•Nov 1, 2022•6 citations

On the Closed-form Weight Enumeration of Polar Codes: 1.5ddd-weight Codewords

Mohammad Rowshan, Vlad Drăgoi, Jinhong Yuan

The weight distribution of error correction codes is a critical determinant of their error-correcting performance, making enumeration of utmost importance. In the case of polar codes, the minimum weight $\wm$ (which is equal to minimum distance $d$) is the only weight for which an explicit enumerator formula is currently available. Having closed-form weight enumerators for polar codewords with weights greater than the minimum weight not only simplifies the enumeration process but also provides valuable insights towards constructing better polar-like codes. In this paper, we contribute towards understanding the algebraic structure underlying higher weights by analyzing Minkowski sums of orbits. Our approach builds upon the lower triangular affine (LTA) group of decreasing monomial codes. Specifically, we propose a closed-form expression for the enumeration of codewords with weight $1.5\wm$. Our simulations demonstrate the potential for extending this method to higher weights.

arXiv:2305.02921•May 4, 2023•4 citations

A Pre-Transformation Method to Increase the Minimum Distance of Polar-Like Codes

Samet Gelincik, Philippe Mary, Anne Savard +1 more

Reed Muller (RM) codes are known for their good minimum distance. One can use their structure to construct polar-like codes with good distance properties by choosing the information set as the rows of the polarization matrix with the highest Hamming weight, instead of the most reliable synthetic channels. However, the information length options of RM codes are quite limited due to their specific structure. In this work, we present sufficient conditions to increase the information length by at least one bit for some underlying RM codes and in order to obtain pre-transformed polar-like codes with the same minimum distance than lower rate codes.The proofs give a constructive method to choose the row triples to be merged together to increase the information length of the code and they follow from partitioning the row indices of the polar encoding matrix with respect to the recursive structure imposed by the binary representation of row indices. Moreover, our findings are combined with the method presented in [2] to further reduce the number of minimum weight codewords. Numerical results show that the designed codes perform close to the meta-converse bound at short blocklengths and better than the polarization-adjusted-convolutional polar codes with the same parameters.

arXiv:2202.04366•Feb 9, 2022•3 citations

Preserving the Minimum Distance of Polar-Like Codes while Increasing the Information Length

Samet Gelincik, Philippe Mary, Anne Savard +1 more

Reed Muller (RM) codes are known for their good minimum distance. One can use their structure to construct polar-like codes with good distance properties by choosing the information set as the rows of the polarization matrix with the highest Hamming weight, instead of the most reliable synthetic channels. However, the information length options of RM codes are quite limited due to their specific structure. In this work, we present sufficient conditions to increase the information length by at least one bit for some underlying RM codes and in order to obtain pre-transformed polar-like codes with the same minimum distance than lower rate codes. Moreover, our findings are combined with the method presented in [1] to further reduce the number of minimum weight codewords. Numerical results show that the designed codes perform close to the meta-converse bound at short blocklengths and better than the polarized adjusted convolutional polar codes with the same parameters.

IEEE International Symposium on Information Theory (ISIT)•Jun 26, 2022•1 citation
Polar codes are usually constructed by ranking synthetic bit-channels according to reliability, which guarantees capacity-achieving behavior but can yield poor low-weight spectra at short and moderate lengths. Recent algebraic results express the contribution of individual bit-channels to the multiplicities of minimum and near-minimum weight codewords in closed form. In this work we combine these insights into a mixed (reliability--weight) bit-channel ordering. We define a per-bit cost whose distance term is derived from orbit enumeration of minimum-weight codewords and scaled by a Bhattacharyya-type factor, and show that the resulting mixed construction minimises a truncated SC/ML union-bound surrogate within a class of decreasing monomial codes. We relate the mixed metric to error events in SCL decoding via a pruning/ML decomposition, and prove that mixed designs act as local perturbations of reliability-based constructions whose asymptotic impact vanishes as code-length approaches infinity. Numerical results for short and moderate lengths on BPSK-AWGN, implemented via Gaussian approximation and closed-form weight contributions, illustrate the trade-off between pure reliability-based and mixed constructions in terms of minimum distance, multiplicity, and union-bound approximations. All proofs are deferred to the appendices.
The weight distribution of an error correction code is a critical determinant of its error-correcting performance. In the case of polar codes, the minimum weight wmin(equal to the minimum distanced) is the only weight for which an explicit enumerator formula is currently available. Having closed-form weight enumerators for polar codewords with weights greater than the minimum weight not only simplifies the enumeration process but also provides valuable insights towards constructing better polar-like codes. In this paper, we contribute towards understanding the algebraic structure underlying higher weights by analyzing Minkowski sums of orbits. Our approach builds upon the lower triangular affine (LTA) group of decreasing monomial codes. Specifically, we propose a closed-form expression for the enumeration of codewords with weight 1.5wmin. The key insight for code design is that the enumeration of codewords with weight wmin and 1.5wminrelies on the set of maximum degree monomials. This set corresponds to the indices of minimum weight rows of the polar transformGNbelonging to the information setI. Consequently, reducing the cardinality of this set can lead to a reduction of the number of codewords in both weight categories.
Polar codes are the first class of error correction codes that achieve the Shannon channel capacity when their code length approaches infinity, but the code lengths of polar codes are limited to powers of 2. Therefore, multi-kernel polar codes are proposed, and their generator matrices are constructed from a variety of polarized kernels with different sizes. In this paper, we study the effect of different kernel orders on the channel polarization in the Gaussian approximation construction for multi-kernel polar codes. Besides, we study the minimum-distance spectrum of polarized kernels and the generator matrices of multi-kernel polar codes, and we calculate the minimum-distance spectrum for some larger kernels. We experimentally compare the advantages and disadvantages of the Gaussian approximation and minimum-distance construction. Finally, we propose a hybrid construction method which considers both the reliability and distance, and experimental results show that the proposed construction method outperforms the construction method that only considers distance and reliability.
Polar codes are normally constructed based on the reliability of the sub-channels in the polarized vector channel. Code construction based on reliability is compatible with successive cancellation decoding. However, due to poor Hamming distance properties, the designed codes cannot perform well with near maximum likelihood decoders. In this work, we propose a new approach that modifies polar codes and PAC codes to significantly lower the number of codewords with minimum distance (a.k.a. error coefficient). This approach is based on the recognition of all the rows of polar transform involved in the formation of the minimum-weight codewords. The numerical results show that the designed codes outperform polar codes and PAC codes under list decoding.
The weight distribution of error correction codes is a critical determinant of their error-correcting performance, making enumeration of utmost importance. In the case of polar codes, the minimum weight $\wm$ (which is equal to minimum distance ddd) is the only weight for which an explicit enumerator formula is currently available. Having closed-form weight enumerators for polar codewords with weights greater than the minimum weight not only simplifies the enumeration process but also provides valuable insights towards constructing better polar-like codes. In this paper, we contribute towards understanding the algebraic structure underlying higher weights by analyzing Minkowski sums of orbits. Our approach builds upon the lower triangular affine (LTA) group of decreasing monomial codes. Specifically, we propose a closed-form expression for the enumeration of codewords with weight $1.5\wm$. Our simulations demonstrate the potential for extending this method to higher weights.
Reed Muller (RM) codes are known for their good minimum distance. One can use their structure to construct polar-like codes with good distance properties by choosing the information set as the rows of the polarization matrix with the highest Hamming weight, instead of the most reliable synthetic channels. However, the information length options of RM codes are quite limited due to their specific structure. In this work, we present sufficient conditions to increase the information length by at least one bit for some underlying RM codes and in order to obtain pre-transformed polar-like codes with the same minimum distance than lower rate codes.The proofs give a constructive method to choose the row triples to be merged together to increase the information length of the code and they follow from partitioning the row indices of the polar encoding matrix with respect to the recursive structure imposed by the binary representation of row indices. Moreover, our findings are combined with the method presented in [2] to further reduce the number of minimum weight codewords. Numerical results show that the designed codes perform close to the meta-converse bound at short blocklengths and better than the polarization-adjusted-convolutional polar codes with the same parameters.
Reed Muller (RM) codes are known for their good minimum distance. One can use their structure to construct polar-like codes with good distance properties by choosing the information set as the rows of the polarization matrix with the highest Hamming weight, instead of the most reliable synthetic channels. However, the information length options of RM codes are quite limited due to their specific structure. In this work, we present sufficient conditions to increase the information length by at least one bit for some underlying RM codes and in order to obtain pre-transformed polar-like codes with the same minimum distance than lower rate codes. Moreover, our findings are combined with the method presented in [1] to further reduce the number of minimum weight codewords. Numerical results show that the designed codes perform close to the meta-converse bound at short blocklengths and better than the polarized adjusted convolutional polar codes with the same parameters.