Leveraging Code Automorphisms for Improved Syndrome-Based Neural Decoding

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Bidan, Raphaël Le, Ismail, Ahmad, Dupraz, Elsa, Nour, Charbel Abdel
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909014453911552
author Bidan, Raphaël Le
Ismail, Ahmad
Dupraz, Elsa
Nour, Charbel Abdel
author_facet Bidan, Raphaël Le
Ismail, Ahmad
Dupraz, Elsa
Nour, Charbel Abdel
contents Syndrome-based neural decoding (SBND) has emerged as a promising deep learning approach for soft-decision decoding of high-rate, short-length codes. However, this approach still has substantial room for improvement. In this paper, we show how to leverage code automorphisms to enhance the ability of existing SBND models to learn and generalize through data augmentation during training and inference. As a result, for the short high-rate codes considered, we obtain models that closely approach MLD performance using small datasets and proper training. Our findings also suggest that many prior results for SBND models in the literature underestimate their true correction capability due to undertraining. Code to reproduce all results is available at: https://github.com/lebidan/sbnd.
format Preprint
id arxiv_https___arxiv_org_abs_2605_03620
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Leveraging Code Automorphisms for Improved Syndrome-Based Neural Decoding
Bidan, Raphaël Le
Ismail, Ahmad
Dupraz, Elsa
Nour, Charbel Abdel
Information Theory
Machine Learning
Syndrome-based neural decoding (SBND) has emerged as a promising deep learning approach for soft-decision decoding of high-rate, short-length codes. However, this approach still has substantial room for improvement. In this paper, we show how to leverage code automorphisms to enhance the ability of existing SBND models to learn and generalize through data augmentation during training and inference. As a result, for the short high-rate codes considered, we obtain models that closely approach MLD performance using small datasets and proper training. Our findings also suggest that many prior results for SBND models in the literature underestimate their true correction capability due to undertraining. Code to reproduce all results is available at: https://github.com/lebidan/sbnd.
title Leveraging Code Automorphisms for Improved Syndrome-Based Neural Decoding
topic Information Theory
Machine Learning
url https://arxiv.org/abs/2605.03620