Leveraging Code Automorphisms for Improved Syndrome-Based Neural Decoding
For researchers working on neural decoding of error-correcting codes, this work provides a simple method to boost SBND performance, though it is an incremental improvement over existing SBND approaches.
The paper shows that leveraging code automorphisms via data augmentation during training and inference significantly improves syndrome-based neural decoding (SBND) for short high-rate codes, achieving near-MLD performance with small datasets. It also suggests that prior SBND results were underestimated due to undertraining.
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.