Doing More With Less: Towards More Data-Efficient Syndrome-Based Neural Decoders

TL;DR

This paper investigates how carefully curated fixed training datasets can improve the data efficiency of syndrome-based neural decoders, reducing the need for large amounts of training data.

Contribution

It introduces heuristics for selecting training samples and demonstrates that fixed datasets can outperform dynamic data generation in neural decoding.

Findings

Curated fixed datasets lead to better neural decoder performance.

Sample selection heuristics improve training efficiency.

Fewer training examples are needed for high accuracy.

Abstract

While significant research efforts have been directed toward developing more capable neural decoding architectures, comparatively little attention has been paid to the quality of training data. In this study, we address the challenge of constructing effective training datasets to maximize the potential of existing syndrome-based neural decoder architectures. We emphasize the advantages of using fixed datasets over generating training data dynamically and explore the problem of selecting appropriate training targets within this framework. Furthermore,we propose several heuristics for selecting training samples and present experimental evidence demonstrating that, with carefully curated datasets, it is possible to train neural decoders to achieve superior performance while requiring fewer training examples. Code to reproduce all results is available at https://github.com/lebidan/sbnd.