RNN Decoding of Linear Block Codes

TL;DR

This paper introduces a recurrent neural network decoder for linear block codes that achieves comparable or better performance than existing neural and belief propagation decoders with fewer parameters and can enhance or simplify the mRRD decoding algorithm.

Contribution

It presents a novel RNN-based decoding architecture for linear block codes, improving efficiency and performance over prior neural and belief propagation methods.

Findings

RNN decoder achieves similar bit error rates as feed-forward neural networks.

The RNN decoder outperforms belief propagation on sparser Tanner graphs.

It can enhance or reduce the complexity of the mRRD decoding algorithm.

Abstract

Designing a practical, low complexity, close to optimal, channel decoder for powerful algebraic codes with short to moderate block length is an open research problem. Recently it has been shown that a feed-forward neural network architecture can improve on standard belief propagation decoding, despite the large example space. In this paper we introduce a recurrent neural network architecture for decoding linear block codes. Our method shows comparable bit error rate results compared to the feed-forward neural network with significantly less parameters. We also demonstrate improved performance over belief propagation on sparser Tanner graph representations of the codes. Furthermore, we demonstrate that the RNN decoder can be used to improve the performance or alternatively reduce the computational complexity of the mRRD algorithm for low complexity, close to optimal, decoding of short BCH codes.