Graph Neural Networks for Channel Decoding

TL;DR

This paper introduces a graph neural network-based decoder for channel coding that learns message passing algorithms, achieving competitive or superior performance to traditional methods across various coding schemes with scalable and efficient training.

Contribution

The authors develop a fully differentiable GNN architecture for channel decoding that outperforms traditional algorithms and previous deep learning approaches, with scalability to arbitrary block lengths.

Findings

Outperforms weighted belief propagation on BCH codes by 0.4 dB

Achieves competitive performance on 5G NR LDPC codes

Uses only 9640 parameters for BCH code decoding

Abstract

In this work, we propose a fully differentiable graph neural network (GNN)-based architecture for channel decoding and showcase a competitive decoding performance for various coding schemes, such as low-density parity-check (LDPC) and BCH codes. The idea is to let a neural network (NN) learn a generalized message passing algorithm over a given graph that represents the forward error correction (FEC) code structure by replacing node and edge message updates with trainable functions. Contrary to many other deep learning-based decoding approaches, the proposed solution enjoys scalability to arbitrary block lengths and the training is not limited by the curse of dimensionality. We benchmark our proposed decoder against state-of-the-art in conventional channel decoding as well as against recent deep learning-based results. For the (63,45) BCH code, our solution outperforms weighted belief propagation (BP) decoding by approximately 0.4 dB with significantly less decoding iterations and even for 5G NR LDPC codes, we observe a competitive performance when compared to conventional BP decoding. For the BCH codes, the resulting GNN decoder can be fully parametrized with only 9640 weights.