Concatenated Classic and Neural (CCN) Codes: ConcatenatedAE
This work addresses error correction for communication systems by extending neural code capabilities, though it appears incremental as it builds on existing small neural network approaches.
The paper tackles the problem of limited code dimension in neural network-based error correction by proposing a concatenated architecture that combines multiple identical neural networks with an outer classic Reed-Solomon code, resulting in significant improvements in block error probabilities for Gaussian noise channels and enhanced robustness to channel model changes.
Small neural networks (NNs) used for error correction were shown to improve on classic channel codes and to address channel model changes. We extend the code dimension of any such structure by using the same NN under one-hot encoding multiple times, then serially-concatenated with an outer classic code. We design NNs with the same network parameters, where each Reed-Solomon codeword symbol is an input to a different NN. Significant improvements in block error probabilities for an additive Gaussian noise channel as compared to the small neural code are illustrated, as well as robustness to channel model changes.