Neural Decoding with Optimization of Node Activations
This work addresses decoding performance in communication systems, but it is incremental as it builds directly on existing neural Belief Propagation decoders.
The paper tackles the problem of improving neural decoders for error-correcting codes by introducing two novel loss terms that optimize node activations, resulting in a performance gain of up to 1.1dB on BCH codes without increasing runtime complexity or model size.
The problem of maximum likelihood decoding with a neural decoder for error-correcting code is considered. It is shown that the neural decoder can be improved with two novel loss terms on the node's activations. The first loss term imposes a sparse constraint on the node's activations. Whereas, the second loss term tried to mimic the node's activations from a teacher decoder which has better performance. The proposed method has the same run time complexity and model size as the neural Belief Propagation decoder, while improving the decoding performance by up to $1.1dB$ on BCH codes.