Channel decoding method and system based on graph neural network

The invention discloses a channel decoding method and system based on a graph neural network, and the core of the channel decoding method is that the iteration of each edge or point of a decoder represented by the graph neural network is equivalent to that a new point or edge vector is output from related information through a multi-layer neural network; and the neural network continuously learns the structure of the graph in message transmission, so that the purpose of channel decoding is realized. Compared with a traditional neural network decoding method, due to the fact that parameters can be shared, the method breaks through the deep learning decoding limitation of a dimension curse, the parameters are reduced in order of magnitude compared with a traditional model, the complexity of the model is reduced, and due to the fact that nodes or edge vectors can be updated in parallel and the parallel capacity is enhanced, the decoding efficiency is improved. Compared with a belief propagation algorithm based o