Читать книгу Artificial Intelligence and Quantum Computing for Advanced Wireless Networks - Savo G. Glisic - Страница 84

In order to analyze the network complexity, we need a more detailed definition of both network topologies and network parameters. For this, we will use the original definitions introduced in [1], one of the first papers in this field. For this reason, in this section we will revisit the description of the GNN with more details on the specific model of the network, computation of the states, the learning process, and specific transition and output function implementations, and to facilitate a better understanding of these processes, we will provide a comparison with random walks and recursive neural nets. When discussing computational complexity issues, we will look at the complexity of instructions and the time complexity of the GNN mode. The illustrations will be based on the results presented in [1].