Graph Neural Networks

This chapter presents a brief overview of the graph neural network (GNN). Several major modifications have been made to the propagation step from the original GNN model, whereas in the output step a simple feedforward neural network setting is the most popular. The chapter examines a number of GNN designs with the very basic description of their principles. It aims to categorize these options into recurrent graph neural networks, convolutional graph neural networks, graph autoencoders, and spatial‐temporal graph neural networks. When discussing computational complexity issues, it looks at the complexity of instructions and the time complexity of the GNN mode. The chapter describes with more details the most complex instructions involved in the learning procedure. In GNNs, the learning phase requires much more time than the test phase, mainly due to the repetition of the forward and backward phases for several epochs.