There are various types of graph convolutions (check out A Comprehensive Survey on Graph Neural Networks), but in this section, we'll discuss the algorithm from Semi-Supervised Classification with Graph Convolutional Networks (https://arxiv.org/abs/1609.02907). We'll denote this convolution with GCN to avoid confusion with the general ConvGNN notation, which refers to graph convolutional networks in general. GCN is a representative of the so-called spectral-based category of ConvGNNs. These algorithms define graph convolutions by introducing filters from the perspective of graph-signal processing, where the graph convolutional operation is interpreted as removing noises from graph signals.
In the Graph neural network ...
