Graph Convolutional Networks (GCN), Kipf and Welling published a paper titled "Graph Convolutional Networks (GCN)" at the ICLR conference in 2017.Semi-Supervised Classification with Graph Convolutional Networks", which deeply discussed the theoretical basis and application of GCN.

GCN uses graph convolution operations to capture the local structural information and feature information of nodes in the graph, thereby achieving effective processing and analysis of graph data. In a graph convolutional network, the features of a node are aggregated with the features of its neighboring nodes to update the state of the node. GCN usually uses the Laplacian matrix of the graph to define the aggregation relationship between nodes and propagate the features of the nodes through this relationship. The key to GCN is that it can handle graph structured data well, which makes it widely used in social network analysis, bioinformatics, recommendation systems and other fields.