Graph convolutional networks (GCNs) are neural networks designed for graph-structured data, where nodes represent entities and edges represent relationships. Unlike traditional CNNs, GCNs handle irregular, non-Euclidean structures, making them suitable for tasks like social network analysis and molecular property prediction. GCNs generalize convolution to graphs by aggregating information from a node’s neighbors, enabling feature propagation across the graph. This chapter details the formulation of GCN. Further, a variant of GAN, graph attention networks (GATs) that introduces attention mechanisms to adaptively weigh neighbors, is also discussed.