This document discusses graph convolutional networks (GCNs), which use graph convolutions to perform semi-supervised learning on graphs. GCNs apply convolutional filters directly on graphs by using a graph Laplacian matrix to capture information from a node's local neighborhood. The document outlines how GCNs perform layer-wise linear models using a graph convolution theorem to aggregate features from neighboring nodes. An example application to a citation network demonstrates how GCNs can be used for semi-supervised node classification tasks on graph-structured data.