#sparse-matrix

True or false: with the adjacency matrix stored as a sparse matrix, the cost of one GCN layer $\hat A X W$ on a graph with $N$ nodes and $|E|$ edges, where $\hat A = \tilde D^{-1/2}\tilde A\tilde D^{-1/2}$, $\tilde A = A + I$ and $\tilde D$ is the degree matrix of $\tilde A$, grows linearly with $|E|$ (for fixed feature sizes), not with $N^2$.