UniGNN Args: args (NamedTuple): global args nfeat (int): dimension of features nhid (int): dimension

Row-normalize matrix

initialises model, optimiser, normalises graph, and features arguments: X, Y, G: the entire dataset (with graph, features, labels)

approximates the E defined by the E Laplacian with/without mediators arguments: V: number of vertices E: dictionary of hyperedges (k

d: initial node-feature dimension h: number of hidden units c: number of classes

computes an sparse adjacency matrix arguments: edges: list of pairs weights: dictionary of edge weights (key: tuple representing edg

row-normalise sparse matrix arguments: M: scipy sparse matrix returns: D^{-1} M where D is the diagonal node-degree matri

converts a scipy sparse matrix (ssm) to a torch sparse tensor (tst) arguments: M: scipy sparse matrix returns: a torch sparse t

symmetrically normalise sparse matrix arguments: M: scipy sparse matrix returns: D^{-1/2} M D^{-1/2} where D is the diagon

updates the weight on {Se,mediator} and {Ie,mediator}

UniGNNII Args: args (NamedTuple): global args nfeat (int): dimension of features nhid (int): dimensi

an l-layer GCN