MEGA: Message Passing Neural Networks for Multigraphs with EdGe Attributes
This work presents MEGA-GNN, a model-agnostic message-passing framework for edge-attributed multigraphs, and shows that MEGA-GNN is permutation equivariant and has the same asymptotic complexity as standard GNNs with edge updates.
H. Bilgi, Kubilay Atasu, Aggregation Aggregation
· 2 citations