Supplementary Material to " Topological Relational Learning on Graphs "

We use the same random seeds across all models (the experiments are repeated 5 times, each with the same random seed).