Learn When and Where to Connect: Adaptive Virtual Nodes for Dynamic Message Passing on Graphs

Benchmark Model Rank Results
graph-classification-on-peptides-funcMAVN-GCN–AP: 0.7238
graph-classification-on-peptides-funcMAVN-GINE–AP: 0.6897
graph-classification-on-peptides-funcMAVN-GatedGCN–AP: 0.7070
graph-regression-on-peptides-structMAVN-GCN–MAE: 0.2432
graph-regression-on-peptides-structMAVN-GINE–MAE: 0.2453
graph-regression-on-peptides-structMAVN-GatedGCN–MAE: 0.2410