Sign is Not a Remedy: Multiset-to-Multiset Message Passing for Learning on Heterophilic Graphs

Benchmark Model Rank Results
node-classification-on-actorM2M-GNN#24Accuracy: 36.72 ± 1.6
node-classification-on-chameleonM2M-GNN#7Accuracy: 75.20 ± 2.3
node-classification-on-cornellM2M-GNN#9Accuracy: 86.48 ± 6.1
node-classification-on-squirrelM2M-GNN#11Accuracy: 63.60 ± 1.7
node-classification-on-texasM2M-GNN#4Accuracy: 89.19 ± 4.5
node-classification-on-wisconsinM2M-GNN#5Accuracy: 89.01 ± 4.1