Federated Node Classification on ogbn-arxiv
68.93AccuracyFedEPD
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FedEPDBackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 68.93 | 43.57 | 46.17 | |
| FedGTABackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 63.96 | 34.42 | 34.52 | |
| FedAvgBackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 61.63 | 31.64 | 31.88 | |
| GraphSMOTEBackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 61.49 | 42.52 | 40.69 | |
| HieTailBackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 59.12 | 25.22 | 24.5 | |
| FedSprayBackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 58.14 | 23.77 | 23.22 | |
| FedProtoBackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 56.3 | 24.09 | 23.23 | |
| LWSBackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 53.99 | 20.63 | 19 | |
| FedLCBackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 52.94 | 20.26 | 18.47 | |
| CReFFBackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 51.74 | 19.28 | 17.63 | |
| GraphFedMigBackbone=two-layer GCN, Hidden Dimension=64, Learning Rate=1e-2, Communication Rounds=200, Local Epochs=3, Number of Clients=102026.06 | 43.37 | 15.88 | 16.97 |