Edge flow prediction on Bike (test)
0.029RMSEFlowSymm
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| FlowSymm2026.01 | 0.029 | 0.014 | 0.95 | |
| Bil-GCNformulation=bilevel, encoder=GCN2026.01 | 0.032 | 0.016 | 0.93 | |
| Bil-MLPformulation=bilevel, encoder=MLP2026.01 | 0.033 | 0.016 | 0.93 | |
| MLP-Divformulation=hybrid2026.01 | 0.037 | 0.019 | 0.9 | |
| GCN-Divformulation=hybrid2026.01 | 0.037 | 0.019 | 0.9 | |
| Divformulation=physics-based2026.01 | 0.039 | 0.018 | 0.91 | |
| GCNformulation=data-driven2026.01 | 0.057 | 0.039 | 0.72 | |
| GINformulation=data-driven2026.01 | 0.059 | 0.04 | 0.71 | |
| EGNNformulation=data-driven2026.01 | 0.063 | 0.038 | 0.75 | |
| MLPformulation=data-driven2026.01 | 0.073 | 0.05 | 0.69 |