Learning to Defer on ImageNet Dog Expert
41.32Error RatePiCCE-CE
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PiCCE-CENumber of Experts=202026.02 | 41.32 | 90.01 | |
| PiCCE-CENumber of Experts=162026.02 | 41.4 | 89.99 | |
| PiCCE-CENumber of Experts=122026.02 | 41.5 | 89.78 | |
| PiCCE-CENumber of Experts=82026.02 | 41.96 | 89.91 | |
| PiCCE-CENumber of Experts=42026.02 | 42.1 | 89.85 | |
| PiCCE-OvANumber of Experts=82026.02 | 42.11 | 89.29 | |
| PiCCE-OvANumber of Experts=162026.02 | 42.17 | 89.12 | |
| PiCCE-OvANumber of Experts=202026.02 | 42.25 | 89.08 | |
| PiCCE-OvANumber of Experts=42026.02 | 42.41 | 89.17 | |
| PiCCE-OvANumber of Experts=122026.02 | 42.45 | 89.2 | |
| CELoss Formulation=Vanilla, Number of Experts=42026.02 | 42.74 | 89.08 | |
| CELoss Formulation=Vanilla, Number of Experts=82026.02 | 44.16 | 88.71 | |
| OvALoss Formulation=Vanilla, Number of Experts=42026.02 | 45.77 | 88.45 | |
| CELoss Formulation=Vanilla, Number of Experts=122026.02 | 46.36 | 88.5 | |
| CELoss Formulation=Vanilla, Number of Experts=162026.02 | 47.01 | 88.45 | |
| CELoss Formulation=Vanilla, Number of Experts=202026.02 | 48 | 88.51 | |
| OvALoss Formulation=Vanilla, Number of Experts=82026.02 | 52.62 | 86.66 | |
| OvALoss Formulation=Vanilla, Number of Experts=122026.02 | 57.66 | 85.74 | |
| OvALoss Formulation=Vanilla, Number of Experts=162026.02 | 62.16 | 84.67 | |
| OvALoss Formulation=Vanilla, Number of Experts=202026.02 | 66.56 | 83.9 |