Positive-Unlabeled Classification on CIFAR-10 (test)
97.22AccuracyWConPU
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| WConPU2025.12 | 97.22 | 96.87 | 96.02 | 96.43 | 99.49 | 99.25 | |
| PiCO2025.12 | 95.64 | 94.89 | 93.97 | 94.75 | 98.67 | 98.22 | |
| puNCE2025.12 | 95.32 | 95.11 | 93.43 | 94.21 | 98.59 | 98.45 | |
| Dist-PU2025.12 | 91.88 | 89.87 | 89.84 | 89.85 | 96.92 | 95.49 | |
| AngularPU2025.12 | 91.39 | 92.26 | 93.49 | 92.87 | 96.61 | 97.38 | |
| Dense-PU2025.12 | 90.59 | 92.68 | 91.25 | 91.96 | 93.22 | 95.03 | |
| PUbN2025.12 | 89.83 | 87.85 | 86.56 | 87.18 | 94.44 | 91.28 | |
| ImbPU2025.12 | 89.43 | 86.72 | 86.91 | 86.77 | 95.53 | 93.33 | |
| Self-PU2025.12 | 89.31 | 86.26 | 87.22 | 86.77 | 95.52 | 93.31 | |
| aPU2025.12 | 89.09 | 86.31 | 86.33 | 86.41 | 95.11 | 92.42 | |
| PUSB2025.12 | 88.97 | 86.15 | 86.22 | 86.18 | 95.15 | 92.44 | |
| nnPU2025.12 | 88.91 | 86.21 | 86.03 | 86.11 | 95.13 | 92.51 | |
| RP2025.12 | 88.74 | 86.02 | 85.72 | 85.93 | 95.21 | 93.01 | |
| uPU2025.12 | 88.41 | 87.21 | 83.02 | 85.12 | 94.98 | 92.71 | |
| Self-PUArchitecture=Neural Network, Self-calibration=false2022.10 | 88.22 | — | — | — | — | — | |
| VPU2025.12 | 87.89 | 86.71 | 82.88 | 84.42 | 94.55 | 92.02 | |
| nnPUArchitecture=Neural Network2022.10 | 81.87 | — | — | — | — | — | |
| PU ETArchitecture=Tree-based, Risk estimator=nonnegative, Loss function=quadratic2022.10 | 79.74 | — | — | — | — | — | |
| uPUArchitecture=Neural Network2022.10 | 62.68 | — | — | — | — | — |