Noisy Label Detection on CIFAR10
0.95AUCKNN
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| KNNN=50,000, d=7682026.05 | 0.95 | — | — | — | |
| LSH-ShapleyN=50,000, d=7682026.05 | 0.899 | — | — | — | |
| G-ShapleyN=50,000, d=7682026.05 | 0.504 | — | — | — | |
| CS-SHAPLEYClassifier=Logistic Regression2022.11 | 0.45 | — | — | — | |
| TMC-ShapleyClassifier=Logistic Regression2022.11 | 0.429 | — | — | — | |
| Beta ShapleyClassifier=Logistic Regression2022.11 | 0.424 | — | — | — | |
| CS-SHAPLEYClassifier=SVM-RBF2022.11 | 0.387 | — | — | — | |
| Beta ShapleyClassifier=SVM-RBF2022.11 | 0.321 | — | — | — | |
| TMC-ShapleyClassifier=SVM-RBF2022.11 | 0.317 | — | — | — | |
| Leave-One-OutClassifier=Logistic Regression2022.11 | 0.275 | — | — | — | |
| Leave-One-OutClassifier=SVM-RBF2022.11 | 0.272 | — | — | — | |
| CleanlabLabel error (%)=5%2026.06 | — | 95.51 | 52.87 | 93.72 | |
| CleanlabLabel error (%)=10%2026.06 | — | 93.76 | 62.5 | 93.96 | |
| CleanlabLabel error (%)=20%2026.06 | — | 89.55 | 67.26 | 93.07 | |
| CleanlabLabel error (%)=40%2026.06 | — | 82.34 | 72.38 | 90.31 |