Image Classification on CIFAR-100 with Symmetric Noise (1k clean samples)
70.36Accuracy (eta=0.2)MW-Net
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MW-NetModel Selection=Best2026.03 | 70.36 | 63.73 | 55.46 | |
| Robust Self-Training with Closed-loop Label CorrectionModel Selection=Best2026.03 | 68.13 | 61.74 | 47.91 | |
| Cross-EntropyModel Selection=Best2026.03 | 67.96 | 63.7 | 53.92 | |
| FasTENModel Selection=Best2026.03 | 67.03 | 55.87 | 51.47 | |
| Robust Self-Training with Closed-loop Label CorrectionModel Selection=Last2026.03 | 65.56 | 58.54 | 43.5 | |
| L2BModel Selection=Best2026.03 | 63.93 | 53.24 | 23.14 | |
| L2BModel Selection=Last2026.03 | 63.74 | 53.08 | 22.98 | |
| MW-NetModel Selection=Last2026.03 | 63.42 | 48.5 | 30.55 | |
| Cross-EntropyModel Selection=Last2026.03 | 62.88 | 48.12 | 29.72 | |
| FasTENModel Selection=Last2026.03 | 61.56 | 41.95 | 28.71 |