Image Classification on CIFAR-10 (test) (Performance Percentiles)
95.5Accuracy (30%)Fully Supervised
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Fully SupervisedLabel Usage=With Labels2026.05 | 95.5 | 95.2 | 93 | 91.4 | 87.1 | |
| Semi-AUM+CutoffLabel Usage=With 10% Labels2026.05 | 95.4 | 95.2 | 93.1 | 91.3 | 87.2 | |
| Semi-DUAL+BetaLabel Usage=With 10% Labels2026.05 | 95.4 | 95 | 92.7 | 90.9 | 86.2 | |
| ELFS (DINO)Label Usage=Without Labels2026.05 | 95.3 | 95.1 | 93 | 91.3 | 86.9 | |
| Score ExtrapolationLabel Usage=With 10% Labels2026.05 | 95.3 | 94.6 | 91.9 | 89.4 | 81.5 | |
| PrototypicalityLabel Usage=Without Labels2026.05 | 95.1 | 93.9 | 87.2 | 79.7 | 59.1 | |
| ELFS (Self-Encoder)Label Usage=Without Labels2026.05 | 94.8 | 93.5 | 91.2 | 89.5 | 84.5 | |
| ZCoreLabel Usage=Without Labels2026.05 | 94.6 | 93.5 | 91 | 89.1 | 84.2 | |
| RandomLabel Usage=Without Labels2026.05 | 94.4 | 93.2 | 90.5 | 88.3 | 83.7 |