Image Classification on MNIST (test) (Accuracy and F1)
96.97AccuracyULHM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ULHMtransfer_direction=Fashion-MNIST to MNIST2026.01 | 96.97 | — | |
| Prototypical Networkstransfer_direction=Fashion-MNIST to MNIST2026.01 | 93.32 | — | |
| Cond. Domain Adversarialtransfer_direction=Fashion-MNIST to MNIST2026.01 | 89.38 | — | |
| Correlation Alignmenttransfer_direction=Fashion-MNIST to MNIST2026.01 | 83.6 | — | |
| ALIGN-FLAlgorithm=ALIGN-FL2025.12 | 74.25 | 74.06 | |
| ContrastAvgSynthAlgorithm=ContrastAvgSynth2025.12 | 62.75 | 59.05 | |
| Min. Class Confusiontransfer_direction=Fashion-MNIST to MNIST2026.01 | 58.87 | — | |
| ContrastSynthOnlyAlgorithm=ContrastSynthOnly2025.12 | 52.5 | 50.66 | |
| FedAvgAlgorithm=FedAvg2025.12 | 43.75 | 40.25 | |
| Max. Classifier Discrepancytransfer_direction=Fashion-MNIST to MNIST2026.01 | 42.4 | — | |
| MOON-BaseAlgorithm=MOON-Base2025.12 | 38.75 | 34.35 | |
| FedProxAlgorithm=FedProx2025.12 | 38.5 | 34.07 | |
| Cross-Domain Self-supervisedtransfer_direction=Fashion-MNIST to MNIST2026.01 | 20.28 | — | |
| Partial Convolutiontransfer_direction=Fashion-MNIST to MNIST2026.01 | 13.8 | — |