Binary Image Classification on PneumoniaMNIST Mild Non-IID (alpha = 1.0) (test)
89.81AccuracyLocal-Best
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Local-BestBackbone=ResNet-50, Resolution=224x224x3, Seeds=52026.05 | 89.81 | 88.3 | 8.41 | |
| ChainLearn-Dropout (Ours-Dropout)Backbone=ResNet-50, Resolution=224x224x3, Seeds=52026.05 | 87.63 | 85.43 | 4.57 | |
| CentralizedBackbone=ResNet-50, Resolution=224x224x3, Seeds=52026.05 | 86.09 | 83 | 10.92 | |
| ChainLearn (Ours)Backbone=ResNet-50, Resolution=224x224x3, Seeds=52026.05 | 84.29 | 80.45 | 10.39 | |
| EqualWt-EnsBackbone=ResNet-50, Resolution=224x224x3, Seeds=5, Weighting=Equal (Wi=1)2026.05 | 83.72 | 79.64 | 10.57 | |
| FedMDBackbone=ResNet-50, Resolution=224x224x3, Seeds=5, Averaging=Temperature-scaled probabilities2026.05 | 83.21 | 78.73 | 10.16 | |
| FedAvgBackbone=ResNet-50, Resolution=224x224x3, Seeds=5, Rounds=52026.05 | 76.67 | 68.56 | 17.89 | |
| FedProxBackbone=ResNet-50, Resolution=224x224x3, Seeds=5, Stabilization=Proximal2026.05 | 73.97 | 62.17 | 19.82 |