| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Image Classification | TinyImageNet (test) | Accuracy90.23 | 440 | |
| Image Classification | TinyImageNet (val) | Accuracy90.65 | 289 | |
| Image Classification | TinyImageNet-200 | Top-1 Accuracy85.04 | 68 | |
| Model Inversion Attack | TinyImageNet-200 | Attack Accuracy37 | 60 | |
| Membership Inference Attack | TinyImageNet 200 | Balanced Accuracy79 | 60 | |
| Image Classification | TinyImageNet (test) | Communication Rounds342 | 56 | |
| Image Classification | TinyImageNet 100 tasks (test) | Accuracy33.2 | 51 | |
| Continual Learning | TinyImageNet 100 tasks (test) | Average Forgetting Rate11.7 | 51 | |
| Multi-class classification | TinyImageNet Dirichlet non-IID setting (test) | Top-5 Acc52.9 | 33 | |
| Online Continual Learning | TinyImageNet 100/2 | Accuracy12.6 | 31 | |
| Generative Image Synthesis | TinyImageNet | FID15.43 | 29 | |
| Generative Image Synthesis | TinyImageNet BigGAN | FID14.15 | 29 | |
| Class-Incremental Learning | TinyImageNet (test) | Accuracy52.77 | 29 | |
| Continual Learning | TinyImageNet 25-split | ACC58.28 | 29 | |
| Accuracy Estimation | TinyImageNet | MAE0.612 | 27 | |
| Image Classification | TinyImageNet 200 | Surrogate Model Accuracy90.88 | 26 | |
| Class-Incremental Learning | TinyImageNet Seq | FAA29.06 | 24 | |
| Continual Learning | TinyImageNet Split 10 sequential tasks (test) | Final Forgetting0 | 24 | |
| Image Classification | TinyImageNet (test) | Accuracy64.23 | 24 | |
| Membership Inference | TinyImageNet | Loss0.51 | 23 | |
| Communication Overhead | TinyImageNet (test) | Upload Size20.48 | 21 | |
| Online Continual Learning | TinyImageNet Sequential (test) | Accuracy15.97 | 21 | |
| Exemplar-Free Class-Incremental Learning | TinyImageNet (test) | Avg Increment36.5 | 18 | |
| Open-set recognition | TinyImageNet 20 closed, 180 open classes 1.0 | AUROC86.7 | 18 | |
| Image Classification | TinyImageNet et = 1/255 (val) | Standard Error74.29 | 18 |