| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Clustering | CIFAR-10 (test) | Accuracy100 | 184 | |
| Image Classification | CIFAR-10 NAS-Bench-201 (test) | Accuracy94.37 | 173 | |
| Out-of-distribution Detection | CIFAR-10 vs SVHN (test) | AUROC99.84 | 101 | |
| Image Classification | CIFAR-10 v1 (test) | Accuracy90.06 | 98 | |
| Neural Architecture Search | CIFAR-10 NAS-Bench-201 (val) | Accuracy94.37 | 86 | |
| Image Classification | CIFAR-10 Long-Tailed 1.0 (test) | Top-1 Error6.77 | 79 | |
| Backdoor Defense | CIFAR-10 | Attack Success Rate6,010 | 78 | |
| Adversarial Robustness | CIFAR-10 (test) | Attack Success Rate (ASR)47.7 | 76 | |
| Out-of-Distribution Detection and Generalization | CIFAR-10 ID LSUN-C semantic OOD & CIFAR-10-C covariate OOD | OOD Accuracy91.08 | 74 | |
| Data Poisoning Defense | CIFAR-10 (test) | Test Accuracy95.33 | 72 | |
| Image Classification | CIFAR-10 (test) | Accuracy96.1 | 68 | |
| Image Classification | CIFAR-10 (250 labels) | Error Rate2.42 | 66 | |
| Image Classification | CIFAR-10 Symmetric Noise (test) | Test Accuracy (Overall)93.6 | 64 | |
| Image Classification | CIFAR-10 (test) | Accuracy94.41 | 63 | |
| Generative Image Synthesis | CIFAR-10 BigGAN | FID8.01 | 62 | |
| Generative Image Synthesis | CIFAR-10 SNGAN | FID10.6 | 62 | |
| Image Classification | CIFAR-10-C (test) | Accuracy (Clean)91.57 | 61 | |
| Image Classification | CIFAR-10 Long Tailed Imbalance Ratio 50 (test) | Top-1 Accuracy89.2 | 57 | |
| Two-sample testing | CIFAR-10 vs CIFAR-10.1 1.0 (test) | Test Power0.993 | 54 | |
| Task-incremental learning | CIFAR-10 Split (test) | Average Accuracy98.31 | 46 | |
| Image Generation | CIFAR-10 32x32 | FID1.96 | 44 | |
| Image Classification | CIFAR-10 Standard data augmentation (test) | Test Error Rate3.38 | 43 | |
| Image Classification | CIFAR-10 10% label | Accuracy91.32 | 42 | |
| Membership Inference Attack | CIFAR-10 (test) | TPR@0.1%FPR1,038 | 42 | |
| Classification | CIFAR-10 | Robust Accuracy94.1 | 40 |