Image Classification on CIFAR-10 (test) (AUCPR)
0.972AUCPRMetricOpt
Evaluation results
| Method | Links | |
|---|---|---|
| MetricOptOptimizer=learned, Surrogate Loss=AUCPR loss, Backbone=Network in [10]2021.04 | 0.972 | |
| MetricOptOptimizer=SGD, Surrogate Loss=AUCPR loss, Backbone=Network in [10]2021.04 | 0.966 | |
| MetricOptOptimizer=learned, Surrogate Loss=cross-entropy, Backbone=Network in [10]2021.04 | 0.962 | |
| MetricOptOptimizer=SGD, Surrogate Loss=cross-entropy, Backbone=Network in [10]2021.04 | 0.957 | |
| Adaptive Loss Alignment (ALA)Backbone=Network in [10]2021.04 | 0.949 | |
| Example weighting (MOEW)Backbone=Network in [10]2021.04 | 0.946 | |
| Black-box differentiationBackbone=Network in [10]2021.04 | 0.944 | |
| Pairwise AUCROC lossBackbone=Network in [10]2021.04 | 0.942 | |
| AUCPR lossBackbone=Network in [10]2021.04 | 0.942 | |
| Cross-entropy finetune + dense metric evalBackbone=Network in [10]2021.04 | 0.908 | |
| Cross-entropy lossBackbone=Network in [10]2021.04 | 0.846 |