Classification on ImageNet-16H noise level 80
6.03Error RateP+L (MAP TS)
Evaluation Results
| Method | Links | |
|---|---|---|
| P+L (MAP TS)Backbone=VGG-19, Number of seeds=25, Labeled data points=5067-5152, Calibration=MAP TS2021.09 | 6.03 | |
| P+L (IMax Binning)Backbone=VGG-19, Number of seeds=25, Labeled data points=5067-5152, Calibration=IMax Binning2021.09 | 6.31 | |
| L+LBackbone=VGG-19, Number of seeds=25, Labeled data points=5067-51522021.09 | 6.78 | |
| P+L (Ensemble TS)Backbone=VGG-19, Number of seeds=25, Labeled data points=5067-5152, Calibration=Ensemble TS2021.09 | 6.79 | |
| P+LBackbone=GoogLeNet, Calibration=MAP TS2021.09 | 6.8 | |
| L+LBackbone=GoogLeNet2021.09 | 7.33 | |
| P+L (Uncalibrated)Backbone=VGG-19, Number of seeds=25, Labeled data points=5067-51522021.09 | 7.52 | |
| P+LBackbone=GoogLeNet, Calibration=IMax Binning2021.09 | 7.66 | |
| P+LBackbone=GoogLeNet, Calibration=Ensemble TS2021.09 | 7.73 | |
| P+LBackbone=GoogLeNet, Calibration=Uncalibrated2021.09 | 8.06 | |
| Human2021.09 | 9.99 | |
| Human2021.09 | 9.99 | |
| VGG-19Backbone=VGG-192021.09 | 11.1 | |
| GoogLeNetBackbone=GoogLeNet2021.09 | 14.48 |