List your company
ResearchFormerly WizWand

Image Classification on ImageNet noise level 125 16H (test)

19.45Error RateP+L (Uncalibrated)
18.637624.121329.60535.0887Sep 29, 2021
Updated 8mo ago

Evaluation results

MethodLinks
P+L (Uncalibrated)Backbone=VGG-19, Number of seeds=25, Labeled data points=5067-51522021.0919.45
P+L (MAP TS)Backbone=VGG-19, Number of seeds=25, Labeled data points=5067-5152, Calibration=MAP TS2021.0919.63
P+L (IMax Binning)Backbone=VGG-19, Number of seeds=25, Labeled data points=5067-5152, Calibration=IMax Binning2021.0920.47
P+L (Ensemble TS)Backbone=VGG-19, Number of seeds=25, Labeled data points=5067-5152, Calibration=Ensemble TS2021.0920.74
L+LBackbone=VGG-19, Number of seeds=25, Labeled data points=5067-51522021.0921.89
VGG-19Backbone=VGG-192021.0922.07
Human2021.0939.76