Clusterability as an Alternative to Anchor Points When Learning with Noisy Labels

Benchmark Model Rank Results
learning-with-noisy-labels-on-cifar-10n-random3CAL#8Accuracy (mean): 90.74
learning-with-noisy-labels-on-cifar-100nCAL#9Accuracy (mean): 61.73
learning-with-noisy-labels-on-cifar-10n-random2CAL#9Accuracy (mean): 90.75
learning-with-noisy-labels-on-cifar-10n-random1CAL#10Accuracy (mean): 90.93
learning-with-noisy-labels-on-cifar-10n-worstCAL#10Accuracy (mean): 85.36
learning-with-noisy-labels-on-cifar-10n-aggregateCAL#12Accuracy (mean): 91.97
image-classification-on-clothing1mHOC#25Accuracy: 73.39%