Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

PathMNIST

Benchmarks

Task NameDataset NameSOTA ResultTrend
Medical Image ClassificationPathMNIST
Accuracy99.38
61
Image ClassificationPathMNIST
Clean Accuracy89.45
60
Medical Image ClassificationPATHMNIST
Clean Accuracy89.45
48
Image ClassificationPathMNIST v1 (test)
Accuracy92.02
36
Image ClassificationPathMNIST v2 (test)
Accuracy89.86
35
Image ClassificationPathMNIST
Accuracy96.65
34
9-class classificationPathMNIST
Accuracy93.73
32
Medical Image ClassificationPathMNIST v2 (test)
Accuracy97.5
22
Medical Image ClassificationPathMNIST 40% Noise
Sensitivity96.8
10
Medical Image ClassificationPathMNIST 20% Noise
Sensitivity95
10
Medical Image ClassificationPathMNIST 0% Noise
Sensitivity88.7
10
Machine UnlearningPathMNIST External (test)
Residual Accuracy (Acc_r)88.63
9
Machine UnlearningPathMNIST Internal (test)
Accuracy (Retained)97.78
9
ClassificationPathMNIST D11
PGD Acc45.7
7
Image ClassificationPathMNIST
BP Top-1 Accuracy88.89
6
Federated UnlearningPathMNIST
Training Time (s)350
6
Image ClassificationPathMNIST-C
Accuracy52
6
Conformal PredictionPathMNIST (test)
Coverage95
5
Machine UnlearningPathMNIST 50% removal binary (test)
Specificity99
5
Machine UnlearningPathMNIST 20% removal binary (test)
Specificity99
5
Failure types identificationBP-MNIST
Learnability92
5
Image ClassificationPATHMNIST MedMNIST (test)
Clean Accuracy84.88
4
Image ClassificationPathMNIST (test)
Accuracy@1 (%)78.88
4
Learning to DeferPathMNIST Specialist
Accuracy87.74
2
Learning to DeferPathMNIST Clean
Accuracy87.31
2
Showing 25 of 26 rows