List your company
ResearchFormerly WizWand

Label Noise Identification on MNIST (train)

0.998AUCacci
0.224240.425120.6260.82688Oct 26, 2022Oct 29, 2022Nov 1, 2022Nov 4, 2022Nov 7, 2022Nov 10, 2022Nov 13, 2022
Updated 8mo ago

Evaluation results

MethodLinks
acciScoring Metric=First-Split Cumulative Accuracy2022.100.998
accfScoring Metric=Second-Split Cumulative Accuracy2022.100.998
JointScoring Metric=Combined learning and forgetting ranks2022.100.998
ssftScoring Metric=Second-Split Forgetting Time2022.100.997
fsltScoring Metric=First-Split Learning Time2022.100.973
confiScoring Metric=Confidence2022.100.965
TMC-ShapleyClassifier=Logistic Regression2022.110.933
CS-SHAPLEYClassifier=Logistic Regression2022.110.877
Beta ShapleyClassifier=Logistic Regression2022.110.845
TMC-ShapleyClassifier=SVM-RBF2022.110.747
CS-SHAPLEYClassifier=SVM-RBF2022.110.674
Beta ShapleyClassifier=SVM-RBF2022.110.51
nfScoring Metric=Number of Forgetting Events2022.100.377
Leave-One-OutClassifier=Logistic Regression2022.110.371
Leave-One-OutClassifier=SVM-RBF2022.110.254