ResearchBenchmarksImage Classification on CIFAR-100 (test) (Symmetric Noise Accuracy 20/50/80/90%)Image Classification on CIFAR-100 (test) (Symmetric Noise Accuracy 20/50/80/90%)Follow77.3Accuracy (Symmetric 20%)DivideMix56.561.967.372.7Mar 13, 2026Evaluation resultsMethodMethodLinksAccuracy (Symmetric 20%)Accuracy (Symmetric 50%)Accuracy (Symmetric 80%)Accuracy (Symmetric 90%)DivideMix2026.0377.374.660.231.5LossModel2026.0373.966.148.224.3JEC2026.0373.164.950.116.2PENCIL2026.0369.457.531.115.3Coteaching2026.0365.651.827.913.7CE2026.036246.719.910.1Forward2026.0361.546.619.910.2FEC2026.0360.558.652.744.1Mixup2026.0357.330.814.6—