Co-teaching (Inception-ResNet-v2) | 61.48 | 84.70 | 63.58 | 85.20 | Co-teaching: Robust Training of Deep Neural Networks with Extremely Noisy Labels | |
DivideMix with C2D (ResNet-50) | 78.57 ± 0.37 | 93.04 ± 0.10 | 79.42 ± 0.34 | 92.32 ± 0.33 | Contrast to Divide: Self-Supervised Pre-Training for Learning with Noisy Labels | |
RTE (Inception-ResNet-v2) | 80.84 | 97.24 | - | - | Robust Temporal Ensembling for Learning with Noisy Labels | - |
Dynamic Loss (Inception-ResNet-v2) | 74.76 | 93.08 | 80.12 | 93.64 | Dynamic Loss For Robust Learning | |
LongReMix (Inception-ResNet-v2) | - | - | 78.92 | 92.32 | LongReMix: Robust Learning with High Confidence Samples in a Noisy Label Environment | |
NCT (Inception-ResNet-v2) | 71.73 | 91.61 | 75.16 | 90.77 | Noisy Concurrent Training for Efficient Learning under Label Noise | |
ROLT+ (Inception-ResNet-v2) | 74.64 | 92.48 | 77.64 | 92.44 | Robust Long-Tailed Learning under Label Noise | - |
D2L (Inception-ResNet-v2) | 57.80 | 81.36 | 62.68 | 84.00 | Dimensionality-Driven Learning with Noisy Labels | |
PGDF (Inception-ResNet-v2) | 75.45 | 93.11 | 81.47 | 94.03 | Sample Prior Guided Robust Model Learning to Suppress Noisy Labels | |
DivideMix (ResNet-50) | 74.42 ±0.29 | 91.21 ±0.12 | 76.32 ±0.36 | 90.65 ±0.16 | DivideMix: Learning with Noisy Labels as Semi-supervised Learning | |
F-Correction (Inception-ResNet-v2) | 57.36 | 82.36 | 61.12 | 82.68 | Making Deep Neural Networks Robust to Label Noise: a Loss Correction Approach | |
Crust (Inception-ResNet-v2) | 67.36 | 87.84 | 72.40 | 89.56 | Coresets for Robust Training of Neural Networks against Noisy Labels | - |
ODD (Inception-ResNet-v2) | 66.7 | 86.3 | 74.6 | 90.6 | Robust and On-the-fly Dataset Denoising for Image Classification | - |