Image Classification on CIFAR-100 (val) (20-80% Noise Robustness)
71.3Accuracy (40% noise)MentorMix
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| MentorMixBackbone=ResNet-322019.11 | 71.3 | 78.6 | 64.6 | 41.2 | |
| MD-DYR-SHArchitecture=PRN-182019.04 | 70.1 | 73.7 | 59.5 | 39.5 | |
| Arazo et al.2019.11 | 70.1 | 73.7 | 59.5 | 39.5 | |
| Huang et al.2019.11 | 69.2 | 74.1 | 39.4 | — | |
| MentorNet2019.11 | 68.5 | 73.5 | 61.2 | 35.5 | |
| Jiang et al. (2018b)Architecture=WRN-1012019.04 | 68 | 73 | — | 35 | |
| Mixup2019.11 | 66.8 | 73.9 | 58.8 | 40.1 | |
| M-DYR-HArchitecture=PRN-182019.04 | 64.4 | 70 | 58.1 | 45.5 | |
| Zhang & Sabuncu2019.11 | 62.6 | 67.6 | 54 | 29.6 | |
| Ren et al. (2018)Architecture=WRN-282019.04 | 61.3 | — | — | — | |
| Ma et al. (2018)Architecture=RN-442019.04 | 52 | 62.2 | 42.3 | — |