Image Classification on MNIST 70% symmetric label noise (test)
91.61AccuracyBARE
Evaluation results
| Method | Links | |
|---|---|---|
| BAREAlgorithm=BAtch REweighting, Backbone=1-hidden layer MLP, Noise Level=70% symmetric2021.06 | 91.61 | |
| CLAlgorithm=Curriculum Learning, Backbone=1-hidden layer MLP, Noise Level=70% symmetric2021.06 | 88.28 | |
| COT+Algorithm=Co-Teaching+, Backbone=1-hidden layer MLP, Noise Level=70% symmetric2021.06 | 87.26 | |
| COTAlgorithm=Co-Teaching, Backbone=1-hidden layer MLP, Noise Level=70% symmetric2021.06 | 87.17 | |
| MRAlgorithm=Meta-Reweighting, Backbone=1-hidden layer MLP, Noise Level=70% symmetric2021.06 | 85.1 | |
| MNAlgorithm=Meta-Net, Backbone=1-hidden layer MLP, Noise Level=70% symmetric2021.06 | 65.52 | |
| CCEAlgorithm=Cross-Entropy Error, Backbone=1-hidden layer MLP, Noise Level=70% symmetric2021.06 | 61.19 |