Unsupervised Domain Adaptation on MNIST to USPS (test)
0.985AccuracyRWOT
Evaluation Results
| Method | Links | |
|---|---|---|
| RWOTArchitecture=DTN, Pre-trained=false2021.03 | 0.985 | |
| CSTArchitecture=DTN, Pre-trained=false2021.03 | 0.985 | |
| Self-Ensembling2017.05 | 0.983 | |
| SBADA-GANClassifier training space=Weighted combination2017.05 | 0.976 | |
| SBADA-GANClassifier training space=Source domain (Cs)2017.05 | 0.971 | |
| SBADA-GANClassifier training space=Target domain (Ct)2017.05 | 0.967 | |
| Target Only2017.05 | 0.965 | |
| Target onlyEvaluation Protocol=Target only2017.11 | 0.963 | |
| UNIT2017.03 | 0.9597 | |
| PixelDA2017.05 | 0.959 | |
| UNIT2017.05 | 0.959 | |
| CoGAN2017.03 | 0.9565 | |
| CyCADAAdaptation Level=pixel only, Backbone=LeNet2017.11 | 0.956 | |
| CyCADAAdaptation Level=pixel+feat, Backbone=LeNet2017.11 | 0.956 | |
| CDANArchitecture=DTN, Pre-trained=false2021.03 | 0.956 | |
| CyCADA2017.05 | 0.948 | |
| GenToAdapt2017.05 | 0.925 | |
| DRCN2017.05 | 0.918 | |
| DSN2017.05 | 0.913 | |
| CoGAN2017.05 | 0.912 | |
| COGAN2017.11 | 0.912 | |
| ADDA2017.05 | 0.894 | |
| ADDA2017.11 | 0.894 | |
| DANN2017.05 | 0.851 | |
| Source onlyEvaluation Protocol=Source only2017.11 | 0.822 | |
| CORAL2017.05 | 0.817 | |
| MMD2017.05 | 0.811 | |
| Source Only2017.05 | 0.789 |