List your company

Unsupervised Domain Adaptation on USPS -> MNIST (test)

99.5AccuracySelf-Ensembling
55.40466.85278.389.748Mar 2, 2017Oct 31, 2017Jul 1, 2018Mar 2, 2019Oct 31, 2019Jun 30, 2020Mar 1, 2021

Evaluation results

MethodLinks
Self-Ensembling2017.0599.5
Target Only2017.0599.2
Target onlyEvaluation Protocol=Target only2017.1199.2
CyCADAAdaptation Level=pixel+feat, Backbone=LeNet2017.1196.5
CyCADAAdaptation Level=pixel only, Backbone=LeNet2017.1196.4
Inference-time Label-Preserving Target ProjectionsLabeled target samples (T)=102021.0396.1
CCSALabeled target samples (T)=102021.0395.71
CyCADA2017.0595.7
SBADA-GANClassifier training space=Weighted combination2017.0595
SBADA-GANClassifier training space=Target domain (Ct)2017.0594.4
Inference-time Label-Preserving Target ProjectionsLabeled target samples (T)=72021.0393.81
M-ADALabeled target samples (T)=102021.0393.67
UNIT2017.0393.58
UNIT2017.0593.5
CoGAN2017.0393.15
M-ADALabeled target samples (T)=72021.0392.33
FADALabeled target samples (T)=72021.0391.5
GenToAdapt2017.0590.8
ADDA2017.0590.1
ADDA2017.1190.1
CoGAN2017.0589.1
COGAN2017.1189.1
SBADA-GANClassifier training space=Source domain (Cs)2017.0587.5
Inference-time Label-Preserving Target ProjectionsLabeled target samples (T)=02021.0374.52
DRCN2017.0573.7
DANN2017.0573
M-ADALabeled target samples (T)=02021.0371.19
Source onlyEvaluation Protocol=Source only2017.1169.6
CORAL2017.0557.7
Source Only2017.0557.1