List your company
ResearchFormerly WizWand

Federated Learning Fairness on Fashion MNIST (test)

1.005Accuracy VarianceDFedReweighting
-16.141299.59565215.3325331.06935Dec 12, 2025
Updated 8mo ago

Evaluation results

MethodLinks
DFedReweightingData Heterogeneity (DH)=IID, T=0.012025.121.00582.46
q-FDFLData Heterogeneity (DH)=IID, q=0.012025.121.15181.757
q-FDFLData Heterogeneity (DH)=IID, q=0.52025.121.15383.947
DFedReweightingData Heterogeneity (DH)=IID, T=0.52025.121.19681.653
q-FDFLData Heterogeneity (DH)=IID, q=0.12025.121.23783.04
DFedReweightingData Heterogeneity (DH)=IID, T=0.12025.121.24382.147
DFedAvgData Heterogeneity (DH)=IID2025.121.45581.597
DFedReweightingData Heterogeneity (DH)=Diri(1.0), T=0.12025.1210.45285.549
DFedReweightingData Heterogeneity (DH)=Diri(1.0), T=0.52025.1211.97683.995
DFedReweightingData Heterogeneity (DH)=Diri(0.1), T=0.12025.1213.42295.39
DFedReweightingData Heterogeneity (DH)=Diri(1.0), T=0.012025.1214.25384.003
DFedReweightingData Heterogeneity (DH)=LabelSkew(4), T=0.12025.1215.87692.182
DFedReweightingData Heterogeneity (DH)=Diri(0.1), T=0.012025.1216.65795.881
DFedReweightingData Heterogeneity (DH)=LabelSkew(4), T=0.012025.1217.58991.773
q-FDFLData Heterogeneity (DH)=Diri(1.0), q=0.012025.1218.01282.923
q-FDFLData Heterogeneity (DH)=Diri(1.0), q=0.12025.1218.0283.204
q-FDFLData Heterogeneity (DH)=Diri(1.0), q=0.52025.1220.12583.686
DFedAvgData Heterogeneity (DH)=Diri(1.0)2025.1222.6183.172
DFedReweightingData Heterogeneity (DH)=LabelSkew(4), T=0.52025.1235.30289.697
q-FDFLData Heterogeneity (DH)=LabelSkew(4), q=0.012025.12147.57572.308
DFedReweightingData Heterogeneity (DH)=Diri(0.1), T=0.52025.12161.93386.862
DFedAvgData Heterogeneity (DH)=LabelSkew(4)2025.12184.04375.315
q-FDFLData Heterogeneity (DH)=LabelSkew(4), q=0.12025.12199.82873.005
q-FDFLData Heterogeneity (DH)=LabelSkew(4), q=0.52025.12215.89174.968
q-FDFLData Heterogeneity (DH)=Diri(0.1), q=0.12025.12335.06676.112
DFedAvgData Heterogeneity (DH)=Diri(0.1)2025.12360.77276.972
q-FDFLData Heterogeneity (DH)=Diri(0.1), q=0.52025.12414.97474.97
q-FDFLData Heterogeneity (DH)=Diri(0.1), q=0.012025.12429.6676.493