List your company
ResearchFormerly WizWand

Hyper-data Cleaning on MNIST (test)

0.9316Test AccuracyVPBGD(+EHG)
0.734520.7856850.836850.888015Feb 11, 2026Feb 12, 2026Feb 14, 2026Feb 15, 2026Feb 17, 2026Feb 18, 2026Feb 20, 2026
Updated 8mo ago

Evaluation results

MethodLinks
VPBGD(+EHG)Architecture=Multilayer Perceptron (MLP)2026.020.9316—90.79
VPBGDArchitecture=Multilayer Perceptron (MLP)2026.020.9181—90.18
NHGD2026.020.9180.2765—
TTSA2026.020.91720.235—
Neumann40iterations=402026.020.9170.2244—
Neumann10iterations=102026.020.91680.2627—
CG10iterations=102026.020.91270.188—
OEHGArchitecture=Multilayer Perceptron (MLP)2026.020.9124—90.89
AmIGO2026.020.91010.1925—
CG40iterations=402026.020.90670.1488—
stocBiO2026.020.9060.4253—
OEHG-Weight TrainingArchitecture=Multilayer Perceptron (MLP)2026.020.9052——
IPATT-GM(+EHG)Architecture=Multilayer Perceptron (MLP)2026.020.901—91.09
T-RHG-Weight TrainingArchitecture=Multilayer Perceptron (MLP)2026.020.8962——
RHG-Weight TrainingArchitecture=Multilayer Perceptron (MLP)2026.020.8952——
IPATT-GMArchitecture=Multilayer Perceptron (MLP)2026.020.8947—90.83
SOBA2026.020.89240.1055—
OEHGArchitecture=Softmax Regression (SR)2026.020.892—89.94
OEHG-Weight TrainingArchitecture=Softmax Regression (SR)2026.020.8905——
IPATT-GM(+EHG)Architecture=Softmax Regression (SR)2026.020.8868—90.78
IPATT-GMArchitecture=Softmax Regression (SR)2026.020.8847—90.73
VPBGD(+EHG)Architecture=Softmax Regression (SR)2026.020.8759—89.86
T-RHGArchitecture=Multilayer Perceptron (MLP)2026.020.8596—88.84
RHGArchitecture=Multilayer Perceptron (MLP)2026.020.8591—89.3
RHG-Weight TrainingArchitecture=Softmax Regression (SR)2026.020.8514——
VPBGDArchitecture=Softmax Regression (SR)2026.020.8512—88.75
T-RHG-Weight TrainingArchitecture=Softmax Regression (SR)2026.020.8505——
RHGArchitecture=Softmax Regression (SR)2026.020.8483—88.45
T-RHGArchitecture=Softmax Regression (SR)2026.020.8481—88.11
Dirty TrainingArchitecture=Softmax Regression (SR)2026.020.8152——
Dirty TrainingArchitecture=Multilayer Perceptron (MLP)2026.020.7421——