List your company
ResearchFormerly WizWand

Image Classification on CIFAR (accuracy)

91.77AccuracyState 0 (uninfected)
55.765265.112674.4683.8074Mar 10, 2022Nov 1, 2022Jun 26, 2023Feb 18, 2024Oct 12, 2024Jun 6, 2025Jan 29, 2026
Updated 6mo ago

Evaluation results

MethodLinks
State 0 (uninfected)Model=Inception, Learning Mode=std2024.0991.771.75
State 0 (uninfected)Model=ResNet, Learning Mode=std2024.0991.711.94
State 1 (infected)Model=ResNet, Learning Mode=std2024.0991.2699.96
State 1 (infected)Model=Inception, Learning Mode=std2024.0990.9199.93
State 0 (uninfected)Model=VGG, Learning Mode=std2024.0990.850.96
State 1 (infected)Model=VGG, Learning Mode=std2024.0990.2899.93
Fine-TuningModel=VGG, Learning Mode=std2024.0984.65100
TaborModel=VGG, Learning Mode=std2024.0984.5885.07
Psycho-PassModel=VGG, Learning Mode=std2024.0983.891.21
State 0 (uninfected)Model=Inception, Learning Mode=adv2024.0983.60.21
Neural CleanseModel=VGG, Learning Mode=std2024.0983.3724.43
Pixel BackdoorModel=VGG, Learning Mode=std2024.0983.325.49
State 1 (infected)Model=Inception, Learning Mode=adv2024.0983.28100
State 0 (uninfected)Model=ResNet, Learning Mode=adv2024.0983.091.62
KDEPData=100%, Epoch=90, Time (/h)=40, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0382.89—
KDEPData=100%, Epoch=90, Time (/h)=43, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0382.73—
KDEPData=10%, Epoch=900, Time (/h)=40, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0382.64—
KDEPData=10%, Epoch=900, Time (/h)=43, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0382.53—
KDEPData=100%, Epoch=18, Time (/h)=8, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0382.47—
KDEPData=10%, Epoch=180, Time (/h)=8, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0382.23—
KDEPData=100%, Epoch=9, Time (/h)=4, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0382.22—
KDEPData=100%, Epoch=18, Time (/h)=8.6, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0382.15—
SP. o.Data=100%, Epoch=90, Time (/h)=39, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0382.08—
KDEPData=10%, Epoch=180, Time (/h)=8.6, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0381.98—
SP. o.Data=100%, Epoch=90, Time (/h)=42, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0381.91—
State 1 (infected)Model=ResNet, Learning Mode=adv2024.0981.63100
KDEPData=10%, Epoch=90, Time (/h)=4, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0381.61—
KDEPData=100%, Epoch=9, Time (/h)=4.3, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0381.47—
SP. b.Data=100%, Epoch=18, Time (/h)=7.8, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0381.43—
KDEPData=10%, Epoch=90, Time (/h)=4.3, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0381.06—
SP. b.Data=100%, Epoch=9, Time (/h)=3.9, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0380.27—
SP. b.Data=100%, Epoch=18, Time (/h)=8.4, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0380.13—
UnicornModel=VGG, Learning Mode=std2024.0979.8981.51
SP. b.Data=10%, Epoch=900, Time (/h)=42, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0379.83—
TaborModel=ResNet, Learning Mode=std2024.0979.8180.69
SP. b.Data=10%, Epoch=180, Time (/h)=8.4, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0379.5—
Fine-TuningModel=Inception, Learning Mode=std2024.0979.5100
SP. b.Data=10%, Epoch=900, Time (/h)=39, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0379.49—
TaborModel=Inception, Learning Mode=std2024.0979.3899.62
SP. b.Data=10%, Epoch=180, Time (/h)=7.8, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0379.34—
State 1 (infected)Model=VGG, Learning Mode=adv2024.0979.3100
SP. b.Data=10%, Epoch=90, Time (/h)=3.9, Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0379.29—
SP. b.Data=100%, Epoch=9, Time (/h)=4.2, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0378.95—
Psycho-PassModel=ResNet, Learning Mode=std2024.0978.9420.37
State 0 (uninfected)Model=VGG, Learning Mode=adv2024.0978.841.04
SP. b.Data=10%, Epoch=90, Time (/h)=4.2, Fine-tuning=true, Pre-trained Backbone=ResNet-50, Student Architecture=MobileNetV22022.0378.6—
Fine-TuningModel=ResNet, Learning Mode=std2024.0978.5673.68
Pixel BackdoorModel=Inception, Learning Mode=std2024.0977.6188.31
rand. init.Teacher Backbone=ResNet-50, Student Backbone=ResNet-18, Evaluation Protocol=fine-tuned2022.0377.34—
Pixel BackdoorModel=ResNet, Learning Mode=std2024.0977.2833.59
rand. init.Fine-tuning=true, Pre-trained Backbone=None, Student Architecture=MobileNetV22022.0376.66—
UnicornModel=Inception, Learning Mode=std2024.0976.4650.66
Psycho-PassModel=Inception, Learning Mode=std2024.0976.157.09
Neural CleanseModel=Inception, Learning Mode=std2024.0976.0495.99
Fine-TuningModel=Inception, Learning Mode=adv2024.0975.13100
Neural CleanseModel=ResNet, Learning Mode=std2024.0974.2774.48
Psycho-PassModel=Inception, Learning Mode=adv2024.0973.955.14
UnicornModel=ResNet, Learning Mode=std2024.0973.4350.4
TaborModel=Inception, Learning Mode=adv2024.0973.16100
Neural CleanseModel=Inception, Learning Mode=adv2024.0972.31100
UnicornModel=Inception, Learning Mode=adv2024.0969.05100
Pixel BackdoorModel=Inception, Learning Mode=adv2024.0968.919.1
CTL-MST (Optimization)T=200, B=20002026.0168.6—
CTL-MST (Feature)T=200, B=20002026.0167.9—
IndivT=200, B=20002026.0167.7—
CTL-RandTreeT=200, B=20002026.0167.3—
CTL-MST (Target)T=200, B=20002026.0167.1—
StarT=200, B=20002026.0166.7—
CTL-MST (Optimization)T=200, B=5002026.0166.6—
CTL-MST (Feature)T=200, B=5002026.0166—
CTL-RandTreeT=200, B=5002026.0165.2—
CTL-MST (Target)T=200, B=5002026.0164.8—
Fine-TuningModel=VGG, Learning Mode=adv2024.0964.7299.91
Fine-TuningModel=ResNet, Learning Mode=adv2024.0964.2297.82
StarT=200, B=5002026.0164—
Psycho-PassModel=VGG, Learning Mode=adv2024.0963.043.53
Psycho-PassModel=ResNet, Learning Mode=adv2024.0963.029.24
TaborModel=ResNet, Learning Mode=adv2024.0962.0495.95
TaborModel=VGG, Learning Mode=adv2024.0961.6391.45
Neural CleanseModel=ResNet, Learning Mode=adv2024.0961.5859.37
IndivT=200, B=5002026.0160.9—
Neural CleanseModel=VGG, Learning Mode=adv2024.0960.4483.28
Pixel BackdoorModel=VGG, Learning Mode=adv2024.0958.999.19
Pixel BackdoorModel=ResNet, Learning Mode=adv2024.0958.0959.9
UnicornModel=ResNet, Learning Mode=adv2024.0957.9799.33
UnicornModel=VGG, Learning Mode=adv2024.0957.1510.02