Image Classification on Fashion MNIST (test)
97.9AccuracySupervised
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| SupervisedClass Prior=0.22025.08 | 97.9 | — | — | — | — | — | — | — | — | |
| SupervisedClass Prior=0.52025.08 | 97.6 | — | — | — | — | — | — | — | — | |
| SconfConfDiff-ABSClass Prior=0.22025.08 | 96.9 | — | — | — | — | — | — | — | — | |
| ConfDiff-ABSClass Prior=0.52025.08 | 96.8 | — | — | — | — | — | — | — | — | |
| ConfDiff-ABSClass Prior=0.22025.08 | 96.7 | — | — | — | — | — | — | — | — | |
| Convex(γ = 0.2)-ABSClass Prior=0.22025.08 | 96.7 | — | — | — | — | — | — | — | — | |
| SconfConfDiff-ReLUClass Prior=0.22025.08 | 96.5 | — | — | — | — | — | — | — | — | |
| Convex(γ = 0.5)-ABSClass Prior=0.22025.08 | 96.5 | — | — | — | — | — | — | — | — | |
| Zhong et al.Training set size=Full2019.04 | 96.35 | — | — | — | — | — | — | — | — | |
| SconfConfDiff-UnbiasedClass Prior=0.52025.08 | 96.3 | — | — | — | — | — | — | — | — | |
| SconfConfDiff-ReLUClass Prior=0.52025.08 | 96.3 | — | — | — | — | — | — | — | — | |
| SconfConfDiff-ABSClass Prior=0.52025.08 | 96.3 | — | — | — | — | — | — | — | — | |
| Convex(γ = 0.8)-ABSClass Prior=0.22025.08 | 96.2 | — | — | — | — | — | — | — | — | |
| CCT-7/3x1# Params=3.76 M, MACs=1.19 G2021.04 | 95.56 | — | — | — | — | — | — | — | — | |
| Sconf-ABSClass Prior=0.22025.08 | 95.5 | — | — | — | — | — | — | — | — | |
| CCT-6/3x1# Params=3.23 M, MACs=1.02 G2021.04 | 95.41 | — | — | — | — | — | — | — | — | |
| CCT-6/3x2# Params=3.33 M, MACs=0.25 G2021.04 | 95.34 | — | — | — | — | — | — | — | — | |
| ResNet110# Params=1.73 M, MACs=0.26 G2021.04 | 95.32 | — | — | — | — | — | — | — | — | |
| CVT-7/4# Params=3.72 M, MACs=0.25 G2021.04 | 95.32 | — | — | — | — | — | — | — | — | |
| MobileNetV2/2.0# Params=91.02, MACs=0.02 G2021.04 | 95.26 | — | — | — | — | — | — | — | — | |
| ResNet56# Params=0.85 M, MACs=0.13 G2021.04 | 95.25 | — | — | — | — | — | — | — | — | |
| ViT-Lite-7/4# Params=3.72 M, MACs=0.26 G2021.04 | 95.16 | — | — | — | — | — | — | — | — | |
| CCT-7/3x2# Params=3.85 M, MACs=0.29 G2021.04 | 95.16 | — | — | — | — | — | — | — | — | |
| ViT-Lite-6/4# Params=3.19 M, MACs=0.22 G2021.04 | 95.14 | — | — | — | — | — | — | — | — | |
| MobileNetV2/1.25# Params=3.47 M, MACs=0.01 G2021.04 | 95.05 | — | — | — | — | — | — | — | — | |
| CVT-6/4# Params=3.19 M, MACs=0.22 G2021.04 | 95 | — | — | — | — | — | — | — | — | |
| ResNet50# Params=23.53 M, MACs=0.08 G2021.04 | 94.99 | — | — | — | — | — | — | — | — | |
| MobileNetV2/1.0# Params=2.21 M, MACs=0.01 G2021.04 | 94.85 | — | — | — | — | — | — | — | — | |
| ResNet18# Params=11.18 M, MACs=0.04 G2021.04 | 94.78 | — | — | — | — | — | — | — | — | |
| ResNet34# Params=21.29 M, MACs=0.08 G2021.04 | 94.78 | — | — | — | — | — | — | — | — | |
| CCT-4/3x2# Params=0.48 M, MACs=0.05 G2021.04 | 94.74 | — | — | — | — | — | — | — | — | |
| DCNet++epochs=100E2018.05 | 94.65 | — | — | — | — | — | — | — | — | |
| DCNetepochs=100E2018.05 | 94.64 | — | — | — | — | — | — | — | — | |
| DCNetepochs=100E, reconstruction_subnetwork=NR2018.05 | 94.59 | — | — | — | — | — | — | — | — | |
| CVT-6/8# Params=3.21 M, MACs=0.05 G2021.04 | 94.53 | — | — | — | — | — | — | — | — | |
| CVT-7/8# Params=3.74 M, MACs=0.06 G2021.04 | 94.5 | — | — | — | — | — | — | — | — | |
| ViT-Lite-7/8# Params=3.74 M, MACs=0.06 G2021.04 | 94.49 | — | — | — | — | — | — | — | — | |
| ViT-Lite-6/8# Params=3.22 M, MACs=0.06 G2021.04 | 94.36 | — | — | — | — | — | — | — | — | |
| LP-1STepsilon=infinity, delta=10^-52021.02 | 94.28 | — | — | — | — | — | — | — | — | |
| LP-2STepsilon=4, delta=10^-52021.02 | 94.1 | — | — | — | — | — | — | — | — | |
| CCT-2/3x2# Params=0.28 M, MACs=0.04 G2021.04 | 94.08 | — | — | — | — | — | — | — | — | |
| MobileNetV2/0.5# Params=0.70 M, MACs=< 0.01 G2021.04 | 93.93 | — | — | — | — | — | — | — | — | |
| WaveMix-128/7Number of parameters (Million)=2.42022.03 | 93.91 | — | — | — | — | — | — | — | — | |
| WaveMix-256/7Number of parameters (Million)=9.62022.03 | 93.78 | — | — | — | — | — | — | — | — | |
| Baseline CapsNetepochs=100E2018.05 | 93.65 | — | — | — | — | — | — | — | — | |
| ViT-12/16# Params=85.63 M, MACs=0.43 G2021.04 | 93.61 | — | — | — | — | — | — | — | — | |
| Classifier Baseline (No Attack)Classifier=B2019.03 | 93.54 | — | — | — | — | — | — | — | — | |
| LP-1STepsilon=4, delta=10^-52021.02 | 93.5 | — | — | — | — | — | — | — | — | |
| Whole Dataset2022.10 | 93.5 | — | — | — | — | — | — | — | — | |
| Classifier Baseline (No Attack)Classifier=A2019.03 | 93.46 | — | — | — | — | — | — | — | — | |
| ResNet-18Number of parameters (Million)=11.22022.03 | 93.35 | — | — | — | — | — | — | — | — | |
| ResNet-34Number of parameters (Million)=21.32022.03 | 93.34 | — | — | — | — | — | — | — | — | |
| ResNet-50Number of parameters (Million)=23.62022.03 | 93.3 | — | — | — | — | — | — | — | — | |
| ViT-Lite-7/16# Params=3.89 M, MACs=0.02 G2021.04 | 93.24 | — | — | — | — | — | — | — | — | |
| LP-2STepsilon=3, delta=10^-52021.02 | 93.18 | — | — | — | — | — | — | — | — | |
| ViT-Lite-6/16# Params=3.36 M, MACs=0.02 G2021.04 | 93.09 | — | — | — | — | — | — | — | — | |
| CNLCU-HNoise type=Asymmetric, Noise ratio=20%2021.06 | 92.6 | — | — | — | — | — | — | — | — | |
| CNLCU-SNoise type=Asymmetric, Noise ratio=20%2021.06 | 92.57 | — | — | — | — | — | — | — | — | |
| Bhatnagar et al.Training set size=Full2019.04 | 92.54 | — | — | — | — | — | — | — | — | |
| LP-1STepsilon=3, delta=10^-52021.02 | 92.52 | — | — | — | — | — | — | — | — | |
| CTM (DC)Drop Clause=true2021.05 | 92.5 | — | — | — | — | — | — | — | — | |
| CNLCU-HNoise type=Symmetric, Noise ratio=20%2021.06 | 92.42 | — | — | — | — | — | — | — | — | |
| SGDModel compression status=Raw2018.01 | 92.4 | — | — | — | — | — | — | — | — | |
| KIP ConvNetIPC=50, Augmentation=false2021.07 | 92.4 | — | — | — | — | — | — | — | — | |
| CNLCU-SNoise type=Symmetric, Noise ratio=20%2021.06 | 92.37 | — | — | — | — | — | — | — | — | |
| CNLCU-HNoise type=Tridiagonal, Noise ratio=20%2021.06 | 92.33 | — | — | — | — | — | — | — | — | |
| CNLCU-SNoise type=Tridiagonal, Noise ratio=20%2021.06 | 92.24 | — | — | — | — | — | — | — | — | |
| VanillaOptimizer=Adam, Nlinear in [%]=100, Nconv in [%]=1002021.05 | 92.1 | — | — | — | — | — | — | — | — | |
| VanillaOptimizer=Adam2021.05 | 92.1 | — | 100 | — | — | — | — | — | — | |
| CNLCU-SNoise type=Pairwise, Noise ratio=20%2021.06 | 92.04 | — | — | — | — | — | — | — | — | |
| JoCorNoise type=Tridiagonal, Noise ratio=20%2021.06 | 92.01 | — | — | — | — | — | — | — | — | |
| JoCorNoise type=Symmetric, Noise ratio=20%2021.06 | 91.97 | — | — | — | — | — | — | — | — | |
| HATModel compression status=Compressed2018.01 | 91.9 | 2.3 | — | — | — | — | — | — | — | |
| CARDArchitecture=LeNet2022.06 | 91.79 | — | — | — | — | — | — | — | — | |
| CNLCU-HNoise type=Pairwise, Noise ratio=20%2021.06 | 91.7 | — | — | — | — | — | — | — | — | |
| CNLCU-SNoise type=Instance-dependent, Noise ratio=20%2021.06 | 91.69 | — | — | — | — | — | — | — | — | |
| CNLCU-HNoise type=Symmetric, Noise ratio=40%2021.06 | 91.6 | — | — | — | — | — | — | — | — | |
| JoCorNoise type=Pairwise, Noise ratio=20%2021.06 | 91.52 | — | — | — | — | — | — | — | — | |
| CNLCU-HNoise type=Instance-dependent, Noise ratio=20%2021.06 | 91.5 | — | — | — | — | — | — | — | — | |
| CTMDrop Clause=false2021.05 | 91.5 | — | — | — | — | — | — | — | — | |
| Co-teachingNoise type=Symmetric, Noise ratio=20%2021.06 | 91.48 | — | — | — | — | — | — | — | — | |
| CNLCU-SNoise type=Symmetric, Noise ratio=40%2021.06 | 91.45 | — | — | — | — | — | — | — | — | |
| JoCorNoise type=Instance-dependent, Noise ratio=20%2021.06 | 91.43 | — | — | — | — | — | — | — | — | |
| Co-teachingNoise type=Tridiagonal, Noise ratio=20%2021.06 | 91.24 | — | — | — | — | — | — | — | — | |
| LP-2STepsilon=2, delta=10^-52021.02 | 91.24 | — | — | — | — | — | — | — | — | |
| IDGP2022.04 | 91.2 | — | — | — | — | — | — | — | — | |
| f (LeNet)Architecture=LeNet2022.06 | 91.12 | — | — | — | — | — | — | — | — | |
| CMV-MF-VIArchitecture=LeNet2022.06 | 91.1 | — | — | — | — | — | — | — | — | |
| Co-teachingNoise type=Asymmetric, Noise ratio=20%2021.06 | 91.03 | — | — | — | — | — | — | — | — | |
| KIP ConvNetIPC=10, Augmentation=false2021.07 | 91 | — | — | — | — | — | — | — | — | |
| KIP ConvNetIPC=50, Augmentation=true2021.07 | 91 | — | — | — | — | — | — | — | — | |
| JoCorNoise type=Asymmetric, Noise ratio=20%2021.06 | 90.95 | — | — | — | — | — | — | — | — | |
| CM-MF-VIArchitecture=LeNet2022.06 | 90.95 | — | — | — | — | — | — | — | — | |
| Co-teachingNoise type=Pairwise, Noise ratio=20%2021.06 | 90.77 | — | — | — | — | — | — | — | — | |
| CM-MF-VI OPTArchitecture=LeNet2022.06 | 90.67 | — | — | — | — | — | — | — | — | |
| Co-teachingNoise type=Instance-dependent, Noise ratio=20%2021.06 | 90.6 | — | — | — | — | — | — | — | — | |
| VALEN2022.04 | 90.58 | — | — | — | — | — | — | — | — | |
| AdaBregStrategy=Bregman, Optimizer=AdaBreg2021.05 | 90.5 | — | 2.3 | — | — | — | — | — | — | |
| MWT-LData Augmentation=None2025.03 | 90.43 | — | — | — | — | — | — | — | — | |
| MentorNetNoise type=Symmetric, Noise ratio=20%2021.06 | 90.37 | — | — | — | — | — | — | — | — |