Image Classification on SVHN (test) (Accuracy and Attack Success Rate)
91.03AccuracyUnMixMatch
Evaluation results
| Method | Links | ||
|---|---|---|---|
| UnMixMatchLabeled Samples=10002023.06 | 91.03 | — | |
| Handcrafted Backdoor AttackNetwork=CNN, Trigger Pattern=Checkerboard, Pre-trained Model Accuracy=89%2021.06 | 91 | 99 | |
| Traditional Poisoning AttackNetwork=CNN, Trigger Pattern=Watermark, Pre-trained Model Accuracy=89%2021.06 | 91 | 100 | |
| Handcrafted Backdoor AttackNetwork=CNN, Trigger Pattern=Watermark, Pre-trained Model Accuracy=89%2021.06 | 91 | 100 | |
| FuseFLM=5, K=22024.10 | 90.87 | — | |
| MixMatchLabeled Samples=10002023.06 | 90.57 | — | |
| LLPFC-uniformArchitecture=Wide ResNet-16-4, Bag Size=322022.03 | 90.12 | — | |
| UDALabeled Samples=10002023.06 | 89.95 | — | |
| FuseFLM=5, K=82024.10 | 89.94 | — | |
| FuseFLM=5, K=42024.10 | 89.34 | — | |
| FuseFLM=10, K=42024.10 | 89.31 | — | |
| ScMatchLabeled Samples=10002023.06 | 89.23 | — | |
| LLPFC-approxArchitecture=Wide ResNet-16-4, Bag Size=322022.03 | 89.03 | — | |
| Traditional Poisoning AttackNetwork=CNN, Trigger Pattern=Checkerboard, Pre-trained Model Accuracy=89%2021.06 | 89 | 98 | |
| Handcrafted Backdoor AttackNetwork=CNN, Trigger Pattern=Random, Pre-trained Model Accuracy=89%2021.06 | 89 | 98 | |
| Pseudo-labelLabeled Samples=10002023.06 | 88.91 | — | |
| FixMatchLabeled Samples=10002023.06 | 88.85 | — | |
| FlexMatchLabeled Samples=10002023.06 | 88.85 | — | |
| CCSSLLabeled Samples=10002023.06 | 88.6 | — | |
| SimMatchLabeled Samples=10002023.06 | 88.57 | — | |
| LLPFC-uniformArchitecture=Wide ResNet-16-4, Bag Size=642022.03 | 88.55 | — | |
| FuseFLM=10, K=22024.10 | 88.52 | — | |
| LLPFC-approxArchitecture=Wide ResNet-16-4, Bag Size=642022.03 | 88.44 | — | |
| LLPFC-approxBackbone=ResNet18, Bag Size=322022.03 | 88.24 | — | |
| LLPFC-uniformBackbone=ResNet18, Bag Size=322022.03 | 88.23 | — | |
| LLPFC-approxArchitecture=Wide ResNet-16-4, Bag Size=1282022.03 | 88.15 | — | |
| LLPFC-approxArchitecture=Wide ResNet-16-4, Bag Size=2562022.03 | 88.08 | — | |
| VATLabeled Samples=10002023.06 | 88.07 | — | |
| ReMixMatchLabeled Samples=10002023.06 | 88.06 | — | |
| LLPFC-uniformBag Size=32, Architecture=LLPGAN architecture2022.03 | 88 | — | |
| Pi-ModelLabeled Samples=10002023.06 | 87.83 | — | |
| LLPFC-approxBag Size=32, Architecture=LLPGAN architecture2022.03 | 87.79 | — | |
| SupervisedLabeled Samples=10002023.06 | 87.74 | — | |
| LLPFC-approxArchitecture=Wide ResNet-16-4, Bag Size=5122022.03 | 87.71 | — | |
| Mean TeacherLabeled Samples=10002023.06 | 87.68 | — | |
| LLPFC-uniformArchitecture=Wide ResNet-16-4, Bag Size=1282022.03 | 87.6 | — | |
| LLPFC-approxArchitecture=Wide ResNet-16-4, Bag Size=20482022.03 | 87.38 | — | |
| LLPFC-uniformArchitecture=Wide ResNet-16-4, Bag Size=2562022.03 | 87.36 | — | |
| LLPFC-uniformArchitecture=Wide ResNet-16-4, Bag Size=20482022.03 | 87.17 | — | |
| LLPFC-uniformArchitecture=Wide ResNet-16-4, Bag Size=10242022.03 | 87.09 | — | |
| LLPFC-approxArchitecture=Wide ResNet-16-4, Bag Size=10242022.03 | 87.01 | — | |
| FuseFLM=20, K=42024.10 | 86.94 | — | |
| LLPFC-uniformArchitecture=Wide ResNet-16-4, Bag Size=5122022.03 | 86.81 | — | |
| LLPFC-approxBackbone=ResNet18, Bag Size=642022.03 | 86.72 | — | |
| CoMatchLabeled Samples=10002023.06 | 86.45 | — | |
| LLPFC-uniformBackbone=ResNet18, Bag Size=642022.03 | 86.44 | — | |
| LLPFC-uniformBag Size=64, Architecture=LLPGAN architecture2022.03 | 85.81 | — | |
| LLPFC-approxBackbone=ResNet18, Bag Size=1282022.03 | 85.7 | — | |
| EnsembleM=52024.10 | 85.7 | — | |
| LLPFC-approxBackbone=ResNet18, Bag Size=20482022.03 | 85.34 | — | |
| LLPFC-approxBag Size=2048, Architecture=LLPGAN architecture2022.03 | 85.27 | — | |
| LLPFC-approxBag Size=64, Architecture=LLPGAN architecture2022.03 | 85.19 | — | |
| FuseFLM=20, K=22024.10 | 85.18 | — | |
| LLPFC-approxBackbone=ResNet18, Bag Size=10242022.03 | 84.98 | — | |
| LLPFC-approxBackbone=ResNet18, Bag Size=5122022.03 | 84.92 | — | |
| LLPFC-approxBackbone=ResNet18, Bag Size=2562022.03 | 84.83 | — | |
| LLPFC-uniformBag Size=128, Architecture=LLPGAN architecture2022.03 | 84.8 | — | |
| LLPFC-approxBag Size=256, Architecture=LLPGAN architecture2022.03 | 84.53 | — | |
| LLPFC-uniformBackbone=ResNet18, Bag Size=1282022.03 | 84.33 | — | |
| LLPFC-approxBag Size=512, Architecture=LLPGAN architecture2022.03 | 84.23 | — | |
| LLPFC-uniformBag Size=256, Architecture=LLPGAN architecture2022.03 | 83.93 | — | |
| LLPFC-uniformBackbone=ResNet18, Bag Size=2562022.03 | 83.9 | — | |
| LLPFC-approxBag Size=1024, Architecture=LLPGAN architecture2022.03 | 83.86 | — | |
| LLPFC-uniformBackbone=ResNet18, Bag Size=5122022.03 | 83.6 | — | |
| LLPFC-uniformBag Size=512, Architecture=LLPGAN architecture2022.03 | 83.47 | — | |
| LLPFC-uniformBag Size=2048, Architecture=LLPGAN architecture2022.03 | 83.27 | — | |
| Traditional Poisoning AttackNetwork=FC, Trigger Pattern=Checkerboard, Pre-trained Model Accuracy=81%2021.06 | 83 | 100 | |
| Traditional Poisoning AttackNetwork=CNN, Trigger Pattern=Random, Pre-trained Model Accuracy=89%2021.06 | 83 | 99 | |
| LLPFC-uniformBag Size=1024, Architecture=LLPGAN architecture2022.03 | 82.58 | — | |
| LLPFC-uniformBackbone=ResNet18, Bag Size=20482022.03 | 81.88 | — | |
| Handcrafted Backdoor AttackNetwork=FC, Trigger Pattern=Square, Pre-trained Model Accuracy=81%2021.06 | 81 | 96 | |
| Handcrafted Backdoor AttackNetwork=FC, Trigger Pattern=Checkerboard, Pre-trained Model Accuracy=81%2021.06 | 81 | 100 | |
| LLPFC-uniformBackbone=ResNet18, Bag Size=10242022.03 | 80.86 | — | |
| UnMixMatchLabeled Samples=2502023.06 | 80.78 | — | |
| CCSSLLabeled Samples=2502023.06 | 80.39 | — | |
| DENSEM=52024.10 | 80.03 | — | |
| FlexMatchLabeled Samples=2502023.06 | 79.72 | — | |
| ReMixMatchLabeled Samples=2502023.06 | 79.57 | — | |
| Fed-ADIM=52024.10 | 78.85 | — | |
| Pseudo-labelLabeled Samples=2502023.06 | 78.04 | — | |
| Fed-DAFLM=52024.10 | 78.03 | — | |
| UDALabeled Samples=2502023.06 | 77.73 | — | |
| MixMatchLabeled Samples=2502023.06 | 77.65 | — | |
| Mean TeacherLabeled Samples=2502023.06 | 76.78 | — | |
| Pi-ModelLabeled Samples=2502023.06 | 76.56 | — | |
| LLPGANBag Size=128, Architecture=LLPGAN architecture2022.03 | 76.3 | — | |
| VATLabeled Samples=2502023.06 | 76.18 | — | |
| FuseFLalpha=0.05, K=42024.10 | 75.53 | — | |
| SupervisedLabeled Samples=2502023.06 | 75.42 | — | |
| ScMatchLabeled Samples=2502023.06 | 75.08 | — | |
| FuseFLalpha=0.05, K=22024.10 | 75.08 | — | |
| FixMatchLabeled Samples=2502023.06 | 74.52 | — | |
| SimMatchLabeled Samples=2502023.06 | 74.48 | — | |
| Traditional Poisoning AttackNetwork=FC, Trigger Pattern=Square, Pre-trained Model Accuracy=81%2021.06 | 74 | 93 | |
| FedDFM=52024.10 | 73.98 | — | |
| EnsembleM=102024.10 | 73.45 | — | |
| CoMatchLabeled Samples=2502023.06 | 73.19 | — | |
| UnMixMatchLabeled Samples=402023.06 | 72.9 | — | |
| FuseFLM=10, K=82024.10 | 72.65 | — | |
| FuseFLM=50, K=22024.10 | 72.25 | — |