Malicious Client Detection on Fashion-MNIST
6Avg Malicious Clients DetectedLabel flipping
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Label flippingM (Number of malicious clients)=6, alpha=30%2024.05 | 6 | — | — | |
| PoisonGANM (Number of malicious clients)=6, alpha=30%2024.05 | 6 | — | — | |
| Heavy noiseM (Number of malicious clients)=6, alpha=30%2024.05 | 6 | — | — | |
| VagueGANM (Number of malicious clients)=6, alpha=30%2024.05 | 5.8 | — | — | |
| SAP noiseM (Number of malicious clients)=6, alpha=30%2024.05 | 5.7 | — | — | |
| Light noiseM (Number of malicious clients)=6, alpha=30%2024.05 | 5.6 | — | — | |
| Label flippingM (Number of malicious clients)=4, alpha=20%2024.05 | 4 | — | — | |
| PoisonGANM (Number of malicious clients)=4, alpha=20%2024.05 | 4 | — | — | |
| Heavy noiseM (Number of malicious clients)=4, alpha=20%2024.05 | 4 | — | — | |
| SAP noiseM (Number of malicious clients)=4, alpha=20%2024.05 | 3.9 | — | — | |
| VagueGANM (Number of malicious clients)=4, alpha=20%2024.05 | 3.9 | — | — | |
| Light noiseM (Number of malicious clients)=4, alpha=20%2024.05 | 3.8 | — | — | |
| Label flippingM (Number of malicious clients)=2, alpha=10%2024.05 | 2 | — | — | |
| PoisonGANM (Number of malicious clients)=2, alpha=10%2024.05 | 2 | — | — | |
| Heavy noiseM (Number of malicious clients)=2, alpha=10%2024.05 | 2 | — | — | |
| SAP noiseM (Number of malicious clients)=2, alpha=10%2024.05 | 2 | — | — | |
| VagueGANM (Number of malicious clients)=2, alpha=10%2024.05 | 2 | — | — | |
| Light noiseM (Number of malicious clients)=2, alpha=10%2024.05 | 1.9 | — | — | |
| Label flippingM (Number of malicious clients)=1, alpha=5%2024.05 | 1 | — | — | |
| PoisonGANM (Number of malicious clients)=1, alpha=5%2024.05 | 1 | — | — | |
| Light noiseM (Number of malicious clients)=1, alpha=5%2024.05 | 1 | — | — | |
| Heavy noiseM (Number of malicious clients)=1, alpha=5%2024.05 | 1 | — | — | |
| SAP noiseM (Number of malicious clients)=1, alpha=5%2024.05 | 1 | — | — | |
| VagueGANM (Number of malicious clients)=1, alpha=5%2024.05 | 1 | — | — | |
| FLANDERSAttack Strategy=GAUSS, Malicious client ratio (r)=0.8, Training rounds (T)=502023.03 | — | 100 | 100 | |
| FLANDERSAttack Strategy=LIE, Malicious client ratio (r)=0.8, Training rounds (T)=502023.03 | — | 100 | 100 | |
| FLANDERSAttack Strategy=OPT, Malicious client ratio (r)=0.8, Training rounds (T)=502023.03 | — | 100 | 100 | |
| FLANDERSAttack Strategy=AGR-MM, Malicious client ratio (r)=0.8, Training rounds (T)=502023.03 | — | 100 | 100 | |
| FLANDERSAttack type=GAUSS, Attack ratio (r)=0.6, Number of rounds (T)=502023.03 | — | 1 | 1 | |
| FLANDERSAttack type=LIE, Attack ratio (r)=0.6, Number of rounds (T)=502023.03 | — | 1 | 1 | |
| FLANDERSAttack type=OPT, Attack ratio (r)=0.6, Number of rounds (T)=502023.03 | — | 1 | 1 | |
| FLANDERSAttack type=AGR-MM, Attack ratio (r)=0.6, Number of rounds (T)=502023.03 | — | 1 | 1 | |
| FLDetectorAttack Strategy=GAUSS, Malicious client ratio (r)=0.8, Training rounds (T)=502023.03 | — | 80 | 80 | |
| FLDetectorAttack Strategy=LIE, Malicious client ratio (r)=0.8, Training rounds (T)=502023.03 | — | 80 | 80 | |
| FLDetectorAttack Strategy=OPT, Malicious client ratio (r)=0.8, Training rounds (T)=502023.03 | — | 79 | 79 | |
| FLDetectorAttack Strategy=AGR-MM, Malicious client ratio (r)=0.8, Training rounds (T)=502023.03 | — | 80 | 80 | |
| FLDetectorAttack type=GAUSS, Attack ratio (r)=0.6, Number of rounds (T)=502023.03 | — | 0.6 | 0.6 | |
| FLDetectorAttack type=LIE, Attack ratio (r)=0.6, Number of rounds (T)=502023.03 | — | 0.59 | 0.59 | |
| FLDetectorAttack type=OPT, Attack ratio (r)=0.6, Number of rounds (T)=502023.03 | — | 0.6 | 0.6 | |
| FLDetectorAttack type=AGR-MM, Attack ratio (r)=0.6, Number of rounds (T)=502023.03 | — | 0.61 | 0.61 |