Clean-label poisoning attacks on federated learning for IoT

Keyphrases

• Internet of Things 100%
• Federated Learning 100%
• Poisoning Attack 100%
• Clean Label 100%
• Attack Strategy 66%
• Clean-label Attack 66%
• Backdoor Attack 66%
• Adversary 33%
• Byzantine Empire 33%
• Security Privacy 33%
• Defense Method 33%
• High Probability 33%
• Industrial Internet of Things (IIoT) 33%
• Concealment 33%
• Global Model 33%
• Poisoning 33%
• Data Privacy 33%
• Internet of Things Applications 33%
• Federated Learning System 33%
• Edge Collaboration 33%
• Model Poisoning Attack 33%
• Edge Cloud 33%
• Small Perturbation 33%
• Attack Success Rate 33%
• Cosine Similarity 33%
• Active Attack 33%
• Data Security 33%
• Adversarial Loss 33%
• Training Loss 33%
• Peak Signal to Noise Ratio 33%

Computer Science

• Federated Learning 100%
• Internet-Of-Things 100%
• Backdoors 50%
• Experimental Result 25%
• Security and Privacy 25%
• Industrial Internet of Things 25%
• Cosine Similarity 25%
• Learning Framework 25%
• Data Privacy 25%
• Data Security 25%
• peak signal to noise ratio 25%
• Structural Similarity 25%
• Application Scenario 25%
• Successful Attack 25%