Un-Fair Trojan: Targeted Backdoor Attacks Against Model Fairness

Keyphrases

• Backdoor Attack 100%
• Model Fairness 100%
• Machine Learning Models 66%
• Prediction Accuracy 33%
• Graph Structure 33%
• Adversary 33%
• Control Model 33%
• Complex Data 33%
• Global Model 33%
• Image Data 33%
• Model Accuracy 33%
• Behavioral Model 33%
• Local Model 33%
• Local Gradient 33%
• Federated Learning 33%
• Data Poisoning Attack 33%
• Backdoor Poisoning Attack 33%
• Demographic Parity 33%
• Model of Affect 33%

Computer Science

• Machine Learning 100%
• Backdoors 100%
• Behavior Model 50%
• Model Accuracy 50%
• Federated Learning 50%