Provably Secure Federated Learning against Malicious Clients

Computer Science

• Federated Learning 100%
• Malicious Client 100%
• Learning Algorithm 28%
• Training Data 14%
• Human Activity Recognition 14%

Keyphrases

• Secure Federated Learning 100%
• Byzantine-robust 16%
• Certified Accuracy 16%
• Local Training 16%
• Robust Federated Learning 16%