Virtual Homogeneity Learning: Defending against Data Heterogeneity in Federated Learning

Keyphrases

• Federated Learning 100%
• Data Heterogeneity 100%
• Generalization Performance 40%
• Virtual Dataset 40%
• Client Drift 40%
• Convergence Rate 20%
• Learning Model 20%
• Heterogeneous Clients 20%
• Private Information 20%
• Model Performance 20%
• Effective Means 20%
• Address Data 20%
• Speed Performance 20%
• Natural Distribution 20%
• Improved Convergence 20%

Computer Science

• Data Heterogeneity 100%
• Federated machine learning 100%
• Generalization Performance 40%
• Convergence Speed 20%
• Performance Model 20%