Improved Information-Theoretic Generalization Bounds for Distributed, Federated, and Iterative Learning †

Keyphrases

• Information Theory 100%
• Distributed Learning 100%
• Generalization Error 100%
• Generalization Bounds 100%
• Iterative Learning 100%
• Federated Learning 100%
• Mutual Information 50%
• Learning Problems 50%
• Statistical Learning 50%
• Node number 50%
• Bregman Divergence 50%
• Information-theoretic Bounds 50%
• Communication Constraints 50%
• Training Data 50%
• Multi-round 50%
• Network Context 50%
• Simple Moving Average 50%
• Generalization Properties 50%
• Lipschitz Loss 50%
• Privacy Constraint 50%
• Centralized Model 50%

Computer Science

• Generalization Error 100%
• Mutual Information 50%
• Learning Problem 50%
• Training Dataset 50%
• Centralized Model 50%