Decentralized Federated Learning for Over-Parameterized Models

Keyphrases

• Overparameterized Models 100%
• Decentralized Federated Learning 100%
• Decentralized Stochastic Gradient Descent 100%
• Strong Convergence 50%
• Convergence Guarantee 50%
• Heterogeneous Data 50%
• Convergence Rate 50%
• Machine Learning 50%
• Time-varying Networks 50%
• Local Update 50%
• Feature Model 50%
• Over-parameterized 50%
• Deep Learning Features 50%

Computer Science

• Gradient Descent 100%
• Federated Learning 100%
• Machine Learning 50%
• Local Network 50%
• Convergence Rate 50%
• Deep Learning 50%
• Heterogeneous Data 50%

Mathematics

• Stochastics 100%
• Strong Convergence 50%
• Convergence Rate 50%
• Deep Learning 50%