Resource-Constrained Decentralized Federated Learning via Personalized Event-Triggering

Keyphrases

• Asynchronous Events 33%
• Central Server 33%
• Connected Information 33%
• Consensus Formation 33%
• Convergence Guarantee 33%
• Convergence Rate 66%
• Cooperative Consensus 33%
• Data Heterogeneity 33%
• Decentralized Federated Learning 100%
• Device-to-device (D2D) Communication 33%
• Edge Devices 33%
• Event-triggered 100%
• Event-triggered Communication 33%
• Federated Learning 66%
• Fully Decentralized 33%
• Globally Optimal 33%
• Gradient Method 33%
• Graph Connectivity 33%
• Graph Structure 33%
• Information Flow 33%
• Local Model 33%
• Local Networks 33%
• Machine Learning 33%
• Model Aggregation 33%
• Network Resources 33%
• Numerical Evaluation 33%
• Personalized Communication 33%
• Popular 33%
• Resource Heterogeneity 33%
• Resource-constrained 100%
• Sporadic Communication 33%
• Trigger Delay 33%

Computer Science

• Connectivity Graph 33%
• Convergence Rate 33%
• Convergence Speed 33%
• Data Heterogeneity 33%
• Decentralized Federated Learning 100%
• Device-To-Device 33%
• Event-triggered 33%
• Federated machine learning 66%
• Gradient Descent Method 33%
• Learning System 33%
• Local Network 33%
• Machine Learning 33%
• Network Resource 33%
• Triggering Condition 33%
• Triggering Event 100%