Client Selection for Generalization in Accelerated Federated Learning: A Bandit Approach

Keyphrases

• Federated Learning 100%
• Bandits 100%
• Client Selection 100%
• Latency 20%
• Reliability Prediction 20%
• Communication Resources 20%
• Learning Performance 20%
• Novel Algorithm 20%
• Selection Problem 20%
• Training Model 20%
• Global Model 20%
• Local Data 20%
• Full Knowledge 20%
• Transmission Time 20%
• GENIE 20%
• Local Model 20%
• Multiple Nodes 20%
• Multi-arm Bandit 20%
• Privacy Considerations 20%
• Machine Learning Paradigms 20%
• Logarithmic Regret 20%
• Privacy Communication 20%
• Learning Latency 20%
• Random Processing Times 20%
• Global Broadcast 20%

Computer Science

• Federated Learning 100%
• Machine Learning 20%
• Processing Time 20%
• Communication Resource 20%
• Problem Selection 20%
• Learning Performance 20%
• Transmission Time 20%