Dynamic Split Federated Learning for resource-constrained IoT systems

Keyphrases

• Accuracy Improvement 25%
• Centralized Data 25%
• Client Selection 25%
• Collaborative Model 25%
• Data Privacy 25%
• Device Heterogeneity 25%
• Device Training 25%
• Dynamic Model 25%
• Dynamic Nature 25%
• Dynamically Adjust 25%
• Efficient Training 25%
• Efficient Use of Resources 25%
• Federated Learning 25%
• Genetic Algorithm 50%
• Heterogeneous Internet of Things 25%
• Heterogeneous Resources 25%
• Image-based 25%
• Internet of Things 25%
• Internet of Things Devices 100%
• Internet of Things System 100%
• Learning Architecture 25%
• Longest Processing Time 25%
• Machine Learning 25%
• Metadata Analysis 25%
• On-chip Learning 25%
• Parallelization 25%
• Point Selection 25%
• Privacy Concerns 25%
• Promising Solutions 25%
• Proposed Architecture 25%
• Resource Availability 25%
• Resource Heterogeneity 25%
• Resource Utilization 25%
• Resource-constrained 100%
• Resource-constrained Devices 25%
• Split Federated Learning 100%
• Split Learning 25%
• Split Point 25%
• Time Resources 25%
• Training Abilities 25%
• Training Model 50%
• Training Tasks 25%
• User Privacy 25%

Computer Science

• Aggregate Model 20%
• Baseline Method 20%
• Constrained Device 20%
• Data Privacy 20%
• Dynamic Nature 20%
• Federated machine learning 100%
• Genetic Algorithm 40%
• Heterogeneous Resource 20%
• Individual Client 20%
• Internet of Things Device 80%
• Internet-Of-Things 100%
• Learning System 20%
• Machine Learning 20%
• Machine Learning Model 20%
• Parallelism 20%
• Privacy Concern 20%
• Processing Time 20%
• Resource Availability 20%
• Resource Utilisation 20%
• User Privacy 20%