.png){kind=icon}
Return
Dynamic Split Federated Learning for resource-constrained IoT systems
.png){kind=icon}
PRE
AI
DOI：10.1016/j.comcom.2025.108275.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Proposes a novel Dynamic Split Federated Learning architecture for IoT systems. • Integrates Federated and Split Learning to optimize training on constrained devices. • Introduces a Genetic Algorithm for efficient client selection during training rounds. • Enables dynamic layer training by predicting optimal cut layers for each client. • Achieves improved accuracy and reduces training overhead compared to baselines.
Keywords:
Dynamic Split Federated Learning
Genetic Algorithm
IoT systems
Client selection
Layer optimization
Journal
IF:
4.3
Papers:
592
Citations:
1.1W
Organization
Cited Papers
ShareSavePrivacy-Preserving Machine Learning With Fully Homomorphic Encryption for Deep Neural Network.png)
IEEE ACCESS
OAAI
IEEE ACCESSIF3.6
OAAIShareSaveInternet of Things intrusion Detection: Centralized, On-Device, or Federated Learning?IEEE NETWORK
IEEE NETWORKIF6.3
PREAI
ShareSaveA Machine Learning Approach for Fall Detection and Daily Living Activity RecognitionIEEE ACCESS
OAAI
IEEE ACCESSIF3.6
OAAIShareSaveShareSave
View more