Multi-Job Intelligent Scheduling With Cross-Device Federated Learning

Recent years have witnessed a large amount of decentralized data in various (edge) devices end-users, while the aggregation remains complicated for machine learning jobs because regulations and laws. As practical approach to handling data, Federated Learning (FL) enables collaborative global model training without sharing sensitive raw data. The servers schedule within process FL. In contrast, device scheduling with multiple FL critical open problem. this article, we propose novel multi-job framework, which parallel. framework is composed system method. parallel jobs, cost based on fairness time diverse during process. We intelligent methods, including an original reinforcement learning-based method Bayesian optimization-based method, corresponds small jobs. conduct extensive experimentation datasets. experimental results reveal that our proposed approaches significantly outperform baseline terms (up 12.73 times faster) accuracy 46.4% higher).