A Workflow Management System Approach To Federated Learning: Application to Industry 4.0

Federated Learning (FL) combined with the Industrial Internet of Things (IIoT) enhances decision-making in industrial settings by leveraging decentralized machine learning (ML) to ensure data privacy, optimize edge computing, and facilitate adaptive model training. However, implementing FL and IIoT presents challenges due to distributed architectures, including communication, data transfer, and file management across wide area networks. This paper addresses these challenges by introducing FL and Clustered FL (CFL) models using Pegasus workflows. Evaluated with real data from airport baggage conveyor systems, it offers practical insights into FL's application in IIoT environments, contributing to advancements in intelligent industrial decision-making.