Architectural Patterns for Software Design Problem-Solving in the Implementation of Federated Learning Structures Within the E-Health Sector

Current research in artificial intelligence, including Machine Learning and Deep Learning, is driving innovation in various fields. In the healthcare sector, where considerable amounts of data are used for studies, early diagnosis and disease monitoring, the importance of addressing security and privacy issues is clear. Cloud Edge and Federated Learning, a more privacy-focused approach, allow algorithms to be trained without actual data exchange, using decentralized models. Recent studies show that prototypes trained with Federated Learning and Cloud Edge paradigms achieve reliable performance, generating robust models while preserving security and privacy. This study proposes a review focused on specific Patterns that, when applied to the design of Cloud Edge architectures within the healthcare sector, solve multiple problems, exploring challenges, implications and potential in this context. In particular, for each of the described Patterns, this paper will focus on the specific problem they face and present the actual solution, specifically in association with Federated Learning approaches.