Challenges, Applications and Design Aspects of Federated Learning: A Survey

Federated Learning (FL) is a new technology that has been a hot research topic. It enables training an algorithm across multiple decentralized edge devices or servers holding local data samples, without exchanging them. There are many application domains where large amounts of properly labeled and complete data are not available in a centralized location, for example, doctors’ diagnosis from medical image analysis. There are also growing concerns over data and user privacy as Artificial Intelligence is becoming ubiquitous in new application domains. As such, very recently, a lot of research has been conducted in several areas within the nascent field of FL. A variety of surveys on different subtopics exist in current literature, focusing on specific challenges, design aspects and application domains. In this paper, we review existing contemporary works in the related areas in order to understand the challenges and topics that are emphasized by each type of FL surveys. Furthermore, we categorize FL research in terms of challenges, design factors and applications, conducting a holistic review of each and outlining promising research directions.