Babylonian Journal of Internet of Things

Federated Learning in IoT: A Survey on Distributed Decision Making

Rasha Talal Hameed

Computer Science, College of Education, Al-Iraqia University, Iraq

Omar Abdulwahabe Mohamad

Computer Science, College of Education, Al-Iraqia University, Iraq

DOI: https://doi.org/10.58496/BJIoT/2023/001

Keywords: Federated Learning, Internet of Things (IoT), Distributed Decision Making, Decentralized Machine Learning, IoT Applications

---

Abstract

The proliferation of Internet of Things (IoT) devices has ushered in an era of unprecedented data generation, necessitating innovative approaches to extract valuable insights while respecting privacy and resource constraints. Federated Learning (FL) has emerged as a promising paradigm, enabling decentralized model training across a network of edge devices. This paper presents a comprehensive survey on the application of Federated Learning in IoT, with a specific focus on the challenges and solutions related to distributed decision making. The survey begins by elucidating the foundational concepts of Federated Learning and IoT, highlighting their convergence and the potential benefits of leveraging FL in decentralized IoT ecosystems. The paper explores the diverse applications of FL in various IoT domains, including smart cities, healthcare, industrial IoT, and smart grid systems. It investigates how FL can address the challenges posed by the distributed nature of IoT data, such as data heterogeneity, privacy concerns, and communication constraints. A significant portion of the survey is dedicated to examining the methodologies and algorithms employed in federated learning for distributed decision making in IoT. This encompasses a discussion on federated optimization techniques, communication-efficient algorithms, and privacy-preserving mechanisms. The survey also delves into the role of edge computing in facilitating efficient FL in IoT, considering the resource constraints inherent in edge devices. Furthermore, the paper reviews the current state-of-the-art federated learning frameworks and platforms tailored for IoT environments. It evaluates their capabilities in handling real-world challenges and providing scalable solutions for distributed decision making. The survey concludes by identifying open research challenges and potential avenues for future developments in federated learning for IoT, emphasizing the need for novel algorithms, robust security measures, and standardized frameworks. In summary, this paper offers a comprehensive overview of Federated Learning in the context of IoT, shedding light on its potential, challenges, and solutions with a specific emphasis on the critical aspect of distributed decision making. The insights provided aim to guide researchers, practitioners, and policymakers in navigating the complex landscape of FL in IoT and fostering advancements in this rapidly evolving field.