Journal Article
Machine Learning

Federated Learning in Mobile Edge Networks: A Comprehensive Survey

Wei Yang Bryan Lim(Nanyang Technological University), Nguyen Cong Luong(Phenikaa (Vietnam)), Dinh Thai Hoang(University of Technology Sydney), Yutao Jiao(Nanyang Technological University), Ying‐Chang Liang(University of Electronic Science and Technology of China), Qiang Yang(Hong Kong University of Science and Technology), Dusit Niyato(Nanyang Technological University), Chunyan Miao(Nanyang Technological University)
January 1, 2020IEEE Communications Surveys & Tutorials2,447 citations
arXivDOI

2.4k

Citations

110

Influential Citations

IEEE Communications Surveys & Tutorials

Venue

2020

Year

Abstract

In recent years, mobile devices are equipped with increasingly advanced sensing and computing capabilities. Coupled with advancements in Deep Learning (DL), this opens up countless possibilities for meaningful applications, e.g., for medical purposes and in vehicular networks. Traditional cloud-based Machine Learning (ML) approaches require the data to be centralized in a cloud server or data center. However, this results in critical issues related to unacceptable latency and communication inefficiency. To this end, Mobile Edge Computing (MEC) has been proposed to bring intelligence closer to the edge, where data is produced. However, conventional enabling technologies for ML at mobile edge networks still require personal data to be shared with external parties, e.g., edge servers. Recently, in light of increasingly stringent data privacy legislations and growing privacy concerns, the concept of Federated Learning (FL) has been introduced. In FL, end devices use their local data to train an ML model required by the server. The end devices then send the model updates rather than raw data to the server for aggregation. FL can serve as an enabling technology in mobile edge networks since it enables the collaborative training of an ML model and also enables DL for mobile edge network optimization. However, in a large-scale and complex mobile edge network, heterogeneous devices with varying constraints are involved. This raises challenges of communication costs, resource allocation, and privacy and security in the implementation of FL at scale. In this survey, we begin with an introduction to the background and fundamentals of FL. Then, we highlight the aforementioned challenges of FL implementation and review existing solutions. Furthermore, we present the applications of FL for mobile edge network optimization. Finally, we discuss the important challenges and future research directions in FL.

Analysis

Why This Paper Matters

This survey addresses a critical tension in modern AI: the need to train powerful deep learning models on vast amounts of data generated at the network edge, while respecting stringent data privacy regulations (e.g., GDPR) and minimizing latency. Traditional cloud-centric approaches are impractical for mobile edge networks due to bandwidth bottlenecks and privacy risks. Federated Learning (FL) offers a paradigm shift by keeping data on-device and only sharing model updates. The paper systematically maps the landscape of FL in mobile edge networks, making it an essential reference for practitioners and researchers.

Published in IEEE Communications Surveys & Tutorials, a top venue for comprehensive reviews, and with over 2,400 citations, this work has shaped the direction of research in distributed machine learning for edge systems. It highlights that while FL is conceptually elegant, its practical deployment in heterogeneous, resource-constrained mobile environments introduces significant engineering challenges.

Technical Contributions

  • Comprehensive Problem Taxonomy: The paper clearly delineates three core challenges for FL at scale: (1) communication costs (frequent model updates over limited bandwidth), (2) resource allocation (heterogeneous device capabilities and energy budgets), and (3) privacy and security (gradient leakage, poisoning attacks).
  • Solution Categorization: It reviews existing approaches for each challenge, such as gradient compression and quantization for communication efficiency, client selection and scheduling for resource management, and differential privacy and secure aggregation for privacy.
  • Application Framework: The survey demonstrates how FL can be applied to optimize mobile edge networks themselves, e.g., for traffic prediction, caching, and resource management, creating a closed loop where FL both runs on and optimizes the network.
  • Future Directions: It identifies open problems including personalization, fairness, and incentive mechanisms, which have since become active research areas.

Results

As a survey, the paper does not present new experimental results. Its primary contribution is a structured synthesis of the field. The paper's impact is evidenced by its citation count (2447), indicating it is a foundational reference. It does not report quantitative metrics like accuracy or convergence speed, but qualitatively argues that FL reduces communication overhead compared to centralized learning and preserves privacy by design.

Significance

This survey has had a profound impact on the AI and networking communities. It provided a clear, structured overview that helped standardize terminology and problem definitions for FL in edge environments. By highlighting the practical challenges of communication efficiency, resource allocation, and privacy, it motivated a wave of subsequent research on topics like asynchronous FL, federated optimization, and secure aggregation. For practitioners, it serves as a roadmap for deploying FL in real-world mobile applications, from healthcare to autonomous driving. The paper's emphasis on future directions has also guided funding and research priorities in distributed AI systems.

federated-learningmobile-edge-computingdeep-learningprivacyresource-allocationcommunication-efficiencysurvey

Related Papers