Communication Efficient Federated Learning for Wireless Networks

This book provides a comprehensive study of Federated Learning (FL) over wireless networks. It consists of three main parts: (a) Fundamentals and preliminaries of FL, (b) analysis and optimization of FL over wireless networks, and (c) applications of wireless FL for Internet-of-Things systems. In particular, in the first part, the authors provide a detailed overview on widely-studied FL framework. In the second part of this book, the authors comprehensively discuss three key wireless techniques including wireless resource management, quantization, and over-the-air computation to support the deployment of FL over realistic wireless networks. It also presents several solutions based on optimization theory, graph theory and machine learning to optimize the performance of FL over wireless networks. In the third part of this book, the authors introduce the use of wireless FL algorithms for autonomous vehicle control and mobile edge computing optimization.

Machine learning and data-driven approaches have recently received considerable attention as key enablers for next-generation intelligent networks. Currently, most existing learning solutions for wireless networks rely on centralizing the training and inference processes by uploading data generated at edge devices to data centers. However, such a centralized paradigm may lead to privacy leakage, violate the latency constraints of mobile applications, or may be infeasible due to limited bandwidth or power constraints of edge devices. To address these issues, distributing machine learning at the network edge provides a promising solution, where edge devices collaboratively train a shared model using real-time generated mobile data. The avoidance of data uploading to a central server not only helps preserve privacy but also reduces network traffic congestion as well as communication cost. Federated learning (FL) is one of most important distributed learning algorithms. In particular, FL enables devices to train a shared machine learning model while keeping data locally. However, in FL, training machine learning models requires communication between wireless devices and edge servers over wireless links. Therefore, wireless impairments such as noise, interference, and uncertainties among wireless channel states will significantly affect the training process and performance of FL. For example, transmission delay can significantly impact the convergence time of FL algorithms. In consequence, it is necessary to optimize wireless network performance for the implementation of FL algorithms.

This book targets researchers and advanced level students in computer science and electrical engineering. Professionals working in signal processing and machine learning will also buy this book.

本書提供了對於無線網絡上聯邦學習（FL）的全面研究。它由三個主要部分組成：（a）FL的基礎和初步研究，（b）無線網絡上FL的分析和優化，以及（c）無線FL在物聯網系統中的應用。特別是在第一部分中，作者詳細介紹了廣泛研究的FL框架。在本書的第二部分中，作者全面討論了三個關鍵的無線技術，包括無線資源管理、量化和無線計算，以支持在現實無線網絡中部署FL。它還提出了基於優化理論、圖論和機器學習的幾種解決方案，以優化無線網絡上FL的性能。在本書的第三部分中，作者介紹了在自動駕駛車控制和移動邊緣計算優化中使用無線FL算法的應用。

機器學習和數據驅動方法最近受到了作為下一代智能網絡的關鍵推動者的廣泛關注。目前，大多數現有的無線網絡學習解決方案依賴於將邊緣設備生成的數據上傳到數據中心，以集中進行訓練和推理過程。然而，這種集中式範式可能會導致隱私洩露，違反移動應用程序的延遲限制，或由於邊緣設備的帶寬或功耗限制而不可行。為了解決這些問題，在網絡邊緣分佈機器學習提供了一個有希望的解決方案，其中邊緣設備通過使用實時生成的移動數據共同訓練一個共享模型。避免將數據上傳到中央服務器不僅有助於保護隱私，還可以減少網絡流量擁塞和通信成本。聯邦學習（FL）是最重要的分佈式學習算法之一。特別是，FL使設備能夠在保留本地數據的同時訓練共享的機器學習模型。然而，在FL中，訓練機器學習模型需要無線設備和邊緣服務器之間的無線鏈路通信。因此，無線干擾，如噪聲、干擾和無線信道狀態的不確定性，將顯著影響FL的訓練過程和性能。例如，傳輸延遲可能會顯著影響FL算法的收斂時間。因此，有必要優化無線網絡性能以實現FL算法的部署。

本書針對計算機科學和電氣工程的研究人員和高級學生。從事信號處理和機器學習的專業人士也會購買本書。

Mingzhe Chen (S'15-M'19) is currently an Assistant Professor with the Department of Electrical and Computer Engineering and Institute of Data Science and Computing at University of Miami. His research interests include federated learning, reinforcement learning, virtual reality, unmanned aerial vehicles, and Internet of Things. He has received four IEEE Communication Society journal paper awards including the IEEE Marconi Prize Paper Award in Wireless Communications in 2023, the Young Author Best Paper Award in 2021 and 2023, and the Fred W. Ellersick Prize Award in 2022, and four conference best paper awards at ICCCN in 2023, IEEE WCNC in 2021, IEEE ICC in 2020, and IEEE GLOBECOM in 2020. He currently serves as an Associate Editor of IEEE Transactions on Mobile Computing, IEEE Wireless Communications Letters, IEEE Transactions on Green Communications and Networking, and IEEE Transactions on Machine Learning in Communications and Networking.Shuguang Cui (S'99-M'05-SM'12-F'14) received his Ph.D in Electrical Engineering from Stanford University, California, USA, in 2005. Afterwards, he has been working as assistant, associate, full, Chair Professor in Electrical and Computer Engineering at the Univ. of Arizona, Texas A&M University, UC Davis, and CUHK at Shenzhen respectively. He has also served as the Executive Dean for the School of Science and Engineering at CUHK, Shenzhen, the Executive Vice Director at Shenzhen Research Institute of Big Data, and the Director for Future Network of Intelligence Institute (FNii). His current research interests focus on the merging between AI and communication neworks. He was selected as the Thomson Reuters Highly Cited Researcher and listed in the Worlds' Most Influential Scientific Minds by ScienceWatch in 2014. He was the recipient of the IEEE Signal Processing Society 2012 Best Paper Award. He has served as the general co-chair and TPC co-chairs for many IEEE conferences. He has also been serving as the area editor for IEEE Signal Processing Magazine, and associate editors for IEEE Transactions on Big Data, IEEE Transactions on Signal Processing, IEEE JSAC Series on Green Communications and Networking, and IEEE Transactions on Wireless Communications. He has been the elected member for IEEE Signal Processing Society SPCOM Technical Committee (2009 2014) and the elected Chair for IEEE ComSoc Wireless Technical Committee (2017 2018). He is a member of the Steering Committee for IEEE Transactions on Big Data and the Chair of the Steering Committee for IEEE Transactions on Cognitive Communications and Networking. He is also the Vice Chair of the IEEE VT Fellow Evaluation Committee and a member of the IEEE ComSoc Award Committee. He was elected as an IEEE Fellow in 2013, an IEEE ComSoc Distinguished Lecturer in 2014, and IEEE VT Society Distinguished Lecturer in 2019. In 2020, he won the IEEE ICC best paper award, ICIP best paper finalist, the IEEE Globecom best paper award. In 2021, he won the IEEE WCNC best paper award. In 2023, he won the IEEE Marconi Best Paper Award, got elected as a Fellow of both Canadian Academy of Engineering and the Royal Society of Canada, and starts to serve as the Editor-in-Chief for IEEE Transactions on Mobile Computing.

Mingzhe Chen（S'15-M'19）目前是邁阿密大學電機與計算機工程系和數據科學與計算研究所的助理教授。他的研究興趣包括聯邦學習、強化學習、虛擬現實、無人機和物聯網。他曾獲得四項IEEE通信學會期刊論文獎，包括2023年IEEE Marconi無線通信獎論文獎、2021年和2023年青年作者最佳論文獎，以及2022年Fred W. Ellersick獎。此外，他還在2023年的ICCCN、2021年的IEEE WCNC、2020年的IEEE ICC和2020年的IEEE GLOBECOM等會議上獲得了四項最佳論文獎。他目前擔任IEEE移動計算交易、IEEE無線通信信函、IEEE綠色通信和網絡交易以及IEEE通信和網絡機器學習交易的副編輯。

Shuguang Cui（S'99-M'05-SM'12-F'14）於2005年在美國加利福尼亞州斯坦福大學獲得電機工程博士學位。之後，他分別在亞利桑那大學、德克薩斯A＆M大學、加州大學戴維斯分校和中國香港中文大學擔任助理教授、副教授、教授和主任教授。他還曾擔任中國香港中文大學深圳研究院科學與工程學院的執行院長，深圳大數據研究院的執行副院長，以及未來智能網絡研究院的主任。他目前的研究興趣集中在人工智能和通信網絡之間的融合。他曾被選為湯森路透高被引研究者，並在2014年被ScienceWatch列為世界上最有影響力的科學家之一。他曾獲得IEEE信號處理學會2012年最佳論文獎。他曾擔任多個IEEE會議的總共同主席和技術範疇共同主席。他還擔任IEEE信號處理雜誌的範疇編輯，以及IEEE大數據交易、IEEE信號處理交易、IEEE JSAC綠色通信和網絡系列、IEEE無線通信交易的副編輯。他曾擔任IEEE信號處理學會SPCOM技術委員會（2009-2014）的當選成員，以及IEEE通信學會無線技術委員會（2017-2018）的當選主席。他是IEEE大數據交易的指導委員會成員，以及IEEE認知通信和網絡交易的指導委員會主席。他還是IEEE VT Fellow評估委員會的副主席，以及IEEE通信學會獎項委員會的成員。他於2013年當選為IEEE院士，2014年當選為IEEE通信學會傑出講師，2019年當選為IEEE VT學會傑出講師。2020年，他獲得了IEEE ICC最佳論文獎、ICIP最佳論文入圍獎和IEEE Globecom最佳論文獎。2021年，他獲得了IEEE WCNC最佳論文獎。2023年，他獲得了IEEE Marconi最佳論文獎，當選為加拿大工程院和加拿大皇家學會院士，並開始擔任IEEE移動計算交易的主編。