Cambridge University Press & Assessment

978-1-108-83298-4 — Machine Learning and Wireless Communications

Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

Machine Learning and Wireless Communications

How can machine learning help the design of future communication networks, and
how can future networks meet the demands of emerging machine learning applica-
tions? Discover the interactions between two of the most transformative and impactful
technologies of our age in this comprehensive book.
First, learn how modern machine learning techniques, such as deep neural networks,
can transform how we design and optimize future communication networks. Acces-
sible introductions to concepts and tools are accompanied by numerous real-world
examples, showing you how these techniques can be used to tackle longstanding
problems. Next, explore the design of wireless networks as platforms for machine
learning applications. An overview of modern machine learning techniques and com-
munication protocols will help you to understand the challenges, while new methods
and design approaches will be presented to handle wireless channel impairments such
as noise and interference, to meet the demands of emerging machine learning applica-
tions at the wireless edge.

Yonina C. Eldar is a professor of electrical engineering at the Weizmann Institute of

Science, where she heads the Center for Biomedical Engineering and Signal Process-
ing. She is also a visiting professor at Massachusetts Institute of Technology and the
Broad Institute and an adjunct professor at Duke University. She is a member of the
Israel Academy of Sciences and Humanities, an IEEE fellow, and a EURASIP fellow.

Andrea Goldsmith is the Dean of Engineering and Applied Science and the Arthur
LeGrand Doty Professor of Electrical Engineering at Princeton University. She is a
member of the US National Academy of Engineering and the American Academy of
Arts and Sciences. In 2020, she received the Marconi Prize.

Deniz Gündüz is a professor of information processing in the Electrical and Electronic

Engineering Department of Imperial College London, where he serves as the Deputy
Head of the Intelligent Systems and Networks Group. He is also a part-time faculty
member at the University of Modena and Reggio Emilia.

H. Vincent Poor is the Michael Henry Strater University Professor at Princeton

University. He is a member of the US National Academy of Engineering and the
US National Academy of Sciences. In 2017, he received the IEEE Alexander Graham
Bell Medal.

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

Machine Learning and Wireless

Communications
Edited by

YONINA C. ELDAR
Weizmann Institute of Science

ANDREA GOLDSMITH
Princeton University

DENIZ GÜNDÜZ
Imperial College

H. VINCENT POOR
Princeton University

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

University Printing House, Cambridge CB2 8BS, United Kingdom

One Liberty Plaza, 20th Floor, New York, NY 10006, USA
477 Williamstown Road, Port Melbourne, VIC 3207, Australia
314–321, 3rd Floor, Plot 3, Splendor Forum, Jasola District Centre, New Delhi – 110025, India
103 Penang Road, #05–06/07, Visioncrest Commercial, Singapore 238467

Cambridge University Press is part of the University of Cambridge.

It furthers the University’s mission by disseminating knowledge in the pursuit of
education, learning, and research at the highest international levels of excellence.

www.cambridge.org
Information on this title: www.cambridge.org/9781108832984
DOI: 10.1017/9781108966559
© Cambridge University Press 2022
This publication is in copyright. Subject to statutory exception
and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2022
Printed in the United Kingdom by TJ Books Limited, Padstow Cornwall
A catalogue record for this publication is available from the British Library.
Library of Congress Cataloging-in-Publication Data
Names: Eldar, Yonina C., editor.
Title: Machine learning and wireless communications / edited by Yonina C. Eldar,
Weizmann Institute of Science, Andrea Goldsmith, Princeton University,
Deniz Gündüz, Imperial College, H. Vincent Poor, Princeton University.
Description: First edition. | Cambridge, United Kingdom ; New York, NY :
Cambridge University Press, 2022. | Includes bibliographical references and index.
Identifiers: LCCN 2021063108 (print) | LCCN 2021063109 (ebook) |
ISBN 9781108832984 (hardback) | ISBN 9781108966559 (epub)
Subjects: LCSH: Wireless communication systems. | Machine learning. |
BISAC: TECHNOLOGY & ENGINEERING / Signals & Signal Processing
Classification: LCC TK5103.2 .M3156 2022 (print) | LCC TK5103.2 (ebook) |
DDC 621.382–dc23/eng/20220318
LC record available at https://lccn.loc.gov/2021063108
LC ebook record available at https://lccn.loc.gov/2021063109
ISBN 978-1-108-83298-4 Hardback
Cambridge University Press has no responsibility for the persistence or accuracy
of URLs for external or third-party internet websites referred to in this publication
and does not guarantee that any content on such websites is, or will remain,
accurate or appropriate.

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

To our families.

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

Contents

List of Contributors page ix

Preface xiii

1 Machine Learning and Communications: An Introduction 1

Deniz Gündüz, Yonina C. Eldar, Andrea Goldsmith, and H. Vincent Poor

Part I Machine Learning for Wireless Networks

2 Deep Neural Networks for Joint Source-Channel Coding 23

David Burth Kurka, Milind Rao, Nariman Farsad, Deniz Gündüz, and Andrea Goldsmith

3 Neural Network Coding 55

Litian Liu, Amit Solomon, Salman Salamatian, Derya Malak, and Muriel Medard

4 Channel Coding via Machine Learning 77

Hyeji Kim

5 Channel Estimation, Feedback, and Signal Detection 110

Hengtao He, Hao Ye, Shi Jin, and Geoffrey Y. Li

6 Model-Based Machine Learning for Communications 145

Nir Shlezinger, Nariman Farsad, Yonina C. Eldar, and Andrea Goldsmith

7 Constrained Unsupervised Learning for Wireless Network Optimization 182

Hoon Lee, Sang Hyun Lee, and Tony Q. S. Quek

8 Radio Resource Allocation in Smart Radio Environments 212

Alessio Zappone and Merouane Debbah

9 Reinforcement Learning for Physical Layer Communications 231

Philippe Mary, Christophe Moy, and Visa Koivunen

10 Data-Driven Wireless Networks: Scalability and Uncertainty 285

Feng Yin, Yue Xu, and Shuguang Cui

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

viii Contents

11 Capacity Estimation Using Machine Learning 317

Ziv Aharoni, Dor Tsur, Ziv Goldfeld, and Haim H. Permuter

Part II Wireless Networks for Machine Learning

12 Collaborative Learning over Wireless Networks: An Introductory Overview 353

Emre Ozfatura, Deniz Gündüz, and H. Vincent Poor

13 Optimized Federated Learning in Wireless Networks with Constrained Resources 385

Shiqiang Wang, Tiffany Tuor, and Kin K. Leung

14 Quantized Federated Learning 409

Nir Shlezinger, Mingzhe Chen, Yonina C. Eldar, H. Vincent Poor, and Shuguang Cui

15 Over-the-Air Computation for Distributed Learning over Wireless Networks 434

Mohammad Mohammadi Amiri and Deniz Gündüz

16 Federated Knowledge Distillation 457

Hyowoon Seo, Jihong Park, Seungeun Oh, Mehdi Bennis, and Seong-Lyun Kim

17 Differentially Private Wireless Federated Learning 486

Dongzhu Liu, Amir Sonee, Osvaldo Simeone, and Stefano Rini

18 Timely Wireless Edge Inference 512

Sheng Zhou, Wenqi Shi, Xiufeng Huang, and Zhisheng Niu

Index 539

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

Contributors

Ziv Aharoni,
Ben-Gurion University of the Negev

Mohammad Mohammadi Amiri,

Princeton University

Mehdi Bennis,
University of Oulu

Mingzhe Chen,
Princeton University

Shuguang Cui,
The Chinese University of Hong Kong, Shenzhen

Mérouane Debbah,
CentralSupélec

Nariman Farsad,
Stanford University

Ziv Goldfeld,
Cornell University

Hengtao He,
Southeast University

Xiufeng Huang,
Tsinghua University

Shi Jin,
Southeast University

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

x List of Contributors

Hyeji Kim,
The University of Texas

Seong–Lyun Kim,
Yonsei University

Visa Koivunen,
Aalto University

David Burth Kurka,

Imperial College London

Hoon Lee,
Pukyong National University

Sang Hyun Lee,

Korea University

Kin K. Leung,
Imperial College London

Geoffrey Y. Li,
Imperial College London

Dongzhu Liu,
King’s College London

Litian Liu,
Massachusetts Institute of Technology

Derya Malak,
EURECOM

Philippe Mary,
Institut National des Sciences Appliquées

Muriel Médard,
Massachusetts Institute of Technology

Christophe Moy,
University of Rennes 1

Zhisheng Niu,
Tsinghua University

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

List of Contributors xi

Seungeun Oh,
Yonsei University

Mehmet Emre Ozfatura,

Imperial College London

Jihong Park,
Deakin University

Haim H. Permuter,
Ben-Gurion University of the Negev

Tony Q. S. Quek,
Singapore University of Technology and Design

Milind Rao,
Stanford University

Stefano Rini,
National Chiao Tung University

Salman Salamatian,
Massachusetts Institute of Technology

Hyowoon Seo,
Kwangwoon University

Wenqi Shi,
Tsinghua University

Nir Shlezinger,
Ben-Gurion University of the Negev

Osvaldo Simeone,
King’s College London

Amit Solomon,
Massachusetts Institute of Technology

Amir Sonee,
National Chiao Tung University

Tiffany Tuor,
Imperial College London

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

xii List of Contributors

Shiqiang Wang,
IBM Thomas J. Watson Research Center

Yue Xu,
Beijing University of Posts and Telecommunications

Hao Ye,
Georgia Institute of Technology

Feng Yin,
The Chinese University of Hong Kong, Shenzhen

Alessio Zappone,
University of Cassino

Sheng Zhou,
Tsinghua University

Dor Tsur,
Ben-Gurion University of the Negev

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

Preface

Machine learning (ML) and wireless communications are two of the most rapidly
advancing technologies of our time. The main premise of ML is to enable computers
to learn and perform certain tasks without being explicitly programmed to do so. This
is achieved by training algorithms on data available for the task to be accomplished.
Although the basic ideas and ambitions of ML go back to the 1950s, there has been
a recent surge in interest and applications in this area, fueled by the availability of
increasingly powerful computers, large amounts of data, and developments in new
learning algorithms as well as their theoretical underpinnings. At the same time, wire-
less communication has evolved, through advances in both theory and supporting
technologies, to encompass a variety of application areas, from high-performance
data transmission tasks such as media distribution to the massive deployment of end-
devices to enable Internet of Things (IoT) tasks such as sensing, inference, and control.
We are now witnessing the confluence of these two fields, with two primary aspects
to this connection. One is the application of ML techniques to the optimization of
wireless networks. This is a natural use of ML, as wireless networks involve many
inferential and control tasks, which often must operate under dynamic or uncertain
conditions, and create many examplars for learning because data transmissions take
place at very high rates. The other aspect of this connection is the use of wireless
networks as ML platforms. This again is a natural application of emerging wireless
networks, such as those supporting IoT applications, because they involve sensing,
inference, and control and provide edge devices with considerable processing power.
Learning at the network edge has advantages in terms of latency and privacy, and it
capitalizes on the fact that many learning tasks, such as those supporting automated
driving, are locality specific.
To realize the promise of these opportunities, significant research in many dimen-
sions is needed. Important issues include the adaptation of existing ML techniques
to wireless system design and the design and development of new techniques that
can meet the constraints and requirements of communication networks, including the
capability to implement at least some of these techniques in low-power chips that can
be used in mobile devices, as well as developing fundamental analytical techniques
and bounds on the performance of distributed ML algorithms operating within the con-
straints of wireless connectivity. This book focuses on these research issues through a
series of 18 chapters written by experts in the field, beginning with an introductory
chapter providing a brief general overview of ML methodology. By presenting a

© in this web service Cambridge University Press & Assessment www.cambridge.org

Cambridge University Press & Assessment
978-1-108-83298-4 — Machine Learning and Wireless Communications
Edited by Yonina C. Eldar , Andrea Goldsmith , Deniz Gündüz , H. Vincent Poor
Frontmatter
More Information

xiv Preface

systematic overview of the most promising aspects of the connection between ML

in wireless networks, this book provides an entry point and a comprehensive overview
of the state of the art for researchers in academia and industry who are interested in
learning and contributing to this growing field.
This book is the culmination of the efforts of many people, including the chapter
authors and the editorial and production staff at Cambridge University Press. We wish
to express our deep gratitude for their contributions.

© in this web service Cambridge University Press & Assessment www.cambridge.org