Machine Learning for Low-Latency Communications

1st Edition - October 10, 2024
Authors: Yong Zhou, Yinan Zou, Youlong Wu, Yuanming Shi, Jun Zhang
Language: English
Paperback ISBN:
9 7 8 - 0 - 4 4 3 - 2 2 0 7 3 - 9
eBook ISBN:
9 7 8 - 0 - 4 4 3 - 2 2 0 7 4 - 6

Machine Learning for Low-Latency Communications presents the principles and practice of various deep learning methodologies for mitigating three critical latency component… Read more

Purchase options

Institutional subscription on ScienceDirect

Request a sales quote

Machine Learning for Low-Latency Communications presents the principles and practice of various deep learning methodologies for mitigating three critical latency components: access latency, transmission latency, and processing latency. In particular, the book develops learning to estimate methods via algorithm unrolling and multiarmed bandit for reducing access latency by enlarging the number of concurrent transmissions with the same pilot length. Task-oriented learning to compress methods based on information bottleneck are given to reduce the transmission latency via avoiding unnecessary data transmission.

Lastly, three learning to optimize methods for processing latency reduction are given which leverage graph neural networks, multi-agent reinforcement learning, and domain knowledge. Low-latency communications attracts considerable attention from both academia and industry, given its potential to support various emerging applications such as industry automation, autonomous vehicles, augmented reality and telesurgery. Despite the great promise, achieving low-latency communications is critically challenging. Supporting massive connectivity incurs long access latency, while transmitting high-volume data leads to substantial transmission latency.

Title of Book
Cover image
Title page
Table of Contents
Copyright
List of figures
List of tables
Biographies
Preface
Acknowledgments
Part 1: Introduction and overview
Chapter 1: Introduction and overview
1.1. Low-latency communication
1.1.1. Grant-free random access
1.1.2. Task-oriented data compression
1.1.3. Large-scale resource allocation
1.2. Machine learning for large-scale optimization
1.2.1. Challenges of optimization-based algorithms
1.2.2. Machine learning models and algorithms
1.2.3. Advantages of machine learning algorithms
1.3. Organization
Part 2: Learning to estimate for access latency reduction
Chapter 2: Learning to estimate via group-sparse based algorithm unrolling
2.1. Background on grant-free random access
2.2. System model and problem formulation
2.2.1. System model
2.2.2. Problem formulation
2.3. Group-sparse based algorithm unrolling
2.3.1. Unrolled neural network architectures
2.3.2. Unrolled neural networks training
2.4. Numerical results
2.4.1. Simulation setting and performance metrics
2.4.2. Analysis validation
2.4.3. Performance evaluation
2.5. Summary
Chapter 3: Learning to detect via multi-armed bandit
3.1. Background
3.2. System model and problem formulation
3.2.1. System model
3.2.2. Problem analysis
3.3. Coordinate descent with Bernoulli sampling
3.3.1. Reward function
3.3.2. Algorithm and analysis
3.4. Coordinate descent with Thompson sampling
3.4.1. Stochastic MAB for optimizing ε
3.4.2. Thompson sampling
3.4.3. CD-Thompson
3.5. Application to massive connectivity with low-precision ADCs
3.6. Simulation results
3.6.1. Simulation settings and performance metric
3.6.2. Convergence rate
3.6.3. Probability of error
3.6.4. Applications in low-precision ADCs
3.7. Conclusions
Part 3: Learning to compress for transmission latency reduction
Chapter 4: Learning to compress via information bottleneck
4.1. Background and motivation
4.2. System model and problem formulation
4.2.1. Task-oriented communication for edge inference
4.2.2. Information bottleneck principle
4.3. Variational feature encoding
4.3.1. Variational information bottleneck reformulation
4.3.2. Redundancy reduction and feature sparsification
4.3.3. Variational pruning on dimension importance
4.4. Variable-length variational feature encoding
4.4.1. Background on dynamic neural networks
4.4.2. Selective activation for dynamic channel conditions
4.4.3. Training procedure for the dynamic neural network
4.5. Simulation results
4.5.1. Experimental setup
4.5.2. Static channel conditions
4.5.3. Dynamic channel conditions
4.6. Summary
Chapter 5: Learning to compress via robust information bottleneck with digital modulation
5.1. Background and motivation
5.2. System model and problem formulation
5.2.1. Task-oriented communication with digital modulation model
5.2.2. Robust information bottleneck (RIB) principle
5.3. Problem description
5.4. Robust encoding for task-oriented communication
5.4.1. Robust encoding with RIB
5.4.2. Variational upper bounds of RIB objective
5.5. Joint source-channel coding with digital modulation (DT-JSCC)
5.5.1. The DT-JSCC framework and algorithm
5.5.2. Practical advantages of DT-JSCC
5.6. Simulation results
5.6.1. Experimental setup
5.6.2. Inference performance
5.6.3. Robustness performance
5.6.4. Ablation experiments
5.7. Summary
Chapter 6: Learning to compress for multi-device cooperative edge inference
6.1. Background and motivation
6.2. System model and problem formulation
6.2.1. Multi-device cooperative edge inference systems
6.2.2. Task-relevant feature extraction
6.2.3. Distributed feature encoding
6.2.4. Distributed information bottleneck principle
6.3. Variational approximation for feature extraction
6.3.1. Variational information bottleneck
6.3.2. DNN parameterization
6.4. Variational distributed feature encoding
6.4.1. Distributed deterministic information bottleneck (VDDIB) reformulation
6.4.2. Variational distributed deterministic information bottleneck
6.5. Distributed feature encoding with selective retransmission
6.5.1. Selective retransmission mechanism
6.5.2. The objective of VDDIB with selective retransmission
6.6. Simulation results
6.6.1. Experimental setup
6.6.2. Multi-view image classification
6.6.3. Multi-view object recognition
6.6.4. Ablation experiments
6.7. Summary
Part 4: Learning to optimize for processing latency reduction
Chapter 7: Learning to optimize via graph neural networks
7.1. Background on graph neural networks
7.2. Graph modeling of wireless networks
7.2.1. Directed graphs and permutation equivariance property
7.2.2. Wireless network as a graph
7.2.3. Graph modeling of multi-user interference channels
7.3. Neural network architecture design for radio resource management
7.3.1. Message passing graph neural networks
7.3.2. Key properties of MPGNNs
7.3.3. An effective implementation of MPGNNs
7.4. Theoretical analysis of MPGNN-based radio resource management
7.4.1. Simplifications
7.4.2. Equivalence of MPGNNs and distributed optimization
7.4.3. Performance and generalization of MPGNNs
7.5. Simulation results
7.5.1. Sum rate maximization
7.5.2. Weighted sum rate maximization
7.5.3. Beamforming design
7.6. Conclusions
Chapter 8: Learning to optimize via knowledge guidance
8.1. Background and motivation
8.2. System model
8.2.1. Federated learning model
8.2.2. Over-the-air aggregation
8.3. Convergence analysis and problem formulation
8.3.1. Convergence analysis
8.3.2. Problem formulation
8.4. Alternating optimization algorithm
8.4.1. Receive factor optimization
8.4.2. Transmit power optimization
8.5. Knowledge-guided learning algorithm
8.5.1. Knowledge-guided learning for transceiver design
8.5.2. Deep neural network design
8.5.3. Deep neural network training
8.6. Simulation results
8.6.1. Simulation setup
8.6.2. Performance comparison
8.7. Conclusion
Chapter 9: Learn to optimize via decentralized multi-agent reinforcement learning
9.1. Introduction
9.2. System model and problem formulation
9.2.1. System model
9.2.2. Problem formulation
9.2.3. Multi-agent system design
9.3. Proposed DEC-MAPC algorithm
9.3.1. Value decomposition technique
9.3.2. Network structure
9.4. Performance evaluation
9.4.1. Simulation setup
9.4.2. Sum-rate maximization
9.5. Conclusions
Part 5: Final part: conclusions
Chapter 10: Conclusions
10.1. Conclusions
Index

Yong Zhou

Yong Zhou received the B.Sc. and M.Eng. degrees from Shandong University, Jinan, China, in 2008 and 2011, respectively, and the Ph.D. degree from the University of Waterloo, Waterloo, ON, Canada, in 2015. From Nov. 2015 to Jan. 2018, he worked as a postdoctoral research fellow in the Department of Electrical and Computer Engineering, The University of British Columbia, Vancouver, Canada. He is currently an Assistant Professor in the School of Information Science and Technology, ShanghaiTech University, Shanghai, China. He was the track co-chair of IEEE VTC 2020 Fall and 2023 Spring, as well as the general co-chair of IEEE ICC 2022 workshop on edge artificial intelligence for 6G. He co-authored the book Mobile Edge Artificial Intelligence: Opportunities and Challenges (Elsevier 2021). His research interests include 6G communications, edge intelligence, and Internet of Things.

Affiliations and expertise

Shandong University, Jinan, China

Yinan Zou

Yinan Zou received the B.E. degree in electronic information engineering from Chongqing University, Chongqing, China, in 2020. He is currently pursuing the master’s degree with the School of Information Science and Technology, ShanghaiTech University, Shanghai, China. His research interests include learning to optimize, compressed sensing, and federated learning.

Affiliations and expertise

ShanghaiTech University, China.

Youlong Wu

Youlong Wu obtained his B.S. degree in electrical engineering from Wuhan University, Wuhan, China, in 2007. He received the M.S. degree in electrical engineering from Shanghai Jiaotong University, Shanghai, China, in 2011. In 2014, he received the Ph.D. degree at Telecom ParisTech, in Paris, France. In December 2014, he worked as a postdoc at the Institute for Communication Engineering, Technical University Munich (TUM), Munich, Germany. In 2017, he joined the School of Information Science and Technology at ShanghaiTech University. He obtained the TUM Fellowship in 2014 and is an Alexander von Humboldt research fellow. His research interests in Communication Theory, Information Theory and its applications e.g., coded caching, distributed computation, and machine learning.

Affiliations and expertise

ShanghaiTech University, China

Yuanming Shi

Yuanming Shi received the B.S. degree in electronic engineering from Tsinghua University, Beijing, China, in 2011. He received the Ph.D. degree in electronic and computer engineering from The Hong Kong University of Science and Technology (HKUST), in 2015. Since September 2015, he has been with the School of Information Science and Technology in ShanghaiTech University, where he is currently a tenured Associate Professor. He visited University of California, Berkeley, CA, USA, from October 2016 to February 2017. His research areas include optimization, machine learning, wireless communications, and their applications to 6G, IoT, and edge AI. He was a recipient of the 2016 IEEE Marconi Prize Paper Award in Wireless Communications, the 2016 Young Author Best Paper Award by the IEEE Signal Processing Society, and the 2021 IEEE ComSoc Asia-Pacific Outstanding Young Researcher Award. He is also an editor of IEEE Transactions on Wireless Communications, IEEE Journal on Selected Areas in Communications, and Journal of Communications and Information Networks.

Affiliations and expertise

ShanghaiTech University, China

Jun Zhang

Jun Zhang received the B.Eng. degree in Electronic Engineering from the University of Science and Technology of China in 2004, the M.Phil. degree in Information Engineering from the Chinese University of Hong Kong in 2006, and the Ph.D. degree in Electrical and Computer Engineering from the University of Texas at Austin in 2009. He is an Associate Professor in the Department of Electronic and Computer Engineering at the Hong Kong University of Science and Technology. His research interests include wireless communications and networking, mobile edge computing and edge AI, and cooperative AI. Dr. Zhang co-authored the book Fundamentals of LTE (Prentice-Hall, 2010). He is a co-recipient of several best paper awards, including the 2021 Best Survey Paper Award of the IEEE Communications Society, the 2019 IEEE Communications Society & Information Theory Society Joint Paper Award, and the 2016 Marconi Prize Paper Award in Wireless Communications. Two papers he co-authored received the Young Author Best Paper Award of the IEEE Signal Processing Society in 2016 and 2018, respectively. He also received the 2016 IEEE ComSoc Asia-Pacific Best Young Researcher Award. He is an Editor of IEEE Transactions on Communications, and was an editor of IEEE Transactions on Wireless Communications (2015-2020). He served as a MAC track co-chair for IEEE Wireless Communications and Networking Conference (WCNC) 2011 and a co-chair for the Wireless Communications Symposium of IEEE International Conference on Communications (ICC) 2021. He is an IEEE Fellow and an IEEE ComSoc Distinguished Lecturer.

Affiliations and expertise

The Hong Kong University of Science and Technology, Clear Water Bay, Kowloon, Hong Kong

Life Sciences

Physical Sciences & Engineering

Social Sciences & Humanities

Health

Machine Learning for Low-Latency Communications

Purchase options

Institutional subscription on ScienceDirect

Yong Zhou

Yinan Zou

Youlong Wu

Yuanming Shi

Jun Zhang