Edge Learning for B5G Networks With Distributed Signal Processing: Semantic Communication, Edge Computing, and Wireless Sensing

You are here

Top Reasons to Join SPS Today!

1. IEEE Signal Processing Magazine
2. Signal Processing Digital Library*
3. Inside Signal Processing Newsletter
4. SPS Resource Center
5. Career advancement & recognition
6. Discounts on conferences and publications
7. Professional networking
8. Communities for students, young professionals, and women
9. Volunteer opportunities
10. Coming soon! PDH/CEU credits
Click here to learn more.

Edge Learning for B5G Networks With Distributed Signal Processing: Semantic Communication, Edge Computing, and Wireless Sensing

Wei Xu; Zhaohui Yang; Derrick Wing Kwan Ng; Marco Levorato; Yonina C. Eldar; Mérouane Debbah

To process and transfer large amounts of data in emerging wireless services, it has become increasingly appealing to exploit distributed data communication and learning. Specifically, edge learning (EL) enables local model training on geographically disperse edge nodes and minimizes the need for frequent data exchange. However, the current design of separating EL deployment and communication optimization does not yet reap the promised benefits of distributed signal processing, and sometimes suffers from excessive signalling overhead, long processing delay, and unstable learning convergence. In this paper, we provide an overview on practical distributed EL techniques and their interplay with advanced communication optimization designs. In particular, typical performance metrics for dual-functional learning and communication networks are discussed. Also, recent achievements of enabling techniques for the dual-functional design are surveyed with exemplifications from the mutual perspectives of “communications for learning” and “learning for communications.” The application of EL techniques within a variety of future communication systems are also envisioned for beyond 5G (B5G) wireless networks. For the application in goal-oriented semantic communication, we present a first mathematical model of the goal-oriented source entropy as an optimization problem. In addition, from the viewpoint of information theory, we identify fundamental open problems of characterizing rate regions for communication networks supporting distributed learning-and-computing tasks. We also present technical challenges as well as emerging application opportunities in this field, with the aim of inspiring future research and promoting widespread developments of EL in B5G.


A. Motivation of Edge Learning

Owing to the massive amount of data traffic for the role-out of the Internet-of-Everything (IoE), machine learning (ML) is envisioned to be an important technology to facilitate the evolution of beyond 5G (B5G) wireless networks [1]. Traditional ML methods need to centrally train data on a specific data center [2][3][4][5][6][7]. However, due to privacy concern and shortened wireless communication resource to support extensive data transfer, edge devices can hardly transmit the data that they have collected to a data center for executing centralized ML algorithms for data processing. This has triggered the fast-growing research field, namely edge learning (EL), which deeply integrates two main directions: wireless communications and ML. Advances in EL are widely expected to provide a platform to implement edge artificial intelligence (AI) in B5G networks [8][9][10][11][12][13].

B. Edge Learning in B5G Networks

The EL framework allows distributed implementation of ML algorithms over numerous edge devices that are controlled through multiple wireless servers to collaboratively train massive AI models utilizing local data and distributed processors, e.g., central processing units (CPUs) and graphic processing units (GPUs) [14][15]. Compared with distributed ML, EL refers to that multiple edge devices cooperatively train the ML model implemented over edge networks. The process of EL necessitates the download and upload of large-dimension ML parameters as well as their frequent updates among multiple edge devices. These new paradigms are expected to generate enormous data traffic, which can impose heavy burden to the already congested wireless communication networks [16]. This challenging issue cannot be addressed by using current wireless techniques aiming at capacity maximization, as they are decoupled from ML. Realizing the goal of EL with high communication efficiency requires advanced techniques of new distributed signal processing and wireless techniques that seamlessly integrate communications and learning approaches.



SPS on Twitter

  • DEADLINE EXTENDED: The 2023 IEEE International Workshop on Machine Learning for Signal Processing is now accepting… https://t.co/NLH2u19a3y
  • ONE MONTH OUT! We are celebrating the inaugural SPS Day on 2 June, honoring the date the Society was established in… https://t.co/V6Z3wKGK1O
  • The new SPS Scholarship Program welcomes applications from students interested in pursuing signal processing educat… https://t.co/0aYPMDSWDj
  • CALL FOR PAPERS: The IEEE Journal of Selected Topics in Signal Processing is now seeking submissions for a Special… https://t.co/NPCGrSjQbh
  • Test your knowledge of signal processing history with our April trivia! Our 75th anniversary celebration continues:… https://t.co/4xal7voFER

SPS Videos

Signal Processing in Home Assistants


Multimedia Forensics

Careers in Signal Processing             


Under the Radar