SPECIAL SECTION ON EDGE COMPUTING AND NETWORKING FOR UBIQUITOUS AI

Date of current version July 1, 2021.

Digital Object Identifier 10.1109/ACCESS.2021.3090143

EDITORIAL
IEEE ACCESS SPECIAL SECTION EDITORIAL:
EDGE COMPUTING AND NETWORKING
FOR UBIQUITOUS AI
With its rapid development recently, edge computing with approach for enhancing CNN kernel sparsity, enables
processing, storage, and networking capabilities has become power-efficient CNN inference on edge devices by reducing
an important solution to break through the bottleneck of power consumption of 3-D convolution operations.
emerging technology development by virtue of its advantages The article ‘‘Towards cost-efficient edge intelligent com-
in reducing data transmission, decreasing service latency, and puting with elastic deployment of container-based microser-
easing cloud computing pressure. Among several application vices,’’ by Zhao et al., formulates the container-based
scenarios such as network optimization, intelligent manufac- microservice deployment as a stochastic optimization prob-
turing, and real-time video analytics, edge computing can lem while developing a cost-aware elastic microservice
work with artificial intelligence (AI) synergistically. There- deployment algorithm to balance the tradeoff between system
fore, many researchers are investigating edge computing with cost and quality of service (QoS). Based on the Lyapunov
AI from two perspectives. One is that the emergence of optimization framework, the proposed algorithm makes
AI solves the optimization problem of edge computing. For real-time decisions based on the current queue backlogs and
example, when network devices need to process some com- system states without predicting the future knowledge.
plex and fuzzy information, the powerful learning and rea- In the article ‘‘Collaborative edge computing and caching
soning ability of AI can help to extract valuable information with deep reinforcement learning decision agents,’’ by
from the massive data and realize intelligent management. Ren et al., the authors investigate the coordination between
Another is how edge computing supports AI in a networking compute and cache resources in multi-level edge computing
environment. For example, AI training and inference can be nodes. This dynamic and complex system relies on double
efficiently enabled by a multitude of computing resources deep Q-learning (DDQN) to make decisions, with the ability
from edge computing. Therefore, edge computing and AI are to maximize long-term gains while making quick decisions.
mutually beneficial in networking. In the article ‘‘Deep learning for edge computing applica-
This IEEE ACCESS Special Section on Edge Computing tions: A state-of-the-art survey,’’ by Wang et al., the authors
and Networking for Ubiquitous AI aims to explore the chal- provide a comprehensive survey of the latest efforts on deep-
lenges of ubiquitous intelligence potentially enabled by both learning-enabled edge computing applications, and partic-
edge computing and machine learning, such as what can AI ularly offer insights on how to leverage advances in deep
do for edge computing, what edge computing can do for AI, learning to facilitate edge applications from four domains.
and novel applications of edge computing with AI. The authors suggest that the convergence of edge computing
The Call for Papers drew wide attention from the research and deep learning will bring new possibilities to both inter-
community and received 46 submissions. Out of these, disciplinary research and industrial applications.
14 articles were accepted for inclusion in the Special The article ‘‘An overview on edge computing research,’’ by
Section after a thorough review process involving at least Cao et al., reviews the related research on edge computing.
two independent expert referees. The 14 accepted articles It systematically introduces the edge computing model from
can be broadly categorized into three groups. The first, with the aspects of basic concepts, architecture, key technologies,
five articles, tackles what AI can do for edge computing. and security and privacy protection, and summarizes the
The second group of four articles address the problems of applications of edge computing.
what edge computing can do for AI. The remaining five The article by Chabbouh et al., entitled ‘‘Novel cloud-RRH
articles discuss novel edge computing/AI applications. architecture with radio resource management and QoS strate-
In the first group, the article ‘‘IKW: Inter-kernel weights gies for 5G HetNets,’’ proposes a cloud radio access network
for power efficient edge computing,’’ by Udupa et al., pro- problem in fifth-generation (5G) heterogeneous networks
poses the Inter-Kernel Weight (IKW) technique, which can (HetNets) architecture, using extra computing and storage
be used to eliminate redundant multiplications for a subset resources in network edges to improve the QoS of mobile
of kernel weights in a convolutional neural network (CNN) users and energy efficiency. The authors further consider
layer. The proposed IKW architecture, as an alternative two challenges, then study a two-stage remote radio-head

This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see [Link]
VOLUME 9, 2021 90933
 IEEE ACCESS SPECIAL SECTION EDITORIAL

clustering mechanism, as well a cost-based scheduling The article by Guo et al., titled ‘‘Intelligent and ubiquitous
scheme. Simulation results highlight superior performance. positioning framework in 5G edge computing scenarios,’’
In the second group of articles, ‘‘Multivariate seawater proposes a 5G positioning simulation scheme to cater to
quality prediction based on PCA-RVM supported by edge this requirement. The implementation of such a simulation
computing towards smart ocean,’’ by Sun et al., develops experiment includes scene generation, signal propagation
a hybrid multivariate prediction model for seawater quality simulation, and position estimation.
assessment in an edge computing environment, considering In the article ‘‘A power-efficient optimizing framework
the combination of principal component analysis (PCA) and FPGA accelerator based on winograd for YOLO,’’ by
relevance vector machine (RVM). Moreover, the case on Bao et al., the authors propose to convert multiplication oper-
real-world multivariate seawater data processing is imple- ations to addition operations in field-programmable gate
mented to evaluate the prediction ability and time consump- arrays (FPGAs), thereby reducing the computational com-
tion of the model. plexity of deep learning models.
‘‘Intelligent search and find system for robotic platform In the article ‘‘Lightweight online profiling-based config-
based on smart edge computing service,’’ by Barnawi et al., uration adaptation for video analytics system in edge com-
proposes a heterogeneous robotic system to facilitate the puting,’’ by Kim et al., the authors design a video analytics
development of advanced robotic applications with minimal system to adapt configurations for optimizing the resource-
human interactions. Concretely, they introduce traversal and accuracy tradeoffs of multiple video streams with respect
task division algorithms to support robots performing tasks to frame rate and resolution under the constraint of limited
in both near field and real time, contributing to intelligent resource capacity of a graphical processing unit (GPU)-
search-and-find as an edge service. enabled edge server, by utilizing the underlying character-
‘‘An intelligent path planning scheme of autonomous vehi- istics of the video objects. The above two articles exploit
cles platoon using deep reinforcement learning on network efficient deep learning acceleration.
edge,’’ by Chen et al., integrates edge computing and pla- In conclusion, we would like to thank all the authors
toon computing, and considers the task deadline and fuel who submitted their research articles to our Special Section.
consumption of each vehicle in the representative driving We highly appreciate the contributions of the reviewers for
pattern of autonomous vehicles. A case study of the proposed their constructive comments and suggestions. We would also
solution is presented, and the fuel consumption of vehicle like to acknowledge the guidance from the IEEE ACCESS
platoons is quantified while guaranteeing task deadlines. Editor-in-Chief, Editorial Board, and Editorial Staff.
‘‘An effective training scheme for deep neural network in
edge computing enabled internet of medical things (IoMTs) VICTOR C. M. LEUNG, Lead Editor
systems,’’ by Pustokhina et al., presents a new effective train- College of Computer Science and Software Engineering
ing scheme to facilitate timely data collection and processing Shenzhen University
to make timely decisions using the patterns that exist in the Shenzhen 518060, China
data. The model incorporates a hybrid modified water wave Department of Electrical and Computer Engineering
optimization (HMWWO) algorithm for tuning the parameters The University of British Columbia
of a deep neural network structure, and incorporates the abil- Vancouver, BC V6T 1Z4, Canada
ity to search a high-dimensional space and the competence in
the discovery of new probable candidate solutions. XIAOFEI WANG, Guest Editor
In the third group, the article by Sun et al., titled ‘‘Edge College of Intelligent Computing
computing in Internet of Things: A novel sensing-data recon- Tianjin University
struction algorithm under intelligent-migration strategy,’’ Tianjin 300350, China
proposes a novel sensing-data reconstruction algorithm under
intelligent-migration strategy (RdS-ImS). RdS-ImS can yield ABBAS JAMALIPOUR, Guest Editor
a data packet loss predictive model based on the compressed School of Electrical and Information Engineering
sensing theory, thus preventing data losses by finding alter- The University of Sydney
native paths for uploading sensing data to the cloud. Sydney, NSW 2006, Australia
In the article ‘‘Points-of-interest recommendation algo-
rithm based on LBSN in edge computing environment,’’ by XU CHEN, Guest Editor
Cao et al., the authors investigate how to accurately obtain School of Computer Science and Engineering
users’ places of interest based on location recommendations Sun Yat-sen University
in location-based social networks (LBSNs). Using the Friend- Guangzhou 510275, China
ship Bayesian Personalized Ranking matrix decomposition
framework to analyze the users’ personal preference func- SAMIA BOUZEFRANE, Guest Editor
tion in the edge server, the authors propose to embed the CEDRIC Lab
geographic information into the framework to get the candi- Conservatoire National des Arts et Métiers
dates’ points of interest. 75141 Paris, France

90934 VOLUME 9, 2021

IEEE ACCESS SPECIAL SECTION EDITORIAL

VICTOR C. M. LEUNG (Life Fellow, IEEE) is currently a Distinguished Professor of Computer

Science and Software Engineering with Shenzhen University, China. He is also an Emeritus
Professor of Electrical and Computer Engineering and the Director of the Laboratory for
Wireless Networks and Mobile Systems, The University of British Columbia (UBC), Van-
couver, BC, Canada. His research interests include wireless networks and mobile systems
and he has widely published in these above areas. He is also a fellow of the Royal Society
of Canada, the Canadian Academy of Engineering, and the Engineering Institute of Canada.
He is named in the current Clarivate Analytics list of ‘‘Highly Cited Researchers.’’ He received
the 1977 APEBC Gold Medal, the 1977–1981 NSERC Postgraduate Scholarships, the IEEE
Vancouver Section Centennial Award, the 2011 UBC Killam Research Prize, the 2017 Canadian
Award for Telecommunications Research, the 2018 IEEE TCGCC Distinguished Technical
Achievement Recognition Award, and the 2018 ACM MSWiM Reginald Fessenden Award.
He coauthored articles that won the 2017 IEEE ComSoc Fred W. Ellersick Prize, the 2017 IEEE
Systems Journal Best Paper Award, the 2018 IEEE CSIM Best Journal Paper Award, and the 2019 IEEE TCGCC Best
Journal Paper Award. He is also serving on the Editorial Boards of the IEEE TRANSACTIONS ON GREEN COMMUNICATIONS
AND NETWORKING, IEEE TRANSACTIONS ON CLOUD COMPUTING, IEEE ACCESS, IEEE Network, and several other journals.

XIAOFEI WANG received the B.S. degree from the Department of Computer Science and
Technology, Huazhong University of Science and Technology, in 2005, and the M.S. and
Ph.D. degrees from the School of Computer Science and Engineering, Seoul National Uni-
versity, in 2008 and 2013, respectively. He is currently a Professor with the Tianjin Key
Laboratory of Advanced Networking, School of Computer Science and Technology, Tianjin
University, China. He has participated in Korean and Canadian national research projects
focusing on future internet and wireless networking systems, e.g., the NESRC 5G SIRAS
Project, the KCA NDN Caching and Routing Project, and the NAP of KRCF Future Internet
Project. His current research interests include social-aware multimedia service in cloud com-
puting, cooperative backhaul caching, and traffic offloading in mobile content-centric networks.
Focusing on caching and offloading, he has published more than 110 technical articles in IEEE
Communications Magazine, IEEE Wireless Communications Magazine, IEEE TRANSACTIONS
ON MULTIMEDIA, ICC, and INFOCOM. He received the ‘‘Scholarship for Excellent Foreign
Students in IT Field’’ by the NIPA, Korea Government, from 2008 to 2014, and the ‘‘Global Outstanding Chinese Ph.D. Student
Award’’ by the Ministry of Education of China, in 2012.

ABBAS JAMALIPOUR (Fellow, IEEE) received the Ph.D. degree in electrical engineering
from Nagoya University, Nagoya, Japan. He is currently a Professor of Ubiquitous Mobile
Networking with The University of Sydney, Sydney, NSW, Australia. He has authored nine
technical books, 11 book chapters, over 550 technical articles, and five patents, all in the area
of wireless communications. He is also a Fellow of the Institute of Electrical, Information, and
Communication Engineers and the Institution of Engineers Australia, an ACM Professional
Member, and an IEEE Distinguished Speaker. He was a recipient of a number of prestigious
awards, such as the 2019 IEEE ComSoc Distinguished Technical Achievement Award in
green communications, the 2016 IEEE ComSoc Distinguished Technical Achievement Award
in communications switching and routing, the 2010 IEEE ComSoc Harold Sobol Award,
the 2006 IEEE ComSoc Best Tutorial Paper Award, and 15 best paper awards. He is also the
President of the IEEE Vehicular Technology Society. He held the position of the Executive
Vice-President and the Editor-in-Chief of VTS Mobile World and has been an Elected Member
of the Board of Governors of the IEEE Vehicular Technology Society, since 2014. He has been the General Chair or the
Technical Program Chair of a number of conferences, including IEEE ICC, GLOBECOM, WCNC, and PIMRC. He was the
Editor-in-Chief of IEEE Wireless Communications Magazine, the Vice President-Conferences, and a member of the Board of
Governors of the IEEE Communications Society. He also serves on the Editorial Board of IEEE ACCESS. He is also an Editor
of IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY and several other journals.

VOLUME 9, 2021 90935

 IEEE ACCESS SPECIAL SECTION EDITORIAL

XU CHEN received the Ph.D. degree in information engineering from The Chinese University
of Hong Kong, in 2012. He worked as a Postdoctoral Research Associate with Arizona State
University, Tempe, AZ, USA, from 2012 to 2014, and a Humboldt Scholar Fellow with the
Institute of Computer Science, University of Goettingen, Germany, from 2014 to 2016. He is
currently a Full Professor with Sun Yat-sen University, Guangzhou, China, and the Vice Director
of the National and Local Joint Engineering Laboratory of Digital Home Interactive Applica-
tions. He received the prestigious Humboldt Research Fellowship awarded by the Alexander von
Humboldt Foundation of Germany, the 2014 Hong Kong Young Scientist Runner-Up Award,
the 2017 IEEE Communication Society Asia-Pacific Outstanding Young Researcher Award,
the 2017 IEEE ComSoc Young Professional Best Paper Award, the Honorable Mention Award
of 2010 IEEE International Conference on Intelligence and Security Informatics (ISI), the Best
Paper Runner-Up Award of 2014 IEEE International Conference on Computer Communications
(INFOCOM), and the Best Paper Award of 2017 IEEE International Conference on Communi-
cations (ICC). He is also an Area Editor of the IEEE OPEN JOURNAL OF THE COMMUNICATIONS SOCIETY, an Associate Editor of
the IEEE TRANSACTIONS ON WIRELESS COMMUNICATIONS, IEEE INTERNET OF THINGS JOURNAL, and IEEE JOURNAL ON SELECTED
AREAS IN COMMUNICATIONS (JSAC) Series on Network Softwarization and Enablers.

SAMIA BOUZEFRANE received the Ph.D. degree in computer science from the University of
Poitiers, France, in 1998. After four years at the University of Le Havre, France, she joined
the CEDRIC Lab of Conservatoire National des Arts et Métiers (Cnam), Paris, in 2002. She
is currently a Professor with Cnam. She is the coauthor of many books (Operating Systems,
Smart Cards, and Identity Management Systems). She has coauthored more than 120 technical
articles. Her current research interests include the Internet of Things, vehicular networks, and
security using AI techniques. From 2017 to 2019, she led a research team on embedded and
mobile systems. Since 2019, she has been partly delegated to the French Ministry of Higher
Education, Research and Innovation.

90936 VOLUME 9, 2021