Skip to main content
Publications lead hero image abstract pattern


Publication Date

Third Quarter 2021

Manuscript Submission Deadline

Special Issue

Call for Papers

With various applications of Internet of Things (IoT) and smart city technologies flourishing nowadays, an enormous amount of data is being generated continually by tens of billions of mobile devices at the network edge, which give rise to challenges and opportunities for sensed data management and utilization in practical systems. Machine learning is a powerful tool widely adopted for distilling knowledge and recognizing useful patterns contained in the sensed data. Conventional machine learning such as deep learning algorithms often adopt a centralized scheme which requires the training data to be aggregated in a single cloud datacenter. Although such a centralize scheme has been applied in many digital transformation applications such as smart city and smart manufacturing, it has challenges due to ethical requirements, privacy concern, and scalability issues.

Recently, the Collaborative Distributed Machine Learning (CD-ML) techniques such as Federated Learning (FL) and partitioned learning have been proposed, which enable the decentralized training of a prediction model in a collaborative way. For example, Federated Learning (FL) is developed as a CD-ML approach to facilitate collaborative machine learning of complex models among distributed devices, where the exchange of raw data with an external server is not required. In FL, instead of uploading raw data to an edge server, devices compute model updates (i.e., updates on the model’s parameters) using local data and send them to the FL server for aggregation. The steps are repeated in multiple rounds until a desirable accuracy is achieved. The advancements of CD-ML can largely reduce the communication cost as well as latency while ensuring privacy protection. Generating, storing and processing data at the edge with global coordination in communication systems is made possible by the advanced technology of edge computing, where edge devices, such as sensors and smart phones, are usually equipped with storage and computation capabilities. Therefore, empowered by edge computing, unleashing the full potential of large-scale machine learning by exploiting data at the edge is without any doubt a promising approach for materializing the vision of “edge intelligence”. In addition, the CD-ML techniques can also be integrated into the edge computing framework to build intelligent edge systems for adaptive and real-time edge maintenance and management, i.e., “intelligent edge”. The realization of “edge intelligence” and “intelligent edge” will provide a platform for supporting low-latency, reliable, and intelligent communications in 5G and beyond.

The fusion of CD-ML and edge computing can benefit each other and realize the applications of “edge computing for CD-ML” and “CD-ML for edge computing”, which will contribute to the achievement of “edge intelligence” and “intelligent edge”, respectively. Both of them will play a key role in future intelligent communication systems. Despite the promising potential of the convergence of CD-ML and edge computing, the research area is still in its nascent stage and there remain many open research challenges to tackle. Relevant research issues range from the practical deployment of CD-ML in edge computing systems with different requirements to new network architectures for supporting edge computing to protocols customized for efficient CD-ML implementation. In addition, the application of CD-ML to optimize different functions of edge computing systems such as edge caching and offloading needs systematic investigations. The objective of this special issue is to introduce recent advances in convergence of CD-ML and edge computing, and its applications to tackle both fundamental and practical challenges in intelligent communication systems. This special issue will bring together leading researchers from both the industry and academia and solicit their original research and practical contributions on novel architectures, applications, and technologies of the convergence of CD-ML and edge computing towards the realization of “edge intelligence” and/or “intelligent edge”. Surveys and state-of-the-art tutorials are also welcome. Potential topics of interest include but are not limited to the following:

  • Architectures and frameworks on the convergence of CD-ML and edge computing
  • Novel concepts, principles, and algorithms on the convergence of CD-ML and edge computing
  • Resource management for “edge intelligence” and/or “intelligent edge”
  • Privacy, trust and security in “edge intelligence” and/or “intelligent edge”
  • Adaptive control management for “edge intelligence” and/or “intelligent edge”
  • Incentive mechanism for “edge intelligence” and/or “intelligent edge”
  • Channel modeling analysis on “edge intelligence” and/or “intelligent edge”
  • Energy-efficiency and service optimization for “edge intelligence” and/or “intelligent edge”
  • Big data analysis and knowledge discovery from “edge intelligence” and/or “intelligent edge”
  • Experimental studies on the convergence of CD-ML and edge computing
  • Use cases that highlight the potentials of the convergence of CD-ML and edge computing
  • Applications of the convergence of CD-ML and edge computing in intelligent communications
  • Emerging technologies for enhancing the convergence of CD-ML and edge computing

Submission Guidelines

Prospective authors should submit their manuscripts that conform to the standard format as indicated in the IEEE TCCN guidelines. All manuscripts to be considered for publication must be submitted to Manuscript Central according to the following schedule:

Important Dates

Submission Deadline: 1 December 2020 (Extended Deadline)
First Reviews Complete: 1 February 2021
Revision Due: 15 March 2021
Final Review Decision: 15 May 2021
Final to Publisher: 1 June 2020
Publication: Third Quarter 2021

Guest Editors

Dusit (Tao) Niyato (Lead)
Nanyang Technological University, Singapore

Kaibin Huang
The University of Hong Kong, Hong Kong

Mehdi Bennis
University of Oulu, Finland

Miao Pan
University of Houston, US

Zehui Xiong
Alibaba-NTU Singapore Joint Research Institute, Singapore

Long Bao Le
University of Quebec, Canada

Dong In Kim
Sungkyunkwan University, Korea

Li-Chun Wang
National Chiao Tun University, Taiwan