Call for Papers

With the rapid development of machine learning and wireless communication technologies, intelligent applications and services, e.g., virtual personal assistants like Siri and Alexa, knowledge representation and reasoning to represent information about the world and solve complex tasks, have gained widespread popularity and large-scale implementation in our daily life: mobile entertainment, automotive, healthcare, education or industrial manufacture. These intelligent applications bring about unparalleled levels of transformation and benefits to human societies and national economies. Although the success of centralized machine learning has laid the foundation for many intelligent applications, the performance of the model usually depends on the availability of data. However, in most intelligent applications (e.g., intelligent transportation, smart finance), a large amount of useful data may be generated on multiple nodes and stored by multiple distributed devices, such as vehicles, smart phones and robots. Collecting such data to a central server for training will incur additional communication overhead, management and business compliance costs, privacy issues, and even regulatory and judicial issues (such as GDPR). Furthermore, it is usually impractical to require all the training data to be uploaded to the remote server with an increasing current network congestion, which hinder the applications of centralized machine learning in wireless networks.

As a distributed learning technology, Collaborative Machine Learning (CML) has been recently introduced to collaboratively train a model among multiple networking agents by using on-device computation. With the help of advanced communication technologies, e.g., 6G, a large number of networking agents can achieve timely communicate to share the latest model updates for obtaining high-performance learning model. Typical CML scenarios mainly include 1) federated learning and split learning that enable each agent performs local model updates and exchanges locally with the central server or neighbor agents to iteratively improve model accuracy, e.g., different users hold different private images to jointly train an image classifier; and 2) edge learning that edge agents perform parallelizing model training and distributed model co-inference with agent synergy and task offloading. By integrating the high-potential CML with advanced emerging technologies, e.g., edge computing, blockchain, 6G communication and networking, and quantum communication, intelligent applications are evolving towards next-generation intelligent applications that provide more efficient, intelligent, and secure services.

Although the next-generation intelligent applications dramatically enhance the life experience of humans and revolutionize modern business, there are still many open challenges that are unsolved when applying the fusion of CML and emerging technologies for next-generation intelligent applications. To achieve next-generation intelligent applications, CML needs significant research efforts on theories, algorithms, architecture, and experiences of system deployment and maintenance. Therefore, this Special Issue aims to offer a platform for researchers from both academia and industry to publish recent research findings and to discuss opportunities, challenges, and solutions related to collaborative machine learning. In particular, this Special Issue solicits original research papers about state-of-the-art approaches, methodologies, and technologies enabling efficient and practical collaborative machine learning towards the realization of next-generation intelligent applications. Potential topics of interest include but are not limited to the following:

• New architectures and frameworks of collaborative machine learning for next-generation intelligent applications
• Novel concept, theory, principles, and algorithms of collaborative machine learning for next-generation intelligent applications
• Resource management for collaborative machine learning in next-generation intelligent applications
• Privacy, trust and security issues in collaborative machine learning for next-generation intelligent applications
• Adaptive control management for “edge intelligence” and/or “intelligent edge”
• Incentive mechanism and crowd behavior study for collaborative machine learning in next-generation intelligent applications
• Channel modeling analysis on collaborative machine learning for next-generation intelligent applications
• Communication/Energy-efficiency or service throughput optimization issues on collaborative machine learning for next-generation intelligent applications
• Big data analysis and knowledge discovery from collaborative machine learning for next-generation intelligent applications
• Experimental studies on the convergence of collaborative machine learning for next-generation intelligent applications
• Use cases that highlight the open issues and/or potentials of collaborative machine learning for next-generation intelligent applications
• Emerging technologies (e.g., edge computing, blockchain, 6G or quantum communication) for collaborative machine learning in next-generation intelligent applications

Submission Guidelines

Prospective authors are invited to submit their manuscripts electronically, adhering to the IEEE Transactions on Network Science and Engineering guidelines. Note that the page limit is the same as that of regular papers. Please submit your papers through the online system and be sure to select the special issue or special section name. Manuscripts should not be published or currently submitted for publication elsewhere. Please submit only full papers intended for review, not abstracts, to the ScholarOne portal. If requested, abstracts should be sent by e-mail directly to the Guest Editors.

Important Dates

Manuscript Submission: 15 May 2021
First Review Round: 1 August 2021
Revision Papers Due: 15 September 2021
Acceptance Notification: 30 October 2021
Final Manuscript Due: 30 November 2021
Publication: 2022

Guest Editors