Wireless Future for 2020 and Beyond

Machine Learning Driven Wireless Networking

Organizers and Chairs:

George C. Alexandropoulos and Stathes Hadjiefthymiades,
National and Kapodistrian University of Athens, Greece

Scope of the Papers

The availability of large volumes of data collected from various wireless networking devices, ranging from on device and environmental sensors to network traffic monitoring nodes, and the recent advances in the computational capabilities of wireless network nodes have lately spurred data driven wireless communications and networking approaches. With the ardent goal to learn from the available data in order to take the appropriate actions given desired objectives, machine learning and artificial intelligence are lately gaining increasing popularity in wireless networks as a family of effective mathematical tools for capturing the underlying structure and regularities of collected data and designing accurate models from them.

The discussions for the upcoming 3GPP Release 17 and for beyond 5G wireless networks advocate the prominent role of machine learning in several layers of the network architecture enabling data processing and learning mechanisms for various purposes. Data processing is provisioned to take place locally (i.e., to or close to the wirelessly connected nodes where data is created) and centrally (i.e., where large volumes of data are integrated) for driving actions both on individual devices and the overall wireless network. In the network layer, machine learning contributes to network management (e.g., diagnostics and fault detection) and operation offering increased levels of autonomicity, while paving the way for efficient network function virtualization. Additionally, model-free and model-assisted machine learning approaches are recently being successfully deployed in the design of physical-layer methodologies profiting from the availability of relevant data and various forms of side information.

This Special Session focuses on the latest advances in machine learning for wireless communications and networking with the aim to contribute new tools, algorithms, and architectures for machine learning driven beyond 5G wireless networks. The special session will provide the opportunity for researchers and experts from academia, industry, and governmental bodies across the globe to share their latest research on the topic. Contribution on various layers of the 5G architecture are solicited with emphasis on both centralized and distributed implementations exhibiting either performance improvement or computational complexity reduction in comparison to the state of the art.

Suitable topics for this special session include, but are not limited to, the following applications of machine learning:

  • Interference management, coordination, and user association
  • Massive, distributed, and extreme MIMO systems
  • Full duplex systems and joint communication and sensing
  • Reconfigurable intelligent surfaces and meta-surfaces
  • Beamforming and beam alignment with hybrid transceiver architectures
  • Physical-layer security and privacy
  • Spectrum management, sensing, and shared access
  • Channel estimation and localization
  • Efficient hardware implementations for wireless nodes
  • Intelligent signal processing algorithms
  • Modeling and performance analysis of wireless networks
  • Traffic prediction and traffic classification
  • Network utility optimization, radio resource management and mobile edge intelligence
  • Multi-agent reinforcement learning for intelligent network control
  • Mobile user behavior analysis and inference
  • Ultra-low latency edge machine learning
  • Congestion control
  • QoS/QoE optimization
  • Wireless network security
  • Protocol reconfiguration
  • Distributed and federated learning for wireless networking
  • Emerging wireless applications, e.g., vehicle to everything (V2X), UAV-enabled communication, Internet of Things (IoT), virtual reality (VR), and augmented reality (AR)

Submission Guidelines

This Special Session will feature full papers only, as defined below.

Full Papers: Full paper submissions of original work (not previously published, or under review at another conference or journal) must not be longer than five pages and will be published in the conference proceedings.

Please use the IEEE template as described here. Accepted papers will be published in the conference proceedings and submitted to IEEE Xplore (approval pending).

Submissions are now accepted through EDAS: [Start a new submission here]

Important Dates

Paper submission deadline:   TBD
Notification of acceptance:     TBD
Camera-ready papers due:    TBD

Contact Us

George C. Alexandropoulos: alexandg@di.uoa.gr
Stathes Hadjiefthymiades: shadj@di.uoa.gr