Future Network Artificial Intelligence and Machine Learning Workshop
27-29 September 2021 | Virtual & Free
This workshop will explore Artificial Intelligence and Machine Learning (AI/ML) algorithms, techniques, systems and standards that can be utilized to optimize performance of 5G and Future Networks. The objective of this workshop is to bring together experts from the globe and create a joint platform for information exchanges, presentation of results, and fruitful discussions to identify gaps and future directions for the IEEE Future Network Initiative (FNI) Artificial Intelligence and Machine Learning Group.
Day 1 - 27 September, 12:00-03:00pm ET US
|12:00-12:20||Future Networks Initiative Working Groups Introduction||Ashutosh Dutta (JHU/APL) and Deepak Kataria (IP Junction, Inc.)|
|12:20-12:50||Cascade: An Extensible Platform for Low-Latency Edge Intelligence||Ken Birman (Cornell University)|
|12:50-13:10||End-to-End Wireless Communications Systems Using Deep Machine Learning||Mohammad Reza Ghavidel Aghdam (University of Tabriz, Iran)|
|13:10-13:30||A GCN-TOPSIS framework to Predict User-specific Beam pattern for Smart 5G-market verticals||Tanu Wadhera (NIT, Jalandhar, India)|
|14:00-14:20||Enabling Intelligence in Future Wireless (Mobile) Networks from Theory to Practice||Md Arifur Rahman (IS-Wireless, Poland)|
|14:20-14:40||Data-Driven Advanced Symbol Modulation in the PHY Layer||Ahmet Serdar Tan (InterDigital Europe, United Kingdom)|
|14:40-15:00||Artificial intelligence based tools and techniques for 5G||Manjunath Ramachandra (Wipro, India)|
Day 2 - 28 September, 12:00-03:00pm ET US
|12:00-12:05||Introduction||Ezabo Baron and Sanjay Pawar|
|12:05-12:35||AI/ML for 5G System Operation and Applications||Aladdin Saleh (Rogers Communications, Canada)|
|12:35-12:55||5G Radio Access Network (RAN) Slicing with Deep Reinforcement Learning||Yalin E Sagduyu and Tugba Erpek (Intelligent Automation, Inc.)|
|12:55-13:15||5G Standards & AI/ML - Ecosystem, architecture, and opportunities||Egemen K. Çetinkaya and Michael Salmon (Verizon, USA)|
|13:15-13:35||Building O-RAN Architected E2E Intelligence in a 5G and Beyond Network||Rajarajan Sivaraj (Mavenir, USA)|
|13:35-13:50||A GCN-TOPSIS framework to Predict User-specific Beam pattern for Smart 5G-market verticals||Tanu Wadhera (NIT, Jalandhar, India)|
|14:00-14:20||AI/ML for Network Life Cycle management - a Holistic Perspective||Deepak Das (Federated Wireless)|
|14:20-14:40||Smart Environment using Iot Based on Machine Leaning and Big Data||Brahim Lejdel (University of EL-Oued, Algeria)|
|14:40-15:00||TinyML: The Enabler of Robust Miniature Machine Learning Ecosystem||Anshul A Prasad (Symbiosis Institute of Technology, India)|
Day 3 - 29 September, 12:00-03:00pm ET US
|12:00-12:05||Introduction||Michael A. Enright (Quantum Dimension, Inc.)|
|12:05-12:35||Hybrid Cloud Data Pipelines to Meet the Requirements of Telco Network
and Customer Analytics and AI
|Ramesh Nagarajan and Krishnamurthy Srinivasan (Google)|
|12:35-12:55||Methods for Enhancement in Security of Digital Image||Sudhanshu S. Gonge (Symbiosis Institute of Technology, India)|
|12:55-13:15||Next generation mobile network security||Ali Abdollahi (Senior Consultant, Iran)|
|13:15-13:35||Edge Automation and Security, Issues and Challenges using AI/ML||Sanjay Pawar (Usha Mittal Institute of Technology, SNDT University)|
|13:35-13:50||Secure Federated Learning with a Zero Trust Architecture||Michael A. Enright (Quantum Dimension, Inc.)|
|14:00-14:20||Securing AI/ML Models in Distributed Environments: Challenges and Methods||Oussama H. Hamid (University of Nottingham, United Kingdom)|
|14:20-14:40||Practical offense and defence against 5G Infrastructure||Ali Abdollahi (Senior Consultant, Iran)|
|14:40-15:00||AI Enabled Predictive Security for 5G Networks||Ashutosh Dutta (JHU/APL)|
Many of the AI/ML techniques are focused on application areas that are not directly linked to 5G networks. For example, AI/ML for computer Vision has been a hot research area since it applies directly to Smart Cars and autonomous driving. Here, Deep Neural Networks (DNNs) are used to detect and classify different objects in an image scene by using spatial correlation and processing. On the other hand, Natural Language Processing (NLP) and similar applications use Recurrent Neural Networks (RNN) to capture the temporal relationships of the input signal. However, 5G networks have a richer set of technology needs that include network automation, security, physical layer optimizations and more. The goal of this track is to describe new approaches and techniques that are targeted for 5G performance improvement, autonomy, etc.
Invited research areas of interest include, but are not limited to:
- Application of AI/ML architectures such as DNN, RNN and Transfer Learning (TL)
- Data collection, sources and potential sharing of such for AI/ML systems
- Autonomous verification methods to ensure that AI/ML models are effective and have not been compromised
- Methods and use-cases for the use of AI/ML in network automation
- Techniques for the use of AI/ML in 5G security and privacy to solve problems such as network intrusion detection and prevention, botnet detection etc.
- Methods to secure the AI/ML models in a distributed environment
- Advances in AI/ML for the physical layer including spectrum awareness, Dynamic Spectrum Access (DSA) and resource allocation.
- Description and analysis of current supervised and/or unsupervised learning methods and the potential application to 5G systems, e.g. the application of network data to existing DNN architectures, such as GoogleNet, AlexNet, YOLO, etc.
- Autonomous orchestration and optimization of virtualized and NFV as applicable for 5G and beyond architectures, use-cases and market verticals (uRLLC, mMTC, eMBB, MEC).
- AI/ML architectures for lower power and complexity devices such as Internet of Things
- Future advances in AI/ML frameworks to support 5G market verticals
- AI/ML techniques and implementations for 5G network slicing and market verticals
In the Industry Applications track, we invite speakers to discuss the industrial deployments of security paradigms and their adoption by stakeholders in the value chain, such as operators, vendors, integrators, and the like. The objective of this track is to understand stakeholders' strategies and plans around 5G AI/ML specifically to support the evolution of 5G deployments and the different 5G market verticals. We encourage contributors to share their experiences and how it may be beneficial to 5G and Future Networks. Descriptions of current work on pre-market 5G testbeds, applications, etc. that have been developed through organizations such as ITU, ETSI, 3GPP 5G PPP, and others are also welcomed in this track.
Invited contributions include but are not limited to the following:
- Description of 5G testbeds and applications that have been developed or are currently under development through organizations such as 5G PPP
- The effect, challenges and options with regard to AI/ML of deploying open systems such as Open RAN, Open Core, etc.
- Existing and under-development AI/ML technologies that are relevant and could benefit 5G and future networks
- Practical approaches for scalable, distributed AI/ML implementation and orchestration across the same and different 5G networks
- Interoperability AI/ML-based security challenges and approaches considering evolution of deployments, interconnectivity with public/private cloud providers
- Results and reports of recent studies and pilot projects considering 5G AI/ML
Standards & Architecture
5G technologies provide ubiquitous connectivity while also addressing the demands of both individual consumers and businesses. In order to support various 5G use-cases and applications, there is a critical need to design autonomous systems using AI/ML. 5G networks need to be flexible, adaptive, scalable and able to dynamically react to the changes in the network quite rapidly which will require advanced automation technologies in future networks. Various standards bodies including 3GPP, IEEE, and ETSI have been looking into AI/ML for 5G networks. To that end, the IEEE FNI AI/ML Working Group is calling presenters to provide and share insights around progress and gaps in 5G AI/ML standards and architectures as relevant to 5G deployments and use-cases. It is the intent of the FNI AI/ML WG to develop a roadmap that will evolve into a set of standards to address the technology gaps in 5G and Future Network using AI/ML architectures and techniques.
To that end, this area will investigate current work that includes:
- Standards bodies (IEEE, ITU, ETSI, 3GPP, etc.) AI/ML results and developments for 5G and future networks
- Discussion of national and international organization plans to implement AI/ML for different applications - security, spectrum efficiency and awareness, etc.
- Standards for AI/ML automation and orchestration that addresses the different components of the 5G architecture and 5G enabled market verticals
- Different AI/ML implementation architectures, options and trade-offs in across the 5G network use-cases and relevant market verticals