Associate Professor, Ionian University, Greece
Title of talk: ML - based Network Slicing in Next Generation Networks
Abstract:
Network slicing (NS) is a cornerstone of 5G, Beyond 5G, and emerging 6G systems, enabling customized,
end-to-end services for diverse performance requirements. Defined by ITU-T TS 28.530, the NS life-cycle
comprises four phases - Preparation, Commissioning, Operation, and Decommissioning - each involving complex
tasks such as slice design, admission control, resource allocation, performance monitoring, and parameter
optimization. This talk surveys recent advances in applying Machine Learning (ML) techniques to automate and
enhance these phases. It presents a structured mapping of ML methods, including supervised, unsupervised, and
reinforcement learning to specific life-cycle tasks, detailing their role in improving efficiency,
adaptability, and Quality of Service (QoS) assurance. By aligning ML applications with the ITU-T life-cycle
model, the talk offers a comprehensive view of how intelligent automation can address the dynamic challenges
of NS in next-generation networks.