Abstract: Currently, there are many research activities concerning the 6G network worldwide, which should be available commercially by 2030. These activities start with defining the 6G vision based on use cases and related requirements. According to many white papers, the 6G network should be deeply integrated with AI, leading to an ‘AI-native’ system in which AI will be used in multiple layers and deeply integrated. The use of AI for control loop-based management is nothing new. Such an approach has already been proposed for LTE-SON, 5G and other networking solutions to cope with management complexity, provide quick management reactions and implement so-called “self-functions” (self-configuration, self-healing, etc.).  There are, however, new requirements.

In the talk, synthetic requirements concerning 6G and their impact on management will be presented with the emphasized role of AI-native management operations, having FCAPS in mind. These new requirements will primarily deal with predictive management to provide high network reliability and with the Cloud Continuum and Network-of-Networks paradigms. The talk will include aspects of the AI-native implementation framework in the multistakeholder environment (towards AIaaS), discussing security and privacy issues. The use of Network Digital Twin and XAI will also be outlined, as well as the role of human-in-the-loop. A list of challenges related to the AI deployment cost (lightweight AI for resource-constrained devices), convergence time, multi-objective optimization, coordination, adaptability of AI to changing topologies and trust will be provided at the end of the talk.

Bio: Sławomir Kukliński

In 1994, Sławomir received a PhD with honours from the Warsaw University of Technology, and since then, he has been an assistant professor. His PhD thesis concerned neural networks (now part of AI). For about 20 years, he also works for Orange Polska as Research Expert. At the beginning of his career, he was involved in digital signal processing for military (radar systems) and non-military applications (voice compression and recognition). Later, he worked on wireless mesh networks (MANET), VANET (V2X), and SDN. For about 15 years, he has been involved in research concerning autonomic network management and mobile networks. He is working on network slicing 5G, 6G and autonomic and cognitive techniques applied to network management and control. As a principal investigator, he was involved in more than ten EU-funded projects, including FP6 MIDAS, FP7 4WARD, FP7 EFIPSANS, Celtic COMMUNE (Cognitive Network Management under Uncertainty), 5G!Pagoda, 5G-DRIVE, 5G!Drones, MonB5G and Hexa-X (6G flagship project in Europe). Actually, he is involved in the Hexa-X follow-up project, Hexa-X-II. He has also been engaged in ITU-T standardization concerning future networks (SG13). Slawomir Kukliński has published more than 90 scientific papers, has been a TPC member of many conferences and has given several invited keynotes.