EBTIC carries out research and development to meet the needs of next generation’s systems and services both on the network and the IT sides of the infrastructure. The research focuses on solutions that serves residential and business customers based on emerging wireline and wireless broadband technologies.
EBTIC Smart Infrastructure theme aims to provide EBTIC’s stakeholders with AI models supporting dynamic resource allocation and sharing at all layers of the 5G and IoT protocol stack. It will focus on AI models supporting intelligent slicing mechanisms, while guaranteeing key non-functional properties like the privacy and fairness of allocation decisions. The rationale is to use AI for automating key operations and processes, collecting and processing in (almost) real time data about the state and level of performance of network resources.
Our AI-assisted network management models will either benefit from trustworthy inputs, acquired by accurate positioning and robust timing/synchronization of network devices, or handle adversarial inputs via robust training. Also, they will use dynamic strategies for distributing model training computations, performing them in the cloud, at the network’s edge, or on mobile devices at the periphery.