Transcending the Usual Rationale for the Future of Ubiquitous NETworks – TURFU-NET

The TURFU-NET project orchestrated by Quentin Bramas at the University of Strasbourg represents a pioneering initiative set to redefine network management and optimization through the integration of cutting-edge neurosymbolic Artificial Intelligence (AI). Spanning four years, this collaborative effort includes esteemed partners from Strasbourg, Mulhouse, and Paris, combining expertise from academia and research to address critical challenges in current network systems.

Neurosymbolic AI, the cornerstone of this project, merges the robust, rule-based reasoning of symbolic AI with the dynamic data processing capabilities of neural networks. This hybrid approach aims to overcome the limitations of conventional generative AI models—namely their lack of deterministic guarantees and explainability in decision-making processes. In network environments where reliability and security are paramount, traditional AI methods fall short, offering only probabilistic outcomes without adequate transparency. TURFU-NET's application of neurosymbolic AI promises not only to enhance decision-making transparency but also to provide assured performance crucial in high-stakes network operations.

The project is structured into three main work packages, each designed to tackle distinct aspects of network management: Continuous Data Collection and Analysis; Automated Remediation; and Model Distillation and Distributed Self-Improving Networks.

The ultimate goal of TURFU-NET is to design a self-managed network capable of automatically detecting performance anomalies and adjusting the network state to restore the desired performance, as defined by the operator.

By aligning with ongoing efforts in digital infrastructure improvement, TURFU-NET contributes to the enhancement of network management practices. The project represents a step forward in applying AI in practical, high-impact ways within the realm of network systems, aiming for incremental improvements in efficiency and security.