Our project partner K3Y Ltd, presented the paper entitled “5GCIDS: An Intrusion Detection System for 5G Core with AI and Explainability Mechanisms” at the IEEE Global Communications Conference (GLOBECOM) 2023 held in Kuala Lumpur, during the second workshop on A4E: AI/ML for Edge/Fog Networks.
K3Y Ltd authored the paper which involves the development of PFCPFlowMeter – a new flow statistics generator which is capable of generating bidirectional flow statistics related to the application-layer attributes of the Packet Forwarding and Control Protocol (PFCP) used for control-plane communication at the N4 interface of the 5G core and for handling user sessions. The work also concerns the detection of PFCP-related attacks on said user sessions, through ML methods such as eXtreme Gradient Boosting (XGBoost) and decision tree. Lastly, the work considers the development of a new AI-powered IDS with an explainability mechanism leveraging the SHAP (SHapley Additive exPlanations) method.
