Semantic Process Tracing for Wireless Intrusion Detection: A Petri Net and Knowledge Graph Approach

Wireless Sensor Networks (WSNs) that power critical infrastructure require accurate intrusion detection systems (IDS), so that any intended rogue activity can be spotted and effectively stopped. However, critical infrastructures also can greatly benefit from transparent, actionable explanations for alerts that happen behind-the-scenes in the intrusion detection mechanisms in  WSNs. Traditional IDS outputs opaque alerts that are hard to trust operationally due to their largely black-box nature [1].

This MSc project offers a solution to this problem. By combining semantic Petri nets with a Knowledge Graph [2] and building an Explainable AI scaffolding, we aim to produce alerts that explain the how and why of decisions made by security mechanisms that underpin IDS.

This project aims to build a semantic reasoning engine for intrusion detection in WSNs. It models attack progression using Petri nets and explains alerts using a domain-specific knowledge graph. In essence, the central goal of the project is to make AI-based security systems more transparent and traceable, so humans can understand the reasoning behind alerts and thereby manage well-informed security operations.

 References:

[1] Ramyavarshini, P., Sriram, G.K., Rajasekaran, U., and Malini, A. 2022. Explainable AI for Intrusion Detection Systems. In Proceedings of the 2022 5th International Conference on Contemporary Computing and Informatics (IC3I) (Madurai, India, Nov. 2022), 1563–1567. DOI: 10.1109/IC3156241.2022.10073356

[2] Liebald, B., Roth, D., Shah, N., and Srikumar, V. 2008. Proactive Intrusion Detection. In Proceedings of the Twenty-Third AAAI Conference on Artificial Intelligence (AAAI-08) (Chicago, Illinois, USA, Jul. 2008), 772–777.