Duration: 9 months
Deadline for applications: 14th March 2021
Starting date: April 1st or soon after
Salary: 50kEuros per year (pro rata)
A massive amount of sensors will be deployed in the Internet of things (IoT) framework and hence, intrusion detection schemes (IDS) in the wireless edge are vital to detect cyber threats. In this direction, one of the most challenging questions to address is the identification of the underlying cause that triggers such IDS alarm (e.g., malicious or accidental action). Root cause analysis (RCA) is the integrated procedure of the detection, localization and identification of the causes of anomalies. To tackle RCA, machine learning (ML) techniques are envisioned to bring a promising direction to build intelligent, adaptive and autonomous RCA systems. Along these lines, we will investigate in this project novel mechanisms to support RCA schemes in IoT networks, focusing on lightweight ML approaches. Furthermore, we will build novel IDS for challenging IoT environments, aiming at real-time schemes using agile, data-driven procedures based on change point (CP) analysis, to enable the detection of previously unseen attacks. Finally, these approaches will be extended to investigating the identification of “patient zero” in network cascades, with possible applications in Covid-19 back propagation analysis.
 Humayed, Abdulmalik, et al. “Cyber-physical systems security – A survey.” IEEE IoT Journal 4.6 (2017).
 Cheolmin Kim, Veena B. Mendiratta, Marina Thottan: Unsupervised Anomaly Detection and Root Cause Analysis in Mobile Networks. COMSNETS 2020: 176-183.
 Shruti Bothe, Usama Masood, Hasan Farooq, Ali Imran: Neuromorphic AI Empowered Root Cause Analysis of Faults in Emerging Networks. BlackSeaCom 2020: 1-6.
 A. Sridhar and H.V. Poor, “Sequential Estimation of Network Cascades”, IEEE Asilomar Conf. (2020).
 A. Das et al. “Root Cause Analysis of Failures in Interdependent Power-Communication Networks,” IEEE MILCOM, 2014, pp. 910-915.