Mohammed M. Alani
Professor of Cybersecurity
Rochester Institute of Technology - Dubai
Mohammed M. Alani is a cybersecurity researcher an professor with over 20 years of experience spanning UAE, Canada, Oman, Bahrain, and Jordan. His main field of research is applications of machine learning in cybersecurity, and IoT security.
Speaker sessions
Adversarial Explainability - Breaking explainable machine learning-based intrusion detection
In this session we will dive into a novel attack that utilizes machine learning explainability to target intrusion detection systems. The attack identifies the most effective features the ML-based intrusion detection system and find the specific values expected in normal traffic. Then, the attack proceeds to disguise the attack traffic to look like normal traffic to bypass detection.Our research and experiments proved that by changing one feature in attack traffic only, we can successfully bypass ML-based intrusion detection systems.
- 18:00
- Tue
- 02 Dec
Stage:
Briefings 1
Sessions Type:
Presentation