EE/CSC 7700: ML for Security in Cyber-Physical Systems - Fall 2024

Instructor: Xugui Zhou, xuguizhou@lsu.edu, Office: ERAD Building 333
Office Hours: Friday 3-4 pm or by appointment

Time: M/W/F 10:30-11:20 am
Location: Patrick F. Taylor Hall 1218

This course provides a general exploration of the intersection between machine learning (ML) techniques and security principles within the context of cyber-physical systems (CPS). Cyber-physical systems integrate computational and physical components to monitor and control various processes, including industrial systems, autonomous vehicles, and medical devices. As these systems become increasingly interconnected and autonomous, ensuring their security and safety against cyber threats becomes paramount.

The course begins with an overview of fundamental concepts in machine learning and cybersecurity, laying the groundwork for understanding their application in CPS environments. Students will then delve into advanced ML algorithms and techniques tailored to address security challenges in CPS, such as anomaly detection, threat prediction, and hazard mitigation.

Throughout the course, students will engage in hands-on exercises, simulations, and projects to reinforce theoretical concepts and develop practical skills in ML-based security for cyber- physical systems. By the end of the course, students will be equipped with the knowledge and tools necessary to design, implement, and evaluate ML-driven security solutions tailored to the unique challenges of CPS environments.

Topics

• Introduction to cyber-physical systems and their security challenges
• Fundamentals of machine learning and its applications in CPS security
• Supervised, unsupervised, and reinforcement learning techniques
• Adversarial machine learning and its implications for CPS security
• Anomaly detection and outlier detection in CPS data streams
• Threat modeling and risk assessment in cyber-physical environments
• Runtime hazard prediction and mitigation for CPS control systems
• Secure machine learning techniques for protecting ML-enabled CPS
• Case studies and real-world applications of ML-based security solutions in CPS domains (e.g., autonomous driving)