Speaker: Md. Delwar Hossain, Ph.D. Asst. Professor, Laboratory for Cyber Resilience Nara Institute of Science and Technology (NAIST), Japan.
Lecture Title: "Deep Learning-based Intrusion Detection for Securing Automotive and Industrial Control Systems"

Synopsis: The lecture covers security issues in modern automobiles and Industrial Control Systems and proposes Deep Learning, Federated Learning-based solutions to address them. The CAN bus system used in modern cars lacks basic security features, making it susceptible to attacks such as DoS, Fuzzing, and Spoofing. Similarly, the Modbus RS-485 protocol used in smart meters lacks authentication and encryption mechanisms, making it vulnerable to attacks. As a countermeasure, an intrusion detection system (IDS) using FL approach can effectively detect malicious activities and ensure data protection from intruders.

The lecture is structured as follows:                                                      

1. Security issues of modern automotive and ICS systems
2. Proposed defense verification platform for the CAN bus system
3. Development of a deep learning-based IDS
4. Development of automotive and Modbus attack datasets
5. Design of an effective pre-processing method to develop an efficient model regarding attack detection.
6. Evaluation of the significance of the hyper-parameter values to develop a robust IDS.

Bio: Md Delwar Hossain received the M.Sc. in Engineering in Information Systems Security degree from the Bangladesh University of Professionals and a Ph.D. degree in information science and engineering from the Nara Institute of Science and Technology (NAIST), Japan. He is currently an Assistant Professor with the Laboratory for Cyber Resilience at NAIST. He is a member of IEEE Communication Society. His research interests include cybersecurity, artificial intelligence, automotive security, smart grid security, industrial control systems security.

Join in Teams

Meeting ID: 332 856 987 988
Passcode: W8b7vr