Event
Proposal Exam: Mahshid Noorani
Monday, December 12, 2022
5:00 p.m.
AVW 1146
Maria Hoo
301 405 3681
mch@umd.edu
ANNOUNCEMENT: Ph.D. Research Proposal Exam
Name: Mahshid Noorani
Committee:
Professor John Baras (Chair)
Professor Sennur Ulukus
Professor Qang Qu
Date/time: Dec 12, 2022 / 5:00 PM - 7:00 PM
Location: AVW 1146
Title: Data-Driven Anomaly Detection for Securing Cyber-Physical Systems
Abstract:
Detecting anomalous patterns provides us with useful and actionable information in a variety of real-world scenarios such as autonomous vehicles and robots. More importantly, detecting malicious anomalies could prevent catastrophic damages (e.g. in the case of a hacked autonomous vehicle). Accurate prediction of such malicious anomalies puts us ahead of hackers and malicious actors thus protecting the integrity, availability, and confidentiality of our systems. Two major issues with intrusion and anomaly detection systems are (1) lack of access to real-world data and (2) small data (lack of attack data). In this work, we investigate data-driven anomaly pattern detection methods for securing cyber-physical systems. We address the problem of real-world data by experimenting with and collecting data from unmanned outdoor field robots. We further focus on investigating solutions for addressing the problem of small data in the security of cyber-physical systems by exploring knowledge representation and reasoning and contrastive learning.
