Continuous Authentication in the Digital Age: An Analysis of Reinforcement Learning and Behavioural Biometrics
Master of Engineering Science
Electrical and Computer Engineering
Dr. Abdelkader Ouda
This thesis focuses on developing a continuous authentication system using behavioral biometrics to recognize users accessing computing devices. Keystroke dynamics captures the user's unique behavioral biometric, and reward-based reinforcement learning concepts are applied to recognize them throughout the session. The study shows that the proposed system offers an additional layer of security to traditional authentication methods, forming a robust continuous authentication system that can be added to static authentication systems.
Feature extraction technique is applied to enhance performance metrics, and the study concludes that behavioral biometrics are effective in identifying each user, with device information contributing significantly. The research has practical implications for the field of cybersecurity, providing real-world scenarios showing the benefits of enhancing the security of user authentication systems. This research contributes significantly to the field of cybersecurity, addressing challenging questions in building secure knowledge-based user authentication systems.
Summary for Lay Audience
Continuous authentication is a way to make sure that someone is who they say they are while they use a computer or other digital device. It's important because there are many bad people who try to steal information or cause problems online.
This study looked at using a type of computer learning called "reinforcement learning" and a way of analyzing a person's behavior called "behavioral biometrics" to make a better continuous authentication system. Reinforcement learning helps a computer learn to make decisions based on what it learns from trying different things, while behavioral biometrics looks at things like how someone types to figure out if it's really them.
The study found that using these methods together was better at detecting when someone was trying to pretend to be someone else than other ways of doing it. This means that it can help keep digital systems safe and secure without being too inconvenient for people to use. It can also help prevent bad people from accessing someone's computer if they leave it unattended.
Bansal, Priya, "Continuous Authentication in the Digital Age: An Analysis of Reinforcement Learning and Behavioural Biometrics" (2023). Electronic Thesis and Dissertation Repository. 9271.
Available for download on Thursday, August 01, 2024