A Risk-Based Continuous Authentication Engine Using a Probabilistic Model around Behavioral Biometrics

Traditional static authentication systems have a fundamental deficiency; it assumes the presence of the validated user through the length of the session. Continuous authentication algorithms periodically validate the identity of a user during the entire session. It relies on information that can be automatically extracted from the user such as biometrics and behavior patterns. A probabilistic approach can naturally model the noise and latent variables present in the data. The probabilistic output of such models is a confidence value. Risk-based authentication makes use of this value to define task-specific requirements. The proposed research project has two components: an anomaly-based Intrusion Detection System (IDS), and a risk-based authentication system that uses biometrics and behavior patterns. If our research in this area proves to be successful, we can hopefully replace the use of password challenges as a primary authentication factor, in exchange for continuous algorithms that are more reliable and less of hindrance to the end-consumer

Faculty Supervisor:

Eugene Fiume

Student:

Partner:

BlackBerry (Ottawa, ON)

Discipline:

Computer science

Sector:

Manufacturing

University:

University of Toronto

Program: