Vir Phoha: Spoof-resistant authentication on mobile & wearable devices

Active authentication offers a non-intrusive second layer of defense to the static password, enabling the verification of users’ identity long after the password entry phase.

January 24, 2017
4:30 pm - 5:30 pm
Location
Kemeny Hall 007
Sponsored by
Computer Science Department
Audience
Public
More information
Sandra Hall

Active authentication offers a non-intrusive second layer of defense to the static password, enabling the verification of users’ identity long after the password entry phase. This second layer may use biometric patterns derived from sensor readings from keystrokes, gait, swiping on the phone, or brain activity patterns, etc.

I will present our recent results on developing principles of spoof-resistant active authentication and attribution using the unique patterns obtained from tactile, physiological, and neurobiological sensory information from mobile and wearable devices, such as smart phones, smart watches, and brain computer interface devices such as NeuroSky Mindwave. 

We have achieved Equal Error Rates (EER: when False Accept Rate and False Reject Rate are equal) in the range of 0.03 to 0.11 with accuracies above 94% in identity verification by fusing walking and swiping patterns, and 84% accuracy by fusing walking and typing patterns on mobile phones.  We have achieved EERs ranging from 0.09 to 0.15 on authentication using EEG data obtained through single channel NeuroSky device for activities that range from watching video to doing addition and subtraction.  We are also examining the vulnerabilities and adversarial exploitation of authentication and attribution using wearable devices and defenses against these exploitation(s).  For example, exploiting the correlation between thought and action we have been able to increase the EER by 300% of an identification system that uses either the thought or the action, thus exposing serious vulnerability of the system. I will present methods to guard against this vulnerability.

We are examining enhancing the attribution within the contexts imposed by communication structure and resource structure in adversarial red networks.  I will share some insights on feature behavior of sensory information obtained from different sensory sources within these contexts.

I will share thoughts on authentication in and vulnerabilities of futuristic augmented and virtual reality environments.

Dr. Vir Phoha is currently a Professor of Computer Science in the College of Engineering and Computer Science at Syracuse University. His current research addresses Cyberspace defense and offense mechanisms, in particular net-centric authentication systems (biometrics, authentication in networks, monitoring, and attribution); Machine Learning (Bayesian, reinforcement, evolutionary), and Data Mining.  Through funding from NSF, FBI, and DARPA he is exploring active or continuous authentication, attribution, and exploitation of vulnerabilities, and the corresponding defense mechanisms using wearable devices. Professor Phoha is an ACM Distinguished Scientist and is a Fellow of SDPS.  

Location
Kemeny Hall 007
Sponsored by
Computer Science Department
Audience
Public
More information
Sandra Hall