Project Opportunities and Risks of Biometric Security in the Age of Generative AI Program DUIRI - Discovery Undergraduate Interdisciplinary Research Internship Term Fall 2024 Status Accepted Research Area Global Security Description With the emergence of the Internet of Things (IoT), smart sensing devices ranging from smart wearables, such as Fitbits or Apple watches, to smartphones are packed with a range of sensors helping us with a range of services from unlocking cars and homes to validating financial transactions, among several other services. But often, these services are delivered based on a user’s sensitive personal information, including demographic identity and various biometric data, ranging from heart rate to breathing patterns. Therefore, researchers have been using this biometric data to validate a user. However, in the age of Generative AI, it is possible to produce synthetic data. Hence, it is important to understand how biometric samples can be fatal in predicting a user’s identity and his/her entire cyber-physical space, which has been secured with those biometrics in order to foster global security. Supervisor Sudip Vhaduri Mentor Type of work required Working with an interdisciplinary research team, in this project, students will first process various types of data, e.g., heart rate, gait, and breathing patterns, among several others obtained from smartphones and smart wearables. Then, students will visualize and compute different features. Finally, students will develop machine learning and AI models to generate synthetic data that can be used to breach an authentication model developed with the user’s original data. During this project, student researchers will be closely guided in every step, including problem formulation, data processing and characterizations, statistical analysis, and data visualization, model development, and interpretation of findings. Participating in this project, our student researchers will achieve technical expertise to solve real-world problems. The students will also get a clear idea of how their scientific discovery contributes to the entire community in terms of securing their cyber-physical space.
Websites https://sudipvhaduri.wordpress.com/user-authentication-project-2020-present/ Qualifications Students should be in good academic standing. Students are expected to have some experience with machine learning and/or AI modeling using Scikit learn, Tensorflow, or similar frameworks. Online Citi training/certification will be needed before starting the internship since this work involves human subjects. Credit Hours 0 Weekly Hours 12 (estimated)