Federal Grant Award Summary

Texas Tech University received a $344,417 National Science Foundation (NSF) Computer and Information Science and Engineering (CFDA 47.070) CAREER Project Grant, effective July 1, 2026 through June 30, 2031, to conduct research on differential privacy mechanisms and their interaction with machine learning utility and performance. The project develops a novel geometric framework for interpreting privacy-protection mechanisms based on noise perturbation through random geometry, random matrix methods, and learning theory. By representing privacy mechanisms as geometric objects such as random triangles and induced point clouds, the research provides visual and quantitative characterizations of privacy loss and utility degradation in high-dimensional spaces, enabling the derivation of tighter analysis tools for privacy amplification and composition while designing improved noise mechanisms that optimize privacy-utility trade-offs.

The funded research addresses the critical challenge of protecting personal data in artificial intelligence and machine learning systems affecting health, finance, and online services by analyzing and improving the trade-offs among model accuracy, convergence speed, and learnability in privacy-preserving learning pipelines. Beyond core research activities, the project supports broader impacts through workforce development via course development, student research involvement, and outreach activities that strengthen the scientific foundation of trustworthy data analysis and advance the practical implementation of privacy-preserving machine learning across critical application domains.