SECURITY AND PRIVACY IN AI

Catalog description

This course studies adversarial attacks on discriminative and generative AI models that undermine system security and privacy, and investigates state-of-the-art mitigation strategies as well as the broader challenges of developing trustworthy AI. Students apply their insights through a semester-long project focused on strengthening AI robustness and reliability. Prerequisites: MATH2860 and MATH2100 or MATH2825 (4 credits) fall