Examine secure inference mechanisms and privacy threats, from protected computation to model-inversion risk.
Explore privacy-preserving machine-learning systems through a progression from a general framework to secure nearest-neighbor classification and quantized neural-network evaluation, then examine model inversion as a concrete privacy threat in collaborative filtering. The course distinguishes computational protection from formal privacy guarantees and asks how threat models, leakage channels, and utility costs should be evaluated. The final attack-focused lesson is treated as security analysis, not as endorsement of extracting personal data. Videos have English audio; accompanying notes and assignments are original and do not reproduce transcripts.
Define an ML privacy threat model in terms of protected assets, adversary capabilities, trust assumptions, and observable outputs. Compare secure-inference approaches by protected computation, supported operations, deployment assumptions, and utility or performance cost. Analyze model inversion as an information-disclosure risk and design a responsible evaluation with safeguards and disclosure boundaries.
Prior familiarity with machine learning inference, basic security concepts, and classification is recommended. Comfort with threat models and performance evaluation will help; cryptography expertise is not required.
Experienced learners who want to deepen and apply advanced AI skills
People who learn best through examples, guided lessons, and hands-on practice
Professionals, creators, and independent builders looking for a repeatable workflow
Lesson 1 is free
4 lessons · Advanced · Full course.
This course uses CC BY licensed material attributed to its content provider or uploader. Lesson playback is available only through authenticated access after publication and licensing checks pass.