Projects
Two projects anchor the hands-on portion of the course: a structured Midterm Project (20%) and a sustained, individual Capstone Project (35%).
Midterm Project — 20%
Weeks 6–8 (proposal out Week 6, report & presentation due Week 8). Centered on Pillar 2 (Security of AI), students select an AI system (real or synthetic) — ideally one of the AI-for-Security detectors they built in HW1–HW4 — build a threat model, and execute one or more adversarial attacks or implement a defense mechanism. The project culminates in a written report (5–10 pages) and a short in-class presentation (5–7 minutes). Graded on technical depth, correctness, clarity of presentation, and integration of threat modeling concepts.
Individual Capstone Project — 35%
Weeks 11–17 (phased). A sustained individual project that integrates across the three pillars: build (or adopt) an ML system that is effective for security (Pillar 1), then secure it against attack (Pillar 2) and make it trustworthy (Pillar 3) — for example, an ML-based detector that is then adversarially hardened and audited for fairness, interpretability, privacy, or certified robustness. Students conduct a comprehensive audit, design and implement improvements, and evaluate the system holistically across pillars. Two deliverables: a written report (15–20 pages) and a presentation (10–15 minutes) with slides and/or live demo, followed by Q&A. Graded on technical rigor, novelty, integration of course concepts across the three pillars, clarity, and reproducibility.
Milestones: Proposal (out Week 11, due Week 12) · Progress report (Week 14) · Final report & presentation (Week 17).
- Project brief (README)
- Grading rubric
- Milestone templates:
All projects follow the code submission standards and the same late-work policy (10% per day, up to 48 hours).