Week 13: Privacy in ML Pipelines and Trade-Offs, plus Exam 3
Dates: Apr 5-9 · Reading: Handout 11: Privacy-Preserving ML and Trade-Offs
Learning Objectives
- Explain privacy-preserving machine learning techniques
- Discuss the tension between security and privacy
- Understand real-world privacy dilemmas in security operations
- Apply ethical reasoning to ML deployment decisions
Monday Session
Privacy-preserving ML: differential privacy, federated learning, and encrypted computation. Security versus privacy: why stronger encryption can make detection harder. Real-world dilemmas.
Wednesday Session
Exam 3 (online, non-cumulative, covers Weeks 10-13). Quiz 10 administered the same week.
Lab
Lab 11: Differential Privacy. Implement a simple differential privacy mechanism that adds noise before training, and see the privacy/accuracy trade-off.
Quiz / This Week
Quiz 10. Privacy-preserving ML; differential privacy; federated learning; security vs. privacy trade-offs.