Schedule and slides (Fall 2022 class):

1. Introduction

2. A quick review on data analytics algorithms

3. Perturbation methods
- Additive perturbation
- Multiplicative perturbation
- Random response and other perturbations

4. Data anonymization
- k-anonymization,
- attacks and enhanced methods

5. Differential privacy 1, 2, local differential privacy

6. Cryptographic methods:
- Threat modeling for confidential computing
- Basic encryption methods, Secret sharing, Garbled circuits
- Partial homomorphic encryption, Somewhat HE/Fully HE: Ring-LWE
- Sample crypto protocols

7. Private information retrieval

8. Oblivious RAM: oram1, oram2

9. Trusted execution environments:
- SGX, Graphene-SGX,
- AMD-SEV,
- side-channel attacks

10. Privacy and security issues with deep learning
- Deep learning with differential privacy
- Confidential deep learning: attacks on models, confidential training, confidential inference