Administrative Information
| Title | Privacy Preserving Machine Learning |
| Duration | 60 min |
| Module | C |
| Lesson Type | Interactive Session |
| Focus | Technical - Future AI |
| Topic | Open Problems and Challenges |
Keywords
Encryption, PPML, Privacy-Preserving, Risk,
Learning Goals
- Understand the concept of Privacy Preserving Fundamentals
- Discuss application of PPML in practical Scenarios
Expected Preparation
Learning Events to be Completed Before
Obligatory for Students
Optional for Students
- Review the article on Robust De-anonymization of Large Sparse Datasets
References and background for students
- For a deeper dive into differential privacy and privacy attacks
- The Algorithmic Foundations of Differential Privacy
- Exposed! A survey of Attacks on Private Data
Lesson materials
- This blog post on Oblivious' site gives a nice introduction to attacks on privacy.
- The PPML playlist created by CeADAR provides a high-level outline of how privacy-preserving techniques can be applied to data.
- PPML challenge problem overview
- PPML challenge problem instructions
Instructions for Teachers
Encourage students to make notes and initiate the discussion on the need for privacy-preserving ML techniques Background information here can be used. $ Collate and share all points raised and discussed by students. $ Provide an overview of the discussion.
| Duration (min) | Description | Concepts |
|---|---|---|
| 5 | Introduction to Privacy Preserving Fundamentals | Differential Privacy |
| 10 | Discussion on the need for the preservation of privacy | PPML |
| 25 | Netflix/IMDB linkage attack can be discussed here | a case stuy |
| 10 | Challenges of PPML techniques | |
| 5 | Conclusion |
Acknowledgements
The Human-Centered AI Masters programme was Co-Financed by the Connecting Europe Facility of the European Union Under Grant №CEF-TC-2020-1 Digital Skills 2020-EU-IA-0068.