Tutorial Reading Group 2024 #1

The first tutorial reading group 2024 will be held on 18th April 2024, Thursday 13:00-16:30 online via Teams. Mr. Zhen Chen will give us a talk on “Membership Inference Attacks and Differential Privacy”.

Teams Link

Meeting ID: 367 165 805 641

Passcode: oMwkcq

Reference:

• [1] Papernot N, Song S, Mironov I, et al. Scalable private learning with pate.

• [2] Papernot N, Abadi M, Erlingsson U, et al. Semi-supervised knowledge transfer for deep learning from private training data.

• [3] S. Yeom, I. Giacomelli, M. Fredrikson and S. Jha. Privacy Risk in Machine Learning: Analyzing the Connection to Overfitting.

• [4] Li Y, Tsai Y L, Yu C M, et al. Exploring the benefits of visual prompting in differential privacy.

• [5] Song L, Shokri R, Mittal P. Privacy risks of securing machine learning models against adversarial examples.

• [6] Song L, Mittal P. Systematic evaluation of privacy risks of machine learning models.

Related Codes:

MIAs:

• [1] https://github.com/inspire-group/membership-inference-evaluation

• [2] https://github.com/inspire-group/privacy-vs-robustness

• [3] https://github.com/fshp971/RPG

PATE and VP:

• [1] https://github.com/EzzzLi/Prom-PATE

Slides:

• Slides is available here!

Challenges:

• [1] More Flexible Ways to Find the Optimal $\tau_y$

• [2] Trade-off between the Training Iteration and Privacy Budget in the Training-based DP Methods

• [3] How to Create Dataset to Replace Public Dataset if No Public Dataset with Similar Distribution Exists

• [4] Evaluation on Visual Prompting-Trained Models

Contact:

If you have any ideas or questions related to this tutorial/presentation, you are most welcome to contact Mr. Zhen Chen (zhen.chen2@liverpool.ac.uk) for further discussion and collaboration.