The paper discusses how Differential Privacy (specifically DPSGD from ) impacts model performance for underrepresented groups.
Configure environment by running:
pip install -r requirements.txt
We use Python3.7 and GPU Nvidia TitanX.
File playing.py allows run the code. It uses
to set parameters from the paper and builds a graph on Tensorboard.
For Sentiment prediction we use
- MNIST (part of PyTorch)
- Diversity in Faces (obtained from IBM here)
- iNaturalist (download from here)
- UTKFace (from here)
- AAE Twitter corpus (from here)
compute_dp_sgd_privacy.py copied from public repo
DP-FedAvg implementation is taken from public repo
Implementation of DPSGD is based on TF Privacy repo and papers:
 M. Abadi, A. Chu, I. Goodfellow, H. B. McMahan, I. Mironov, K. Talwar, and L. Zhang. Deep learning with differential privacy. In CCS, 2016.
 H. B. McMahan and G. Andrew. A general approach to adding differential privacy to iterative training procedures. arXiv:1812.06210, 2018
 H. B. McMahan, D. Ramage, K. Talwar, and L. Zhang. Learning differentially private recurrent language models. In ICLR, 2018