Privacy-preserving AI for mental disease detection
We are trying to build an end-to-end encrypted deep learning pipeline for mental health disease detection using private patient data.
Technologies used so far:
- PyTorch
- PySyft - for federated learning
- diffprivlib - for differential privacy
- Quail - for synthetic data generation with noise