A major open research challenge is developing privacy-preserving machine learning methods that both achieves high performance and privacy guarantees even though the original training data contains sensitive personal information. This talk outlines the challenges and present synthetic data generation as one solution. A mobility data case study will be presented.
- What is privacy-preserving machine learning is all about?
- Synthetic data generation as a privacy preserving method to make sensitive data available
- Learning generative models to create synthetic data sets
- Mobility data case study