Personal data is being collected at an unprecedented scale by businesses and public organizations, driven by the progress of data science and AI. While such data can be turned into useful knowledge about the global population by computing aggregate statistics or training machine learning models, this can also lead to undesirable (sometimes catastrophic) disclosure of sensitive information. We must therefore deal with two conflicting objectives : maximizing the utility of data while protecting the privacy of individuals whose data is used in the analysis.

In this talk, I will present differential privacy (DP), a mathematical definition of privacy which comes with rigorous guarantees as well as an algorithmic framework that allows the design of practical privacy preserving algorithms for data analysis. In recent years, DP has become the gold standard in various fields and has recently seen several real-world deployments by companies and government agencies. Focusing on the central model of DP where a trusted curator wants to release the result of an analysis, I will introduce key algorithmic building blocks for privately answering simple queries, and briefly illustrate how they can be leveraged to construct private machine learning algorithms. Finally, if time permits, I will present recent contributions on estimating sample means and U-statistics in the decentralized model of differential privacy, where individuals or data owners do not trust a curator to handle their private data.