Repository logo

Differentially Private Database Release via Kernel Mean Embeddings

Accepted version


Conference Object

Change log


Tolstikhin, Ilya 
Schölkopf, Bernhard 


We lay theoretical foundations for new database release mechanisms that allow third-parties to construct consistent estimators of population statistics, while ensuring that the privacy of each individual contributing to the database protected. The proposed framework rests on two main ideas. First, releasing (an estimate of) the kernel mean embedding of the data generating random variable instead of the database itself still allows third-parties to construct consistent estimators of a wide class of population statistics. Second, the algorithm can satisfy the definition of differential privacy by basing the released kernel mean embedding on entirely synthetic data points, while controlling accuracy through the metric available in a Reproducing Kernel Hilbert Space. We describe two instantiations of the proposed framework, suitable under different scenarios, and prove theoretical results guaranteeing differential privacy of the resulting algorithms and the consistency of estimators constructed from their outputs.



stat.ML, stat.ML

Journal Title

Proceedings of Machine Learning Research Volume 80:

Conference Name

International Conference on Machine Learning

Journal ISSN

Volume Title



MIR Press
EPSRC (1626332)