Title: Differentially private sub-Gaussian location estimators
Authors: Victor-Emmanuel Brunel - ENSAE ParisTech (France) [presenting]
Marco Avella-Medina - Columbia University (United States)
Abstract: The focus is on the problem of estimating a location parameter with differential privacy guarantees and sub-Gaussian deviations. Namely, we propose estimators that achieve an error with sub-Gaussian tails and satisfy the standard differential privacy constraint, even when the data only have a few finite moments. Moreover, our method does not require the unknown location parameter to be bounded in a known region, unlike previous results.