The Discrete Gaussian for Differential Privacy

    Clément L. Canonne         Gautam Kamath          Thomas Steinke
    IBM Research, Almaden       University of Waterloo     IBM Research, Almaden
  ccanonne@cs.columbia.edu        g@csail.mit.edu      dgauss@thomas-steinke.net

                        Abstract
      A key tool for building differentially private systems is adding Gaussian noise to
      the output of a function evaluated on a sensitive dataset. Unfortunately, using a
      cont ...