Undergraduate researcher Daniel Muise and Senior Research Fellow Kobbi Nissim have published "Differential Privacy in CDFs," available for download.
The ultimate aim of this presentation is to familiarize social scientists with the errors introduced by differential privacy (DP), and to explain how to manage DP’s random noise. In this document, Undergraduate Daniel Muise and Senior Research Fellow Kobbi Nissim explain the effect of random noise introduced in DP-computations by making analogies to sampling error. We focus on the case of cumulative density functions (CDFs) and histograms.
 | dpcdf_user_manual_aug_2016.pdf | 5.65 MB |