Fairness Certificates for Differentially Private Classification

10/28/2022
by   Paul Mangold, et al.
0

In this work, we theoretically study the impact of differential privacy on fairness in binary classification. We prove that, given a class of models, popular group fairness measures are pointwise Lipschitz-continuous with respect to the parameters of the model. This result is a consequence of a more general statement on the probability that a decision function makes a negative prediction conditioned on an arbitrary event (such as membership to a sensitive group), which may be of independent interest. We use the aforementioned Lipschitz property to prove a high probability bound showing that, given enough examples, the fairness level of private models is close to the one of their non-private counterparts.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
03/05/2022

The Impact of Differential Privacy on Group Disparity Mitigation

The performance cost of differential privacy has, for some applications,...
research
09/10/2020

Neither Private Nor Fair: Impact of Data Imbalance on Utility and Fairness in Differential Privacy

Deployment of deep learning in different fields and industries is growin...
research
07/27/2021

Statistical Guarantees for Fairness Aware Plug-In Algorithms

A plug-in algorithm to estimate Bayes Optimal Classifiers for fairness-a...
research
06/26/2023

Balanced Filtering via Non-Disclosive Proxies

We study the problem of non-disclosively collecting a sample of data tha...
research
08/27/2018

Downstream Effects of Affirmative Action

We study a two-stage model, in which students are 1) admitted to college...
research
06/22/2022

Input-agnostic Certified Group Fairness via Gaussian Parameter Smoothing

Only recently, researchers attempt to provide classification algorithms ...
research
05/15/2021

Fairly Private Through Group Tagging and Relation Impact

Privacy and Fairness both are very important nowadays. For most of the c...

Please sign up or login with your details

Forgot password? Click here to reset