Exponential Randomized Response: Boosting Utility in Differentially Private Selection

01/11/2022
by   Gonzalo Munilla Garrido, et al.
0

A differentially private selection algorithm outputs from a finite set the item that approximately maximizes a data-dependent quality function. The most widely adopted mechanisms tackling this task are the pioneering exponential mechanism and permute-and-flip, which can offer utility improvements of up to a factor of two over the exponential mechanism. This work introduces a new differentially private mechanism for private selection and conducts theoretical and empirical comparisons with the above mechanisms. For reasonably common scenarios, our mechanism can provide utility improvements of factors significantly larger than two over the exponential and permute-and-flip mechanisms. Because the utility can deteriorate in niche scenarios, we recommend our mechanism to analysts who can tolerate lower utility for some datasets.

READ FULL TEXT
research
10/23/2020

Permute-and-Flip: A new mechanism for differentially private selection

We consider the problem of differentially private selection. Given a fin...
research
04/28/2016

Two Differentially Private Rating Collection Mechanisms for Recommender Systems

We design two mechanisms for the recommender system to collect user rati...
research
05/15/2021

The Permute-and-Flip Mechanism is Identical to Report-Noisy-Max with Exponential Noise

The permute-and-flip mechanism is a recently proposed differentially pri...
research
03/02/2021

DPlis: Boosting Utility of Differentially Private Deep Learning via Randomized Smoothing

Deep learning techniques have achieved remarkable performance in wide-ra...
research
11/19/2018

Private Selection from Private Candidates

Differentially Private algorithms often need to select the best amongst ...
research
09/08/2022

Majority Vote for Distributed Differentially Private Sign Selection

Privacy-preserving data analysis has become prevailing in recent years. ...
research
11/08/2018

Private Continual Release of Real-Valued Data Streams

We present a differentially private mechanism to display statistics (e.g...

Please sign up or login with your details

Forgot password? Click here to reset