Locally Differentially Private Frequency Estimation with Consistency
This work addresses privacy-preserving data analysis for users and data collectors, offering incremental improvements to existing frequency oracle protocols.
The paper tackles the problem of improving accuracy in locally differentially private frequency estimation by adding post-processing steps that enforce non-negativity and sum-to-one constraints, achieving significantly better accuracy across various query tasks.
Local Differential Privacy (LDP) protects user privacy from the data collector. LDP protocols have been increasingly deployed in the industry. A basic building block is frequency oracle (FO) protocols, which estimate frequencies of values. While several FO protocols have been proposed, the design goal does not lead to optimal results for answering many queries. In this paper, we show that adding post-processing steps to FO protocols by exploiting the knowledge that all individual frequencies should be non-negative and they sum up to one can lead to significantly better accuracy for a wide range of tasks, including frequencies of individual values, frequencies of the most frequent values, and frequencies of subsets of values. We consider 10 different methods that exploit this knowledge differently. We establish theoretical relationships between some of them and conducted extensive experimental evaluations to understand which methods should be used for different query tasks.