The paper presents an algorithm that relaxes the Differential Privacy (DP) guarantee of a randomized response gradually. The output from each relaxation maintains the same probability distribution as a standard randomized response with the equivalent DP guarantee. The algorithm is adaptable to any Local Differential Privacy (LDP) mechanisms that rely on randomized response. It has been integrated into RAPPOR, an LDP crowdsourcing string-collecting tool to enhance utility. It also facilitates the gradual relaxation of the DP guarantee for mean estimation based on randomized response. Numerical experiments validate the utility and DP guarantee of the algorithm.
Publication date: 26 Jan 2024
Project Page: Not provided
Paper: https://arxiv.org/pdf/2401.13952
