Abstract:Local Differential Privacy (LDP) can provide each user with strong privacy guarantees under untrusted data curators while ensuring accurate statistics derived from privatized data. Due to its powerfulness, LDP has been widely adopted to protect privacy in various tasks (e.g., heavy hitters discovery, probability estimation) and systems (e.g., Google Chrome, Apple iOS). In particular, <math>(∊,δ)</math>-LDP has been studied in related statistical tasks like private learning and hypothesis testing, but is mainly achieved by using Gaussian mechanism, leading to the limited data utility. In this paper, we investigate several novel mechanisms that achieve <math>(∊,δ)</math>-LDP with higher data utility in collecting and analyzing users' data. Specifically, we first design two <math>(∊,δ)</math>-LDP algorithms for mean estimations on multi-dimensional numeric data, which can ensure higher accuracy than the optimal Gaussian mechanism. Then, we investigate different local protocols for frequency estimations on categorical attributes under <math>(∊,δ)</math>-LDP. Based on the proposed mechanisms, we further study on <math>(∊,δ)</math>-LDP-compliant stochastic gradient descent algorithms for machine learning models. Besides, the theoretical analysis of the error bound and the variance of the proposed algorithms are also presented in the paper. We have conducted extensive experiments on both real-world and synthetic datasets and demonstrated the high data utility of our proposed algorithms in the perspectives of simple data statistics tasks and complex machine learning tasks. The experimental results have shown that our proposed algorithms can effectively improve the data utility in different tasks while alleviating the privacy concerns of each individual.

Randomized Response Mechanisms for Differential Privacy Data Analysis: Bounds and Applications

Randomized Response Mechanisms for Differential Privacy Data Analysis: Bounds and Applications

A Statistical Framework for Differential Privacy

Optimal Differentially Private Mechanisms for Randomised Response

A Simple Algorithm for Estimating Distribution Parameters from $n$-Dimensional Randomized Binary Responses

Mechanisms for Robust Local Differential Privacy

Universal Exact Compression of Differentially Private Mechanisms

Local Differential Privacy for data collection and analysis

Statistical Inference in the Differential Privacy Model

The Cost of Privacy: Optimal Rates of Convergence for Parameter Estimation with Differential Privacy

Inferential Privacy Guarantees for Differentially Private Mechanisms

Constrained Differential Privacy for Count Data

Privacy Profiles for Private Selection

A New Analysis of Differential Privacy's Generalization Guarantees

Differential Privacy at Risk: Bridging Randomness and Privacy Budget

Differentially private and decentralized randomized power method

Bounded and Unbiased Composite Differential Privacy

A Unifying Privacy Analysis Framework for Unknown Domain Algorithms in Differential Privacy

Statistical Quantification of Differential Privacy: A Local Approach

Private measures, random walks, and synthetic data

AAA: an Adaptive Mechanism for Locally Differential Private Mean Estimation