Differential Privacy Mechanisms: A State-of-the-Art Survey

Abstract

The advantage of digitization is the availability of enormous data that make decision-making efficient and accurate. However, the data that improve the decision-making create a wide range of privacy concerns for users. The privacy-preserving data analysis is becoming a crucial research topic in the domain of computer science. One of the popular procedure used to ensure the privacy of data is anonymization where the identifiable information related to the users are removed before using the data for analysis. However, there are several issues associated with anonymization. In this article, we discuss the differential privacy mechanisms used to ensure the privacy of data. We formally discuss the definition of differential privacy, and then provide the seminal algorithms in the domain of differential privacy that enables the privacy-preserving data analysis. We discuss the applications of differential privacy. In addition, we present the state-of-the-art issues (or research gaps) in the domain of differential privacy and provide the future research directions.