ASurveyofDifferentialPrivacyinDataPublicationandAnalysis.pdfVIP

面向数据发布和分析的差分隐私保护研究 张啸剑, 孟小峰 ( 中国人民大学信息学院 北京 100872) 摘 要 随着数据分析和发布等应用需求的出现和发展，如何保护隐私数据和防止敏感信息泄露成为当前面临的重大挑战。 基于k-匿名或者划分的隐私保护方法，只适应特定背景知识下的攻击而存在严重的局限性。差分隐私作为一种新出现的隐私 保护框架，能够防止攻击者拥有任意背景知识下的攻击并提供有力的保护。文中对差分隐私保护领域已有的研究成果进行了 总结，对该技术的基本原理和特征进行了阐述，重点介绍了当前该领域的研究热点：差分隐私下基于直方图的发布技术、基 于划分的发布技术以及回归分析技术。在对已有技术深入对比分析的基础上，指出了差分隐私保护技术的未来发展方向。 关键词 差分隐私；数据发布；隐私保护；数据分析 中图法分类号 **** DOI 号：*投稿时不提供DOI 号* A Survey of Differential Privacy in Data Publication and Analysis* ZHANG Xiao-Jian, MENG Xiao-Feng (School of Information, Renmin University of China, Beijing 100872) Abstract As the emergence and development of application requirements such as data analysis and data publication, a challenge to those applications is to protect private data and prevent sensitive information from disclosure. However, most existing methods based on k-anonymity or partition-based have serious limitations because they only preserve individuals’ privacy under special assumption of adversary’s background knowledge. Differential privacy has emerged as a new paradigm for privacy protection with strong privacy guarantees against adversaries with arbitrary background knowledge. This paper surveys the state of the art of differential privacy for data publication and analysis. The mechanisms and properties of this model are described, while our focuses are put on private data releases in terms of histogram and partition techniques, and analysis based on regression skills. Following the comprehensive comparison and analysis of existing works, future research directions are put forward. Key words Differential Privacy; Data Publication; Privacy-preserving; Data Analysis 着敏感信息的数据，而不采用适当数据保护技术， 1 引 言