arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2411.03351cs.CRcs.AIcs.DB
收藏

带差分隐私的表格数据合成:综述

Tabular Data Synthesis with Differential Privacy: A Survey

CSIRO, Data61(联邦科学与工业研究组织 Data61);Nanyang Technological University(南洋理工大学);Xidian University(西安电子科技大学);Chongqing Normal University(重庆师范大学)

Mengmeng Yang, Chi-Hung Chi, Kwok-Yan Lam, Jie Feng, Taolin Guo, Wei Ni

AI总结:

本文综述了差分隐私表格数据合成方法,将其分为统计与深度学习两类,评估其在效用、隐私和计算复杂度上的优劣,并讨论数据质量评估方法及未来研究方向。

摘要:

数据共享是协作创新的前提,使组织能够利用多样化的数据集获得更深入的见解。在金融科技和智能制造等实际应用中,通常以表格形式存在的交易数据被生成和分析以产生见解。然而,此类数据集通常包含敏感的个人/商业信息,引发隐私担忧和监管风险。数据合成通过生成保留真实数据统计特征、去除与个人直接联系的人工数据集来解决这一问题。然而,攻击者仍可利用背景知识推断敏感信息。差分隐私通过提供可证明且可量化的隐私保护提供了一种解决方案。因此,差分隐私数据合成已成为一种有前景的隐私感知数据共享方法。本文全面概述了现有的差分隐私表格数据合成方法,重点介绍了在差分隐私约束下每种生成模型用于生成表格数据所面临的独特挑战。我们根据生成模型将这些方法分为基于统计的方法和基于深度学习的方法,并在集中式和分布式环境中对它们进行讨论。我们评估并比较了每个类别中的这些方法,强调了它们在效用、隐私和计算复杂度方面的优缺点。此外,我们提出并讨论了用于评估合成数据质量的各种评估方法,指出了该领域的研究空白和未来研究方向。

英文摘要:

Data sharing is a prerequisite for collaborative innovation, enabling organizations to leverage diverse datasets for deeper insights. In real-world applications like FinTech and Smart Manufacturing, transactional data, often in tabular form, are generated and analyzed for insight generation. However, such datasets typically contain sensitive personal/business information, raising privacy concerns and regulatory risks. Data synthesis tackles this by generating artificial datasets that preserve the statistical characteristics of real data, removing direct links to individuals. However, attackers can still infer sensitive information using background knowledge. Differential privacy offers a solution by providing provable and quantifiable privacy protection. Consequently, differentially private data synthesis has emerged as a promising approach to privacy-aware data sharing. This paper provides a comprehensive overview of existing differentially private tabular data synthesis methods, highlighting the unique challenges of each generation model for generating tabular data under differential privacy constraints. We classify the methods into statistical and deep learning-based approaches based on their generation models, discussing them in both centralized and distributed environments. We evaluate and compare those methods within each category, highlighting their strengths and weaknesses in terms of utility, privacy, and computational complexity. Additionally, we present and discuss various evaluation methods for assessing the quality of the synthesized data, identify research gaps in the field and directions for future research.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑