Privacy-Preserving Social Media Data Publishing.
arXiv: Social and Information Networks(2017)
摘要
User-generated social media data are exploding and also of high demand in public and private sectors. The disclosure of complete and intact social media data exacerbates the threats to user privacy. In this paper, we first identify a text-based user-linkage attack on current social media data publishing practices, in which the real users of anonymous IDs in a published dataset can be pinpointed based on the usersu0027 unprotected text data. Then we propose a framework for differentially privacy-preserving social media data publishing for the first time in literature. Within our framework, social media data service providers can publish perturbed datasets to provide differential privacy to social media users while offering high data utility to social media data consumers. Our differential privacy mechanism is based on a novel notion of $epsilon$-text indistinguishability, which we propose to thwart the text-based user-linkage attack. Extensive experiments on real-world and simulated datasets confirm that our framework can enable high-level differential privacy protection and also high data utility at the same time.
更多查看译文
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络