Revealing Information while Preserving Privacy保护隐私的同时透露信息.pptVIP

下载本文档

59
0
约 25页
2017-03-08 发布于上海
举报
版权申诉

Revealing Information while Preserving Privacy保护隐私的同时透露信息.ppt

1、原创力文档（book118）网站文档一经付费（服务费），不意味着购买了该文档的版权，仅供个人/单位学习、研究之用，不得用于商业用途，未经授权，严禁复制、发行、汇编、翻译或者网络传播等，侵权必究。。
2、本站所有内容均由合作方或网友上传，本站不对文档的完整性、权威性及其观点立场正确性做任何保证或承诺！文档内容仅供研究参考，付费前请自行鉴别。如您付费，意味着您自己接受本站规则且自行承担风险，本站不退款、不进行额外附加服务；查看《如何避免下载的几个坑》。如果您已付费下载过本站文档，您可以点击这里二次下载。
3、如文档侵犯商业秘密、侵犯著作权、侵犯人身权等，请点击“版权申诉”（推荐），也可以打举报电话：400-050-0827(电话支持时间：9:00-18:30)。
4、该文档为VIP文档，如果想要下载，成为VIP会员后，下载免费。
5、成为VIP后，下载本文档将扣除1次下载权益。下载后，不支持退款、换文档。如有疑问请联系我们。
6、成为VIP后，您将拥有八大权益，权益包括：VIP文档下载权益、阅读免打扰、文档格式转换、高级专利检索、专属身份标志、高级客服、多端互通、版权登记。
7、VIP文档为合作方或网友上传，每下载1次，网站将根据用户上传文档的质量评分、类型等，对文档贡献者给予高额补贴、流量扶持。如果你也想贡献VIP文档。上传文档

Revealing Information while Preserving Privacy保护隐私的同时透露信息

Privacy-Preserving Datamining on Vertically Partitioned Databases Kobbi Nissim Microsoft, SVC Privacy and Usability in Large Statistical Databases Kobbi Nissim Microsoft, SVC Data Privacy The Data Privacy Game: an Information-Privacy Tradeoff Private functions: E.g ?kobbi(DB)=dkobbi Information functions: want to reveal f (q, DB) for queries q Explicit definition of private functions The question: which information functions may be allowed? Crypto: secure function evaluation want to reveal f() want to hide all functions ?() not computable from f() Implicit definition of private functions Model: Statistical Database (SDB) Perturbation (Randomization Approach) Exact answer to query (q, f): aq,f=??i?q f(d1…dk) Actual SDB answer: aq,f Perturbation E: For all q,f: | aq,f – aq,f| ≤ E Questions: Does perturbation give any privacy? How much perturbation is needed for privacy? Usability Previous Work [Dinur, N] considered 1-attribute SDBs: Unlimited adversary: Perturbation of magnitude ?(n) required Polynomial-time adversary: Perturbation of magnitude ?(sqrt(n)) required In both cases, adversary may reconstruct a good approximation for the database Disallows even very week notions of privacy These results hold also for our model! Bounded adversary, restricted to T n queries (SuLQ): [D, Dwork, N] privacy preserving access mechanism with perturbation magnitude sqrt(n) Chance for usability Reasonable model as database grow larger and larger Previous Work - Privacy Definitions (1) X – data, Y – (noisy) observation of X [Agrawal, Srikant ‘00] Interval of confidence Let Y = X+noise (e.g. uniform noise in [-100,100]). Intuition: the larger the interval, the better privacy is preserved. Problematic when knowledge about how X is distributed is take into account [AA] [Agrawal, Aggarwal ‘01] Mutual information Intuition: the smaller I(X;Y) is, the better privacy is preserved Example where privacy is not preserved but mutual information does not show any trouble [EGS] Previ