Control, Confidentiality, and the Right to be Forgotten

Control, Confidentiality, and the Right to be Forgotten
复制标题

控制、保密和被遗忘的权利

DOI:
10.1145/3576915.3616585
复制
发表时间:
2023
期刊:
2023
影响因子:
--
通讯作者:
Vasudevan, Prashant Nalini
Vasudevan, Prashant Nalini
中科院分区:
--
文献类型:
--
作者:
Cohen, Aloni;Smith, Adam;Swanberg, Marika;Vasudevan, Prashant Nalini

文献摘要

参考文献

被引文献

相似文献

最近的数字权利框架给予用户从存储和处理他们的个人信息的系统中删除他们的数据的权利(例如,GDPR中的“被遗忘权”)。在与许多用户交互并存储衍生信息的复杂系统中,删除应该如何正式化?我们认为,以前的方法是不够的。机器非学习的定义[6]范围太窄,不适用于一般的交互式设置。删除即保密的自然方法[15]限制太多:通过要求删除数据的保密性,它排除了社交功能。我们提出了一种新的形式主义:删除即控制。它允许用户在删除前自由使用数据,同时在删除后提出有意义的要求,从而给予用户更多的控制权。“删除即控制”提供了在不同设置中实现删除的新方法。我们将其应用于社会功能,并从文献中给出了各种机器学习定义的新的统一视图。这是通过一种新的自适应泛化的历史独立性。删除控制也提供了一种新的方法来实现机器学习的目标,即在尊重用户删除请求的同时维护模型。我们表明,发布一系列的更新模型,是私人差异下不断释放满足删除控制。这种算法的准确性不依赖于删除点的数量,与机器学习文献相反。
Recent digital rights frameworks give users the right to delete their data from systems that store and process their personal information (e.g., the "right to be forgotten" in the GDPR).How should deletion be formalized in complex systems that interact with many users and store derivative information? We argue that prior approaches fall short. Definitions of machine unlearning[6] are too narrowly scoped and do not apply to general interactive settings. The natural approach of deletion-as-confidentiality[15] is too restrictive: by requiring secrecy of deleted data, it rules out social functionalities.We propose a new formalism: deletion-as-control. It allows users' data to be freely used before deletion, while also imposing a meaningful requirement after deletion--thereby giving users more control.Deletion-as-control provides new ways of achieving deletion in diverse settings. We apply it to social functionalities, and give a new unified view of various machine unlearning definitions from the literature. This is done by way of a new adaptive generalization of history independence.Deletion-as-control also provides a new approach to the goal of machine unlearning, that is, to maintaining a model while honoring users' deletion requests. We show that publishing a sequence of updated models that are differentially private under continual release satisfies deletion-as-control. The accuracy of such an algorithm does not depend on the number of deleted points, in contrast to the machine unlearning literature.
DOI: --
发表时间: 2021-03
期刊: ArXiv
影响因子: --
作者:
Ayush Sekhari;Jayadev Acharya;Gautam Kamath;A. Suresh
通讯作者: Ayush Sekhari;Jayadev Acharya;Gautam Kamath;A. Suresh
弥合计算机科学与隐私法律方法之间的差距
DOI: --
发表时间: 2018
期刊: Harvard Journal of Law & Technology
影响因子: --
作者:
Kobbi Nissim;A. Bembenek;Alexandra Wood;Mark Bun;Marco Gaboardi;Urs Gasser;David O'Brien;S. Vadhan;T. Steinke
通讯作者: T. Steinke
DOI: --
发表时间: 2018
影响因子: --
作者:
A. Cohen;Kobbi Nissim
通讯作者: Kobbi Nissim
关于差异隐私的“语义”:贝叶斯公式
DOI: 10.29012/jpc.v6i1.634
发表时间: 2008
期刊: J. Priv. Confidentiality
影响因子: --
作者:
S. Kasiviswanathan;Adam D. Smith
通讯作者: Adam D. Smith
DOI: 10.1136/ebmh.11.4.102
发表时间: 2008-10
期刊: Evidence Based Mental Health
影响因子: --
作者:
P. Cochat;L. Vaucoret;J. Sarles
通讯作者: P. Cochat;L. Vaucoret;J. Sarles