• 主题
高级搜索  >
历史搜索 清空历史
首页> 外文会议>Discovery science >Differentially Private Empirical Risk Minimization with Input Perturbation
【24h】

Differentially Private Empirical Risk Minimization with Input Perturbation

机译：具有输入扰动的差异化私人经验风险最小化

获取原文
获取原文并翻译 | 示例
页面导航
  • 摘要
  • 著录项
  • 相似文献
  • 相关主题

摘要

We propose a novel framework for the differentially private ERM, input perturbation. Existing differentially private ERM implicitly assumed that the data contributors submit their private data to a database expecting that the database invokes a differentially private mechanism for publication of the learned model. In input perturbation, each data contributor independently randomizes her/his data by itself and submits the perturbed data to the database. We show that the input perturbation framework theoretically guarantees that the model learned with the randomized data eventually satisfies differential privacy with the prescribed privacy parameters. At the same time, input perturbation guarantees that local differential privacy is guaranteed to the server. We also show that the excess risk bound of the model learned with input perturbation is O(1) under a certain condition, where n is the sample size. This is the same as the excess risk bound of the state-of-the-art.
机译：我们为差分私人企业风险管理提出了一种新颖的框架，即输入扰动。现有的差异私有E​​RM隐式假定数据提供者将其私有数据提交给数据库，期望该数据库调用差异私有机制来发布学习的模型。在输入扰动中，每个数据提供者自己独立地将其数据随机化，并将扰动的数据提交给数据库。我们表明，输入扰动框架从理论上保证了使用随机数据学习的模型最终满足了具有指定隐私参数的差分隐私。同时，输入扰动可确保为服务器保证本地差分隐私。我们还表明，在一定条件下，通过输入扰动学习的模型的额外风险边界为O（1 / n），其中n是样本大小。这与最新技术的超额风险界限相同。

著录项

  • 来源
    《Discovery science》|2017年|82-90|共9页
  • 会议地点 Kyoto(JP)
  • 作者

    Kazuto Fukuchi; Quang Khai Tran; Jun Sakuma;

  • 作者单位

    Department of Computer Science, Graduated School of System and Information Science, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8577, Japan;

    Intelligent Systems Laboratory, Secom Co., Ltd., 10-16, Shimorenjaku 8-chome, Mitaka City, Tokyo 181-8528, Japan;

    Department of Computer Science, Graduated School of System and Information Science, University of Tsukuba, 1-1-1 Tennodai, Tsukuba, Ibaraki 305-8577, Japan ,JST CREST, Ks Gobancho 6F, 7, Gobancho, Chiyoda-ku, Tokyo 102-0076, Japan;

  • 会议组织
  • 原文格式 PDF
  • 正文语种 eng
  • 中图分类
  • 关键词

    Differential privacy; Empirical risk minimization; Local privacy; Linear regression; Logistic regression;

    机译：差异性隐私；经验风险最小化；当地隐私；线性回归；逻辑回归;

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱：kefu@zhangqiaokeyan.com

京公网安备：11010802029741号 ICP备案号：京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有

  • 客服微信

  • 服务号