多级本地化差分隐私算法推荐框架

doi:10.11959/j.issn.1000-436x.2022106

通信学报 ›› 2022, Vol. 43 ›› Issue (8): 52-64.doi: 10.11959/j.issn.1000-436x.2022106

多级本地化差分隐私算法推荐框架

王瀚仪¹^,², 李效光³, 毕文卿¹^,², 陈亚虹¹^,², 李凤华¹^,², 牛犇¹

¹ 中国科学院信息工程研究所，北京 100093
² 中国科学院大学网络空间安全学院，北京 100049
³ 西安电子科技大学网络与信息安全学院，陕西西安 710071

修回日期:2022-04-18 出版日期:2022-08-25 发布日期:2022-08-01
作者简介:王瀚仪（1994- ），女，吉林省吉林市人，中国科学院信息工程研究所博士生，主要研究方向为隐私计算
李效光（1995- ），男，陕西西安人，西安电子科技大学博士生，主要研究方向为差分隐私
毕文卿（1997- ），女，山东菏泽人，中国科学院信息工程研究所硕士生，主要研究方向为隐私计算
陈亚虹（1995- ），女，福建泉州人，中国科学院信息工程研究所博士生，主要研究方向为隐私计算
李凤华（1966- ），男，湖北浠水人，博士，中国科学院信息工程研究所研究员、博士生导师，主要研究方向为网络与系统安全、信息保护、隐私计算
牛犇（1984- ），男，陕西西安人，博士，中国科学院信息工程研究所副研究员、博士生导师，主要研究方向为隐私计算、网络安全防护
基金资助:
国家重点研发计划基金资助项目(2021YFB3100300);国家自然科学基金资助项目(61872441);国家自然科学基金资助项目(61932015)

Multi-level local differential privacy algorithm recommendation framework

Hanyi WANG¹^,², Xiaoguang LI³, Wenqing BI¹^,², Yahong CHEN¹^,², Fenghua LI¹^,², Ben NIU¹

¹ Institute of Information Engineering, Chinese Academy of Sciences, Beijing 100093, China
² School of Cyber Security, University of Chinese Academy of Sciences, Beijing 100049, China
³ School of Cyber Engineering, Xidian University, Xi’an 710071, China

Revised:2022-04-18 Online:2022-08-25 Published:2022-08-01
Supported by:
The National Key Research and Development Program of China(2021YFB3100300);The National Natural Science Foundation of China(61872441);The National Natural Science Foundation of China(61932015)

摘要/Abstract

摘要：

本地化差分隐私（LDP）算法通常为不同用户分配相同的保护机制及参数，却忽视了不同用户终端设备资源与隐私需求的差异。为此，提出一种多级 LDP 算法推荐框架。该框架考虑服务商以及用户的需求，通过服务商和用户的多级管理实现多用户差异化隐私保护。将框架应用至频数统计场景形成 LDP 算法推荐方案，改进LDP算法以保证统计结果的可用性，设计协同机制保护用户的隐私偏好。实验结果证明了所提方案的可用性。

关键词: 本地化差分隐私, 资源自适应, 个性化隐私预算

Abstract:

Local differential privacy (LDP) algorithm usually assigned the same protection mechanism and parameters to different users.However, it ignored the differences among the device resources and the privacy requirements of different users.For this reason, a multi-level LDP algorithm recommendation framework was proposed.The server and the users’ requirements were considered in the framework, and the multi-users’ differential privacy protections were realized by the server and the users’ multi-level management.The framework was applied to the frequency statistics scenario to form an LDP algorithm recommendation scheme.LDP algorithm was improved to ensure the availability of statistical results, and a collaborative mechanism was designed to protect users’ privacy preferences.The experimental results demonstrate the availability of the proposed scheme.

Key words: local differential privacy, resource adaptation, personalized privacy budget

中图分类号:

TP309

王瀚仪, 李效光, 毕文卿, 陈亚虹, 李凤华, 牛犇. 多级本地化差分隐私算法推荐框架[J]. 通信学报, 2022, 43(8): 52-64.

Hanyi WANG, Xiaoguang LI, Wenqing BI, Yahong CHEN, Fenghua LI, Ben NIU. Multi-level local differential privacy algorithm recommendation framework[J]. Journal on Communications, 2022, 43(8): 52-64.

图/表 10

表1

系统参数"

参数	含义
ALG	LDP算法集合
${alg}_{i}$	服务商为用户i选择的LDP算法集合
$r_{1}, r_{2}, \dots, r_{o}$	LDP算法的各类资源开销，其中，r₁为电量，r₂为流量
$r_{var}$	LDP算法的可用性
$a_{1}, a_{2}, \dots, a_{h}$	影响LDP算法资源开销与可用性的因素
$w_{1}, w_{2}, \dots, w_{o}$	用户i对o类资源开销的权重
$w_{var}$	用户i对可用性的权重
${alg}_{opt}$	用户选择的最优LDP算法
$ε_{i}$	用户i的个性化隐私预算
D	位置PoI集合
D_i	用户所选扰动范围，D_i?D
n	用户个数
g	后处理环节中的第g 个群组
m_g	群组g 内的用户数
G_g	群组g 内用户统一的扰动结果集合
${G^{'}}_{g}$	群组g 内用户统一的扰动范围集合
$F^{d_{l}}$	所有用户访问位置d_l 的估计频数

表1

图1

图2

图3

图4

表2

图5

图6

图7

图8

参考文献 27

[1]	DWORK C , . Differential privacy[C]// Proceedings of 2006 International Colloquium on Automata,Languages and Programming (ICALP). Berlin:Springer, 2006: 1-12.
[2]	DWORK C , MCSHERRY F , NISSIM K ,et al. Calibrating noise to sensitivity in private data analysis[C]// Theory of Cryptography. Berlin:Springer, 2006: 265-284.
[3]	MCSHERRY F , TALWAR K . Mechanism design via differential privacy[C]// Proceedings of the 48th Annual IEEE Symposium on Foundations of Computer Science. Piscataway:IEEE Press, 2007: 94-103.
[4]	WANG T , ZHANG X F , FENG J Y ,et al. A comprehensive survey on local differential privacy toward data statistics and analysis[J]. Sensors, 2020,20(24): 7030.
[5]	KAIROUZ P , OH S , VISWANATH P . Extremal mechanisms for local differential privacy[J]. Journal of Machine Learning Research, 2016,17(17): 1-51.
[6]	KAIROUZ P , BONAWITZ K , RAMAGE D . Discrete distribution estimation under local privacy[C]// Proceedings of 2016 International Conference on Machine Learning (ICML). New York:ACM Press, 2016: 2436-2444.
[7]	WANG T H , BLOCKI J , LI N H ,et al. Locally differentially private protocols for frequency estimation[C]// Proceedings of 2017 USENIX Security Symposium (USENIX Security). Berkeley:USENIX Association, 2017: 729-745.
[8]	BASSILY R , SMITH A . Local,private,efficient protocols for succinct histograms[C]// Proceedings of the 47th Annual ACM Symposium on Theory of Computing. New York:ACM Press, 2015: 127-135.
[9]	WANG T H , LI N H , JHA S . Locally differentially private frequent itemset mining[C]// Proceedings of 2018 IEEE Symposium on Security and Privacy. Piscataway:IEEE Press, 2018: 127-143.
[10]	YE Q Q , HU H B , MENG X F ,et al. PrivKV:key-value data collection with local differential privacy[C]// Proceedings of 2019 IEEE Symposium on Security and Privacy. Piscataway:IEEE Press, 2019: 317-331.
[11]	DUCHI J C , JORDAN M I , WAINWRIGHT M J . Privacy aware learning[J]. Journal of the ACM, 2014,61(6): 1-57.
[12]	NGUYêN T T , XIAO X , YANG Y ,et al. Collecting and analyzing data from smart device users with local differential privacy[J]. arXiv Preprint,arXiv:1606.05053, 2016.
[13]	YILMAZ E , AL-RUBAIE M , CHANG J M . Locally differentially private naive bayes classification[J]. arXiv Preprint,arXiv:1905.01039, 2019.
[14]	MAHAWAGA A P C , BERTOK P , KHALIL I ,et al. Local differential privacy for deep learning[J]. IEEE Internet of Things Journal, 2020,7(7): 5827-5842.
[15]	SHIN H , KIM S , SHIN J ,et al. Privacy enhanced matrix factorization for recommendation with local differential privacy[J]. IEEE Transactions on Knowledge and Data Engineering, 2018,30(9): 1770-1782.
[16]	ERLINGSSON ú , PIHUR V , KOROLOVA A . RAPPOR:randomized aggregatable privacy-preserving ordinal response[C]// Proceedings of the 2014 ACM SIGSAC Conference on Computer and Communications Security. New York:ACM Press, 2014: 1054-1067.
[17]	WANG N , XIAO X K , YANG Y ,et al. Collecting and analyzing multidimensional data with local differential privacy[C]// Proceedings of 2019 IEEE 35th International Conference on Data Engineering. Piscataway:IEEE Press, 2019: 638-649.
[18]	ZHAO Y , ZHAO J , YANG M ,et al. Local differential privacy-based federated learning for Internet of things[J]. IEEE Internet of Things Journal, 2021,8(11): 8836-8853.
[19]	CHEN R , LI H R , QIN A K ,et al. Private spatial data aggregation in the local setting[C]// Proceedings of 2016 IEEE 32nd International Conference on Data Engineering. Piscataway:IEEE Press, 2016: 289-300.
[20]	NIE Y W , YANG W , HUANG L S ,et al. A utility-optimized framework for personalized private histogram estimation[J]. IEEE Transactions on Knowledge and Data Engineering, 2019,31(4): 655-669.
[21]	XIA C , HUA J Y , TONG W ,et al. Distributed k-means clustering guaranteeing local differential privacy[J]. Computers ＆ Security, 2020,90:101699.
[22]	SHEN Z X , XIA Z H , YU P P . PLDP:personalized local differential privacy for multidimensional data aggregation[J]. Security and Communication Networks,2021, 2021:6684179.
[23]	NALDI M , D’ACQUISTO G , . Differential privacy:an estimation theory-based method for choosing epsilon[J]. arXiv Preprint,arXiv:1510.00917, 2015.
[24]	SHAHANI S , ABRAHAM J , VENKATESWARAN R . Selection and verification of privacy parameters for local differentially private data aggregation[C]// Proceedings of the 5th International Conference on Information System and Data Mining. New York:ACM Press, 2021: 84-89.
[25]	LI F H , LI H , NIU B ,et al. Privacy computing:concept,computing framework,and future development trends[J]. Engineering, 2019,5(6): 1179-1192.
[26]	叶青青, 孟小峰, 朱敏杰 ,等. 本地化差分隐私研究综述[J]. 软件学报, 2018,29(7): 1981-2005.
	YE Q Q , MENG X F , ZHU M J ,et al. Survey on local differential privacy[J]. Journal of Software, 2018,29(7): 1981-2005.
[27]	NIU B , LI Q H , WANG H Y ,et al. A framework for personalized location privacy[J]. IEEE Transactions on Mobile Computing,2021:doi.org/10.1109/TMC.2021.3055865.

电量状态	流量状态	D_i	ε_i	最优算法
未充电，剩余电量5	未连接Wi-Fi，剩余流量0 MB，套餐价格0.3元/MB	10	4.0	DE
充电中，电量充足	未连接Wi-Fi，剩余流量0 MB，套餐价格0.3元/MB	128	4.0	OUE
未充电，剩余电量5	连接Wi-Fi，流量充足	4 096	4.0	OLH
未充电，剩余电量5	连接Wi-Fi，流量充足	1 024	0.5	THE

多级本地化差分隐私算法推荐框架

Multi-level local differential privacy algorithm recommendation framework

在线阅读

PDF下载

可视化

摘要/Abstract

引用本文

使用本文

图/表 10

参考文献 27

相关文章 3

Metrics

推荐阅读 0

[1]	冯涛, 陈李秋, 方君丽, 石建明. 基于本地化差分隐私和属性基可搜索加密的区块链数据共享方案[J]. 通信学报, 2023, 44(5): 224-233.
[2]	刘杨, 李珺, 陈文韵, 彭木根. 面向6G的雾无线接入网内生安全数据共享机制研究[J]. 通信学报, 2021, 42(1): 67-78.
[3]	周异辉,鲁来凤,吴振强. 随机响应机制效用优化研究[J]. 通信学报, 2019, 40(6): 74-81.