Differentially Private Network Data Release via Structural Inference

被引:100
|
作者
Xiao, Qian [1 ]
Chen, Rui [2 ]
Tan, Kian-Lee [1 ,3 ]
机构
[1] Natl Univ Singapore, NGS, Singapore, Singapore
[2] Hong Kong Baptist Univ, Dept Comp Sci, Hong Kong, Peoples R China
[3] Natl Univ Singapore, Sch Comp, Singapore, Singapore
关键词
Network data; differential privacy; structural inference;
D O I
10.1145/2623330.2623642
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Information networks, such as social media and email networks, often contain sensitive information. Releasing such network data could seriously jeopardize individual privacy. Therefore, we need to sanitize network data before the release. In this paper, we present a novel data sanitization solution that infers a network's structure in a differentially private manner. We observe that, by estimating the connection probabilities between vertices instead of considering the observed edges directly, the noise scale enforced by differential privacy can be greatly reduced. Our proposed method infers the network structure by using a statistical hierarchical random graph (HRG) model. The guarantee of differential privacy is achieved by sampling possible HRG structures in the model space via Markov chain Monte Carlo (MCMC). We theoretically prove that the sensitivity of such inference is only O(log n), where n is the number of vertices in a network. This bound implies less noise to be injected than those of existing works. We experimentally evaluate our approach on four real-life network datasets and show that our solution effectively preserves essential network structural properties like degree distribution, shortest path length distribution and influential nodes.
引用
收藏
页码:911 / 920
页数:10
相关论文
共 50 条
  • [31] PrivTDSI: A Local Differentially Private Approach for Truth Discovery via Sampling and Inference
    Zhang, Pengfei
    Cheng, Xiang
    Su, Sen
    Zhu, Binyuan
    IEEE TRANSACTIONS ON BIG DATA, 2023, 9 (02) : 471 - 484
  • [32] Private data release via learning thresholds
    Hardt, Moritz
    Rothblum, Guy N.
    Servedio, Rocco A.
    Proceedings of the Annual ACM-SIAM Symposium on Discrete Algorithms, 2012, : 168 - 187
  • [33] EdgeSanitizer: Locally Differentially Private Deep Inference at the Edge for Mobile Data Analytics
    Xu, Chugui
    Ren, Ju
    She, Liang
    Zhang, Yaoxue
    Qin, Zhan
    Ren, Kui
    IEEE INTERNET OF THINGS JOURNAL, 2019, 6 (03) : 5140 - 5151
  • [34] Differentially-Private Release of Check-in Data for Venue Recommendation
    Riboni, Daniele
    Bettini, Claudio
    2014 IEEE INTERNATIONAL CONFERENCE ON PERVASIVE COMPUTING AND COMMUNICATIONS (PERCOM), 2014, : 190 - 198
  • [35] Boosting Accuracy of Differentially Private Continuous Data Release for Federated Learning
    Cai, Jianping
    Ye, Qingqing
    Hu, Haibo
    Liu, Ximeng
    Fu, Yanggeng
    IEEE TRANSACTIONS ON INFORMATION FORENSICS AND SECURITY, 2024, 19 : 10287 - 10301
  • [36] A Federated Learning Framework Based on Differentially Private Continuous Data Release
    Cai, Jianping
    Liu, Ximeng
    Ye, Qingqing
    Liu, Yang
    Wang, Yuyang
    IEEE TRANSACTIONS ON DEPENDABLE AND SECURE COMPUTING, 2024, 21 (05) : 4879 - 4894
  • [37] On the Complexity of Differentially Private Data Release Efficient Algorithms and Hardness Results
    Dwork, Cynthia
    Naor, Moni
    Reingold, Omer
    Rothblum, Guy N.
    Vadhan, Salil
    STOC'09: PROCEEDINGS OF THE 2009 ACM SYMPOSIUM ON THEORY OF COMPUTING, 2009, : 381 - 390
  • [38] Differentially Private Multi-Party Data Release for Linear Regression
    Wu, Ruihan
    Yang, Xin
    Yao, Yuanshun
    Sun, Jiankai
    Liu, Tianyi
    Weinberger, Kilian Q.
    Wang, Chong
    UNCERTAINTY IN ARTIFICIAL INTELLIGENCE, VOL 180, 2022, 180 : 2128 - 2137
  • [39] Differentially Private Data Release: Improving Utility with Wavelets and Bayesian Networks
    Xiao, Xiaokui
    WEB TECHNOLOGIES AND APPLICATIONS, APWEB 2014, 2014, 8709 : 25 - 35
  • [40] Data Synthesis via Differentially Private Markov Random Fields
    Cai, Kuntai
    Lei, Xiaoyu
    Wei, Jianxin
    Xiao, Xiaokui
    PROCEEDINGS OF THE VLDB ENDOWMENT, 2021, 14 (11): : 2190 - 2202