中国科学院机构知识库网格
Chinese Academy of Sciences Institutional Repositories Grid
首页 机构 成果 学者
  
  1. CAS IR Grid
  2. 自动化研究所
  3. 中国科学院自动化研究所
  4. 多模态人工智能系统全国重点实验室
  5. 互联网大数据与安全信息学研究中心
epsilon-k anonymization and adversarial training of graph neural networks for privacy preservation in social networks

文献类型：期刊论文

作者Tian, Hu1,2; Zheng, Xiaolong1,2; Zhang, Xingwei1,2; Zeng, Daniel Dajun1,2
刊名ELECTRONIC COMMERCE RESEARCH AND APPLICATIONS
出版日期2021-11-01
卷号50页码:17
关键词Privacy preservation Anonymization Graph neural networks Social network
ISSN号1567-4223
DOI10.1016/j.elerap.2021.101105
通讯作者Zheng, Xiaolong(xiaolong.zheng@ia.ac.cn)
英文摘要With the explosive growth of social networks, privacy preservation as a social good has been one common concern. Graph neural networks (GNNs) have been utilized by social network service providers to improve business service. However, traditional anonymization techniques of social networks cannot satisfy the desired privacy preservation of node attribute and graph structure and introduce information disturbance from the anonymization, leading to the performance degradation of GNNs in social network analysis. To protect sensitive user data and persist GNNs' performance in social network analysis, we propose a two-stage privacy-preserving method of graph neural networks in the social network domain. During the first stage, we design a novel e-k anonymization method that can achieve e-local differential privacy (e-LDP) and k-degree anonymity by incorporating the classical LDP and k-degree anonymization (k-DA) while retaining as much network community information as possible. At the second stage, we develop an adversarial training mechanism for GNNs to resist the disturbance from e-k anonymization and retain as much task performance as possible on anonymous social network data. Comprehensive experiments on several real-world social network datasets demonstrate the effectiveness of the proposed method for privacy-preserving node classification, link prediction, and graph clustering in social networks. The proposed method represents an interesting and important combination of classical anonymous technologies and recent GNNs and can preserve user privacy while providing business service.
资助项目Ministry of Health of China, China[2017ZX10303401-002] ; Ministry of Science and Technology of China, China[2020AAA0108401] ; Natural Science Foundation of China, China[71472175] ; Natural Science Foundation of China, China[71602184] ; Natural Science Foundation of China, China[71621002] ; National Key Research and Development Program of China, China[2016QY02D0305]
WOS研究方向Business & Economics ; Computer Science
语种英语
WOS记录号WOS:000722136200001
出版者ELSEVIER
资助机构Ministry of Health of China, China ; Ministry of Science and Technology of China, China ; Natural Science Foundation of China, China ; National Key Research and Development Program of China, China
源URL[http://ir.ia.ac.cn/handle/173211/46533]  
专题自动化研究所_复杂系统管理与控制国家重点实验室_互联网大数据与安全信息学研究中心
通讯作者Zheng, Xiaolong
作者单位1.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing, Peoples R China
2.Chinese Acad Sci, Inst Automat, State Key Lab Management & Control Complex Syst, Beijing 100190, Peoples R China
推荐引用方式
GB/T 7714		 Tian, Hu,Zheng, Xiaolong,Zhang, Xingwei,et al. epsilon-k anonymization and adversarial training of graph neural networks for privacy preservation in social networks[J]. ELECTRONIC COMMERCE RESEARCH AND APPLICATIONS,2021,50:17.
APA Tian, Hu,Zheng, Xiaolong,Zhang, Xingwei,&Zeng, Daniel Dajun.(2021).epsilon-k anonymization and adversarial training of graph neural networks for privacy preservation in social networks.ELECTRONIC COMMERCE RESEARCH AND APPLICATIONS,50,17.
MLA Tian, Hu,et al."epsilon-k anonymization and adversarial training of graph neural networks for privacy preservation in social networks".ELECTRONIC COMMERCE RESEARCH AND APPLICATIONS 50(2021):17.

入库方式： OAI收割

来源：自动化研究所

浏览0
下载0
收藏0
其他版本

除非特别说明，本系统中所有内容都受版权保护，并保留所有权利。