1. Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) http://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn 2011-12-05 6:00PM 1 Speaker : XuBo Personalized Privacy Protection in Social Networks

2. 2 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Privacy Protection

3. 3 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Privacy Protection in Social Networks  more and more people join multiple social networks on the Web  as a service provider, it is essential to protect users’ privacy and at the same time provide “useful” data

4. 4 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Previous Works  Clustering-based approaches  Graph editing methods  Drawback Different users may have different privacy preferences Same level may not be fair and data may useless

5. 5 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Personalized Privacy Protection in Social Networks  define different levels of protection for users and incorporate them into the same published social network  A user can have a clear estimation of the knowledge that an attacker can know about him  The knowledge an attacker uses to find the privacy information of a user is called the background knowledge

6. 6 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) background knowledge

7. 7 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) A simple example  protection objectives the probability that an attacker finds a person P is node u in the published graph should be less than 50% the probability that an attacker finds person P1 and P2 have a connection is less than 50%

8. 8 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Personalized Protection

9. 9 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Method  For Level 1 protection, we use node label generalization  For Level 2 protection, we combine the noise node/edge adding methods based on the protection at Level 1  For Level 3 protection, we further use the edge label generalization to achieve the protection objectives

10. 10 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Abstract  PROBLEM DEFINITION  LEVEL 1 PROTECTION  LEVEL 2 PROTECTION  LEVEL 3 PROTECTION  EXPERIMENTS

11. 11 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) PROBLEM DEFINITION  G(V,E), Given a constant k

12. 12 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) NP-hard problem  Here the cost stands for the sum of label generalization levels. For example, in Figure 1(d), if let the labels in the group {a,d} to be the same, they become [(Africa, American), 2*], thus the node label generalization cost of this group is 4.

13. 13 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) LEVEL 1 PROTECTION  Level 1’s background knowledge is about node label list  generalization is to group nodes and make the nodes within each group have one generalized node label list

14. 14 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Achieve Level 1 protection  if each group’s size is at least k, the first objective is achieved.  To satisfy objectives (2) and (3), the following condition must be satisfied: The first condition constrains that any two nodes in the same group do not connect to a same node The second condition constrains that no edges within a group

15. 15 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) The Safety Grouping Condition (SGC)

16. 16 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Algorithm 1: Generate safe groups

17. 17 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) LEVEL 2 PROTECTION  Adding nodes/edges into GL1  Label Assignment

18. 18 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Construct a k-degree anonymous graph  The first step is to set a target degree for each node  the second step is to randomly create edges between the nodes which need to increase their degrees under SGC  we add noise nodes to enable all the nodes in GL1 have their target degrees by connecting them with noise nodes under SGC.  we need to hide the noise nodes by making their degree to a preselected value degree target Degree target is the degree that the maximum number of groups of size k in GL1 have

19. 19 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Algorithm 2: Degree Anonymizing Algorithm with X

20. 20 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) A running example of Algorithm 2

21. 21 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Label Assignment  No con (A1,A2, l)  a node/edge label assignment with less No con (A1,A2, l) changes is preferred  heuristic method first assign labels to the noise edges decide the labels of noise nodes

22. 22 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Assign labels to the noise edges

23. 23 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Decide the labels of noise nodes

24. 24 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) LEVEL 3 PROTECTION  degree label sequence  a group contains at least one node that needs Level 3 protection, for any nodes in it, change their degree label sequence to be the same by generalizing edge labels.

25. 25 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Generate degree label anonymous graph

26. 26 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) EXPERIMENTS

27. 27 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Datasets  Speed Dating Data 552 nodes and 4194 edges Each node has 19 labels Edge label is “match”/”unmatch”  ArXiv Data 19835 nodes and 40221 edges Each node denotes an author set label ”seldom” to edges with weight 1, ”normal” to edges with weight 2 - 5 and ”frequent” to edges with weight larger than 5  ArXiv Data with uniform labels

28. 28 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Results  higher utility  average relative error increases with the increasing of k

29. 29 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Results(2)

30. 30 2011-12-05 6:00PM Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDANhttp://gdm.fudan.edu.cn 徐波 Email : bolang1988@yahoo.com.cn Graph Data Management Lab, School of Computer Science GDM@FUDANGDM@FUDAN Personalized Privacy Protection in Social Networks (VLDB2011) Results(3)