Attacks on Rich Graphs

For more info about our services contact : help@bestpfe.com

Table of contents

List of Figures
1 Introduction
1.1 General Context
1.2 Graph Data Anonymization Issue
1.3 Contributions
1.4 Detailed Content
2 State of the Art
2.1 Introduction
2.2 Privacy Protection for Graph Data
2.2.1 What to protect?
2.2.2 De-anonymization Techniques
2.3 Anonymization Techniques for Graphs
2.3.1 Structural Modification Based on k-Anonymity Concept
2.3.2 Randomization Techniques
2.3.3 Generalization Techniques
2.4 Utility Loss Evaluation
2.4.1 Graph Topological Properties
2.4.2 Graph Spectral Properties
2.4.3 Network Queries Aggregation
2.5 Complex Graphs Anonymization
2.5.1 Hypergraphs
2.5.2 Temporal Graphs
2.6 Differential Privacy for Graph Data Anonymization
2.6.1 Principle
2.6.2 Differential Privacy for Data Release
2.7 Machine Learning in Data Anonymization Process
2.7.1 Machine Learning used for Data De-Anonymization
2.7.2 Machine Learning used for Data Anonymization
2.7.3 Exploring the Privacy-Utility Tradeoff
2.8 Conclusion
3 Temporal Graphs Anonymization Issue
3.1 Introduction
3.2 Graphs Risk for De-Anonymization based on Subgraphs Partitioning
3.3 Anonymization by Data Partitioning
3.4 System Architecture
3.5 Conclusion
4 Anonymization Methodology Based on Machine Learning
4.1 Introduction
4.2 Methodology
4.3 Notations and Definitions
4.3.1 Anonymization Function
4.3.2 Utility Loss
4.3.3 Privacy Risk
4.4 Optimization Problem: Balance between Utility Loss and Privacy Risk
4.5 Summary
4.6 Optimization Methods
4.6.1 Estimation of Distribution Algorithm
4.6.2 Genetic Algorithms
4.7 Conclusion
5 Simple Graphs Anonymization
5.1 Introduction
5.2 Optimization Problem for Simple Graphs
5.3 Anonymization Method
5.4 Privacy Risks
5.4.1 k-Degree Anonymity
5.4.2 k-Neighborhood Anonymity
5.5 Utility Loss Evaluation
5.5.1 Clustering Coefficient Based Utility Loss (CC)
5.5.2 Page Rank Based Utility Loss (PR)
5.5.3 Two-hop neighborhood based utility loss (THN)
5.6 Experiments
5.6.1 Baseline (BL)
5.6.2 Datasets
5.6.3 Results
5.7 Conclusion
6 Adaptive Temporal Graphs Anonymization for Data Publishing
6.1 Introduction
6.2 Machine Learning for Call Detail Records Anonymization
6.2.1 Notations and Definitions
6.2.2 Learning Problem
6.3 Anonymization Method
6.4 Privacy Risks in Call Logs
6.4.1 Privacy Attack by Communication Sequence Generation (CSG)
6.4.2 Privacy attack by Neighborhood Degree Distribution (NDD) .
6.5 Utility Loss Evaluation
6.5.1 Changes Performed by the Anonymization Algorithm in the Graph (CHG)
6.5.2 Query Based Measures: Call Distribution Distance (CDD)
6.5.3 Graph Topological Properties: Vertices In/Out Degrees (DE)
6.6 Experiments
6.6.1 Baseline: Random data perturbation
6.6.2 Datasets
6.6.3 Results
6.7 Conclusion
7 Conclusion
7.1 Contributions
7.2 Perspectives
Bibliography