Adversary for Social Good: Leveraging Attribute-Obfuscating Attack to Protect User Privacy on Social Networks

Keyphrases

• User Privacy 100%
• Obfuscation 100%
• Adversary 100%
• Social Networks 50%
• Graph Neural Network 50%
• User Attributes 37%
• Attacker 25%
• Graph Structure 25%
• Adversarial Attack 25%
• Inference Attack 25%
• User Activity 12%
• Vulnerability 12%
• Daily Life 12%
• Neural Network 12%
• Social Information 12%
• Misclassification 12%
• User Information 12%
• Optimal Training 12%
• Attribute Value 12%
• Structure Feature 12%
• People's Daily 12%
• Private Values 12%
• Attribute Privacy 12%
• Feature Attribute 12%
• Bilevel Optimization 12%
• Network Inference 12%
• Privacy Disclosure 12%
• Attributes Annotation 12%
• Adversarial Process 12%
• Black-box Settings 12%
• Node Features 12%
• Attribute Inference Attack 12%

Computer Science

• User Privacy 100%
• Social Network 100%
• Graph Neural Network 80%
• Obfuscation 40%
• Attackers 40%
• Adversarial Machine Learning 40%
• Annotation 20%
• User Information 20%
• Structured Data 20%
• Attribute Value 20%
• Bilevel Optimisation 20%
• Private Attribute 20%