Preserving Privacy in GANs Against Membership Inference Attack

Keyphrases

• Adversary 10%
• Bhattacharyya Coefficient 10%
• Defense Mechanisms 10%
• Defense Strategy 30%
• Differential Privacy 10%
• Fano's Inequality 10%
• General Measure 10%
• Generated Samples 10%
• Generating Synthetic Data 10%
• Generative Adversarial Networks 100%
• Heuristic Model 10%
• Loss Function 10%
• Maximum Entropy 10%
• MEGAN 10%
• Membership Inference Attack 100%
• Memorization 10%
• Mutual Information 10%
• Mutual Information Minimization 10%
• Overfitting 30%
• Preserving Privacy 100%
• Privacy Concerns 10%
• Simple Modification 10%
• Synthetic Data 20%
• Synthetic Samples 10%
• Training Data 40%
• Variational Representation 10%

Computer Science

• Bhattacharyya 10%
• Defense Strategy 30%
• Differential Privacy 10%
• Discriminator 10%
• Generative Adversarial Networks 100%
• Information Leak 10%
• Maximum Entropy 10%
• Membership Inference Attack 100%
• Mutual Information 20%
• Privacy Preserving 100%
• Synthetic Data 30%
• Training Data 10%
• Training Data Point 10%
• Training Dataset 20%