Defending Against Adversarial Samples Without Security through Obscurity

Keyphrases

• Deep Neural Network 100%
• Adversarial Samples 100%
• Learning Algorithm 50%
• Cybersecurity 33%
• Publicly Available 16%
• Classification Performance 16%
• Adversary 16%
• Security Applications 16%
• Defense Mechanisms 16%
• Data Transformation 16%
• Synthetic Instances 16%
• Generic Approach 16%
• Learning Techniques 16%
• Attack Strategy 16%
• Model Forcing 16%
• Fool 16%
• Robustness to Attacks 16%
• Classification Robustness 16%
• MNIST 16%

Computer Science

• Deep Neural Network 100%
• Learning Algorithm 42%
• Neural Network Model 28%
• Classification Performance 14%
• Malware 14%
• Learning Technique 14%
• Data Transformation 14%
• Critical Application 14%
• Cyber Security Application 14%