Security analysis and enhancement of model compressed deep learning systems under adversarial attacks

Keyphrases

• Adversarial Attack 100%
• Input Perturbation 33%
• Adversarial Examples 22%
• Software-based Approach 11%
• Model Computing 11%
• Marginal Accuracy 11%
• Deep Compression 11%
• Adversarial Input 11%
• Defense Techniques 11%
• Weight Compression 11%
• Optimized Deep Learning 11%
• Generating Method 11%
• Network Reshaping 11%
• Uncertainty Factors 11%

Computer Science

• Adversarial Machine Learning 100%
• Adversarial Example 18%