Detecting Adversarial Samples Using Influence Functions and Nearest Neighbors

Keyphrases

• Activation Layer 33%
• Adversarial Attack 66%
• Adversarial Detection 33%
• Adversarial Detectors 33%
• Adversarial Examples 33%
• Adversarial Input 33%
• Adversarial Samples 100%
• Attack Strategy 33%
• Classification Framework 33%
• Deep Neural Network 33%
• Influence Function 100%
• Influence Measure 33%
• K-nearest 66%
• Nearest Neighbor 100%
• Neighborhood Model 33%
• Neural Network Classifier 33%
• Pre-trained Networks 33%
• Robust Classification 33%
• Set Data 33%
• Small Perturbation 33%
• Training Samples 100%
• Validation Set 33%
• Vulnerability 33%

Computer Science

• Adversarial Example 66%
• Adversarial Machine Learning 66%
• Classification Framework 33%
• Deep Neural Network 66%
• Influence Function 100%
• Trained Neural Network 33%
• Training Sample 100%
• Validation Set 33%