A Self-supervised Approach for Adversarial Robustness

Keyphrases

• Deep Neural Network 100%
• Vision System 100%
• Adversarial Training 100%
• Supervised Approach 100%
• Adversarial Robustness 100%
• Adversarial Examples 100%
• Vulnerability 50%
• Neural Network 50%
• By Design 50%
• Unseen 50%
• Transferability 50%
• Success Rate 50%
• Object Detection 50%
• Input Space 50%
• Plug-and-play 50%
• Training Mechanism 50%
• Adversarial Attack 50%
• Real-world Deployment 50%
• Take-the-first 50%
• Classification Detection 50%
• Input Processing 50%
• Translation Invariant 50%
• Cross-task 50%

Computer Science

• Adversarial Machine Learning 100%
• Adversarial Robustness 100%
• Deep Neural Network 66%
• Vision System 66%
• Adversarial Example 66%
• Object Detection 33%
• Generalizability 33%
• Input Processing 33%
• Translation Invariant 33%