Vulnerability of deep reinforcement learning to policy induction attacks

Keyphrases

• Vulnerability 100%
• Deep Reinforcement Learning (deep RL) 100%
• Deep Q-network 100%
• Adversarial Examples 75%
• Transferability 50%
• Network Model 25%
• Learning Process 25%
• Attack Strategy 25%
• Learning Scenario 25%
• Game-based Learning 25%
• Reinforcement Learning Algorithm 25%
• Novel Class 25%
• Input Perturbation 25%
• Deep Learning Classifier 25%
• Adversarial Input 25%

Computer Science

• Deep Reinforcement Learning 100%
• Deep Q-Network 100%
• Adversarial Example 75%
• Learning Process 25%
• Reinforcement Learning 25%
• Deep Learning 25%
• Learning Technique 25%