Characterizing Attacks on Deep Reinforcement Learning

Keyphrases

• Deep Reinforcement Learning (deep RL) 100%
• Deep Reinforcement Learning Model 60%
• Attacker 40%
• Adversarial Attack 40%
• Real-world Application 20%
• Computational Demand 20%
• Imperfection 20%
• Vulnerability 20%
• Small Perturbation 20%
• Learning Settings 20%
• Robotics 20%
• Physical World 20%
• Temporal Consistency 20%
• Physical Robot 20%
• Black Hole Attack 20%
• Online Sequential 20%
• Autonomous Driving 20%
• Environment Dynamics 20%
• Physical Perturbation 20%
• Atari Games 20%
• Efficient Attack 20%
• Real-world Robotics 20%
• Reinforcement Learning System 20%
• Sequential Attack 20%

Computer Science

• Deep Reinforcement Learning 100%
• Robot 25%
• Attackers 25%
• Adversarial Machine Learning 25%
• Learning Systems 12%
• Physical World 12%
• World Application 12%
• Dynamic Environment 12%
• Baseline Approach 12%
• Autonomous Driving 12%
• Black-Box Attack 12%