Defending Observation Attacks in Deep Reinforcement Learning via Detection and Denoising.
Zikang XiongJoe EappenHe ZhuSuresh JagannathanPublished in: CoRR (2022)
Keyphrases
- denoising
- reinforcement learning
- function approximation
- detection algorithm
- image denoising
- false positives
- attack detection
- object detection
- automatic detection
- false alarms
- multi agent
- ddos attacks
- noisy images
- detection rate
- detection accuracy
- countermeasures
- image processing
- detection method
- anomaly detection
- natural images
- multiscale
- intrusion detection system
- total variation
- wavelet domain
- state space
- gaussian noise
- model free
- noise removal
- temporal difference
- denoising algorithm
- machine learning
- normal traffic
- detecting malicious