PolicyCleanse: Backdoor Detection and Mitigation for Competitive Reinforcement Learning.
Junfeng GuoAng LiLixu WangCong LiuPublished in: ICCV (2023)
Keyphrases
- reinforcement learning
- object detection
- detection method
- detection accuracy
- automatic detection
- false alarms
- function approximation
- multi agent
- state space
- markov decision processes
- detection algorithm
- anomaly detection
- learning algorithm
- temporal difference
- robotic control
- reinforcement learning algorithms
- detection scheme
- database
- learning classifier systems
- optimal policy
- learning process
- object recognition
- machine learning