BACKDOORL: Backdoor Attack against Competitive Reinforcement Learning.
Lun WangZaynah JavedXian WuWenbo GuoXinyu XingDawn SongPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- function approximation
- countermeasures
- robotic control
- temporal difference
- optimal policy
- multi agent
- learning algorithm
- reinforcement learning algorithms
- state space
- supervised learning
- learning process
- detection mechanism
- malicious users
- optimal control
- attack detection
- multi agent reinforcement learning
- reinforcement learning methods
- function approximators
- learning agent
- data sets
- evaluation function
- special case
- machine learning