Robustness Verification of Deep Reinforcement Learning Based Control Systems Using Reward Martingales.
Dapeng ZhiPeixin WangCheng ChenMin ZhangPublished in: AAAI (2024)
Keyphrases
- reinforcement learning
- control system
- eligibility traces
- function approximation
- model checking
- machine learning
- state space
- model free
- multi agent
- control strategies
- reward function
- learning algorithm
- temporal difference
- control policies
- markov decision processes
- optimal policy
- reinforcement learning algorithms
- learning agent
- signature verification
- average reward
- probability theory
- initially unknown
- verification method
- markov decision problems
- control law
- learning process
- transfer learning
- dynamic environments