Risk Sensitive Dead-end Identification in Safety-Critical Offline Reinforcement Learning.
Taylor W. KillianSonali ParbhooMarzyeh GhassemiPublished in: CoRR (2023)
Keyphrases
- risk sensitive
- safety critical
- reinforcement learning
- optimal control
- model free
- markov decision processes
- dead end
- formal methods
- fault tolerant
- control policies
- markov decision problems
- optimal policy
- function approximation
- reinforcement learning algorithms
- agent architecture
- utility function
- real time
- support systems
- embedded systems
- search algorithm
- temporal difference
- action space
- dynamic programming
- state space
- multi agent
- finite state
- policy iteration
- expert systems
- markov decision process
- average cost
- average reward
- expected utility
- infinite horizon
- machine learning
- data mining