End-to-End Safe Reinforcement Learning through Barrier Functions for Safety-Critical Continuous Control Tasks.
Richard ChengGábor OroszRichard M. MurrayJoel W. BurdickPublished in: CoRR (2019)
Keyphrases
- end to end
- safety critical
- reinforcement learning
- congestion control
- formal methods
- fault tolerant
- safety analysis
- control system
- nuclear power plant
- embedded systems
- learning algorithm
- control policy
- agent architecture
- learning process
- information systems
- markov decision processes
- intelligent agents
- optimal policy
- intelligent systems
- decision support
- peer to peer
- learning environment
- real world