SAUTE RL: Almost Surely Safe Reinforcement Learning Using State Augmentation.
Aivar SootlaAlexander Imani Cowen-RiversTaher JafferjeeZiyan WangDavid MguniJun WangHaitham Bou-AmmarPublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- state space
- function approximation
- action space
- hidden state
- markov decision processes
- model free
- partially observable
- reinforcement learning algorithms
- state abstraction
- multi agent
- optimal control
- transition model
- autonomous learning
- markov decision process
- optimal policy
- learning algorithm
- learning capabilities
- learning problems
- multi agent systems
- multi agent reinforcement learning
- machine learning