Provably Efficient Partially Observable Risk-Sensitive Reinforcement Learning with Hindsight Observation.
Tonghe ZhangYu ChenLongbo HuangPublished in: CoRR (2024)
Keyphrases
- partially observable
- reinforcement learning
- risk sensitive
- markov decision processes
- markov decision problems
- optimal control
- partial observability
- model free
- state space
- infinite horizon
- decision problems
- dynamical systems
- partially observable environments
- optimal policy
- reinforcement learning algorithms
- reward function
- function approximation
- dynamic programming
- control policies
- machine learning
- heuristic search
- action selection
- learning algorithm
- multi agent
- average cost
- computational complexity
- average reward
- action space
- markov decision process
- transfer learning