Pessimism Meets Risk: Risk-Sensitive Offline Reinforcement Learning.
Dake ZhangBoxiang LyuShuang QiuMladen KolarTong ZhangPublished in: CoRR (2024)
Keyphrases
- risk sensitive
- reinforcement learning
- model free
- optimal control
- markov decision processes
- control policies
- markov decision problems
- reinforcement learning algorithms
- risk neutral
- function approximation
- utility function
- state space
- dynamic programming
- optimal policy
- finite state
- infinite horizon
- partially observable
- markov decision chains
- policy iteration
- temporal difference
- real time
- reward function
- average cost
- control strategy
- linear programming
- expected utility
- control strategies
- multi agent