Optimal Uniform OPE and Model-based Offline Reinforcement Learning in Time-Homogeneous, Reward-Free and Task-Agnostic Settings.
Ming YinYu-Xiang WangPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- model free
- dynamic programming
- average reward reinforcement learning
- average reward
- control policy
- total reward
- markov decision processes
- function approximation
- reward function
- optimal policy
- optimal control
- optimal design
- learning process
- optimal solution
- worst case
- approximate dynamic programming
- real time
- partially observable environments
- state and action spaces
- reinforcement learning algorithms
- machine learning
- state space
- learning capabilities
- learning agent
- temporal difference
- closed form
- transfer learning
- eligibility traces
- markov chain
- robotic control
- genetic algorithm