Cost-aware Offline Safe Meta Reinforcement Learning with Robust In-Distribution Online Task Adaptation.
Cong GuanRuiqi XueZiqian ZhangLihe LiYi-Chen LiLei YuanYang YuPublished in: AAMAS (2024)
Keyphrases
- reinforcement learning
- real time
- function approximation
- online learning
- machine learning
- cost reduction
- computationally efficient
- confidence weighted
- robust estimation
- spatial distribution
- random variables
- learning algorithm
- data distribution
- markov decision processes
- partial occlusion
- estimation error
- optimal policy
- probability distribution
- meta level
- model free
- batch mode
- online environment
- objective function