Ensemble-based Offline-to-Online Reinforcement Learning: From Pessimistic Learning to Optimistic Exploration.
Kai ZhaoYi MaJinyi LiuYan ZhengZhaopeng MengPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- learning algorithm
- online learning
- active exploration
- learning process
- exploration exploitation tradeoff
- action selection
- state space
- learning problems
- autonomous learning
- prior knowledge
- supervised learning
- knowledge acquisition
- machine learning methods
- function approximation
- hybrid learning
- reinforcement learning algorithms
- exploration strategy
- balancing exploration and exploitation
- neural network