Learning on Abstract Domains: A New Approach for Verifiable Guarantee in Reinforcement Learning.
Peng JinMin ZhangJianwen LiLi HanXuejun WenPublished in: CoRR (2021)
Keyphrases
- reinforcement learning
- learning process
- learning algorithm
- learning problems
- knowledge acquisition
- learning tasks
- learning systems
- learning capabilities
- state space
- online learning
- temporal difference
- background knowledge
- robot control
- learning agents
- temporal difference learning
- reinforcement learning methods
- partially observable domains
- function approximation
- transfer learning
- neural network
- active learning
- decision trees
- machine learning
- real world