Learning Pseudometric-based Action Representations for Offline Reinforcement Learning.
Pengjie GuMengchen ZhaoChen ChenDong LiJianye HaoBo AnPublished in: ICML (2022)
Keyphrases
- reinforcement learning
- learning process
- learning algorithm
- learning problems
- supervised learning
- function approximation
- autonomous learning
- inductive inference
- partially observable domains
- reinforcement learning methods
- multiple representations
- action selection
- learning tasks
- learning systems
- real time
- learning classifier systems
- reinforcement learning algorithms
- action sequences
- knowledge acquisition
- learning agents
- action models
- online learning
- continuous state
- mobile robot
- learning from experience
- machine learning