CtrlFormer: Learning Transferable State Representation for Visual Control via Transformer.
Yao Mark MuShoufa ChenMingyu DingJianyu ChenRunjian ChenPing LuoPublished in: ICML (2022)
Keyphrases
- visual representation
- learning process
- learning algorithm
- knowledge acquisition
- learning problems
- visual information
- active learning
- visual learning
- reinforcement learning
- control rules
- fuzzy logic
- online learning
- neural network
- real time
- learning tasks
- control method
- image representation
- optimal control
- visual perception
- perceptual information
- control strategy
- spatial relations
- learning experience
- control system
- video sequences
- machine learning
- data sets