Variational Empowerment as Representation Learning for Goal-Conditioned Reinforcement Learning.
Jongwook ChoiArchit SharmaHonglak LeeSergey LevineShixiang Shane GuPublished in: ICML (2021)
Keyphrases
- reinforcement learning
- learning process
- learning algorithm
- learning problems
- learning systems
- learning capabilities
- supervised learning
- online learning
- knowledge acquisition
- machine learning
- active learning
- inductive inference
- robot control
- learning tasks
- action selection
- multiscale
- temporal difference learning
- continuous state