On the Importance of Feature Decorrelation for Unsupervised Representation Learning in Reinforcement Learning.
Hojoon LeeKoanho LeeDongyoon HwangHyunho LeeByungkun LeeJaegul ChooPublished in: CoRR (2023)
Keyphrases
- reinforcement learning
- learning process
- unsupervised learning
- learning algorithm
- feature hierarchies
- supervised learning
- learning problems
- feature representation
- knowledge acquisition
- autonomous learning
- face recognition
- function approximation
- learning capabilities
- background knowledge
- evolutionary learning
- eligibility traces
- active learning