Information asymmetry in KL-regularized RL.
Alexandre GalashovSiddhant M. JayakumarLeonard HasencleverDhruva TirumalaJonathan SchwarzGuillaume DesjardinsWojciech M. CzarneckiYee Whye TehRazvan PascanuNicolas HeessPublished in: CoRR (2019)
Keyphrases
- information asymmetry
- reinforcement learning
- kullback leibler
- least squares
- function approximation
- multi agent
- total least squares
- optimal policy
- lead time
- reinforcement learning algorithms
- learning process
- regularized least squares
- learning agents
- kl divergence
- action selection
- model free
- learning classifier systems
- action space
- optimal control
- learning algorithm