A functional mirror ascent view of policy gradient methods with function approximation.
Sharan VaswaniOlivier BachemSimone TotaroRobert MuellerMatthieu GeistMarlos C. MachadoPablo Samuel CastroNicolas Le RouxPublished in: CoRR (2021)
Keyphrases
- function approximation
- natural actor critic
- policy gradient methods
- policy gradient
- reinforcement learning
- function approximators
- actor critic
- reinforcement learning problems
- model free
- learning tasks
- temporal difference
- temporal difference learning
- radial basis function
- decision problems
- reinforcement learning algorithms
- approximation methods
- learning experience
- learning process
- genetic algorithm