Compatible Natural Gradient Policy Search.
Joni PajarinenHong Linh ThaiRiad AkrourJan PetersGerhard NeumannPublished in: CoRR (2019)
Keyphrases
- policy search
- natural gradient
- policy gradient
- reinforcement learning
- gradient method
- function approximation
- reinforcement learning algorithms
- optimal control
- blind source separation
- independent component analysis
- approximation methods
- learning rate
- partially observable markov decision processes
- variance reduction
- average reward
- temporal difference
- reinforcement learning methods
- learning tasks
- objective function