Variance Reduction Techniques for Gradient Estimates in Reinforcement Learning.
Evan GreensmithPeter L. BartlettJonathan BaxterPublished in: NIPS (2001)
Keyphrases
- variance reduction
- gradient estimation
- policy gradient
- reinforcement learning
- importance sampling
- confidence intervals
- monte carlo
- actor critic
- sample size
- function approximation
- quasi monte carlo
- reinforcement learning algorithms
- state space
- markov chain
- bias variance decomposition
- learning algorithm
- text mining
- probabilistic model
- dynamic programming