Bayesian Inference for Least Squares Temporal Difference Regularization.
Nikolaos TziortziotisChristos DimitrakakisPublished in: ECML/PKDD (2) (2017)
Keyphrases
- bayesian inference
- temporal difference
- least squares
- policy evaluation
- prior information
- reinforcement learning
- td learning
- function approximation
- evaluation function
- policy iteration
- monte carlo
- step size
- probabilistic model
- statistical inference
- hyperparameters
- model free
- action selection
- parameter estimation
- reinforcement learning algorithms
- particle filter
- prior knowledge
- expectation propagation
- supervised learning
- convergence speed
- radial basis function
- function approximators
- pairwise
- training data
- wavelet transform
- data sets
- em algorithm