Inferring Smooth Control: Monte Carlo Posterior Policy Iteration with Gaussian Processes.
Joe WatsonJan PetersPublished in: CoRR (2022)
Keyphrases
- monte carlo
- gaussian processes
- gaussian process
- policy iteration
- policy evaluation
- temporal difference
- markov chain
- markov decision processes
- importance sampling
- finite state
- optimal control
- bayesian framework
- markov chain monte carlo
- regression model
- semi supervised
- fixed point
- hyperparameters
- infinite horizon
- model free
- variance reduction
- approximate inference
- control strategy
- latent variables
- optimal policy
- least squares
- multi task
- model selection
- linear programming
- reinforcement learning
- posterior distribution
- decision trees
- maximum likelihood
- markov random field