Bayesian Learning of Noisy Markov Decision Processes
This work addresses the problem of estimating the optimal value function in a MarkovDecision Process from observed state-action pairs. We adopt a Bayesian approach toinference, which allows both the model to be estimated and predictions about actions tobe made in a unified framework, providing a principled approach to mimicry of a controlleron the basis of observed data. A new Markov chain Monte Carlo (MCMC) sampler isdevised for simulation from the posterior distribution over the optimal value function.This step includes a parameter expansion step, which is shown to be essential for goodconvergence properties of the MCMC sampler. As an illustration, the method is appliedto learning a human controller.
Year of publication: |
2010
|
---|---|
Authors: | Singh, Sumeetpal S. ; Chopin, Nicolas ; Whiteley, Nick |
Institutions: | Centre de Recherche en Économie et Statistique (CREST), Groupe des Écoles Nationales d'Économie et Statistique (GENES) |
Saved in:
freely available
Saved in favorites
Similar items by person
-
Chopin, Nicolas, (2013)
-
Bayesian learning of noisy Markov decision processes
Singh, Sumeetpal S., (2010)
-
A Sequential Particle Filter Method for Static Models
Chopin, Nicolas, (2000)
- More ...