Probabilistic inference for determining options in reinforcement learning.
Christian DanielHerke van HoofJan PetersGerhard NeumannPublished in: Mach. Learn. (2016)
Keyphrases
- probabilistic inference
- reinforcement learning
- graphical models
- weighted model counting
- influence diagrams
- bayesian networks
- conditional probabilities
- belief networks
- approximate inference
- message passing
- context specific independence
- efficient inference
- probabilistic reasoning
- exact inference
- state space
- bayesian belief networks
- belief propagation
- elimination algorithm
- optimal policy
- bucket elimination
- dynamic programming
- probabilistic model
- learning process
- supervised learning
- feature vectors
- logic programming