Learning nonlinear state-space models using smooth particle-filter-based likelihood approximations.
Andreas SvenssonFredrik LindstenThomas B. SchönPublished in: CoRR (2017)
Keyphrases
- state space
- prior knowledge
- reinforcement learning
- learning process
- learning algorithm
- accurate models
- learning tasks
- neural network
- structured prediction
- hidden variables
- dynamic model
- computational models
- supervised learning
- learned models
- learning models
- bayesian framework
- online learning
- markov decision processes
- dynamical systems
- statistical models
- object tracking
- learning systems
- particle filter
- markov chain