Learning from demonstration with partially observable task parameters using dynamic movement primitives and Gaussian process regression.
Tohid AlizadehMilad S. MalekzadehSoheila BarzegariPublished in: AIM (2016)
Keyphrases
- gaussian process regression
- partially observable
- markov decision processes
- state space
- gaussian processes
- decision problems
- dynamical systems
- partial observability
- reinforcement learning
- gaussian process
- parameter estimation
- maximum likelihood
- partial observations
- infinite horizon
- belief state
- orders of magnitude
- expectation maximization