Learning When to Trust a Dynamics Model for Planning in Reduced State Spaces.
Dale McConachieThomas PowerPeter MitranoDmitry BerensonPublished in: IEEE Robotics Autom. Lett. (2020)
Keyphrases
- prior knowledge
- learning algorithm
- decision theoretic
- supervised learning
- learning scheme
- probabilistic model
- active learning
- experimental data
- heuristic search
- mathematical model
- computational model
- learning process
- artificial neural networks
- learning mechanism
- high level
- learning phase
- trust model
- neural network
- planning problems
- dynamical systems
- domain independent
- learning systems
- state space
- objective function
- reinforcement learning
- similarity measure