No substantial change in the balance between model-free and model-based control via training on the two-step task.
Elmar D. GrosskurthDominik R. BachMarcos EconomidesQuentin J. M. HuysLisa HolperPublished in: PLoS Comput. Biol. (2019)
Keyphrases
- model free
- reinforcement learning
- impedance control
- reinforcement learning algorithms
- temporal difference
- function approximation
- policy iteration
- control system
- training set
- supervised learning
- control method
- control strategy
- policy evaluation
- average reward
- optimal control
- neural network
- robotic systems
- data sets
- closed loop
- training examples
- labeled data
- graph cuts
- markov chain
- text mining
- rl algorithms
- machine learning