Curious Meta-Controller: Adaptive Alternation between Model-Based and Model-Free Control in Deep Reinforcement Learning.
Muhammad Burhan HafezCornelius WeberMatthias KerzelStefan WermterPublished in: IJCNN (2019)
Keyphrases
- model free
- impedance control
- reinforcement learning
- adaptive control
- reinforcement learning algorithms
- optimal control
- temporal difference
- function approximation
- control policy
- control system
- policy iteration
- control strategies
- visual servoing
- control method
- adaptive neural
- learning algorithm
- reinforcement learning methods
- control strategy
- robotic manipulator
- state space
- inverted pendulum
- control algorithm
- neural network
- average reward
- policy evaluation
- actor critic
- temporal difference learning
- feature selection
- supervised learning
- closed loop