Control Frequency Adaptation via Action Persistence in Batch Reinforcement Learning.
Alberto Maria MetelliFlavio MazzoliniLorenzo BisiLuca SabbioniMarcello RestelliPublished in: ICML (2020)
Keyphrases
- reinforcement learning
- action selection
- control problems
- optimal control
- function approximation
- control policy
- control system
- robot control
- dynamic programming
- robotic control
- action space
- partially observable domains
- real time
- reinforcement learning algorithms
- learning algorithm
- neural network
- control algorithm
- model free
- control strategies
- low frequency
- control strategy
- incremental learning
- high frequency
- frequency domain
- batch mode
- transition model
- machine learning