Transfer of samples in batch reinforcement learning.
Alessandro LazaricMarcello RestelliAndrea BonariniPublished in: ICML (2008)
Keyphrases
- reinforcement learning
- transfer learning
- batch mode
- function approximation
- state space
- knowledge transfer
- training set
- control policy
- small sample
- active learning
- model free
- neural network
- action selection
- learning problems
- markov decision processes
- labeled data
- supervised learning
- optimal control
- text classification
- dynamic programming
- reinforcement learning algorithms
- learning algorithm
- genetic algorithm
- sample set
- data sets
- previously learned