Model Learning for Multistep Backward Prediction in Dyna-Q Learning.
Kao-Shing HwangWei-Cheng JiangYu-Jen ChenIris HwangPublished in: IEEE Trans. Syst. Man Cybern. Syst. (2018)
Keyphrases
- learning algorithm
- reinforcement learning
- prior knowledge
- prediction model
- learning process
- cooperative
- learning scheme
- bi directional
- objective function
- model free
- decision theoretic
- learning mechanism
- online learning
- active learning
- action selection
- predictive model
- learning phase
- neural network
- temporal difference learning
- learning tasks
- mathematical model
- learning systems
- prediction accuracy
- probabilistic model
- multi agent