LC-Learning: Phased Method for Average Reward Reinforcement Learning - Preliminary Results.
Taro KondaShinjiro TensyoTomohiro YamaguchiPublished in: PRICAI (2002)
Keyphrases
- prior knowledge
- learning scheme
- neural nets
- reinforcement learning
- average reward reinforcement learning
- learning algorithm
- objective function
- high precision
- support vector machine
- unsupervised learning
- input data
- cost function
- dynamic programming
- high accuracy
- supervised learning
- feature set
- em algorithm
- synthetic data
- machine learning methods
- significant improvement
- active learning
- learning process
- artificial neural networks
- genetic algorithm
- neural network
- learning mechanism
- feature extraction
- similarity measure
- learning tasks
- preprocessing
- training set
- classification method
- detection method
- theoretical analysis
- computationally efficient
- computational cost