Learning by switching generation and reasoning methods - acquisition of meta-knowledge for switching with reinforcement learning.
Masahiro TomaruMotohide UmanoYuji MatsumotoKazuhisa SetaPublished in: FUZZ-IEEE (2008)
Keyphrases
- reinforcement learning
- meta knowledge
- learning algorithm
- reinforcement learning methods
- learning process
- machine learning
- meta learning
- prior knowledge
- knowledge base
- active learning
- supervised learning
- logic programs
- machine learning methods
- learning tasks
- development process
- inductive learning
- feature vectors
- reinforcement learning algorithms
- function approximators
- data sets