Speeding-up reinforcement learning through abstraction and transfer learning.
Marcelo Li KogaValdinei Freire da SilvaFábio Gagliardi CozmanAnna Helena Reali CostaPublished in: AAMAS (2013)
Keyphrases
- transfer learning
- reinforcement learning
- state abstraction
- learning tasks
- knowledge transfer
- cross domain
- function approximation
- machine learning
- manifold alignment
- state space
- labeled data
- semi supervised learning
- collaborative filtering
- text categorization
- multi task learning
- target domain
- reinforcement learning agents
- transferring knowledge
- active learning
- multi task
- transfer knowledge
- learning algorithm
- text classification
- learning process
- model free
- unlabeled data
- cross domain learning
- markov decision processes
- optimal policy
- machine learning algorithms
- previously learned
- unsupervised learning
- supervised learning