Merging with Extraction Method for Transfer Learning in Actor-Critic.
Toshiaki TakanoHaruhiko TakaseHiroharu KawanakaShinji TsuruokaPublished in: J. Adv. Comput. Intell. Intell. Informatics (2011)
Keyphrases
- transfer learning
- actor critic
- reinforcement learning
- learning tasks
- temporal difference
- policy gradient
- reinforcement learning algorithms
- function approximation
- approximate dynamic programming
- neuro fuzzy
- cross domain
- optimal control
- gradient method
- machine learning
- model free
- labeled data
- multi task
- state space
- learning algorithm
- policy iteration
- structure learning
- transfer knowledge
- data mining
- learning problems
- machine learning algorithms
- semi supervised learning
- collaborative filtering
- active learning
- convergence rate
- optimal policy
- text categorization
- reward function
- single agent
- text mining
- average reward
- semi supervised
- learning process
- data sets