Feature selection and policy optimization for distributed instruction placement using reinforcement learning.
Katherine E. CoonsBehnam RobatmiliMatthew E. TaylorBertrand A. MaherDoug BurgerKathryn S. McKinleyPublished in: PACT (2008)
Keyphrases
- reinforcement learning
- feature selection
- optimal policy
- optimal placement
- multi agent
- policy search
- partially observable environments
- optimization algorithm
- distributed systems
- action selection
- function approximation
- state space
- markov decision process
- cooperative
- feature selection algorithms
- distributed environment
- markov decision processes
- function approximators
- learning algorithm
- text classification
- learning process
- parameter optimization
- policy iteration
- text categorization
- optimization problems
- support vector
- dynamic programming
- reinforcement learning algorithms
- machine learning
- multimedia
- state and action spaces
- actor critic
- policy gradient
- multi agent systems
- feature space
- partially observable markov decision processes
- model free
- classification accuracy
- feature set
- mutual information
- peer to peer