Simultaneous feature selection and parameter optimization for training of dialog policy by reinforcement learning.

Teruhisa Misu Hideki Kashioka

Published in: SLT (2012)

Keyphrases

parameter optimization
reinforcement learning
feature selection
genetic algorithm ga
optimal policy
parameter settings
genetic algorithm
parameter estimation
support vector machine
differential evolution
pid controller
state space
supervised learning
training set
learning process
knn
fitness function