Neural Fitted Q Iteration - First Experiences with a Data Efficient Neural Reinforcement Learning Method
Abstract
This paper introduces NFQ, an algorithm for efficient and effective training of a Q-value function represented by a multi-layer perceptron. Based on the principle of storing and reusing transition experiences, a model-free, neural network based Reinforcement Learning algorithm is proposed. The method is evaluated on three benchmark problems. It is shown empirically, that reasonably few interactions with the plant are needed to generate control policies of high quality.
Cite
Text
Riedmiller. "Neural Fitted Q Iteration - First Experiences with a Data Efficient Neural Reinforcement Learning Method." European Conference on Machine Learning, 2005. doi:10.1007/11564096_32Markdown
[Riedmiller. "Neural Fitted Q Iteration - First Experiences with a Data Efficient Neural Reinforcement Learning Method." European Conference on Machine Learning, 2005.](https://mlanthology.org/ecmlpkdd/2005/riedmiller2005ecml-neural/) doi:10.1007/11564096_32BibTeX
@inproceedings{riedmiller2005ecml-neural,
title = {{Neural Fitted Q Iteration - First Experiences with a Data Efficient Neural Reinforcement Learning Method}},
author = {Riedmiller, Martin A.},
booktitle = {European Conference on Machine Learning},
year = {2005},
pages = {317-328},
doi = {10.1007/11564096_32},
url = {https://mlanthology.org/ecmlpkdd/2005/riedmiller2005ecml-neural/}
}