Viewing Classifier Systems as Model Free Learning in POMDPs
Abstract
Classifier systems are now viewed disappointing because of their prob(cid:173) lems such as the rule strength vs rule set performance problem and the credit assignment problem. In order to solve the problems, we have de(cid:173) veloped a hybrid classifier system: GLS (Generalization Learning Sys(cid:173) tem). In designing GLS, we view CSs as model free learning in POMDPs and take a hybrid approach to finding the best generalization, given the total number of rules. GLS uses the policy improvement procedure by Jaakkola et al. for an locally optimal stochastic policy when a set of rule conditions is given. GLS uses GA to search for the best set of rule conditions.
Cite
Text
Hayashi and Suematsu. "Viewing Classifier Systems as Model Free Learning in POMDPs." Neural Information Processing Systems, 1998.Markdown
[Hayashi and Suematsu. "Viewing Classifier Systems as Model Free Learning in POMDPs." Neural Information Processing Systems, 1998.](https://mlanthology.org/neurips/1998/hayashi1998neurips-viewing/)BibTeX
@inproceedings{hayashi1998neurips-viewing,
title = {{Viewing Classifier Systems as Model Free Learning in POMDPs}},
author = {Hayashi, Akira and Suematsu, Nobuo},
booktitle = {Neural Information Processing Systems},
year = {1998},
pages = {989-995},
url = {https://mlanthology.org/neurips/1998/hayashi1998neurips-viewing/}
}