ML Anthology

Learning to Select a Model in a Changing World

Mieczyslaw M. Kokar, Spyros A. Reveliotis
ICML 1991 pp. 313-317
doi:10.1016/B978-1-55860-200-7.50065-9 /icml/1991/kokar1991icml-learning/

Abstract

This paper presents initial results of an investigation of the feasibility of using hierarchical reinforcement learning methods in a restructurable control scenario. In restructurable control the plant's behavior at different times needs to be described by different sets of variables and relationships. Two main problems addressed in this research are: (1) the ability to learn two policy functions (model selection and action selection), and (2) the space complexity for representing such policies in the case of continuous physical systems.

ICML Semantic Scholar

Cite

Text

Kokar and Reveliotis. "Learning to Select a Model in a Changing World." International Conference on Machine Learning, 1991. doi:10.1016/B978-1-55860-200-7.50065-9

Markdown

[Kokar and Reveliotis. "Learning to Select a Model in a Changing World." International Conference on Machine Learning, 1991.](https://mlanthology.org/icml/1991/kokar1991icml-learning/) doi:10.1016/B978-1-55860-200-7.50065-9

BibTeX

@inproceedings{kokar1991icml-learning,
  title     = {{Learning to Select a Model in a Changing World}},
  author    = {Kokar, Mieczyslaw M. and Reveliotis, Spyros A.},
  booktitle = {International Conference on Machine Learning},
  year      = {1991},
  pages     = {313-317},
  doi       = {10.1016/B978-1-55860-200-7.50065-9},
  url       = {https://mlanthology.org/icml/1991/kokar1991icml-learning/}
}