ML Anthology

A Hybrid Approach to Guaranteed Effective Control Strategies

Jonathan Gratch, Gerald DeJong
ICML 1991 pp. 509-513
doi:10.1016/B978-1-55860-200-7.50104-5 /icml/1991/gratch1991icml-hybrid/

Abstract

Machine learning has demonstrated success in enhancing the effectiveness of domain–independent planning. Unfortunately the conditions for which success can be realized are difficult to articulate. Promising empirical results can be overturned by subtle changes to the input. This paper describes amethod which generates strategies guaranteed to enhanceplanning performance to an arbitrary level of certainty. The problem is naturally approached by a combination of theory–based and empirical learning techniques.

ICML Semantic Scholar

Cite

Text

Gratch and DeJong. "A Hybrid Approach to Guaranteed Effective Control Strategies." International Conference on Machine Learning, 1991. doi:10.1016/B978-1-55860-200-7.50104-5

Markdown

[Gratch and DeJong. "A Hybrid Approach to Guaranteed Effective Control Strategies." International Conference on Machine Learning, 1991.](https://mlanthology.org/icml/1991/gratch1991icml-hybrid/) doi:10.1016/B978-1-55860-200-7.50104-5

BibTeX

@inproceedings{gratch1991icml-hybrid,
  title     = {{A Hybrid Approach to Guaranteed Effective Control Strategies}},
  author    = {Gratch, Jonathan and DeJong, Gerald},
  booktitle = {International Conference on Machine Learning},
  year      = {1991},
  pages     = {509-513},
  doi       = {10.1016/B978-1-55860-200-7.50104-5},
  url       = {https://mlanthology.org/icml/1991/gratch1991icml-hybrid/}
}