ML Anthology

KBG : A Knowledge Based Generalizer

Gilles Bisson
ICML 1990 pp. 9-15
doi:10.1016/B978-1-55860-141-3.50005-5 /icml/1990/bisson1990icml-kbg/

Abstract

Fundamentally, generalization can be seen as a technique for performing compression of information; the result of this process provides the basis to learn new knowledge. However, in all domains, the use of an elementary process requires the implementation of both fast and efficient algorithms. In this paper, we present a new generalization mechanism, inspired by the structural matching algorithm. The principle of our method consists in using the domain theory in order to perform the saturation of the examples given by an expert. We will see that this “rough” approach provides some advantages both for generalization quality and for construction rapidity.

ICML Semantic Scholar

Cite

Text

Bisson. "KBG : A Knowledge Based Generalizer." International Conference on Machine Learning, 1990. doi:10.1016/B978-1-55860-141-3.50005-5

Markdown

[Bisson. "KBG : A Knowledge Based Generalizer." International Conference on Machine Learning, 1990.](https://mlanthology.org/icml/1990/bisson1990icml-kbg/) doi:10.1016/B978-1-55860-141-3.50005-5

BibTeX

@inproceedings{bisson1990icml-kbg,
  title     = {{KBG : A Knowledge Based Generalizer}},
  author    = {Bisson, Gilles},
  booktitle = {International Conference on Machine Learning},
  year      = {1990},
  pages     = {9-15},
  doi       = {10.1016/B978-1-55860-141-3.50005-5},
  url       = {https://mlanthology.org/icml/1990/bisson1990icml-kbg/}
}