ML Anthology

A Domain-Independent Framework for Effective Experimentation in Planning

Yolanda Gil
ICML 1991 pp. 13-17
doi:10.1016/B978-1-55860-200-7.50007-6 /icml/1991/gil1991icml-domain/

Abstract

This papers develops a general method for acquiring domain knowledge for planning by experimenting with their environment. When the expectations suggested by the domain knowledge and the observations differ, there is need and opportunity for learning. Since there are usually several possible ways to correct the domain, the system must experiment to gather additional information. This paper describes how to exploit the characteristics of planning domains in order to search the space of plausible hypotheses without the need for additional background knowledge to build causal explanations. Common features of planning domains are used in our system as heuristics to identify the most plausible hypotheses and avoid costly experiments. This work has been implemented in the PRODIGY planning architecture.

ICML Semantic Scholar

Cite

Text

Gil. "A Domain-Independent Framework for Effective Experimentation in Planning." International Conference on Machine Learning, 1991. doi:10.1016/B978-1-55860-200-7.50007-6

Markdown

[Gil. "A Domain-Independent Framework for Effective Experimentation in Planning." International Conference on Machine Learning, 1991.](https://mlanthology.org/icml/1991/gil1991icml-domain/) doi:10.1016/B978-1-55860-200-7.50007-6

BibTeX

@inproceedings{gil1991icml-domain,
  title     = {{A Domain-Independent Framework for Effective Experimentation in Planning}},
  author    = {Gil, Yolanda},
  booktitle = {International Conference on Machine Learning},
  year      = {1991},
  pages     = {13-17},
  doi       = {10.1016/B978-1-55860-200-7.50007-6},
  url       = {https://mlanthology.org/icml/1991/gil1991icml-domain/}
}