Knowledge-Based Validity Maintenance for Production Systems
Abstract
In many problem domains, an action may be taken by an expert, which, due to new inferences or a changing domain situation, should be retracted. To this end, an effective problem solver will need to use some kind of validity-maintenance system, so that it can gracefully recover from invalid previous decisions. Unfortunately, the standard IF/THEN paradigm often used to encode expert behavior does not readily allow the expression and processing of this validity knowledge. We present a new extension to that rule paradigm which can be used to augment production-rule-based systems with validity maintenance capabilities, and demonstrate a straightforward algorithm for its interpretation.
Cite
Text
Schaefer et al. "Knowledge-Based Validity Maintenance for Production Systems." AAAI Conference on Artificial Intelligence, 1986.Markdown
[Schaefer et al. "Knowledge-Based Validity Maintenance for Production Systems." AAAI Conference on Artificial Intelligence, 1986.](https://mlanthology.org/aaai/1986/schaefer1986aaai-knowledge/)BibTeX
@inproceedings{schaefer1986aaai-knowledge,
title = {{Knowledge-Based Validity Maintenance for Production Systems}},
author = {Schaefer, Philip and Bozma, H. Isil and Beer, Randall D.},
booktitle = {AAAI Conference on Artificial Intelligence},
year = {1986},
pages = {918-922},
url = {https://mlanthology.org/aaai/1986/schaefer1986aaai-knowledge/}
}