DMCA
Refinement of Approximate Domain Theories by Knowledge-Based Neural Networks (1990)
| Venue: | In Proceedings of the Eighth National Conference on Artificial Intelligence |
| Citations: | 195 - 15 self |
BibTeX
@INPROCEEDINGS{Towell90refinementof,
author = {Geoffrey G. Towell and Jude W. Shavlik and Michiel O. Noordewier},
title = {Refinement of Approximate Domain Theories by Knowledge-Based Neural Networks},
booktitle = {In Proceedings of the Eighth National Conference on Artificial Intelligence},
year = {1990},
pages = {861--866}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Standard algorithms for explanation-based learning require complete and correct knowledge bases. The KBANN system relaxes this constraint through the use of empirical learning methods to refine approximately correct knowledge. This knowledge is used to determine the structure of an artificial neural network and the weights on its links, thereby making the knowledge accessible for modification by neural learning. KBANN is evaluated by empirical tests in the domain of molecular biology. Networks created by KBANN are shown to be superior, in terms of their ability to correctly classify unseen examples, to randomly initialized neural networks, decision trees, "nearest neighbor" matching, and standard techniques reported in the biological literature. In addition, KBANN's networks improve the initial knowledge in biologically interesting ways. Introduction Explanation-based learning (EBL) (Mitchell et al. 1986; DeJong & Mooney 1986) provides a way of incorporating pre-existing knowledge i...
Keyphrases
approximate domain theory knowledge-based neural network neural learning decision tree neighbor matching pre-existing knowledge correct knowledge interesting way kbann system unseen example standard algorithm biological literature artificial neural network dejong mooney molecular biology initial knowledge empirical test standard technique introduction explanation-based learning explanation-based learning require correct knowledge base initialized neural network
