A general method is described for translating an expert system into a functionally equivalent neural network. The neural network may be retranslated back into a functionally equivalent expert system. An example of the translation methodology is applied to the expert system shell M.1. The primary motivation for the development of the translation process is to aid in knowledge acquisition, however, there may be other applications such as the migration of a finished expert system to a neural network chip. The expert system and neural network knowledge base may be modified by either traditional knowledge engineering methods, associative learning techniques, or a synthesis of the two methods. Two learning algorithms are described that are appropriate for learning in the hybrid system--a modified backpropagation method and Goal-Directed Monte Carlo learning. These methods are used to demonstrate the repair of damaged certainty factors in several small rule bases. / Source: Dissertation Abstracts International, Volume: 52-03, Section: B, page: 1547. / Major Professor: R. C. Lacher. / Thesis (Ph.D.)--The Florida State University, 1991.
| Identifer | oai:union.ndltd.org:fsu.edu/oai:fsu.digital.flvc.org:fsu_76376 |
| Contributors | Kuncicky, David Carl., Florida State University |
| Source Sets | Florida State University |
| Language | English |
| Detected Language | English |
| Type | Text |
| Format | 98 p. |
| Rights | On campus use only. |
| Relation | Dissertation Abstracts International |
Page generated in 0.0905 seconds