Automated Knowledge Generation
-
Gonzalez, A. J. and et al., "Validation
of an Automated System Model Generator," Institute
of Electrical and Electronic Engineers, 1994.
[
ABSTRACT
|
FULL DOCUMENT ] (Full document will be loaded onto a new page.)
Modeling and simulation have always been highly useful techniques to use when designing, analyzing, or diagnosing an engineered system. Development of an appropriate model In terms of accuracy, granularity, and coniplestity, has typically been the burden of the designer, analyst, or troubleshooter. It would nutw'ally be advantageous If a model could Lie developed automatically, with the user supplying only some final minor refinements. The fact that most modern systems are designed in a cosnputer.aided design (CAD) environment makes this a realistic prospect, because much of the data necessary forthe model is already in electronic form. This article outlines a system called the automated knowledge generator (AKG), which embodies techniques that automatically create a model of an engineered system directly from its CAD representation, and describes the extensive testing process followed In order to validate its performance.
-
Gonzalez, A. J., Myler, H. R., and Kladke, R. R., "Identification
of Unconstrained Item Description Using String-Matched Hueristics," Journal
of Expert Systems - Research Applications, 1991.
[
ABSTRACT
|
FULL DOCUMENT ] (Full document will be loaded onto a new page.)
The objective of the research described is to identify unknown items by using heuristics to compare their unconstrained description with the exact labels of known items in an external database composed of generic elements found in the domain of interest. The string match-ing heuristics return a numerical measure of the similarity between the unknown description and the various candidate identifications in the external database. An examination of the functional constraints of the unknown item as compared to the candidate matches also contributes to the aforementioned numerical measure. The system described, called the Heuristic String Identifier (HSI), also employs auxiliary processes to optimize the search through the database and ensure that the best matches will always result. The goal of the HSI is to return an optimal set of those possible identifications, correctly ordered by string similarity and functional relatedness. The HSI attempts to identify the item within the domain data-base that exactly matches the target item. Failing this, it must determine which items in the database are most closely related descriptionally and conceptually to the unidentified item. Such a process entails discrimination of the most plausible hypotheses under the given constraints from the entire set of possibilities (i.e., the entire database). The approach can be viewed as a rudimentary analog to concept acquisition (Rendell, 1987).
-
Gonzalez, A. J. and Myler, H. R., "Issues
in Automating the Extraction of a System Model from CAD Databases
for the Use in Model-based Reasoning," Journal
of Expert Systems - Research Applications, 1991.
[
ABSTRACT
|
FULL DOCUMENT ] (Full document will be loaded onto a new page.)
The knowledge acquistion task has proven to be one of the greatest abstacles to the widespread use of knowledge-based systems. Traditional methods for eliciting knowledge from the various sources (usually human experts) are generally labor-intensive. This has made development of such systems quite expensive in many cases. One attractive alternative for breaking the bottleneck is to automate knowledge acquisition task.