No abstract

This paper presents a symbol level account of some of the representation and reasoning structures within the LOOM knowledge representation system. Reasoning in LOOM centers around a classifier whose primary function is to construct a taxonomy of all descriptions that have been entered into the system. The LOOM classifier is unique in that it constructs a separate taxonomy for each of seven kinds of non-composite descriptions, and uses a marker passing algorithm to replace the quadratic time subsumption test found in most classifiers with a linear time test. We briefly illustrate how the selection of data structures within LOOM impacts the completeness of the classification algorithm, and we describe the LOOM option that allows concepts to be reasoned with in either a forward-chaining or a backward-chaining mode. IntroductionObject-centered languages (class-slot-object syntax) and rule-based languages (relational syntax) represent two distinct frameworks for representing knowledge. Objectcentered representations can be found in a variety of settings, including programming languages, database systems, and knowledge representation (KR) systems (where the objectcentered language is often referred to as the "frame" language). In each of these settings, the object-centered representation frequently enables the development of applications that exhibit increased modularity and extensibility, and are easier to comprehend by application builders. Many KR systems attempt to combine object-centered and rule based languages to obtain the benefits of both. These efforts have in general been less than successful because the object-centered paradigm does not mesh well with the rule-based paradigm--in particular, representations that exploit syntactic unification (e.g., Prolog-style unification) are antithetical to the object-centered approach. Terminological logics[4] (we prefer the term description logics) represent a logic-based paradigm that is compatible with an object-centered approach, and present an alternative to rule-based languages. In this paper we introduce description logics, focusing on the description language implemented for the LOOM [6] KR system. LOOM defines a vocabulary of features that constitute the basic building blocks for constructing descriptions. We introduce a reasoner for descriptions, called a classifier, and present a glimpse of how the LOOM classifier reasons with descriptions and features. We provide a very informal complexity analysis of the the LOOM classification algorithm, and present some indications of how altering the feature vocabulary might improve the efficiency and/or completeness of *This research was sponsored by the Defense Advanced Research Projects Agency under contract MDA903-87-C-0641. the reasoner. Finally, we present LOOM's mixed forward and backward chaining architecture, and illustrate how it offers a user a trade-off between efficiency and inferential completeness. Descriptions and FeaturesIn a description logic, the notion of a "description" takes the place of a class or f...

No abstract

No abstract