Loom(ID:3991/loo007)Knowledge Representation Language with Frames, Objects, Constraints and PRS Raymond Bates and Robert MacGregor USC/ISI 1987 Related languages
References: knowledge representation formalisms that employ a formal language with a formal semantics for the definition of terms (more commonly referred to as concepts or classes) and that deduce whether one term subsumes (is more general than) another. These formalisms generally descend from the ideas presented in KL-One (Brachman and Schmolze 1985). TSLs are a generalization of both semantic networks and frames. One result of the workshop was to standardize use of the term terminological logics to describe these formalisms; term subsumption languages was chosen as a neutral term for describing the workshop. In the last few years, many knowledge representation systems have been built using TSLs, including Krypton (Brachman et al. 1985), KLTwo (Vilain 1984), NIKL (Robbins 1986; Kaczmarek, Bates, and Robbins 1986), Back (Peltason et al. 1989; Nebel and vonLuck 1988), Meson (Edelmann and Owsnicki 1986), SBOne (Kobsa 1990), Loom (MacGregor and Bates 1987), Quirk (Bergmann and Gerlach 1987), and Classic (Borgida et al. 1989). These systems go beyond a bare TSL in various ways: Almost all of them incorporate assertional languages that enable the systems to reason about instances of terms, some of them allow for retraction of told facts, and so on. The workshop not only concerned TSLs but also TSL-based knowledge representation systems and their use in larger AI systems. Outline of the Workshop The workshop was designed to encourage discussion. To aid this approach, no formal talks were presented, and no proceedings is being produced. For a large portion of the workshop, the attendees were divided into working groups of 7 to 15 participants. Each working group was devoted to in-depth discussion of particular topics. Moderators were chosen to keep the discussions flowing and on track and were assisted by preselected discussants who presented short position statements. Ample time was left for intensive discussion, although several of the discussions could not be completed within their allotted time and had to be continued in the evening. Moderators reported the results of the working groups in plenary sessions that also allowed for further discussion of the topics covered. in AI Magazine Summer 1990 view details in AI Magazine Summer 1990 view details in John Sowa, ed., Principles of Semantic Networks: Explorations in the representation of knowledge , Morgan-Kaufmann: San Mateo, California, 1991 view details in ACM SIGART Bulletin 2(3) June 1991 Special issue on implemented knowledge representation and reasoning systems view details Introduction In its inception, the LOOM system [MacGregor 88j was designed as a self-contained, logic-based knowledge representation (KR) system. From observing early applications based on LOOM we concluded that this "black box" approach to building a KR system was wrong?a great deal of programming effort was expended in developing useful programming interfaces to the KR system. Our response was ro look for programming paradigms which couid be directly hooked into the KR system. We extended the scope of the language to incorporate several of these paradigms, so that now instead of being an inscrutable oracle. LOOM represents an environment within which application programs can be written. The present LOOM language is designed to capture the best features among the following paradigms: 1) Object-oriented programming (message passing) 2) Data-driven programming (production systems) 3) Problem solving 4) Constraint programming Our intent was to design a language such that the several paradigms blend together and complement one another. This paper describes the approach taken and some of the motivations behind that approach. in ACM SIGART Bulletin 2(3) June 1991 Special issue on implemented knowledge representation and reasoning systems view details in ACM SIGART Bulletin 2(3) June 1991 Special issue on implemented knowledge representation and reasoning systems view details |