Douglas R. Skuce

CODE4: a unified system for managing conceptual knowledge

Abstract CODE4 is a general-purpose knowledge management system, intended to assist with the common knowledge processing needs of anyone who desires to analyse, store, or retrieve conceptual knowledge in applications as varied as the specification, design and user documentation of computer systems; the construction of term banks, or the development of ontologies for natural language understanding. This paper provides an overview of CODE4 as follows: We first describe the general philosophy and rationale of CODE4 and relate it to other systems. Next, we discuss the knowledge representation, specifically designed to meet the needs of flexible, interactive knowledge management. The highly-developed user interface, which we believe to be critical for this type of system, is explained in some detail. We finally describe how CODE4 is being used in a number of applications.

Towards a new generation of terminological resources: an experiment in building a terminological knowledge base

This paper describes a project to construct a terminological knowledge base, called COGNITERM. First, we position our research framework in relationship to recent developments in computational lexicology and knowledge engineering. Second, we describe the COGNTTERM prototype and discuss its advantages over conventional term banks. Finally, we outline some of the methodological issues that have emerged from our work.

A logic-based knowledge source system for natural language documents

Abstract A methodology is described for creating and structuring a knowledge base system for storing and intelligently retrieving the types of knowledge found in regulations, specifications and other similar documents. The approach involves translating the document into a simple, formal, English-like rule-based language. The language combines production rules with constraints, hierarchies and scripts. Our methodology structures and guides the process of translation and verification using several software tools. The retrieval process is based on associative indexing of rules and facts. Though our current domain is government travel regulations, the method is general enough to be useful for most technical material, such as software descriptions, design or service specifications, or parts of training manuals.

Concept analysis and terminology: a knowledge-based approach to documentation

The central concern of terminology, a component of the general documentation process, is concept analysis, an activity which is becoming recognized as fundamental as term banks evolve into knowledge bases. We propose that concept analysis can be facilitated by knowledge engineering technology, and describe a generic knowledge acquisition tool called CODE (Conceptually Oriented Design Environment) that has been successfully used in two terminology applications: 1) a bilingual vocabulary project with the Terminology Directorate of the Secretary of State of Canada, and 2) a software documentation project with Bell Northern Research. We conclude with some implications of computer-assisted concept analysis for terminology.

Integrating Web‐Based Documents, Shared Knowledge Bases, and Information Retrieval for User Help

We describe a prototype system, IKARUS, with which we investigated the potential of integrating web‐based documents, shared knowledge bases, and information retrieval for improving knowledge storage and retrieval. As an example, we discuss how to implement both a user manual and an online help system as one system. The following technologies are combined: a web‐based design, a frame‐based knowledge engine, use of an advanced full‐text search engine, and simple techniques to control terminology. We have combined graphical browsing with several unusual forms of text retrieval—for example, to the sentence and paragraph level.

Beyond hypertext: knowledge management for technical documentation

We describe the use of our knowledge management system, CODE2, as an aid to documenters of a complex software system called Telos. CODE2 was first used by the designers of Telos to clarify their design concepts and terminology. CODE2 served the following purposes: 1) to acquire the knowledge about the system; 2) to check the terminology in natural language documents; 3) as an on-line knowledge resource for documenters and end-users, and 4) to automatically generate parts of the printed documentation. We describe some features that make CODE particularly useful to documenters: its sophisticated user interface, its ability to handle both formal and informal knowledge, and its support for language. We also describe our vision for the future of such knowledge-based technology.

BRIDGING THE GAP BETWEEN NATURAL AND COMPUTER LANGUAGE

Very few people possess the skill to express themselves concisely and unambiguously in natural language (NL). Modern society nevertheless can expect a continuing increase in demand for the production of NL text, for the concomitant comprehension of this information, and for converting NL information into computer languages (CLs). The semantic errors and confusion which usually result from the imprecise use of NL will therefore be multiplied, particularly in the latter case. A solution to this problem is proposed, based on a language LESK (Language for Exactly Stating Knowledge), which is intermediate between NLs and CLs. LESK looks like NL, but employs simple mathematical, linguistic, and computer principles. An example in which LESK is used to clarify some of a universitys academic rules is presented. It is argued that both developed and developing nations could benefit greatly from widespread use of a lanugage like LESK.

A FRAME-LIKE KNOWLEDGE REPRESENTATION INTEGRATING ABSTRACT DATA TYPES AND LOGIC

Abstract This paper describes a very general, formal knowledge representation based on algebraic ideas from abstract data types, i.e., on ideas originating outside the AI literature in research on foundations for programming and specification languages. Nevertheless, the representation incorporates practically necessary features found in inheritance systems such as AI frame systems used for natural language understanding, while offering a precise algebraic semantics. We term the approach conceptually oriented description. The contribution of this chapter is (1) to reformulate and simplify these ideas for AI applications, incorporating the useful features found in many practical AI inheritance systems, while retaining the theoretical foundation, and (2) to show how the approach is valuable in natural language semantics applications. This chapter will use some difficult examples motivated by natural language applications, but the formalism is very general and could be used for other applications, such as software requirements specification. The approach is based mainly on ideas from the language LOTOS, intended for the formal specification of concurrent systems, and also on similar work by Goguen and Meseguer [1987] on algebraically based functional specification. LOTOS adds to basic ADT concepts additional concepts for defining the notions like state, event, and temporal relationships including causality and synchronization. The main components of a software system for creating and debugging conceptual definitions using the formalism have been implemented and are briefly mentioned.

A language and system for making definitions of technical concepts

Abstract A simple English-like language is described in which one can make clear definitions of certain types of technical concepts needed in computer systems design. The semantics are based on conceptual graphs, as introduced by Sowa [25], which we briefly review. An example using real concepts from commercial fourth generation software is discussed. In addition, a Smalltalk-based tool is described which permits the rapid acquisition, editing, and graphical display of definitions, to aid in conceptual design and the choice of terminology. A major point of the paper is that such a process of clarifying terminology should become an essential component of software design. This should become recognized as influences from artificial intelligence become more common in software design.

Expressing qualitative biomedical knowledge exactly using the language LESK

A language (LESK) is described which allows one to make mainly qualitative declarations of the essential terminology and facts in some technical subject. LESK is essentially first order predicate calculus, made to look like natural language. A LESK system, now partially implemented in PROLOG, allows one to enter declarations and assertions, check consistency, and deductively answer predominately generic, universally quantified questions. An extensive example is presented in which LESK is used to describe neurophysiological knowledge about the stretch reflex.