Showing papers in "Knowledge Based Systems in 1990"

Mapping rule-based systems into neural architecture

Abstract: A novel approach has been developed that maps a rule-based expert system into the neural architecture in both the structural and behavioural aspects. Under this approach, the knowledge base and the inference engine are mapped into an entity are called 'conceptualization', where a node represents a concept and a link represents a relation between two concepts. A concept node is designated by a small number of language symbols. In the neural system transformed from a knowledge-based system, the inference behaviour is characterized by propagating and combining activations recursively through the network, and the learning behaviour is based upon a mechanism called 'back-propagation', which allows proper modification of connection strengths in order to adapt the system to the environment. This approach is based on the analogies observed between a belief network and a neural network, and its validity has been demonstrated by experiments. Finally, the advantages and disadvantages of this approach are discussed with respect to inference and learning.

Knowledge-support systems

Abstract: The social role of information technology is analysed in order to provide a framework for reasonable requirements for knowledge support systems. The many different scientific and engineering communities targeted on the development of specific knowledge support technologies are noted, and the problems of integrating developments from different communities are highlighted. The architectures of current integrated knowledge support systems are analysed within the social framework and the strengths and weaknesses of current modules are evaluated. Major system developments in recent years are used to exemplify the analysis, and expected directions for future development are outlined. The framework, analyses and examples are used to define a structured requirements specification for future research and development related to knowledge support systems.

Framework of knowledge-based systems: Multiple meta-level architecture for representing problems and problem-solving processes

Abstract: Currently available expert systems have a performance limit because of the lack of capability to describe problems and problem-solving methods. It is closely related with knowledge representation language, but this is not the only concern with this issue. Real world problems and problem-solving methods are not so simple as to be represented always in the same way by the same language. Their representations must be different depending on various factors involved in the problems themselves and the situations these problems are surrounded with. In this paper, the author discusses first the intrinsic nature of problem representation and problem-solving process representation. The requirements for and the conceptual framework of a knowledge-based system that is suited for dealing with various problems then become apparent quite naturally. The author asserts that a multiple meta-level architecture is necessary as well as a knowledge-representation language that can describe complex data structures as the basic framework of knowledge-based systems.

An architecture for knowledge-based front ends

Abstract: The problem of constructing knowledge-based front-ends to existing software systems is considered. An architecture for such a front-end is proposed that has, as its central component, a module known as the Harness. The Harness provides the communication facilities required by the various elements of the knowledge-based front-end as well as handling the interactions with the user. It, therefore, extends the notion of a user-interface management system to be applicable to the problem of integrating diverse and unmodified software systems to provide a single user service together with the necessary support.

A survey of methods used in probabilistic expert systems for knowledge integration

Abstract: A number of papers (and recently even a book) dealing with probabilistic methods for expert systems have appeared during the last five years. Nevertheless, they do not present many different philosophical view-points and methodological tools. This paper surveys the main approaches used to solve the following basic questions connected with the knowledge integration process: What is a relevant criterion for selection of an appropriate probability distribution complying with all the input knowledge? How should this distribution be constructed and represented in a computer memory?

Back-end manager: an interface between a knowledge-based front end and its application subsystems

Abstract: Front Ends for Open and Closed User Systems (FOCUS) is an ESPRIT/2 (no. 2620) project aimed at designing tools and techniques for the construction of knowledge-based front ends (KBFEs) for open-user systems (reusable software components, libraries, etc) and closed-user systems (free-standing software, packages, etc). An important part of the project involves the establishment of an architecture for KBFEs and the specification of the KBFE/back-end interface. This paper describes the properties and related issues of such an interface, known as the back-end manager (BEM), and its relationship to the proposed KBFE architecture.

Argument representation for editorial text

Abstract: This paper presents a theory of argument representation for computer comprehension of editorial text. The theory has been implemented in a prototype system called OpEd, which reads politico-economic editorials and answers questions about their contents. The theory characterizes five major classes of knowledge structures: beliefs, attack relationships, support relationships, argument units (AUs), and meta-argument units (meta-AUs). During editorial comprehension, OpEd represents explicitly the beliefs of the editorial writer and his/her implicit opponents. Beliefs involve predictions about goals, plans, events, and states. Three types of predications are distinguished: evaluations about plans, causal relationships, and beliefs about beliefs. Beliefs in an editorial are involved in support relationships or attack relationships. An attack is a relationship between two beliefs whose contents involve mutually-exclusive planning situations or opposite effects of a plan on interrelated goals. A support is a relationship that consists of a belief, the justification for the belief, and a warrant (itself a belief) that connects a belief to its justifications. Belief justifications are based on refinements of plan evaluations, refinements of plan-goal relationships, analogies, and examples. Beliefs, attack relationships and support relationships are organized by AUs, which encode language-free and domain-free knowledge about argument structure and content. With the aid of domain-specific knowledge, AUs can be instantiated to model arguments in which an arguer refutes his/her opponent's position that a given plan should/should not be used. Arguments about arguments are represented as meta-AUs, which specify argument errors that result from either inconsistencies between actions and beliefs, or support strategies involving plausibilities, circularities, self-contradictions, or shifts on the burden of proof. Meta-AUs are represented in OpEd as attacks on warrants and are used to model discussions about the nature of valid reasoning.

Systems development life-cycle for expert systems

Abstract: Existing life-cycles for the development of traditional information systems are shown to be inadequate for addressing expert system requirements. A life cycle for expert systems is constructed, outlining the tasks and activities to be performed at each stage of system development. The life cycle highlights the role of alternative development paradigms and the importance of social and organizational characteristics in system transfer to users. The operationalization of the life-cycle is illustrated through a case study, describing the development process and major architectural components of an expert system that configures air-conditioning units. Several important issues encountered in transferring the system from test to production mode and in system evaluation are highlighted. Such a life-cycle approach can enhance project management for expert systems.

Designers' activities examined at three levels: organization, strategies & problem-solving

Abstract: Through examples of data from three empirical design studies, this paper presents the type of analysis which cognitive psychology makes of the mental activities involved in design. These activities are analyzed at three levels: the way in which designers organize their activity, the main strategies they adopt, and the problem-solving processes they use. Different types of design tasks are presented: functional specification, software design and composite-structure design tasks. The relevance of the results for A.I. is discussed from an assistance viewpoint.

Argument comprehension and retrieval for editorial text

Abstract: OpEd is a prototype editorial comprehension and question-answering system in the domain of politico-economic protection. Editorial comprehension in OpEd involves the application of: domain-specific knowledge, abstract knowledge of argumentation, and strategies for mapping input editorial text into conceptual structures. OpEd's model of domain-specific knowledge includes four major elements: authority triangles and social acts, to represent explicitly all the information associated with conflicts in international trade, including beliefs, goals, and conflict-resolution methods: goals and plans, to represent political and economic actions in terms of desired economic states; a trade graph, to represent causal relationships among the economic quantities associated with products and consumers; and reasoning scripts, to represent common chains of cause-effect relationships in editorials. OpEd's central, organizing constructs of argument knowledge are termed argument units (AUs). AUs consist of configurations of support and attack relationships among beliefs, where the content of each belief refers to goal/plan situations. The knowledge of argumentation encoded by AUs is highly abstract and independent of any particular domain. As a result, argument comprehension in OpEd requires instantiating AUs with the aid of politico-economic knowledge. Associated with each knowledge construct in OpEd are one or more processing strategies. These strategies are invoked to recognize knowledge constructs that are not explicitly stated in the text, but which must be inferred to understand the argument, planning and causal structure of the text. The result of processing strategies is the construction of an argument graph, organized in terms of beliefs, belief relationships, and AUs. Question answering in OpEd requires retrieving appropriate beliefs, belief relationships, or AUs from the editorial's argument graph. Intial entry to the argument graph is provided by indexing structures associated with argument participants, plans and goals. To answer belief-related questions, OpEd analyses the contents of each question into one of six conceptual question categories. Each conceptual question category leads to the selection of a specific search and retrieval strategy.

Designing deontic knowledge bases from regulation texts

Abstract: In any organization, large amounts of knowledge are stored in various textual forms — informative texts, norms, procedures, regulations, etc. This research explores the possibility of creating knowledge bases by exploiting information contained in such texts. The documents analysed correspond to 'prescriptive texts' which can be found in companies, such as norm books, regulations, instruction books, etc. The current work concentrates on French versions of regulation texts from the Government of Quebec, and presents details of a knowledge acquisition system that can analyse prescriptive texts using a text-structuring approach which distinguishes the macrostructure, the microstructure and the domanial component of a text. The main characteristics of a regulation text microstructure are described, and a model to represent deontic knowledge in a knowledge base is proposed. The general architectures of a knowledge acquisition system and of a consultation system which enables the manipulation of deontic knowledge bases are presented.

A knowledge-based approach for checking software information using a non-monotonic reasoning system

Abstract: In this paper, the knowledge representation of software component relations using non-monotonic logic to assist the validity and integrity checking of software information is presented. Software components and their interconnection information are represented by axioms. These axioms exhibit the structure and behaviour of the software system. Another set of axioms represents the basic truisms about a software system in general. These axioms can easily be extended to cover a wide variety of software systems architectures. In the software development and maintenance phases, information about the software system can be derived from these axioms using an automated reasoning system, and the software system itself can easily be checked against a specification of the system and compared for validity. The developed knowledge-based system will be able to provide programmers useful software information and assist the software development and maintenance process.

Propa: a papaya management expert system

Abstract: Professor Papaya (Propa) is an expert system that solves disease, nutrition, insect and other problems related to papaya management. A moderator manages the interaction between the user and a panel of experts. During a session, the moderator periodically chooses a lead expert to work with the user. The role of the moderator and the knowledge of each member of the panel of experts are implemented as software routines. Dynamic relevance factors aid in the identification of the lead expert. Blackboards serve as communication devices between the user and panel. Propa can diagnose problems, recommend solutions and display and print articles related to papaya crop management and postharvest handling problems.

Knowledge acquisition mechanisms for a logical knowledge base including hypotheses

Abstract: A hypothesis-based reasoning system handles a knowledge base including complete (fact) and incomplete (hypothesis) knowledge. The handling of the incomplete knowledge plays an important role in realizing advanced AI functions such as commonsense, flexible matching, learning, etc. From both theoretical and practical viewpoints, the hypothesis-based logical reasoning system can be considered to be an important start point towards a next-generation knowledge base architecture. This paper describes an enhanced knowledge representation of the hypothetical reasoning system designed particularly for interactive diagnosis problems. The mechanisms of two knowledge acquisition modules developed for this knowledge base including hypotheses are then described. The first module, which is based on an inductive inference mechanism for the knowledge including hypotheses, enables multiple-concept formation from given examples. The key technology of this mechanism is a minimum generalization extended to the knowledge including hypotheses. The second module provides knowledge assimilation and management functions for the frame knowledge base constructed on the hypothetical reasoning system. That is, this module enables the following functions: (1) the system can assimilate new knowledge while maintaining the consistency and non-redundancy of the knowledge base; (2) the system rearranges the knowledge base when existing knowledge is deleted; and (3) the system can control the inheritance link of the frame knowledge base in response to an input indicating that the property inheritance from an upper frame is denied as an exceptional case. The nonmonotonicity of the hypothetical reasoning system is considered in these mechanisms. Shapiro's logical debugging algorithm is employed effectively to identify the knowledge which causes inconsistency. The whole system has been implemented using the metaprogramming of Prolog.

Advanced knowledge-base environments for large database systems

Abstract: The functionality and architecture of the KIWIS system is described. This is an advanced knowledge-base environment for large database systems, which is currently being developed by a consortium of industrial and research organzations within the framework of the ESPRIT programme (P2424).

Using errors to direct design

Abstract: Within many of the large IT companies, systems are being developed iteratively. This iterative development is often guided by the results of usability trials, where systems are tested with user in 'usability laboratories'. Much of the analysis of these trials is concentrated upon a qualitative assessment of the errors that occur during interaction. There is, however, no generally accepted means by which errors are analysed. Consequently, an evaluative scheme for classifying user-system errors has been developed, and is outlined in this paper. The results of an initial limited study of the use of the scheme are described, together with some example recommendations for changing the system that was analysed during the study.

Music composition as hypothesis formation: a blackboard concept of musical creativity

Abstract: A theoretical sketch of music composition as hypothesis formation in a blackboard environment in presented; it is detailed both with regard to the nature of knowledge residing in the knowledge base (competence), and to the activation of that knowledge through a scheduling mechanism (performance) in the pursuit of specific goals. Both the competence and the performance model are kept completely general (unspecific with regard to particular tasks). The intent of the sketch is to clarify the notion of 'music composition'. Emphasis is placed on two points: first, to conceive of music knowledge without an inclusion of world knowledge is unsatisfactory. Second, to explicate music knowledge without simultaneous attention to its use is insufficient. The sketch outlines a pragmatic approach to studying musical creativity.

Computers, choreography and creativity

Abstract: A discussion of the computer as an artist's tool in the domain of choreography is presented. It is shown why and how computers can both expand creativity and illuminate the process of dance composition. The actual working procedures of experienced dance artists need to be taken into account in planning and designing computer programs to assist dance composition. It is demonstrated that a parametrical view of dance movement, based upon Laban concepts, provides an excellent theoretical foundation for the use of the computer as a problem-solving environment for the dance-making process.

Goal-driven blackboard control architecture based on extending partially complete general goal trees

Abstract: A control architecture for goal-driven blackboard systems is introduced. The basic elements of the architecture are goals, policies, strategies, methods and knowledge sources. The basic control loop employs a bidding mechanism to determine the knowledge source to be executed at the current cycle. The architecture employs separate control and domain blackboards, and separate knowledge sources for the control problem and for representing the domain knowledge. The major characteristics of the architecture are that it has a simple and uniform structure, and that its basic control loop is based on a formal basis, namely, extending a partially complete general goal tree. The architecture is implemented in Smalltalk and tested on a multiple-task planning problem.

Logic of design actions

Abstract: The way in which knowledge about design can be incorporated into knowledge-based design systems is discussed and demonstrated within the framework of an overall logical/linguistic model of the design process. The technique of hierarchical planning is discussed. within this framework. The domain under consideration is that of spatial layout in buildings.

Knowledge-base reconstruction

Abstract: The aim of this paper is the reconstruction of a knowledge base from observational data. The method is concerned with the simplest but current case of relations — binary relations. The knowledge base is then represented by a tree-type oriented graph. Both the structure and the node type are identified, the structure by means of mean mutual information and the node type according to the channel matrix. The relative entropy is used to determine the edges of orientation.

Grouping knowledge-base data into distributable clusters

Abstract: Knowledge-base data are constructed from domains and subdomains of specific problem or application areas. Usually, the domains are not necessarily distinct because the original problem may have many interrelated components. As a result, the processing of the data becomes lengthy and windable. However, KB data can be reorganized into groups or clusters using some common relational information of the data objects. The reorganization process isolates the data and localizes the interdependency within the clusters, leaving weak linkages between clusters. Therefore, the clusters offer opportunities for mapping the data into distributed or parallel processing environments to facilitate a computational efficiency. The focus of this paper is on methods for structuring, partitioning, and clustering KB data for distributed computations.

Hermeneutics in expert systems

Abstract: Hermeneutics belongs to a sphere of sciences that were traditionally developed within a framework of theology; its origin dates back to the second century BC. Techniques and methods developed by hermeneutics stem from an attempt to reconstruct the correct interpretation of a text through structural analysis of the text itself. The techniques of hermeneutics have important applications today in expert systems. This paper considers hermeneutics under the original meaning of the word.

Knowledge-based management support systems: G I Doukidis, F Land and G Miller (Eds) Ellis Horwood, UK (1989) £39.95, ISBN 0-745805310, 356pp

Abstract: This book is primarily a collection of papers from a conference held at the London Business School in March 1987, plus a number of other papers selected for their special relevance. It deals with knowledgebased management support systems, decision support systems, and knowledge-based systems, with a separate section for each of these subjects. Each section has a useful introduction, written by the editors Doukidis, Land, and Miller respectively. In a book of papers, balance is all important, and the editors have selected 22 papers (almost all by academics), covering a comprehensive range of applications, case studies, research projects, methodology, and models. The editors point out how AI has evolved separately to, but in parallel with, decision support systems, and say how much can be gained from integrating the two areas, 'creating a powerful computer-based system (KBMSS) that has the potential to considerably improve managerial decision making'. They go on to say how this integration has to be taken a step further: the technology of KBMSS must be integrated with the human activity systems they support. The papers, they say, not only emphasize this integration, but also sketch 'the result of the synergy--KBMSS which can be demonstrated to enhance the effectiveness of managerial decision making'. Alas, the book doesn't quite take us into this promised land. There are tantalizing hints of the benefits of KBMSS to Citibank, RollsRoyce, and Distillers Company, for example, but closer reading uncovers no hard stuff about the details or quantified effectiveness of their management systems; but then, perhaps one wouldn't really expect that. On the other hand, there are some accounts of the benefits to real users of slightly more mundane systems, solving knotty problems using such systems. Practical examples include DEC and their XSEL configuring management ES, which the DSS used in deciding the annual financial relationship between the Federal and State Governments in Malaysia; a 'DBMS vs ES' study concerning the issuing of insolvency permits by the Institute of Chartered Accountants of Scotland; and a 'who's using what ES in finance' survey. Although theory is not neglected, the hard world of experience shows through. 'Gaining access to top managers for research purposes is a difficult and time consuming affair' writes one author. (Which probably explains why the book is thin in its coverage of KBMSS as used by Top Persons.) There is also an air of frankness about much of the discussion; it is acknowledged, for instance, that many failures in DSS have been in the design and application methodology rather than in the systems per se. The book also serves to show how uneven is the progress in KBS. For example, one paper, which in trying to move us out of the 'computers for computation' era says that 'a machine that works with people, a collaborator, would be far more appropriate than one which hitherto has been protrayed as a replacement to human thinking (namely artificial intelligence)'. Recent developments in humancomputer cooperation confirm the validity of that particular thrust. On the other hand, some of the 'how to' papers exude quiet confidence which could lull an innocent reader into believing that progress has been made on all fronts, whereas the reality is that much work has yet to be done in even the most fundamental areas. For instance, in late 1989, the Journal of Applied Systems Analysis published a paper by Clarke and Finlay proposing a typology of DSS; this clearly shows that even the phase of definition in KBS is not yet completed• The book is generally wellpresented, with a good spread of figures, tables, extracts from protocois, and examples of production rules. There are plenty of good, up-to-date references• The quality of the figures varies from 'black india-ink on white paper', through dot-matrix graphics to blurred screen dumps. (The latter shows that there is a pressing need to be able to reproduce 'screens' in print.) There are a number of irritating spelling mistakes, and instances of unreferenced author's names in the text. These are small criticisms of an otherwise interesting and rewarding book which, although a pricey £39.95, carries the general message that the workers in the field are producing useful results in conjunction with managers and other users in a (for the most part) multidisciplinary approach, and that the self-evident progress in both technique and technology has a firm basis in achievement. The readership is intended to be • . . managers throughout the world of commerce and industry, seeking information on new technologies• Were this to be qualified by saying 'managers with responsibility for seeking information etc.' then it would be more accurate, for the book is not likely to be read by the non-technical 'average' manager, certainly not in the UK anyway. The topics said to be covered include financial services, public and governmental administration, industry, and the professions e.g. legal, accounting, and medical. This is the case, although it is likely that only the keen seekers after truth in those categories would read it. As the jacket says, 'the book should be in computer department libraries in universities, polytechnics, and schools of business and management'.

Knowledge represented by mathematical models for fault diagnosis in chemical processing units

Abstract: The automation of fault diagnos in the field of chemical processing has been thwarted, in the past, chiefly by the inability to represent quantitative causal models and process constraints. This paper describes a system which uses mathematical model-based reasoning for fault diagnosis in a chemical plant. This approach appears to give somewhat better results in comparison with past attempts at using expert systems, which typically carried out the diagnosis on the basis of empirical knowledge alone. The latter systems usually suffer from the disadvantage of being too process-specific, inflexible and cumbersome.