Showing papers on "Domain knowledge published in 1993"

A framework for the continual improvement of health care: building and applying professional and improvement knowledge to test changes in daily work.

Abstract: We seem to lack a well-defined, comprehensive, and shared understanding of what is required for the continual improvement of health care--at the organizational and the industry levels. This article presents a framework that defines the new body of knowledge which, when joined with the professional knowledge of health care workers, can make continual improvement possible; and gives requirements for building and applying this knowledge to bring about improvement in health care organizations.

Expertise transfer and complex problems: using AQUINAS as a knowledge-acquisition workbench for knowledge-based systems

Abstract: Acquiring knowledge from a human expert is a major problem when building a knowledge-based system. Aquinas, an expanded version of the Expertise Transfer System (ETS), is a knowledge-acquisition workbench that combines ideas from psychology and knowledge-based systems research to support knowledge-acquisition tasks. These tasks include eliciting distinctions, decomposing problems, combining uncertain information, incremental testing, integration of data types, automatic expansion and refinement of the knowledge base, use of multiple sources of knowledge and providing process guidance. Aquinas interviews experts and helps them analyse, test, and refine the knowledge base. Expertise from multiple experts or other knowledge sources can be represented and used separately or combined. Results from user consultations are derived from information propagated through hierarchies. Aquinas delivers knowledge by creating knowledge bases for several different expert-system shells. Help is given to the expert by a dialog manager that embodies knowledge-acquisition heuristics. Aquinas contains many techniques and tools for knowledge acquisition; the techniques combine to make it a powerful testbed for rapidly prototyping portions of many kinds of complex knowledge-based systems.

Knowledge acquisition tools based on personal construct psychology

Abstract: Knowledge acquisition research supports the generation of knowledge-based systems through the development of principles, techniques, methodologies and tools. What differentiates knowledge-based system development from conventional system development is the emphasis on in-depth understanding and formalization of the relations between the conceptual structures underlying expert performance and the computational structures capable of emulating that performance. Personal construct psychology is a theory of individual and group psychological and social processes that has been used extensively in knowledge acquisition research to model the cognitive processes of human experts. The psychology takes a constructivist position appropriate to the modelling of human knowledge processes, but develops this through the characterization of human conceptual structures in axiomatic terms that translate directly to computational form. In particular, there is a close correspondence between the intensional logics of knowledge, belief and action developed in personal construct psychology, and the intensional logics for formal knowledge representation developed in artificial intelligence research as term subsumption, or KL-ONE-like, systems. This paper gives an overview of personal construct psychology and its expression as an intensional logic describing the cognitive processes of anticipatory agents, and uses this to survey knowledge acquisition tools deriving from personal construct psychology.

Knowledge-based connectionism for revising domain theories

Abstract: A knowledge-based connectionist model for machine learning referred to as KBCNN is presented. In the KBCNN learning model, useful domain attributes and concepts are first identified and linked in a way consistent with initial domain knowledge, and then the links are weighted properly so as to maintain the semantics. Hidden units and additional connections may be introduced into this initial connectionist structure as appropriate. Then, this primitive structure evolves to minimize empirical error. The KBCNN learning model allows the theory learned or revised to be translated into the symbolic rule-based language that describes the initial theory. Thus, a domain theory can be pushed onto the network, revised empirically over time, and decoded in symbolic form. The domain of molecular genetics is used to demonstrate the validity of the KBCNN learning model and its superiority over related learning methods. >

Learning by Experimentation: Incremental Refinement of Incomplete Planning Domains

Abstract: : Building a knowledge base requires iterative refinement to correct imperfections that keep lurking after each new version of the system. This paper concentrates on the automatic refinement of incomplete domain models for planning systems, presenting both a methodology for addressing the problem and empirical results obtained from an implemented system in several domains when initial domain knowledge is up 50% incomplete. Planning knowledge may be refined automatically through direct interaction with their environmental. Missing conditions cause unreliable predictions of action outcomes. Missing effects cause unreliable predictions of facts about the state. The paper shows that, contrary to popular belief, missing information is not necessarily associated with and selective interaction with the environment that pinpoints the type of fault in the domain knowledge that causes any unexpected behavior of the environment, and resorts to experimentation when additional information is needed to correct fault. Our approach has been implemented in EXPO, a system that uses PRODIGY as a baseline planner and improves its domain knowledge in several domains. The empirical results presented show that EXPO dramatically improves its prediction accuracy and reduces the amount of unreliable action outcomes. Planning, Learning, Experimentation, Theory, Refinement, Incomplete theories.

Eliciting knowledge and transferring it effectively to a knowledge-based system

Abstract: The knowledge acquisition bottleneck impeding the development of expert systems is being alleviated by the development of computer-based knowledge acquisition tools. These work directly with experts to elicit knowledge, and structure it appropriately to operate as a decision support tool within an expert system. However, the elicitation of expert knowledge and its effective transfer to a useful knowledge-based system is complex and involves diverse activities. The complete development of a decision support system using knowledge acquisition tools is illustrated. The example is simple enough to be completely analyzed but exhibits enough real-world characteristics to give significant insights into the processes and problems of knowledge engineering. >

Engineering Knowledge: The Construction of Knowledge in Artificial Intelligence

Abstract: This paper presents an anthropological study of knowledge production in the expert systems community within AI. Expert systems are built by knowledge engineers, specialists in the task known as `knowledge acquisition'. This is a complex process of interpretation and translation; not surprisingly (to an anthropologist, at least), it presents a troublesome `bottleneck'. However, knowledge engineers have a different perspective on why this is so. Typically positivist in approach, they see knowledge acquisition as conceptually straightforward. In their view, it is difficult, not because of the nature of knowledge or the complexity of the process, but rather because it requires extended face-to-face interaction between knowledge engineer and expert. Believing that automation will `get around' the inexact and uncontrollable nature of this interaction, they seek to automate it. Drawing on ethnographic material, the paper explores the knowledge engineers' epistemological stance, noting its characteristic deletion...

Theories underlying requirements engineering: an overview of NATURE at Genesis

Abstract: NATURE is a collaborative basic research project on theories underlying requirements engineering funded by the ESPRIT III program of the European communities. Its goals are to develop a theory of knowledge representation that embraces subject, usage and development worlds surrounding the system, including expressive freedoms; a theory of domain engineering that facilitates the identification, acquisition and formalization of domain knowledge as well as similarity-based matching and classifying of software engineering knowledge; and a process engineering theory that promotes context and decision-based control of the development process. These theories are integrated and evaluated in a prototype environment constructed around an extended version of the conceptual modeling language Telos. >

Explanation-Driven Case-Based Reasoning

Abstract: Problem solving in weak theory domains should compensate for the lack of strong theories by combining the various other knowledge types involved. Such methods should be able to effectively combine general domain knowledge with specific case knowledge. A method is described that utilises a presumably extensive and dense model of general domain knowledge as explanatory support for case-based problem solving and learning. A generic reasoning method — captured in what is called the Activate-explain-focus cycle — is able to utilise a rich knowledge model in producing context-dependent explanations. A specialisation of this method for each of the main subprocesses of case-based reasoning is presented, and illustrated with examples.

Cognitive Task Analysis of Expertise in Air Traffic Control

Abstract: A cognitive task analysis was performed to analyze knowledge structures, mental models, skills, and strategies of en route controllers to provide an understanding of the key cognitive components of the air traffic controller's job. This article presents the results of three procedures as they contributed to an understanding of controller expertise: paper problem solving, performance modeling, and structured problem solving. The procedures resulted in the identification of (a) 13 primary tasks, (b) a mental model representing expert controller's organization of domain knowledge, (c) three categories of controller strategies, and (d) a hierarchy of goals. These results are being used to specify the instructional content and sequencing for the new Federal Aviation Administration en route air traffic control curriculum.

Embedding computer-based critics in the contexts of design

Abstract: Computational critiquing mechanisms provide an effective form of computer-human interaction supporting the process of design. Critics embedded in domain-oriented design environments can take advantage of additional knowledge residing in these environments to provide less intrusive, more relevant critiques. Three classes of embedded critics have been designed, implemented, and studied: Generic critics use domain knowledge to detect problematic situations in the design construction. Specific critics take advantage of additional knowledge in the partial specification to detect inconsistencies between the design construction and the design specification. Interpretive critics are tied to perspective mechanisms that support designers in examining their artifact from different viewpoints.

Species Intelligibilis: From Perception to Knowledge

Abstract: The main purpose of this book is to offer a comprehensive historical analysis of the discussions on a crucial problem for the early modern theory of knowledge: the formal mediation of sensible reality in intellectual knowledge.

Knowledge acquisition as a constructive modeling activity

Abstract: Knowledge acquisition is a constructive modeling process, not simply a matter of “expertise transfer.” Consistent with this perspective, we advocate knowledge acquisition practices and tools that facilitate active collaboration between expert and knowledge engineer, that exploit a serviceable theory in their application, and that support knowledge-based system development from a life-cycle perspective. A constructivist theory of knowledge is offered as a plausible theoretical foundation for knowledge acquisition and as an effective practical approach to the dynamics of modeling. In this view, human experts construct knowledge from their own personal experiences while interacting with their social constituencies (e.g., supervisors, colleagues, clients patients) in their niche of expertise. Knowledge acquisition is presented as a cooperative enterprise in which the knowledge engineer and expert collaborate in constructing an explicit model of problem solving in a specific domain. From this perspective, the agenda for the knowledge acquisition research community includes developing tools and methods to aid experts in their efforts to express, elaborate, and improve their models of the domain. This functional view of expertise helps account for several problems that typically arise in practical knowledge acquisition projects, many of which stem directly from the inadequacies of representations used at various stages of system development. to counter these problems, we emphasize the use of mediating representations as a means of communication between expert and knowledge engineer, and intermediate representations to help bridge the gap between the mediating representations themselves, as well as between the mediating representations and a particular implementation formalism. © 1993 John Wiley & Sons, Inc.

Investigating the applicability of Petri nets for rule-based system verification

Abstract: It is suggested that as rule-based system (RBS) technology gains wider acceptance, the need to create and maintain large knowledge bases will assume greater importance. Demonstrating a rule base to be free from error remains one of the obstacles to the adoption of this technology. An approach to RBS verification in which the system is modeled as a Petri net on which error detection is performed is presented. A set of propositions is formulated to locate errors of redundancy, conflict, circularity, and gaps in domain knowledge. Rigorous proofs of these propositions are provided. Difficulties in implementing a Petri net-based verifier and the potential restrictions of the applicability of this approach are discussed. >

Acquiring, representing, and evaluating a competence model of diagnostic strategy

Abstract: NEOMYCIN is a computer program that models one physician''s diagnostic reasoning within a limited area of medicine. NEOMYCIN''s diagnostic procedure is represented in a well-structured way, separately from the domain knowledge it operates upon. We are testing the hypothesis that such a procedure can be used to simulate both expert problem-solving behavior and a good teacher''s explanations of reasoning. The model is acquired by protocol analysis, using a framework that separates an expert''s causal explanations of evidence from his descriptions of knowledge relations and strategies. The model is represented by a procedural network of goals and rules that are stated in terms of the effect the problem solver is trying to have on his evolving model of the world. The model is evaluated for sufficiency by testing it in different settings requiring expertise, such as providing advice and teaching. The model is evaluated for plausibility by arguing that the constraints implicit in the diagnostic procedure are imposed by the task domain and human computational capability. This paper discusses NEOMYCIN''s diagnostic procedure in detail, viewing it as a memory aid, as a set of operators, as proceduralized constraints, and as a grammar. This study provides new perspectives on the nature of "knowledge compilation" and how an expert-teacher''s explanations relate to a working program.

Writing quality, reading skills, and domain knowledge as factors in text comprehension.

Abstract: Subjects listened to and recalled three passages. Each subject was also given a general reading comprehension test. The passages were presented either in such a way that subjects could use their general knowledge to help understand them, or in such a way that no specific world knowledge seemed applicable. This was achieved by giving the passages a helpful title, versus no title or an unhelpful title. The passages were written in two different versions, preserving their content but varying their style. In one version, the language was as helpful as we could make it in signalling to the listener discourse importance, while in the other version the language was as unhelpful as we could make it while still writing an English text. All three factors--domain knowledge, writing style, and skill--significantly affected reproductive recall, and there were no interactions between these factors. However, while good writing was sufficient to improve the reproduction of the texts, an analysis of the recall elaborations subjects made revealed that the correctness of their elaborations depended strongly on the availability of appropriate domain knowledge. Thus, good writing and domain knowledge are not simply substitutable, but affect comprehension in somewhat different ways.

Using Qualitative Models to Guide Inductive Learning

Abstract: This paper presents a method for using qualitative models to guide inductive learning. Our objectives are to induce rules which are not only accurate but also explainable with respect to the qualitative model, and to reduce learning time by exploiting domain knowledge in the learning process. Such explainability is essential both for practical application of inductive technology, and for integrating the results of learning back into an existing knowledge-base. We apply this method to two process control problems, a water tank network and an ore grinding process used in the mining industry. Surprisingly, in addition to achieving explainability the classificational accuracy of the induced rules is also increased. We show how the value of the qualitative models can be quantified in terms of their equivalence to additional training examples, and finally discuss possible extensions.

From “folklore” to “living design memory”

Abstract: We identify an important type of software design knowledge that we call community specific folklore and show problems with current approaches to managing it. We built a tool that serves as a living design memory for a large software development organization. The tool delivers knowledge to developers effectively and is embedded in organizational practice to ensure that the knowledge it contains evolves as necessary. This work illustrates important lessons in building knowledge management systems, integrating novel technology into organizational practice, and managing research-development partnerships.

Towards a unification of knowledge modelling approaches

Abstract: In this article we present a coherent framework for modelling reasoning processes in knowledge based systems. The aim of the framework is to integrate different lines of research and in particular, though not exclusively, the KADS approach and the Components of Expertise framework. We are especially concerned with enhanced facilities for domain modelling and with the notion of problem solving method. The resulting modelling framework, called the CommonKADS modelling framework, fits into a comprehensive methodology, called CommonKADS, that covers all aspects of knowledge based applications. In this article we first present a set of principles on which our modelling framework in founded. These are derived from a careful study of the different approaches. We then describe the modelling framework itself, illustrating it with an example. We also discuss various approaches to building models for a particular application using this framework.

Differences in the problem solving of stronger and weaker novices in physics: Knowledge, strategies, or knowledge structure?

Abstract: This study investigates the extent to which differences in the problem-solving performance of stronger and weaker novices in physics arise from: (a) differences in amount of domain knowledge, (b) differences in how domain knowledge is organized, and (c) differences in the strategic application of domain knowledge. Ten first-year university physics students attempted to solve one easy and one difficult problem involving Newton's second law. Clear differences in the protocols of stronger and weaker students for the difficult problem, combined with successful performance by all students on the easy problem, were interpreted as evidence for differences in the organization of relevant knowledge held by more versus less successful first-year physics students. Some differences in procedural knowledge were also observed, but all students used the working forward strategy that had been presented to them in lectures.

CADSYN: A case-based design process model

Abstract: In solving a new design problem, the case-based reasoning paradigm provides a process model where previous experience in the form of multiple, individual design situations can be used in a new design context. Design synthesis presents challenges to current methodologies of CBR in the application of the various approaches to case memory organization, indexing, selection and transformation. The focus of this paper is on the transformation process. Multiple types of design knowledge are essential to derive a new design solution. A hybrid case-based design process model, CADSYN, is proposed to integrate specific design situations and generalized domain knowledge, where specific cases are represented as attribute-value pairs and domain knowledge is represented by generalized design concepts and constraints. Case transformation is treated as a constraint satisfaction problem, where a specific design case provides a starting point for a new design problem and constraints are used to revise the case for consistency with the new context.

Knowledge-based approaches to database design

Abstract: Database design is often described as an intuitive, even artistic, process. Many researchers, however, are currently working on applying techniques from artificial intelligence to provide effective automated assistance for this task. This article presents a summary of the current state of the art for the benefit of future researchers and users of this technology. Thirteen examples of knowledge-based tools for database design are briefly described and then compared in terms of the source, content, and structure of their knowledge bases; the amount of support they provide to the human designer; the data models and phases of the design process they support; and the capabilities they expect of their users. The finding show that there has apparently been very little empirical verification of the effectiveness of these systems. In addition, most rely exclusively on knowledge provided by the developers themselves and have little ability to expand their knowledge based on experience. Although such systems ideally would be used by application specialists rather than database professionals, most of these systems expect the user to have some knowledge of database technology.

Domain modelling with hierarchies of alternative viewpoints

Abstract: It is shown how domain modelling can be used within requirements engineering to reveal the conceptual models used by the participants, and relate these to one another. Existing elicitation techniques used in AI adopt a purely cognitive stance, in that they model a single problem-cognitive stance, and ignore the social and organizational context. A framework for representing alternative, conflicting viewpoints in a single domain model is described. The framework is based on the development of a hierarchy of viewpoint descriptions, where lower levels of the hierarchy contain the conflicts. The hierarchies can be viewed in a number of ways, and allow the participants to develop an understanding of each other's perspective. The framework is supported by a set of tools for developing and manipulating these hierarchies. >

Constructing knowledge-based systems

Abstract: Vital, a four-and-a-half-year ESPRIT II research and development project that involves nine organizations in five countries is discussed. It addresses the problems of effective process modeling for knowledge-based systems, providing guidelines on when to use various knowledge-engineering methods and techniques, and reducing the bottleneck in acquiring expert knowledge by providing both methodological and software support for developing large, industrial, knowledge-based system applications. The project goals, approach, and workbench are outlined, and a case study is described. >

Structuring interviews with experts during knowledge elicitation

Abstract: We have developed interviewing techniques for knowledge elicitation based on (1) relevant research and techniques from the social sciences, (2) the nature of expertise, (3) a desire to assist a knowledge engineer to avoid reductive bias, one of the pitfalls associated with the acquisition of highly complex concepts, and (4) the desirability of de-coupling elicitation from implementation. the approach consists of four phases with guidelines for questions suited to each stage. First, the descriptive elicitation stage is intended to reveal the important entities and concepts in the domain as reflected in the terms and specialized language used by the expert. A second stage, structured expansion, is designed to probe the relationships between the domain concepts and the organization of the expert's knowledge using the terminology uncovered in the previous stage. the scripting phase relies on the declarative knowledge found through the two previous stages to discover procedural knowledge, and the final component, validation, is important throughout the process of knowledge elicitation to ensure that the knowledge being elicited is correct and adequate to enable a system in which it is implemented to solve the class of problems with which it is concerned. © 1993 John Wiley & Sons, Inc.