Skip to main content
Fachgebiete
Automobil + Motoren Bauwesen + Immobilien Business IT + Informatik Elektrotechnik + Elektronik Energie + Nachhaltigkeit Finance + Banking Management + Führung Marketing + Vertrieb Maschinenbau + Werkstoffe Versicherung + Risiko
DE
EN
Bücher
Zeitschriften
Themenseiten
Organisationspsychologie Projektmanagement Marketing Smart Manufacturing
Weitere Formate
Podcasts Webinare Technik Webinare Wirtschaft Kongresse Veranstaltungskalender Awards
Jetzt Einzelzugang starten
Zugang für Unternehmen
Referenzkunden
SLX-Digitalkonferenz: Zukunftswerkstatt 2024

Mehr Umsatz mit Top-Kontakten

Am 7. Mai erfahren Sie, wie Vertriebsteams Leadgenerierung, Leadnurturing und Leadstrategien effizient steuern können.
Erweiterte Suche
Anmelden
nach oben

1988 | Buch

Introduction Kapitel lesen Erstes Kapitel lesen

Automating Knowledge Acquisition for Expert Systems

herausgegeben von: Sandra Marcus

Verlag: Springer US

Buchreihe : The International Series in Engineering and Computer Science

Enthalten in: Professional Book Archive

Einloggen, um Zugang zu erhalten
insite
SUCHEN

Über dieses Buch

In June of 1983, our expert systems research group at Carnegie Mellon University began to work actively on automating knowledge acquisition for expert systems. In the last five years, we have developed several tools under the pressure and influence of building expert systems for business and industry. These tools include the five described in chapters 2 through 6 -­ MORE, MOLE, SALT, KNACK and SIZZLE. One experiment, conducted jointly by developers at Digital Equipment Corporation, the Soar research group at Carnegie Mellon, and members of our group, explored automation of knowledge acquisition and code development for XCON (also known as R1), a production-level expert system for configuring DEC computer systems. This work influenced the development of RIME, a programming methodology developed at Digital which is the subject of chapter 7. This book describes the principles that guided our work, looks in detail at the design and operation of each tool or methodology, and reports some lessons learned from the enterprise. of the work, brought out in the introductory chapter, is A common theme that much power can be gained by understanding the roles that domain knowledge plays in problem solving. Each tool can exploit such an understanding because it focuses on a well defined problem-solving method used by the expert systems it builds. Each tool chapter describes the basic problem-solving method assumed by the tool and the leverage provided by committing to the method.

Inhaltsverzeichnis

Frontmatter
1.. Introduction
Abstract
Expert systems (knowledge-based programs that use a large body of domain facts to solve problems) are coming into widespread use in buisiness and industry. There is a potential for us to learn from our growing experience in building expert systems. Yet the creation of an individual system is often performed as a one-of-a-kind experiment, with costly interviewing of experts and analysis of the problem, resulting in systems that are sometimes difficult to maintain. This book describes a set of studies in automating knowledge acquisition that attempt to capture some of the expertise gained by knowledge engineers. The studies aim for creation of tools or methodologies designed to reduce the cost of building and maintaining expert systems.
Sandra Marcus
2.. MORE: From Observing Knowledge Engineers to Automating Knowledge Acquisition
Abstract
MORE interviews experts for the information they use to solve diagnostic problems, demonstrating that some knowledge-engineering skills can be simulated. These interviews are guided by very specific problem-solving and knowledge-acquisition strategies. As MORE’s knowledge base becomes more complete, MORE shows an improved ability to ask relevant questions and to accurately detect errors in the knowledge base. Knowledge elicited by MORE is represented in an event (or qualitative causal) model. This model is used to generate rules and to recognize inconsistencies in the confidence factors assigned to rules by domain experts. MORE applies these rules to solve diagnostic problems.
Gary Kahn
3.. MOLE: A Knowledge-Acquisition Tool for Cover-and-Differentiate Systems
Abstract
MOLE is a knowledge-acquisition tool for generating expert systems that do heuristic classification. More specifically, MOLE assumes that the task can be performed using a cover-and-differentiate problem-solving method. Using this method, the expert system generated by MOLE proposes a set of candidate explanations for the events or states that need to be explained (or covered) and then differentiates among the candidates, picking the candidates that best explain the specified events or states. The problem-solving method presupposed by MOLE makes several heuristic assumptions about the space of covering hypotheses that MOLE is able to exploit when acquiring knowledge. In particular, by distinguishing between covering and differentiating knowledge and by using this distinction to help it refine the expert’s preferences, MOLE is able to disambiguate an under-specified knowledge base and to interactively refine an incomplete knowledge base.
Larry Eshelman
4.. SALT: A Knowledge-Acquisition Tool for Propose-and-Revise Systems
Abstract
SALT6 is a knowledge-acquisition tool for generating expert systems that use a propose-and-revise problem-solving strategy. The SALT-assumed method constructs a design incrementally by proposing values for design parameters, identifying constraints on design parameters as the design develops, and revising decisions in response to constraint violations in the proposal. This problem-solving strategy provides the basis for SALT’s knowledge representation. SALT uses its knowledge of the intended problem-solving strategy in identifying relevant domain knowledge, in detecting weaknesses in the knowledge base in order to guide its interrogation of the domain expert, in generating an expert system that performs the task and explains its line of reasoning, and in analyzing test case coverage. The strong commitment to problem-solving strategy that gives SALT its power also defines its scope.
Sandra Marcus
5.. KNACK: Sample-Driven Knowledge Acquisition for Reporting Systems
Abstract
KNACK is a specialized knowledge-acquisition tool that generates expert systems for reporting tasks. The tool derives its power from exploiting the presupposed acquire-and-present problem-solving method used by the expert systems it generates. The method incrementally acquires relevant information and produces a report. It can also be combined with other problem-solving methods. An important goal in the development of KNACK is to create a tool that elicits knowledge from domain experts without requiring knowledge-engineering skills on their part. To reach that goal, KNACK’s approach to knowledge acquisition uses existing AI techniques to derive a general description of how to acquire and present information from a specific sample description.
Georg Klinker
6.. SIZZLE: A Knowledge-Acquisition Tool Specialized for the Sizing Task
Abstract
SIZZLE is a prototype knowledge-acquisition tool for building sizers: expert systems that solve sizing problems. SIZZLE uses an extrapolate­from-a-similar-case problem-solving method. Using this strategy, a sizer produces a solution by first becoming reminded of a source sizing case similar to a target sizing problem to be solved, and then adjusting the solution of the source case to account for the differences between the source and the target. The problem-solving strategy assumed by SIZZLE makes strong assumptions about the problem domain. SIZZLE assumes that knowledge about sizing can be organized as a collection of validated cases (each case is a problem-description/solution pair) and that similarities among problem descriptions imply similarities among solutions.
Daniel Offutt
7.. RIME: Preliminary Work Toward a Knowledge-Acquisition Tool
Abstract
When analyzing a task for potential use of a knowledge-acquisition tool, it may not be clear whether to use an existing tool or build a new one. If the latter is indicated, it is even less apparent how to proceed. XCON14 (also known as R1) [McDermott 82, Bachant 84], an expert system application that configures DEC computer systems, has evolved over time and expanded in scope. At one level, its task is understood, as the program is used extensively on a daily basis and performs well. However, this understanding appears incomplete when considering a knowledge-acquisition tool for XCON’s task. Additional groundwork needs to be covered and foundations set before it will be feasible to design such a tool. RIME is an attempt to establish some of that foundation. It is a programming methodology that takes a step toward understanding the nature of a potential automated tool and, in so doing, helps human knowledge engineers design and develop an expert system.
Judith Bachant
8.. Preliminary Steps Toward a Taxonomy of Problem-Solving Methods
Abstract
Although efforts, some successful, to develop expert systems (application systems that can perform knowledge-intensive tasks) have been going on now for almost 20 years, we are not yet very good at describing the variations in problem-solving methods that these systems use, nor do we have much of an understanding of how to characterize the methods in terms of features of the types of tasks for which they are appropriate. This chapter takes a few steps toward creating a taxonomy of methods -- a taxonomy that identifies some of the discriminating characteristics of the methods expert systems use and that suggests how methods can be mapped onto tasks.
John Mcdermott
Backmatter
Metadaten
Titel
Automating Knowledge Acquisition for Expert Systems
herausgegeben von
Sandra Marcus
Copyright-Jahr
1988
Verlag
Springer US
Electronic ISBN
978-1-4684-7122-9
Print ISBN
978-1-4684-7124-3
DOI
https://doi.org/10.1007/978-1-4684-7122-9