All articles are up for ftp in PostScript at ftp://ksi.cpsc.ucalgary.ca/KAW/KAW96.
Or click on the file name shown at the end of each abstract.
All these documents have been checked and print OK. Let me know if you have problems.
Amedeo Cesta* and Angelo Oddi**
*IP-CNR, National Research Council of
Italy, Viale Marx 15, I-00137 Rome, Italy
**Dipartimento di Informatica e
Sistemistica, Universita di Roma "La Sapienza", Via Salaria 113, I-00198 Rome,
Italy
amedeo@pscs2.irmkant.rm.cnr.it, oddi@assi.dis.uniroma1.it
This paper describes a domain description language DDL.1 able to represent physical domains to solve planning and scheduling problems. DDL.1 uses a representation, inspired by classical control theory, based on state-variables to represent the relevant features of a domain. Each state variable is meant to represent a set of plausible temporal evolutions those features may have. DDL.1 allows to specify constraints on the sequence of values that a state variable may assume over time. For the language a syntactic specification, and a model theoretic semantic are given. The problem of temporal planning using a DDL.1 specification is also addressed and a planning algorithm named TP-SV introduced. The paper tries to show how this kind of description languages may generate a methodology to gracefully model the relevant constraints in physical domains, and how a formally specified planner may be associated to this description.
Track: Planning and temporal reasoning. 03cesta.ps.Z (91075)
Yolanda Gil and Eric Melz
USC/Information Sciences Institute
{gil,
melz}@ISI.EDU
Role-limiting approaches support knowledge acquisition (KA) by centering knowledge base construction on common types of tasks or domain-independent problem-solving strategies. Within a particular problem-solving strategy, domain-dependent knowledge plays specific roles. A KA tool then helps a user to fill these roles. Although role-limiting approaches are useful for guiding KA, they are limited because they only support users in filling knowledge roles that have been built in by the designers of the KA system. EXPECT takes a different approach to KA by representing problem-solving knowledge explicitly, and deriving from the current knowledge base the knowledge gaps that must be resolved by the user during KA. This paper contrasts role-limiting approaches and EXPECT's approach, using the propose-and-revise strategy as an example. EXPECT not only supports users in filling knowledge roles, but also provides support in 1) adapting the problem-solving strategy, 2) changing the types of information to be acquired about a knowledge role, 3) adding new knowledge roles, and 4) acquiring additional background information about the domain needed by the knowledge-based system. EXPECT's guidance changes as the knowledge base changes, providing a more flexible approach to knowledge acquisition. This work provides evidence supporting the need for explicit representations in building knowledge-based systems.
Track: Shareable and reusable problem-solving methods. 04gil.ps.Z (111671)
Udo Hahn, Manfred Klenner, Klemens Schnattinger
Freiburg University,
Computational Linguistics Group, Europaplatz 1, D-79085 Freiburg,
Germany
{hahn, klenner, schnattinger}@coling.uni-freiburg.de
We introduce a methodology for knowledge acquisition and concept learning from texts that relies upon a quality-based model of terminological reasoning. Concept hypotheses which have been derived in the course of the text understanding process are assigned specific "quality labels'' (indicating their significance, reliability, strength). Quality assessment of these hypotheses accounts for conceptual criteria referring to their given knowledge base context as well as linguistic indicators (grammatical constructions, discourse patterns), which led to their generation. We advocate a metareasoning approach which allows for the quality-based evaluation and a bootstrapping-style selection of alternative concept hypotheses as text understanding incrementally proceeds. We also provide a preliminary empirical evaluation of our approach, with focus on the learning rates and the learning accuracy that can be achieved.
Track: Knowledge Acquisition from Natural Language. 05hahn.ps.Z (113291)
M.V. Nagendra Prasad and Enric Plaza
Dept. of Computer Sc., LGRC,
University of Massachusetts, Amherst, MA 01003. USA
nagendra@cs.umass.edu
Rising operating costs and structural transformations such as resizing and globalization of companies all over the world have brought into focus the emerging discipline of knowledge management that is concerned with making knowledge pay off. Corporate memories form an important part of such knowledge management initiatives in a company. In this paper, we discuss how, viewing corporate memories as distributed case libraries can benefit from existing techniques for distributed case-based reasoning for resource discovery and exploitation of previous expertise. We present two techniques developed in the context of multi-agent case-based reasoning for accessing and exploiting past experience from corporate memory resources. The first approach, called Negotiated Retrieval, deals with retrieving and assembling "case pieces'' from different resources in a corporate memory to form a good overall case. The second approach, based on Federated Peer Learning, deals with two modes of cooperation called DistCBR and ColCBR that let an agent exploit the experience and expertise of peer agents to achieve a local task.
Track: Corporate memory and enterprise modeling 06prasad.ps.Z (103033)
Renaud Lecoeuche and Olivier Catinaud
INSA Rouen, Place Emile Blondel, BP
08, 76131 Mont-Saint-Aignan Cedex, France
lecoeuch@insa-rouen.fr,
catinaud@insa-rouen.fr
Second generation expert systems are designed to overcome first generation expert system deficiencies in various domains such as knowledge acquisition or explanation. Following those efforts, we will focus our attention on competence and try to find the cause of brittleness. We will first define competence in individuals and organisations. Then, we will study competence in expert systems, and we will point out some causes of insufficiency. We will particularly stress the fact that systems are ignorant of the way their knowledge is structured. We will conclude by providing some clues for improvements and we especially advocate the creation of a competence assessment module.
Track: Corporate memory and enterprise modeling. 08lecoeuche.ps.Z (853555)
*Rudi Studer, **Henrik Eriksson, #John Gennari, #Samson Tu, *##Dieter Fensel
and #Mark Musen
* Institute AIFB, University of Karlsruhe, D-76128
Karlsruhe
** Department of Computer and Information Science, Linkoping
University, S-58183 Linkoping
#Section on Medical Informatics, Knowledge
Systems Laboratory, Stanford, University School of Medicine, Stanford, CA
94305-5479, USA
##Department SWI, University of Amsterdam, NL-1018 WB
Amsterdam
studer@aifb.uni-karlsruhe.de, her@ida.liu.se,
{gennari,tu,musen}@camis.stanford.edu, dieter@swi.psy.uva.nl
Problem-solving methods model the problem-solving behavior of knowledge-based systems. The PROTÉGÉ-II framework includes a library of problem-solving methods that can be viewed a reusable components. For developers to use these components as building blocks in the construction of methods for new tasks, they must configure the components to fit with each other and with the needs of the new task. As part of this configuration process, developers must relate the ontologies of the generic methods to the ontologies associated with other methods and submethods. We present a model of method configuration that incorporates the use of several ontologies in multiple levels of methods and submethods, and we illustrate the approach by providing examples of the configuration of the board-game method using an extension of the Knowledge Acquisition and Representation Language KARL.
Track: Shareable and reusable problem-solving methods. 13studer.ps.Z (66691)
Annette ten Teije
Department of Social Science Informatics, University of
Amsterdam, Roetersstraat 15, NL 1018 WB Amsterdam
annette@swi.psy.uva.nl
The literature contains a number of approaches for configuring or selecting problem solvers. Some of these approaches are based on indexing and selecting a PSM from a library, others are more based on a KA process, yet others are based on search-strategies. None of these approaches sees configuring of a PSM as a configuration task that could be solved with an appropriate configuration method. We regard the configuration of PSMs as a configuration problem and give a method for solving this configuration problem.
Track: Shareable and reusable problem-solving methods. 14tenteije.ps.Z (68649)
Carlos A. Iglesias, Mercedes Garijo, José C. Gonzàlez and Juan
R. Velasco
Dep. de Ingeniería de Sistemas Telemáticos,
E.T.S.I. Telecomunicación, Universidad Politécnica de
Madrid
cif@gsi.dit.upm.es
The application of agent technology to real applications needs the development of a methodology which support all the software development life cycle (SDLC) of the agent based system including its management. This paper proposes an extension of CommonKADS for fitting the characteristics of the agent approach in the SDLC and the definition of a new model, the coordination model, for describing the coordination protocols.
Track: Agent-Oriented approaches to Knowledge engineering. 16iglesias.ps.Z (120843)
*Martin Molina and **Yuval Shahar
*Department of Artificial Intelligence,
Technical University of Madrid, Campus de Montegancedo S/N, Boadilla del Monte
28660, Madrid, Spain
** Section on Medical Informatics, School of Medicine,
Stanford University, Stanford, CA 94305, USA
mmolina@dia.fi.upm.es,
shahar@camis.stanford.edu
In the knowledge engineering field, models of expertise present interesting properties for reuse. In particular, Problem-Solving Methods (PSM), that identify generic lines of reasoning with specific uses of knowledge types, are good reusable components for knowledge modeling. In this paper we present a case study of PSM reuse, where we focus on two particular issues: (1) how to reuse a PSM by generalizing some of its task assumptions, and (2) the formulation of a new PSM by assembling simpler PSMs. The case study starts with a PSM called Knowledge Based Temporal Abstraction that was originally defined for the clinical domain. We show how to reuse it for a new task in the traffic domain: first, the temporal dimension on which the method is referred is generalized into a linear dimension in order to make abstractions for both temporal and spatial dimensions and, second, a new method for spatiotemporal abstraction is def ned by assembling two different PSMs. In summary, the paper underlines and illustrates some of the sign)ficant steps for PSM reuse, and it proposes a modular organization of ontologies (using is-a and part-of relations) to facilitate this activity.
Track: Shareable and reusable problem solving methods. 17molina.ps.Z (89973)
Gian Piero Zarri
Centre National de la Recherche Scientifique, EHESS -
CAMS, 54, boulevard Raspail, 75270 PARIS Cedex 06,
France
zarri@cams.msh-paris.fr
The problem of dealing with the dispersed know-how thatexist in a corporation ("corporate memory") has been normally dealt with, on this side of the ocean, as a problem of "enterprise modelling", see KADS or EM. One of the main hypothesis of this modelling approachconcerns the possibility of detecting, within the global, "expert" behaviour of a corporation, a set of elementary tasks,independent from a particular application domain. Once the tasks discovered and formalised, theycan be used to set up libraries of basic building blocks, to be reused forthe description of a large number of intellectual processes in a company.This endeavour is really very ambitious -- which explains why so many studies in this domain limitthemselves to a purely theoretical approach -- and, moreover, it meets all sort of practical problems which range fromthe difficulties in defining the building blocks in a really general way to the ambiguities concerning which aspects (the model or the code) of theblocks can really be reused. In this paper, we suggest that a more modest,but better defined and pragmatically useful approach to the practical utilisation of corporate memory canbe found in the construction and use of corporate knowledge repositories.They can be defined as on-line, computer-based storehouses of expertise, knowledge, experience and documentation about particular aspects of a corporation. We consider here only the"textual component" of corporate knowledge -- i.e., all sort of economically valuable, natural language (NL) documentslike news stories, telex reports, internal documentation (memos, policystatements, reports and minutes), normative texts, intelligence messages, etc. In this case, the construction of effectively usable corporateknowledge repositories can be achieved with the translation of the originaldocuments into some type of conceptual format. The "metadocuments" obtained in this way can then be stored into a knowledge repository (knowledge base) and, given their role ofadvanced document models, all the traditional functions of informationretrieval, e.g., searching, retrieving and producing an answer (and other functions like the intelligent navigation inside the repository) can be directly executed on them. We thenillustrate here, in some details, the architecture of a prototypical systemdesigned to exploit a knowledge repository of metadocuments.
Track: Corporate memory and enterprise modeling. 18zarri.ps.Z (438176)
Gaële Simon
CRIN/CNRS, Université Henri Poincaré -
Nancy 1, B.P 239 54506 Vandoeuvre-lès-nancy cedex
simon@loria.fr
In this paper, we describe important steps of the knowledge capitalisation process we are working on, in a metallurgical domain. From this particular and practical experience, our purpose is to focus on general characteristics which seem to be reusable for other knowledge capitalisation systems. We would like, in particular, to put emphazise on specific constraints linked to the design of this kind of systems, in order to make general methods, techniques and tools emerge allowing to answer to these requirements.
Track: Corporate memory and enterprise modeling. 20simon.ps.Z (499145)
*Pim Borst, *# Hans Akkermans and **Jan Top
*University of Twente,
Information Systems Department INF/IS, P.O. Box 217, NL-7500 AE Enschede, The
Netherlands.
**Agro-Technological Research Organization ATO-DLO, P.O. Box
17, NL-6700 AA Wageningen, The Netherlands.
#Netherlands Energy Research
Foundation ECN, P.O. Box 1, NL-1755 ZG Petten (NH), The Netherlands.
{borst,
akkerman}@cs.utwente.nl, J.L.Top@ato.dlo.nl, akkermans@ecn.nl
We analyze the construction as well as the role of ontologies in knowledge sharing and reuse for complex industrial applications. In this article, the practical use of ontologies in large-scale applications not restricted to knowledge-based systems is demonstrated, for the domain of engineering systems modelling, simulation and design. A general and formal ontology, called PhysSys, for dynamic physical systems is presented and its structuring principles are discussed. It has been applied to develop the conceptual database schema of the Olmeco library of reusable engineering model components and to a full-scale numerical simulation experiment on this basis pertaining to an existing large hospital heating installation. From the application scenario, several general guidelines and experiences emerge. It is possible to identify various viewpoints that are seen as natural within a large domain: broad and stable conceptual distinctions that give rise to a categorization of concepts and properties. This provides a first mechanism to break up ontologies into smaller pieces with strong internal coherence but relatively loose coupling, thus reducing ontological commitments. Secondly, we show how general and abstract ontological super-theories, for example mereology, topology and systems theory, can be used and reused as generic building blocks in ontology construction. We believe that this is an important element in knowledge sharing across domains. Thirdly, we introduce {\em ontology projections} as a flexible means to connect different base ontologies. Ontology projections can occur in simple forms such as include-and-extend and include-and-specialize, but are in their richest form very knowledge-intensive, being in fact themselves full-blown ontological theories.
Track: Shareable and reusable ontologies. 21borst.ps.Z (170932)
Jérôme Euzenat
INRIA Rhne-Alpes Q IMAG, 655 avenue de
lUEurope, 38330 Montbonnot Saint-Martin, France
Jerome.Euzenat@inrialpes.fr
The Co4 system is dedicated to the representation of formal knowledge in an object and task based manner. It is fully interleaved with hyper-documents and thus provides integration of formal and informal knowledge. Moreover, consensus about the content of the knowledge bases is enforced with the help of a protocol for integrating knowledge through several levels of consensual knowledge bases. Co4 is presented here as addressing three claims about corporate memory: (1) it must be formalised to the greatest possible extent so that its semantics be clear and manipulation can be automated; (2) it cannot be totally formalised and thus formal and informal knowledge must be organised such that they can refer to each other; (3) in order to be useful, it must be accepted by the involved people (providers and users) and thus must be non contradictory and consensual.
Track: Corporate memory and enterprise modeling. 22euzenat.ps.Z (109871)
Tim Menzies, Monash Uni and Simon Goss
Dept. Software Development, Monash
University, Caulfield, Melbourne, Australia,
3145.
timm@insect.sd.monash.edu.au
Standard software engineering methodologies are typically prescriptions on how to develop some initial system. Here we formalise the process of using an existing, possibly poorly understood, system. Informal vague causal diagrams are a common technique for illustrating and sharing intuitions about such poorly understood systems. Normally, such diagrams are viewed as precursors to other modeling techniques. Here, we take another approach and explore what we can do with these vague diagrams without requiring precise analysis. Vague models can contain inaccuracies and must be tested. In vague domains, if we can't test it then we shouldn't model it. That is, the computational properties of the test engine constrains the modeling process.
Track: Knowledge engineering. 23menzies.ps.Z (105887)
Tim Menzies, Monash Uni
Dept. Software Development, Monash University,
Caulfield, Melbourne, Australia, 3145.
timm@insect.sd.monash.edu.au
The dominant knowledge modeling paradigm in the KA field assumes that old knowledge expressed symbolically (i.e. problem-solving strategies or ontologies) is a productivity tool for building new knowledge bases. That is, it assumes that knowledge is context-independent. Researchers of situated cognition (SC) claims that knowledge is mostly context-dependent and that concepts elicited prior to direct experience are less important than functional units developed via direct experience with the current problem. We argue that there is sufficient evidence to support some of the SC view; i.e. we may need to modify the knowledge modeling approach. Symbolic approaches exist which could be said to handle SC; e.g. abduction, verification & validation tools, repitory grids, certain frameworks for decision support systems, expert critiquing systems, ripple-down-rules, and ripple-down-models. However, such approaches need careful assessment in order to test their appropriateness as a response to the challenge of SC. Hence we propose a metrics program to assess the impact of SC and the success of our symbolic KA tools for handling SC. Interestingly, some of the symbolic tools we review for handling SC also assist us in the metrics program.
Track: Knowledge engineering. 24menzies.ps.Z (125331)
*Gertjan van Heijst, **Rob van der Spek, **Eelco Kruizinga
*University of
Amsterdam, Social Science Informatics, Roetersstraat 15, NL-1018 WB Amsterdam,
The Netherlands
**CIBIT, Arthur van Schendelstraat 570, PO-box 573, NL-3500
AN Utrecht, The Netherlands
gertjan@swi.psy.uva.nl,
{rvdspek,ekruizinga}@cibit.hvu.nl
This paper presents some preliminary thoughts on how corporate memories should be organized in such a way that they maximally contribute to the competitiveness of an organization. We argue that a corporate memory should support three types of organizational learning, which are described. Then we formulate functional requirements and present an architecture for corporate memories that would satisfy these requirements. The paper end with pointers to related work an future research issues.
Track: Corporate memory and enterprise modeling. 25vanheijst.ps.Z (88635)
Juergen Angele, Stefan Decker, Rainer Perkuhn, and Rudi Studer
Institute
AIFB, University of Karlsruhe (TH), D-76128 Karlsruhe,
Germany
{angele|decker|perkuhn|studer}@aifb.uni-karlsruhe.de
In this paper it is shown how the language KARL (Knowledge Acquisition and Representation Language) which has been developed for representing the model of expertise has been modified and extended to New KARL to better meet the needs for the representation, the retrieval, the adaption and the configuration of problem-solving methods (PSM). Based on a conceptual structure of PSMs and based on the way methods are configured to a complex model new language primitives are introduced for KARL to specify such a conceptual structure and to support the configuration of methods. New KARL allows to describe all parts of a PSM formally which may be exploited for verification purposes and for supporting the retrieval, the adaption and the configuration process of PSMs.
Track: Shareable and reusable problem-solving methods. 26angele.ps.Z (72333)
Josefina Sierra Santibáñez
Computer Science Department.
Stanford University, CA 94305-9020.
jsierra@cs.stanford.edu
Telecommunication technologies have transformed the world of computing. They have laid out a computation environment through which it is possible interoperate with remote programs, get access to spatially distributed resources and communicate, in a wide variety of modalities, with an almost unrestricted number of people. This paper describes an agent oriented approach to knowledge systems, which addresses interaction with software and human agents using a mathematical model. We present several scenarios demonstrating the enhanced capabilities of knowledge agents to deal with interaction with human and software agents, reusability, and distributed knowledge acquisition.
Track: Agent-oriented approaches to knowledge engineering. 29santibanez.ps.Z (106043)
Eliana Coelho and Guy Lapalme
Département d'Informatique et
Recherche OpérationnelleUniversité de
Montréal
coelho@iro.umontreal.ca
To reuse a problem-solving method, a knowledge engineer must know the assumptions about the structure of the knowledge roles inherent to the method. We propose to construct an ontology of roles in order to formalize these assumptions. This ontology defines formally the roles played by the knowledge in the inference process, and serves as an interface for the knowledge engineer to evaluate if an application domain conforms to the structure of knowledge used by the method. In addition, we define the inferences of the problem-solving method according to its ontology of roles. This definition is declarative and abstracts from the implementation details. We illustrate our approach using the problem-solving method Propose&Revise.
Track: Shareable and reusable problem-solving methods. 34coelho.ps.Z (125639)
Henrik Eriksson and Niklas Frost
Department of Computer and Information
Science, Linkoping University, S-581 83 Linkoping, Sweden
her@ida.liu.se
The use of tools that support the knowledge-engineering process can be challenging. Developers often make modeling mistakes that could have been avoided in hindsight. Design-critiquing systems can assist the developer by highlighting potential problems with modeling decisions, and with the use of knowledge-engineering tools. Naturally, automated systems cannot detect all potential problems, but they help developers avoid many common mistakes, and they improve the quality of the target knowledge-based system.
Track: Knowledge engineering. 35eriksson.ps.Z (169872)
*Jintae Lee, **Michael Gruninger, ***Yan Jin, #Thomas Malone, ##Austin Tate,
###Gregg Yost and other members of the PIF Working Group
*University of
Hawaii, Department of Decision Sciences.
**University of Toronto, Department
of Industrial Engineering
***Stanford University, Department of Civil
Engineering
#MIT, Center for Coordination Science
##University of
Edinburgh, Artificial Intelligence Applications Institute
###Digital
Equipment Corporation
jl@celine.ics.Hawaii.Edu
This document provides the specification of the Process Interchange Format (PIF) version 1.1. The goal of this work is to develop an interchange format to help automatically exchange process descriptions among a wide variety of business process modeling and support systems such as workflow software, flow charting tools, planners, process simulation systems, and process repositories. Instead of having to write ad hoc translators for each pair of such systems, each system will only need to have a single translator for converting process descriptions in that system into and out of the common PIF format. Then any system will be able to automatically exchange basic process descriptions with any other system. This document describes the PIF-CORE 1.1, i.e. the core set of object types (such as activities, agents, and prerequisite relations) that can be used to describe the basic elements of any process. The document also describes a framework for extending the core set of object types to include additional information needed in specific applications. These extended descriptions are exchanged in such a way that the common elements are interpretable by any PIF translator and the additional elements are interpretable by any translator that knows about the extensions. The PIF format was developed by a working group including representatives from several universities and companies and has been used for experimental automatic translations among systems developed independently at three of these sites. This document is being distributed in the hopes that other groups will comment upon the interchange format proposed here and that this format (or future versions of it) may be useful to other groups as well. The PIF Document 1.0 was released in December 1994, and the current document reports the revised PIF that incorporate the feedback received since then.
Track: Sharable and Reusable Ontologies. 36lee.ps.Z (568695)
*Dieter Fensel, **Arno Schönegge, ***Rix Groenboom, *Bob
Wielinga
*University of Amsterdam, Department SWI, Roetersstraat 15, 1018 WB
Amsterdam, the Netherlands.
**University of Karlsrnhe, Institut fur Logik,
Komplexität und Deduktionssysteme, 76128 Karlsrnhe,
Germany.
***University of Groningen, Department of Computing Science, P.O.
Box 800, 9700 AV Groningen, the Netherlands.
fensel@swi.psy.uva.nl,
schoeneg@ira.uka.de, rix@cs.rug.nl, wielinga@swi.psy.uva.nl
The paper introduces a formal approach for the specification and verification of knowledge-based systems. We identify different elements of such a specification: a task definition, a problem-solving method, a domain model, an adapter, and assumptions that relate these elements. We present abstract data types and a variant of dynamic logic as formal means to specify these different elements. Based on our framework we can distinguish several verification tasks. In the paper, we discuss the application of the Karlsruhe Interactive Verifier (KIV) for this purpose. KIV was originally developed for the verification of procedural programs but it fits well for our approach. We illustrate the verification process with KIV and show how KIV can be used as an exploration tool that helps to detect assumptions necessary to close the gap between the task definition and the competence of a problemsolving method.
Track: Shareable and reusable problem solving methods. 37fensel.ps.Z (181577)
Frank Maurer
University of Kaiserslautern, P.O. Box 3049, 67653
Kaiserslautern
maurer@informatik.uni-kl.de
In this paper we describe how explicit models of software or knowledge engineering processes can be used to guide and control the distributed development of complex systems. The paper focuses on: techniques which automatically infer dependencies between decisions from a process model; and methods which allow to integrate planning and execution steps. Managing dependencies between decisions is a basis for improving the traceability of development processes. Switching between planning and execution of subprocesses is an inherent need in the development of complex systems. The paper concludes with a description of the CoMo-Kit system which implements the technologies mentioned above and which will use WWW technology to coordinate development processes.
Track: Distributed knowledge modeling over the Internet. 38maurer.ps.Z (97505)
D. Fensel and R. Benjamins
Department of Social Science Informatics,
University of Amsterdam, Roetersstraat 15, NL 1018 WB
Amsterdam
fensel@swi.psy.uva.nl
Most papers in model-based diagnosis and most papers in knowledge engineering on problem-solving methods focus on the description of reasoning strategies and discuss their underlying assumptions as a side-aspect. We take a complementary point of view by focusing on these underlying assumptions as these assumptions play an important role: They are necessary to characterise the precise competence of a problem-solving method in terms of the tasks that can be solved by it and in terms of the domain knowledge that is required by it. They are necessary to enable tractable problem solving for complex problems. Their introduction and the refinement of existing assumptions can be used to develop new problem-solving methods or to adapt existing ones according to task and domain-specific circumstances of a given application. For this purpose, one require a framework for dealing with these assumptions. The paper makes a step in this direction by summarizing the assumptions that can be found in the literature on component-based diagnosis. The main contribution of the paper is to collect these assumptions, to make their role for the reasoning process explicit, and to systematize these assumptions by making explicit their relationships.
Track: Problem-solving methods. 39fensel.ps.Z (71289)
Takahira Yamaguchi and Daiki Kishimoto
School of Information, Shizuoka
University, 3-5-1 Johoku Hamamatsu, 432 Japan
{yamaguti,
kisimoto}@cs.inf.shizuoka.ac.jp
In order to model distributed expert systems at a proper level of granularity of knowledge, focusing on a common task ontology, inference primitive structure templates has been proposed so that they can interoperate with other expert systems to solve problems that cannot be solved alone. Furthermore, in order to modularize such expert-system deep interoperation at a semantic level, the cooperation method has been presented, which pays attention to the context of the correspondence between inference primitives from an originator and a recipient. The wrapper with conversion facilities has been also provided, using a common domain ontology. After designing and implementing such an interoperation facilities, the experiment has been done between the following model-based like diagnostic expert systems in real task-domains, a trouble-shooting expert system and an enterprise diagnosis (financial management) expert system. The results have shown us that the former helps the latter get a refined inference structure and find some way to perform a given task better.
Track: Agent-oriented approaches to knowledge engineering. 41yamaguchi.ps.Z (221295)
*Rodrigo Martinez-Bejar, **Richard Benjamins, ***Fernando Martin and *Victor
Castillo
*Spanish Council for Scientific Research - CEBAS, Avda. La Fama, 1,
C.P. 30080, Murcia, Spain.
**Department of Social Science Informatics,
University of Amsterdam, Roetersstraat 15, 1018 WB Amsterdam, The
Netherlands.
***Department of Computing Sciences, University of Murcia,
Murcia, Spain.
rodrigo@natura.cebas.csic.es, richard@swi.psy.uva.nl,
fmartin@dif.um.es, victor@natura.cebas.csic.es
The knowledge elicitation process has a considerable influence on the quality of the knowledge-based system to be developed. Moreover, because knowledge elicitation is an important cost-determining factor, a good elicitation technique should reveal the relevant knowledge in the minimum amount of time possible. In this paper, we derive a set of formal parameters, based on mathematical functions, for comparing knowledge elicitation techniques. We use the landscape study task (a sub-task of environmental planning) as an illustrating example throughout.
Track: Knowledge Acquisition from Natural Language. 42martinez.ps.Z (580924)
Ashok Goel, Kavi Mahesh, Justin Peterson, and Kurt Eiselt
College of
Computing, Georgia Institute of Technology, Atlanta, GA 30332-0280
USA
goel@cc.gatech.edu
Cognitive agents often acquire knowledge of how devices work by reading a book. We describe a computational theory of understanding a natural language description of a device, comprehending how the device works, and acquiring a device model. The theory posits a complex interplay between language, memory, comprehension, problem-solving and learning faculties. Long-term memory contains cases of previously encountered devices and associated structure-behavior-function (SBF) models that explain how the known device works. Language processing is both bottom-up and top-down. Bottom-up processing is done through spreading-activation networks, where the semantics of the nodes and links in the network arises from the SBF ontology. The comprehension process constructs a SBF model for the new device by adapting the known device models - we call this process adaptive modeling. This multi-faculty computational theory is instantiated in an operational computer system called KA that (i) reads and understands English language descriptions of devices from David Macaulay's popular science book The Way Things Work, (ii) comprehends how the described device works, and (iii) acquires a SBF model for the device.
Track: Knowledge Acquisition from Natural Language. 43goel.ps.Z (38291)
*Frances Brazier, **Barbara Dunin-Keplicz, *Jan Treur, *Rineke
Verbrugge
*Vrije Universiteit Amsterdam, Department of Mathematics and
Computer Science, Artificial Intelligence Group, De Boelelaan 1081a, 1081 HV
Amsterdam, The Netherlands
** Warsaw University, Institute of Informatics,
ul. Banacha 2, 02-097 Warsaw, Poland
{frances, treur, rineke}@cs.vu.nl,
keplicz@mimuw.edu.pl
A declarative compositional modelling framework, DESIRE, designed to model multi-agent systems, is shown to provide a means to model agents' motivational attitudes. Thus desires, goals, intentions, commitments and plans, as well as the static and dynamic relations between them, are modelled. The specification in the framework DESIRE is sufficient to generate an implementation.
Track: Agent-oriented approaches to knowledge engineering. 44brazier.ps.Z (40095)
Frances Brazier, Jan Treur
Universiteit Amsterdam, Department of
Mathematics and Computer Science, De Boelelaan 1081a, 1081 HV Amsterdam, The
Netherlands.
{frances, treur}@cs.vu.nl
A formal approach to the design of (meta-level) compositional architectures for multi-agents systems is presented. A structure for reflective agents is proposed within which reasoning about observation and communication, an agent's own information state and reasoning processes, other agents' information states and reasoning processes, and combinations of these types of reflective reasoning are explicitly modelled. To illustrate the approach the wise men's puzzle has been modelled using different types of reflection.
Track: Agent-oriented Approaches to Knowledge Engineering. 45brazier.ps.Z (29002)
Frances Brazier, Jan Treur, Niek Wijngaards and Mark Willems
Intelligence
Group, Department of Mathematics and Computer Science, Vrije Universiteit
Amsterdam, De Boelelaan 1081a, 1081 HV Amsterdam, The Netherlands.
{frances,
treur, niek, willems}@cs.vu.nl
A formal approach to modelling and specifying complex dynamic tasks is introduced. A basic assumption is that complex (reasoning) tasks can be modelled by a compositional architecture wherein (sub-)components correspond to (sub-)tasks. Dynamic aspects are essential when modelling complex reasoning tasks. A modelling framework is presented including a language for the specification of compositional architectures. The semantics is a description of a compositional systems behaviour, and a temporal approach provides a means to describe the dynamics involved. Temporal models are used to formalise this semantics. The compositional structure of information states, transitions and reasoning traces provides a transparant model of the systems behaviour, both conceptually and formally.
Track: KA for temporal reasoning and planning. 46brazier.ps.Z (65391)
Stefan Decker, Michael Erdmann and Rudi Studer
Institut AIFB, University
of Karlsruhe (TH), D-76128 Karlsruhe, Germany
mer@aifb.uni-karlsruhe.de
In recent years the demand on business process modeling (BPM) became apparent in many different communities. To provide a unifying framework we propose an approach to integrate (well known) knowledge engineering techniques into a business process modeling context. We see knowledge based systems as one possibility (among others) to implement (re-engineered) business processes. As an example we will compare and integrate the knowledge engineering approach MIKE (Modelbased and Incremental Knowledge Engineering) with ARIS (ARchitecture of Information Systems), a prominent representative of business process modeling. We will answer a number of questions, that arise when unifying both approaches.
Track: Corporate memory & enterprise modelling. 47decker.ps.Z (79229)
H. Paul de Greef
Social Science Informatics, University of Amsterdam,
Roetersstraat 15, 1018 WB Amsterdam, The Netherlands.
degreef@swi.psy.uva.nl
This paper presents a high-level language for interaction protocols or cooperation methods involving many agents. The language concentrates on the social behaviour of agents. It mediates between high level languages such as Winograd and Flores' networks and low level implementation languages such as agent-process languages. It can serve as a tool for analysis and modelling of cooperation and collaboration and it can be used in model-based design of CSCW or Multi-Agent systems. In this paper first a set of criteria is established to evaluate languages for modelling collaborative workprocesses. In an interdisciplinary survey, a number of existing approaches are reviewed. From these, we synthesize a modelling language that meets the criteria. The language plays an important role in a model-based design method comprising three steps. The first is to draft a Winograd and Flores diagram. The second is to transform it to a specification in the proposed language and to add detail. Finally, this specification is transformed to a set of agent programs.
Track: Agent-oriented Approaches to Knowledge Engineering. 48degreef.ps.Z (103517)
John H. Gennari and Mark A. Musen.
Section on Medical Informatics,
Stanford University School of Medicine, Stanford, CA 94305-5479,
USA.
gennari@camis.stanford.edu, musen@camis.stanford.edu
For many years, researchers in knowledge-based systems area have worked toward the development of sharable and reusable problem-solving methods and knowledge bases. The aim is to reduce development and maintenance costs, and to build flexible, component-based systems that can be adapted to fit changing environments. Unfortunately, although there has been progress with the conceptual ideas necessary for building and connecting components, there has been little success with large-scale implementations of sharable component libraries. Technological problems have made it extremely difficult or impossible to share executable components across platforms and environments. Recently, the CORBA 2.0 specification has made exactly this sort of component reuse possible. The software industry is rapidly moving to solve exactly those implementation problems that have hindered work in reuse for knowledge-based systems. Thus, researchers interested in reusable components for knowledge- based systems should implement their ideas with these emerging standards. Although CORBA provides an implementation-level mechanism for reuse, the challenge remains to build frameworks for libraries of components that share some semantics and can inter-operate to be used to build systems that solve large, real-world problems. We illustrate how CORBA might be used to implement knowledge-based systems reuse with a number of scenarios drawn from our experience with reuse in the PROTÉGÉ-II environment.
Track: Distributed knowledge modeling over the internet. 49gennari.ps.Z (144061)
Mildred L. G. Shaw and Brian R. Gaines
Knowledge Science Institute,
University of Calgary, Alberta, Canada T2N 1N4
{mildred,
gaines}@cpsc.ucalgary.ca
The development of knowledge-based systems involves the management of a diversity of knowledge sources, computing resources and system users, often geographically distributed. The knowledge acquisition, modeling and representation communities have developed a wide range of tools relevant to the development and management of large-scale knowledge-based systems, but the majority of these tools run on individual workstations and use specialist data formats making system integration and knowledge interchange very problematic. The World Wide Web is a distributed hypermedia system available internationally through the Internet. It provides general-purpose client-server technology which supports interaction through documents with embedded graphic user interfaces. This paper reports on the development of knowledge open architecture knowledge management tools operating through the web to support knowledge acquisition, representation and inference through semantic networks and repertory grids. It illustrates how web technology provides a new knowledge medium in which knowledge-based system methodologies and tools can be integrated with hypermedia technologies to provide a new generation of knowledge management facilities.
Track: Distributed knowledge modeling over the Internet . 51shaw.ps.Z (554063)
Lee Li-Jen Chen and Brian R. Gaines
Knowledge Science Institute,
University of Calgary, Alberta, Canada T2N 1N4
{lchen,
gaines}@cpsc.ucalgary.ca
With the growth of usage of List Severs and the World Wide Web the Internet has become a major resource for the acquisition of knowledge, and it has given new prominence to human discourse as a continuing source of knowledge. The society of distributed intelligent agents that is the Internet community at large provides an `expert system' with a scope and scale well beyond that yet conceivable with computer-based systems alone. It is important to model and support the processes by which knowledge is acquired through the net. In developing new support tools is one asks "what is the starting point for the person seeking information, the existing information that is the basis for their search." A support tool is then one that takes that existing information and uses it to present further information that is likely to be relevant. Such information may include relevant concepts, text, existing documents, people, sites, list servers, news groups, and so on. The support system may provide links to further examples of all of these based on content, categorization or linguistic or logical inference. The outcome of the search may be access to a document but it may also be email to a person, a list or a news group. This articles develops a model of services and knowledge processes on the Internet, describes various forms of support tool, and categorizes them in terms of the model.
Track: Distributed knowledge modeling over the Internet . 52chen.ps.Z (119357)
Brian R. Gaines
Knowledge Science Institute, University of Calgary,
Alberta, Canada T2N 1N4
gaines@cpsc.ucalgary.ca
A model is developed of the emergence of the knowledge level in a society of agents where agents model and manage other agents as resources, and manage the learning of other agents to develop such resources. It is argued that any persistent system that actively creates the conditions for its persistence is appropriately modeled in terms of the rational teleological models that Newell defines as characterizing the knowledge level. The need to distribute tasks in agent societies motivates such modeling, and it is shown that if there is a rich order relationship of difficulty on tasks that is reasonably independent of agents then it is efficient to model agents competencies in terms of their possessing knowledge. It is shown that a simple training strategy of keeping an agent's performance constant by allocating tasks of increasing difficulty as an agent adapts optimizes the rate of learning and linearizes the otherwise sigmoidal learning curves. It is suggested that this provides a basis for assigning a granularity to knowledge that enables learning processes to be managed simply and efficiently.
Track: Agent-oriented Approaches to Knowledge Engineering. 53gaines.ps.Z (278295)
Brian R. Gaines, Douglas H. Norrie, Andrew Z. Lapsley and Mildred L.G.
Shaw
Knowledge Science Institute and Division of Manufacturing Engineering,
University of Calgary, Alberta, Canada T2N 1N4
gaines@cpsc.ucalgary.ca,
norrie@enme.ucalgary.ca, drewl@insource.com, mildred@cpsc.ucalgary.ca
The GNOSIS project in the Intelligent Manufacturing Systems international research program is concerned with the use of advanced information technology for knowledge systematization to support the complex intellectual and managerial processes involved in the manufacturing life cycle. It has developed technologies to coordinate distributed manufacturing enterprises, and these technologies have also proved useful in supporting the similar intellectual and managerial processes involved in distributed collaborative research. This articles gives the background to the project, and illustrates its use of information technology to provide a corporate memory, its use of knowledge acquisition and modeling tools to model the project objectives and conceptual structures, and the architecture of the Mediator system to support knowledge processes in distributed enterprises.
Track: Corporate memory and enterprise modeling. 54gaines.ps.Z (384417)
Klas Orsvärn
Swedish Institute of Computer Science, Intelligent
Systems Laboratory, Box 1263, 164 28 Kista, Sweden
klasorsv@sics.se
Chandrasekaran and Steels proposed several years ago that libraries of reusable problem solving methods, for use in model-driven knowledge acquisition, should be organized as hierarchies of task decomposition methods, rather than as collections of complete methods. One of the most comprehensive examples to date is Benjamins' library of methods for diagnosis tasks. In a case-study of using Benjamins' library, to model a specific diagnosis application, the most suitable model generated by the library had to be modified in several ways, despite the fact that the application is relatively simple and mainstream. This caused significant difficulties, both in identifying the modification requirements, and in creating the necessary adaptations. This paper proposes a set of general principles which libraries of task decomposition methods can be evaluated against, in order to prevent unnecessary adaptations. The principles concern method correctness, specialization of selection criteria, and method generality.
Track: Sharable and reusable problem solving methods. 55orsvarn.ps.Z (84683)
Dickson Lukose
Distributed Artificial Intelligence Centre (DAIC),
Department of Mathematics, Statistics, and Computing Science, The University of
New England, Armidale, 2351, N.S.W., Australia.
lukose@peirce.une.edu.au
In the last five years, there has been much effort towards developing knowledge modelling languages. These languages can be classified on two main dimensions: non-executable vs executable, and algorithmic vs graphical. In this paper, the author describes the graphical executable conceptual modelling language called MODEL-ECS, for operationalising the Knowledge Analysis and Design Support (KADS) structured methodology. The major contribution of this paper is in the development of all the necessary modelling constructs into MODEL-ECS, to enable it to be liberally exploited for formal modelling process in building executable Problem Solving Methods (PSMs). Primitive modelling constructs in MODEL-ECS is made up of Actor Graphs (i.e., a synergy between conceptual graphs and actor formalism). Actor Graphs form the most primitive executable conceptual structures. They can be partially temporally ordered to form complex executable structures called the Problem Maps. These Problem Maps form the PSMs. Complex Problem Maps are constructed using nested Problem Maps. Knowledge passing between Knowledge Sources within a context, and between contexts, in a nested Problem Map is achieved using coreference links, and line of identity. Complex modelling constructs available in MODELECS include: conditional construct; while loop; repeat loop; and case construct. The advantage of using MODEL-ECS comes from the expressibility provided by Conceptual Graphs, and the executability that is provided by the Executable Conceptual Structures.
Track: Knowledge Modelling using Conceptual Graphs. 56lukose.ps.Z (834960)
Florence Sellini and Catalina Vargas
PSA Peugeot Citroën, 62 Bd
Victor Hugo, 92208 Neuilly/Seine, France
101346.2215@CompuServe.com,
as13@calvanet.calvacom.fr
In this article, we present firsts considerations about validation issue from experiences in KBS design generated by collected know how at PSA Peugeot Citroën. Based on several projects an particularly on the methodology results of European DEKLARE Project, our survey highlighted that our approach of KBS design is often incomplete. In fact, the validation aspect is not addressed in a global view. Therefore, we shall demonstrate the need of setting our validation approach at the core of the design methodology of KBS.
Track: Corporate memory and enterprise modeling. 57sellini.ps.Z (215199)
Aldo Gangemi and Geri Steve
Medical Informatics Department, Institute of
Biomedical Technologies, National Research Council, Rome
{aldo,
geri}@color.irmkant.rm.cnr.it
A description of the ontological commitment assumed in a medical ontology produced by the ONIONS methodology is presented. The choices for individuating meta-level categories and their assignment to ontological items is described. This approach is constrained by the assumptions of ONIONS, which provides a feasible procedure to integrate structured linguistic repositories, which is based on a theory of semiology. The application of ONIONS to medicine is described as far as commitment choices and reuse issues are concerned. A-posteriori, a formal characterization of the commitment is shown to be obtainable by means of Guarino's ontology of meta-level categories.
Track: Shareable and Reusable Ontologies. 58gangemi.ps.Z (115547)
Sabina Falasconi, Giordano Lanzola and Mario Stefanelli
University of
Pavia, Dept. of Informatics and Systems Science, Medical Informatics
Laboratory, Via Ferrata, 1 - 27100 Pavia, Italy
sabina@ipvaimed6.unipv.it
Within knowledge engineering a new research paradigm is emerging based on the Multi-Agent System (MAS) architectural framework, allowing human and software agents to interoperate and thus cooperate within common application areas. Within a MAS, the different "views of the world" of knowledgeable agents are to be bridged through their commitment to common ontologies and terminologies. In our vision, ontological and terminological services are entrusted to dedicated agents, namely ontology and terminology servers providing other agents with the common semantic foundation required for effective interoperation, and allowing the configuration of suitable application ontologies for distributed applications.
Track: Shareable and Reusable Ontologies. 59falasconi.ps.Z (106921)
S. Mahe, C. Rieu and D. Beauchene
LLP / CESALP, Laboratoire de Logiciels
pour la Productique, 41, av. de la Plaine, BP 806, 74016 Annecy Cedex,
France
mahe@esia.univ-savoie.fr, rieu@esia.univ-savoie.fr,
beauchene@esia.univ-savoie.fr
Know-how capitalisation is now bringing a large interest in enterprises. After explaining what are the raisons of this brand new interest, we characterize the know how of the enterprise and we show how the Olympios organization model can help us in acquiring and structuring pieces of know-how of the enterprise. We show how attaching knowledge and know-how inside the representation of the organisation of the enterprise is relevant, particularly because that gives a part of the context of know-how.Then we present directions on which we are currently working for the design and structure of a tool which aim is to help knowledge identification and formulation and which contain locally a KB for expertise, communication aided tools with a capture module performing automatic knowledge discovering and a shared know-how KB allowing enterprises to share parts of their know-how, which is very important in the context of benchmarking. This research takes place in a large cooperation project between France (Region Rhone-Alpes) and Switzerland which aim is to realize a benchmarking between 15 enterprises in both part of the franco-swiss border. First part of this project is to give diagnosis tool enabling performance evaluation (presented in the Modelling Techniques, Business process and Benchmarking International Workshop, Bordeaux, April 1996). We wish to present here the second part concerning the necessary exchanges of know-how, and know-how capitalisation for benchmarking process, and what we propose for that.
Track: Corporate memory and enterprise modeling. 60mahe.ps.Z (751323)
*Yuval Shahar, **Silvia Miksch, *Peter Johnson
*Section on Medical
Informatics, Medical School Office Building, x215, Stanford University,
Stanford, CA 94305-5479, USA
**Knowledge System Laboratory (KSL), M/C 9020,
Gates Computer Science, Building 2A, Stanford, California 94305,
USA
shahar@camis.stanford.edu, miksch@hpp.stanford.edu,
pdj@camis.stanford.edu
Skeletal plans are a powerful way to reuse existing domain-specific procedural knowledge while leaving room for execution-time flexibility. The plan schemata are instantiated and refined dynamically by the executing agent over significant periods of time and in highly dynamic environments. In the ASGAARD project, we are investigating a set of tasks that support the execution of skeletal plans by a human executing agent, other than the original plan designing agent. We are developing task-specific problem-solving methods that perform these tasks in multiple clinical domains, given an instance of a clinical guideline plan and an electronic medical patient record. We point out the precise domain-specific knowledge roles required by each problem-solving method and present a sharable, text-based, machine-readable language, called ASBRU, to represent and to annotate execution plans in standardized form. Finally, we introduce an automated knowledge-acquisition tool for clinical guidelines based on the task-specific ontology, which was developed using the PROTÉGÉ-II framework's suite of tools.
Track: KA for temporal reasoning and planning. 62shahar.ps.Z (1166029)
*#Pascal Beys, *Richard Benjamins and *Gertjan van Heijst
*Department of
SWI, University of Amsterdam, Roetersstraat 15, 1018 WB Amsterdam, the
Netherlands
# LAFORIA-IBP, University Pierre et Marie Curie-CNRS, Case
courrier 169, Tour 46-00, 2eme etage 4, Place Jussieu 75 252 Paris cedex
05-France
pascal@swi.psy.uva.nl, richard@swi.psy.uva.nl,
gertjan@swi.psy.uva.nl
We expect reuse of parts of knowledge-based systems to become more and more important in the near future, as opposed to building systems from scratch every time. It is, however, well known that there exists a tradeoff between usability and reusability: the more components become reusable, the less usable they become. This is due to the fact that, in order to have a working system, the gap has to be bridged between the general reusable part and the specific application. In this paper, we elaborate on a recent approach to enhance the reusability of problem-solving methods and we propose a solution for resulting usability problem.
Track: Sharable and reusable problem solving methods. 64beys.ps.Z (121271)
*Richard Benjamins, ** Leliane Nunes de Barros, ***Andre
Valente
*University of Amsterdam, Social Science Informatics (SWI),
Roetersstraat 15, 1018 WB Amsterdam, the Netherlands.
**University of
Maryland, Department of Computer Science, College Park, MD 20742,
USA.
***University of Southern California,Information Sciences Institute,
4676 Admiralty Way, Marina del Rey, CA 90292, USA.
richard@swi.psy.uva.nl,
leliane@cs.umd.edu, valente@isi.edu
Constructing a planner for a particular application is a difficult job, for which currently not much concrete support is available. The literature on planning is overwhelming and there does not exist a clear and synthetic understanding of the various planning methods around, that could be of direct concrete use for a knowledge engineer. In this paper, we show how a general knowledge-level framework for conceptually specifying knowledge-based systems, can be of concrete use to support knowledge acquisition for planning systems. The framework encompasses three inter-related components: (1) problem-solving methods, (2) their assumptions and (3) domain knowledge. The presented analysis of planning performed in the framework can be considered as a library with reusable components based on which planners can be configured. Two experiments are presented that illustrate the use of the library in knowledge engineering.
Track: KA for temporal reasoning and planning. 65bejamins.ps.Z (107997)
*I. Laresgoiti, **A. Anjewierden, *A. Bernaras, ***J. Corera, **A.Th.
Schreiber and **B.J. Wielinga
*LABEIN, Parque Tecnologico, Ed. 101, 48016 -
Zamudio Bickaia, Spain.
**University of Amsterdam, Department of Social
Science Informatics, Roetersstroat 15, NL-1018 WB Amsterdam, The
Netherlands.
***IBERDROLA, Gardoqui, 8, 48008 - Bilboo,
Spain.
lares@labein.es, anjo@swi.psy.uva.nl, amaia@labein.es,
jose.corera@iberdrola.es, schreiber@swi.psy.uva.nl, wielinga@swi.psy.uva.nl
KACTUS is a European ESPRIT project in which a group of industrial and academic partners set out to work on methods and tools to enable reuse and sharing of technical knowledge. The technical basis of this project was founded on insights in knowledge-engineering research, where he emphasis has shifted over the last few years from modeling reasoning processes to modeling "ontologies": reusable pieces of domain knowledge. In the view of KACTUS the area of reusing and sharing knowledge and complex data will become a prime focus of software-engineering practice in the near future. To be able to build a new generation of intelligent, interacting IT systems, it is necessary to get some grip on a number of standard ways of "viewing the world", albeit in a limited and pragmatic, engineeringoriented fashion. KACTUS is trying to provide a linlc between CIME (Computer-Integrated Manufacturing and Engineering) research on standard product models for engineering (the ISO STEP standard) and emerging standards in knowledge engineering, in particular the CommonKADS approach.
Track: Shareable and Reusable Ontologies. 66laresgoiti.ps.Z (227059)
*Martin Molina, **Yuval Shahar, *Jose Cuena, **Mark A. Musen
* Department
of Artificial Intelligence, Technical University of Madrid, Campus de
Montegancedo S/N, Boadilla del Monte 2866O, Madrid, Spain.
**Section on
Medical Informatics, School of Medicine, Stanford University, Stanford, CA
94305, USA.
mmolina@dia.fi.upm.es, shahar@camis.stanford.edu,
jcuena@dia.fi.upm.es, musen@camis.stanford.edu
The paper presents a case study that compares two of the existing knowledge modeling platforms: PROTÉGÉ-II and KSM. These two software environments allow a developer to build a knowledge level model and to create the final operational version. Both environments have been used to develop real world models (e.g., for medical and traffic domains respectively). In the paper, we first describe a knowledge model following the KADS methodology that has been defined to carry out real-time decision support tasks in domains such as traffic control. Then, we present how the model is defined and operationalized using, first, the PROTÉGÉ-II environment and, second, the KSM environment. Finally, we discuss similarities and differences between both approaches.
Track: Shareable and reusable problem solving methods. 67molina.ps.Z (231523)
Nicola Guarino
LADSEB-CNR, National Research Council
Corso Stati Uniti
4, I-35127 Padova, Italy.
guarino@ladseb.pd.cnr.it
I will defend here the thesis of the independence of domain knowledge. This thesis should not be intended in a rigid sense, since it is clear that - more or less - ontological commitments always reflect particular points of view; rather, what I would like to stress is the fact that reusability across multiple tasks or methods should be systematically pursued even when modeling knowledge related to a single task or method: the more this reusability is pursued, the closer we get to the intrinsic, task-independent aspects of a given piece of reality (at least, in the commonsense perception of a human agent).
Track: Shareable and Reusable Ontologies. 69guarino.ps.Z (70027)
Adam Farquhar, Richard Fikes and James Rice
Knowledge Systems Laboratory,
Stanford University, Gates Building 2A MC9020, Stanford, CA
94305
{afarquhar,fikes,rice}@ksl.stanford.edu
Reusable ontologies are becoming increasingly important for tasks such as information integration, knowledge-level interoperation, and knowledge-base development. We have developed a set of tools and services to support the process of achieving consensus on common shared ontologies by geographically distributed groups. These tools make use of the world-wide web to enable wide access and provide users with the ability to pub-lish, browse, create, and edit ontologies stored on an ontology server. Users can quickly assemble a new ontology from a library of modules. We discuss how our system was constructed, how it exploits existing protocols and browsing tools, and our experience supporting hundreds of users. We describe applications using our tools to achieve con-sensus on ontologies and to integrate information.
Track: Distributed knowledge modeling over the Internet or Shareable and Reusable Ontologies. 70farquhar.ps.Z (138036)
*P. Compton, *P. Preston, **G. Edwards, ***B. Kang
*School of Computer
Science and Engineering, University of New South Wales, Sydney 2052,
Australia
**Western Diagnostics, Perth, Australia
***Hitachi Advanced
Research Laboratories, Saitama Japan
compton@spectrum.cs.unsw.oz.au
A major problem with Knowledge Based Systems (KBS) is that they do not know when they have reached the limits of their knowledge and so are likely to make stupid conclusions. One possible approach to this is have the KBS keep some sort of record of every type of case it has seen and issue a warning when it is dealing with a case outside its range of experience. In these studies we have used the simple technique of issuing a warning if any attribute has a value outside the range of those seen previously for cases satisfying the same sequence of rules during inference. This technique has been evaluated using standard machine learning data bases and a simulated expert (built by machine learning) rather than a human expert. The conclusions from this study are that single attribute warnings cannot be guaranteed to pick up all potential errors, but go a long way towards a knowledge acquisition methodology whereby a KBS will fairly reliably prompt that it needs more knowledge.
Track: Knowledge Engineering. 72compton.ps.Z (105567)
*Andre Valente and **Joost Breuker
*Information Sciences Institute,
University of Southern California, 4676 Admiralty Way, Marina del Rey, CA
90292, USA
**Dept. of Social Science Informatics (SWI), University of
Amsterdam, Roetersstraat 15, 1018 WB Amsterdam, the
Netherlands
valente@isi.edu, breuker@swi.psy.uva.nl
One of the important issues in the newborn discipline of ontological engineering is the construction of libraries of ontologies which are built to achieve maximum reusability. [vanHeijst et al., 1996] suggested that a central part of ontology libraries is the definition of what they called a core ontology that contains elements as generic and method-independent as possible. However, their specification of how these core ontologies should be built is highly pragmatical, and leaves many open problems. In this article we propose and discuss a number of specific principles for the construction of core ontologies. We demonstrate the advantages of these principles using as an example a core ontology we have built for the domain of law. Several conclusions about the construcion of ontology libraries based on core ontologies are drawn.
Track: Shareable and Reusable Ontologies. 73valente.ps.Z (110351)
Steve A. Chien
Jet Propulsion Laboratory, California Institute of
Technology, 4800 Oak Grove Drive, M/S 525-3660, Pasadena, CA 91109-8099,
USA
steve.chien@jpl.nasa.gov
A key obstacle hampering fielding of AI planning applications is the considerable expense of developing, verifying, updating, and maintaining the planning knowledge base (KB). Planning systems must be able to compare favorably in terms of software lifecycle costs to other means of automation such as scripts or rule-based expert systems. Consequently, in order to field real systems, planning practitioners must be able to provide: 1. tools to allow domain experts to create and debug their own planning knowledge bases; 2. tools for software verification, validation, and testing; and 3. tools to facilitate updates and maintenance of the planning knowledge base. This paper begins by describing a planning application of automated image processing and our overall approach to knowledge acquisition for this application. This paper then describes two types of tools for planning knowledge base development: static KB analysis techniques to detect certain classes of syntactic errors in a planning knowledge base; and completion analysis techniques, to interactively debug the planning knowledge base. We describe these knowledge development tools and describe empirical results documenting the usefulness of these tools.
Track: Knowledge Acquisition for Planning and Temporal reasoning. 74chien.ps.Z (171804)
Zdenek Zdrahal and Enrico Motta
Knowledge Media Institute, The Open
University, Walton Hall, Milton Keynes, MK7 6AA, UK
{z.zdrahal,
e.motta}@open.ac.uk
We analyse the behaviour of a Propose & Revise architecture in the VT elevator design problem and we show that this problem solving method cannot solve all possible cases covered by the available domain knowledge. We investigate this problem and we show that this limitation is caused by the restricted search regime employed by the method and that the competence of the method cannot be improved by acquiring additional domain knowledge. We therefore propose an alternative design problem solver, which integrates case-based reasoning and heuristic search techniques and overcomes the competence-related limitations exhibited by the Propose & Revise architecture, while maintaining the same level of efficiency. We describe four algorithms for case-based design, which exploit both general properties of parametric design tasks and application specific heuristic knowledge.
Track: Shareable and reusable problem solving methods. 75zdrahal.ps.Z (36553)
Piet-Hein Speel and Manfred Aben
Unilever Research Laboratory
Vlaardingen
Piet-Hein.Speel@Unilever.com, Manfred.Aben@Unilever.com
This paper describes the application of a library of Problem Solving Methods (PSMs) for model-based diagnosis (Benjamins, 1993) on a real-life task. PSMs have been proposed as a comprehensive and feasible approach to the construction of conceptual models for knowledge-based systems (KBSs). A number of evaluations of this approach have been reported in the literature. Most of these evaluations have involved toy problems, or reverse engineering of existing KBSs. In this paper we have applied Benjamin's library to a real-life task. In this paper we discuss how we have applied the PSMs, and how we had to adapt them to suit our task. We have found that the library of PSMs was extremely helpful in our project. Thanks to this reuse, the complete scope and quality of the "prime diagnostic method" was incorporated in our application, ambiguities in the task and domain were brought to light, and the functional specifications of the target system were structured transparently. Since the reuse of library components in our real-life application was not trivial, we emphasize on the need for guidelines that support knowledge engineers to apply libraries of PSMs during the development of real-life KBSs.
Track: Shareable and reusable problem solving methods.
76speel.ps.Z (not yet up, awaiting Unilever approval for distribution)
Rob Kremer
Knowledge Science Institute, University of Calgary, Calgary,
Alberta Canada T2N 1N4
kremer@cpsc.ucalgary.ca
Concept maps are used in a wide variety of disciplines for their ability to make complex information structures explicit. The Web is seen as an excellent medium in which to make concept maps available for real-time, multi-user interaction for planning, decision making, documenting, and designing tasks, or merely as a navigational aid to information. KSIMapper is a program that can run as a Netscape plug-in to allow this kind of flexible interaction. Concept maps can be used informally or formally, where the graphical "syntax" of the maps are tightly controlled. Both forms are needed. Graphs is a program in which users can constrain arbitrary graphs to conform to any of a wide variety of graphical formalisms. The Graphs program is combined with KSIMapper to bring it a graphical user interface; the two can function as a multi-user, interactive concept mapping system on the Web that can transcend informal concept mapping and at least several concept mapping formalisms.
Track: Distributed knowledge modeling over the Internet. 77kremer.ps.Z (249033)
Bill Swartout, Ramesh Patil, Kevin Knight and Tom Russ
USC/Information
Sciences Institute, 4676 Admiralty Way, Marina del Rey, CA 90292,
USA
swartout@isi.edu, ramesh@isi.edu, knight@isi.edu, tar@isi.edu
construct. If we could share knowledge across systems, costs would be reduced. However, because knowledge bases are typically constructed from scratch, each with their own idiosyncratic structure, sharing is difficult. Recent research has focused on the use of ontologies to promote sharing. An ontology is a hierarchically structured set of terms for describing a domain that can be used as a skeletal foundation for a knowledge base. If two knowledge bases are built on a common ontology, knowledge can be more readily shared, since they share a common underlying structure. This paper outlines a set of desiderata for ontologies, and then describes how we have used a large-scale (50,000+ concept) ontology to develop a specialized, domain-specific ontology semi-automatically. We then discuss the relation between ontologies and the process of developing a system, arguing that to be useful, an ontology needs to be created as a "living document", whose development is tightly integrated with the system's. We conclude with a discussion of Web-based ontology tools we are developing to support this approach.
Track: Shareable and Reusable Ontologies. 78swartout.ps.Z (585808)
Enrico Motta & Zdenek Zdrahal
Knowledge Media Institute, The Open
University, Walton Hall, Milton Keynes, MK7 6AA, UK
{e.motta,
z.zdrahal}@open.ac.uk
Parametric design problems constitute a class of design problems where a valid design can be characterised in terms of a consistent and complete assignment of values to parameters. In most applications, however, a valid design does not necessarily provide an acceptable solution. Different valid designs can have different costs, and for this reason an acceptable solution design typically has to meet, or at least approximate, some optimization criterion. The aim of this paper is to understand what is involved in parametric design problem solving. In order to achieve this goal, in this paper i) we identify and detail the conceptual elements defining a parametric design task specification; ii) we illustrate how these elements are interpreted and operationalised during the design process; and iii) we produce a generic model of parametric design problem solving, characterised at the knowledge level, which generalizes from existing methods for parametric design. We then re-describe a number of problem solving methods in terms of the proposed generic model and we show that such a re-description enables us to provide a more precise account of the different competence behaviours expressed by the methods in question.
Track: Shareable and reusable problem solving methods. 79motta.ps.Z (58387)
Hugh Cottam and Nigel Shadbolt
University of Nottingham, AI Group, Dept
of Psychology, University Park, Nottingham NG7 2RD, UK.
{hdc,
nrs}@psyc.nott.ac.uk
The field of knowledge level modelling has achieved great success when applied to various domains, yet has thus far largely neglected the generic areas of planning, scheduling and resource allocation. In this paper we outline the development of a knowledge level modelling approach within the domain of planning for Search and Rescue. Existing problem solving models for planning are almost exclusively derived from the analysis of systems. We argue that this makes their suitability for directly assisting knowledge acquisition debatable. Our approach makes a clear distinction between domain derived knowledge level models and those derived from systems. We describe how the combination of these two types of model can achieve definite benefits within the course of KBS development. The paper also highlights important aspects of modelling expert workflow, which reveal key requirements for any effective knowledge intensive system.
Track: KA for temporal reasoning and planning. 80cottam.ps.Z (152530)
* Samson W. Tu, Mark A. Musen
Section on Medical Informatics, Stanford
University School of Medicine, Stanford, CA 94305-5479
{tu,
musen}@smi.stanford.edu
This paper describes the reformulation of the episodic skeletal-plan refinement (ESPR) problem-solving method (PSM) in a new framework that seeks to integrate knowledge-based applications with a temporal data-abstraction and data-management system. In this framework, both applications and temporal-data managers are encapsulated as CORBA objects. We found that we need to reformulate the method ontology, mapping relations, and control structure of the PSM in this framework. Our experience suggests that PSMs are not necessarily fixed structures that can be plugged into arbitrary application environments, and that we need to develop a flexible configuration environment and expressive mapping formalisms to accommodate the requirements of application environments. These requirements include the ways data are made available and the ways software components interact with each other.
Track: KA for temporal reasoning and planning. 81tu.ps.Z (255117)
Jeffrey M. Bradshaw
The Boeing Company
jbrad@redwood.rt.cs.boeing.com
Current trends have made it clear that automation of dynamic, real-time environments will dramatically increase in the coming decades. The complexity, real-time constraints, and distributed nature of such tasks require that software not merely respond to requests for information but intelligently anticipate, adapt, and actively seek ways to support users. Not only must these systems assist in coordinating tasks among humans, they must also help manage cooperation among distributed programs. Software agents have been proposed as one way to help people better cope with the increasing volume and complexity of information and computing resources. Researchers are hopeful that this approach will help restore the lost dimension of individual perspective to the content-rich, context-poor world of the next decade. What will such agents do? At the user interface, they will work in conjunction with compound document frameworks and document management tools to select the right data, assemble the needed components, and present the information in the most appropriate way for a specific user and situation. Behind the scenes, agents will take advantage of distributed object management, database, workflow, messaging, transaction, searching, indexing, and networking capabilities to discover, link, and securely access the appropriate data and services.
Track: Agent-oriented approaches to knowledge engineering. 82bradshaw.ps.Z (210895)
.
gaines@cpsc.ucalgary.ca 31-Jul-96