Sixth International Protégé Workshop
Monday 7th - Wednesday 9th July 2003
Manchester, England


Presentation Abstracts
Note: Brian Gaines, editor-in-chief of The International Journal of Human-Computer Studies, has invited us to produce a special issue of the journal based on selected presentations from our workshop.


Ameen Abu-Hanna, R. Cornet, M. Crubezy, S. Tu :
"
On the Correspondence between Protégé and Conceptual Frameworks for Understanding Terminological Systems":

"Terminological systems (TSs) continue to form a hot research topic in medical informatics. Frameworks for understanding TSs have been suggested which provide a typology of TSs along with conceptual and formal specification of TS desiderata. These Frameworks provide guidance in the analysis and design phases of developing a TS but the actual implementation of TSs is still left as an exercise to the TS developer. However, software development environments such as Protégé have been devised for the flexible construction of knowledge based-systems from components. Features of Protégé include its conformance to the OKBC knowledge model; the logical separation between ontology and application; and its componential architecture. In this talk we investigate the compatibility of Protégé with the conceptual framework for understanding Terminological Systems suggested in (de Keizer and Abu-Hanna, 2000). We address the compatibility issue by a constructive approach that provides mappings between the conceptual components of the TS framework and Protégé constructs. The development of an intensive care terminology system,
called DICE, is used as a case study.
N. deKeizer, A. Abu-Hanna.
Understanding Terminological Systems (II):
Experience with Conceptual and Formal Representation of Structure.
Methods of Information in Medicine, 39:22-29, 2000.


Harith Alani & Bo Hu :
"
Protégé-based Knowledge Services in the AKT Project":

This presentation will describe some of the Protégé based knowledge management applications and services developed within the AKT project at Southampton University. We will introduce TGVizTab, a new plugin for visualising large ontologies, and MIAKTab, a domain specific plugin for visualising, navigating and maintaining a breast cancer image knowledge base. A brief description of Ontocopi will also be given, which is a plugin to identify communities of practice using ontology network analyses techniques. We hope to gather some feedback and user requirements from this workshop with respect to visualising ontologies and knowledge bases.


Polly M. Allen :
"
Visualizating Academic Group Knowledge":

Academic groups, especially in university settings, tend to experience high turnover rates of personnel. Whether it is formal knowledge about particular research topics, or informal knowledge about the social network within the group, such knowledge is closely tied to the specific student, and often leaves with them; consequently, new additions to the group are forced to re-acquire this knowledge through experience.
We have observed these problems in our own research group - the CHISEL (Computer-Human Interaction and Software Engineering Lab) group at the University of Victoria, Canada. In this presentation, we will describe how we used Protégé to create a knowledge-based recommender system that captures and shares knowledge about group members and the papers they have read in certain research areas. This system includes a web interface for knowledge capture, and is integrated with a commercial, off-the-shelf tool called EndNote for integrating bibliographic references with popular word-processing tools. We show how,using a visualization plug-in for Protégé called Jambalaya, we were better able to share this knowledge with the group and discover new properties of the group knowledge base. We will also report on our experiences using Protégé to create such a system.


Rizwan Ameer & Krishnakumar Pooloth :
"Embedding CLIPS Engine in
Protégé":

This presentation is intended to share our experience in embedding NASA CLIPS Engine into Protégé and providing a CLIPS command line interface as a Protégé plugin. Since Protégé is written in Java and NASA CLIPS is entirely in C, there are certain issues involved in embedding CLIPS in Protégé. This presentation will discuss the basics of embedding CLIPS Engine in other applications, issues involved in embedding CLIPS engine in Protégé and how those issues were resolved. A working demo of Protégé CLIPS plugin (Command line interface to the embedded CLIPS Engine) is also planned.
We believe that this integration would help the Protégé users to develop complete KBSs' (Ontology+Rules) from within Protégé Platform.


Alvaro Arenas, Krzysztof Krawczyk, Mariusz Dziewierz & Michal Laclavik :

"Experiences with Using Protégé in a Knowledge Management Application":

Pellucid is a project, funded by the European IST programme, concerned with knowledge management for public employees, specifically for those who are organisationally mobile, moving from one department or post to another. This presentation illustrates our experience in using Protégé in Pellucid.
The architecture of the Pellucid system consists of a set of cooperating agents, taking care of monitoring, reasoning and interfacing, implemented in JADE; an organisational memory, containing the “history” of the organisation; and a workflow system, modelling the processes and activities within the organisation. Ontologies are used to support agent communication and reasoning. Two kind of ontologies were modelled: generic ontologies, representing part of the organisational model used in the workflow system, and domain specific ontologies, representing particular functionalities of the system such as contact, document and time-critical management, and to be extended into the particular domains of work of the organizations.
Ontologies have been implemented using Protégé. The Ontoviz Tab plug-in was employed to visualise ontology elements and the Bean Generator plug-in was employed to represent ontologies in the Java format used by JADE. We also use Protégé in the deployment phase as a mechanism for populating the organisational memory. We explore some challenges presented in the use of Protégé and describe future work.


M. Naveed Baqir & A. Rashid Kausar :
"Use of Protégé in the Scientific, Academic and Commercial Areas of Pakistan":

The talk shall focus on the current projects being developed and investigated in the scientific and academic circles of Pakistan. The talk shall also outline the past experience with ontology based project development. Excitements and problems faced by the researchers, students and professionals will also be discussed.


Reiner Borchert :
"
Executing Processes by a Knowledgebase":

Since we are planning to execute data evaluations and simulations of environmental processes which are controlled by a process ontology I have looked for a proper way to link java classes and methods to Protégé frames. The goal was to enable the frames to execute what they describe. The presentation shows a slot-widget plug-in which introduces a “run button” to start the process linked to a frame, and a tab plug-in to configure the connection between Protégé and Java archives and classes. To illustrate the functions of the plug-ins I will present a small example project as work in progress.


Sébastien Brachais, Amedeo Napoli, Jean Lieber & Mathieu d'Aquin:
"
Tentative: Protégé-2000 and the Kasimir Decision-support System for Breast Cancer Treatment":

Kasimir is a decision-support system currently under development for breast cancer treatment. It relies on knowledge bases (implemented within a specific XML format) representing clinical guidelines, and on a inference engine that classifies concepts representing domain knowledge. As in a description logic framework, concepts are divided into primitive and defined concepts.
The increasing size of Kasimir knowledge bases led us to study and use a knowledge editing tool. The use of such a tool has become a necessity regarding the Kasimir system evolution. The choice of Protégé-2000 to fulfil this task was made after a number of Web search for adequate tools, comparison, evaluation, and discussion among the team developing the Kasimir system. Thanks to the metaclass architecture in Protégé-2000, the Kasimir knowledge representation model can be described and, according to this model, Protégé generates forms suitable to the editing process. Then, Protégé classes (corresponding to domain concepts in Kasimir) can be checked by the Kasimir knowledge representation engine in the Kasimir plug-in before being exported to XML files.
Visualisation of huge hierarchies also led to the development of graphical plug-ins helping the user in the editing phase. One of these
tools gives the possibility to see changes (added, deleted, modified concepts) within Protégé project during the editing session.


Eleni Christopoulou & Achilles Kameas :
"
GAS Ontology: An Ontology for Collaboration among Ubiquitous Computing Devices":

This paper presents research that has been carried out during the “extrovert-Gadgets” project (http://www.extrovert-gadgets.net), a research project funded in the context of EU IST/FET proactive initiative “Disappearing Computer” (http://www.disappearing-computer.net). In this project everyday tangible (physical) objects, referred to as eGadgets, are enhanced with sensing, acting, processing and communication abilities, and are able to communicate and collaborate, constituting eGadgetworlds. The Gadgetware Architectural Style (GAS) defines the concepts and mechanisms that will allow people to compose eGadgetworlds by associating eGadgets and use them in a consistent and intuitive way. Each eGadget runs a small operating system (GAS-OS), which locally manages its resources and associations.
GAS Ontology aims to conceptualize GAS by describing the semantics of its basic terms and defining the relations among these terms. Thus GAS Ontology provides a shared means for the communication and collaboration among eGadgets, even though they may be produced by different manufacturers. Protégé-2000 was selected as the suitable tool for the development of GAS Ontology. For this we have developed an evaluation framework for ontology development tools. For the management of GAS Ontology as well as for the facilitation of knowledge sharing and reuse, we are developing the Ontology Manager module. This will be an integral part of the GAS-OS kernel, as every association among eGadgets starts with a negotiation on commonly understood protocols, functions and data types. We intend to develop the Ontology Manager in such a way that it will use and enhance techniques and methods for querying of Protégé-2000.


Ernesto Compatangelo:
"Interoperability Issues in Protégé-2000":

We review a series of issues which could possibly affect the interoperability of Protégé with some of the tools developed within the framework of the UK Interdisciplinary Research Collaboration in Advanced Knowledge technologies (AKT). We group these issues in three major themes, namely expressiveness of the knowledge model, mapping and translation mechanisms, and automated reasoning services.


Oscar Corcho:
"Achieving Interoperability between Protégé-2000 and WebODE":

We will present how to achieve interoperability between the two ontology tools, Protégé-2000 and WebODE, by means of translators that are able to preserve knowledge in the transformations, thus ensuring that the knowledge formalized with both tools is not lost in cyclic transformations.


Monica Crubézy :
"The PSM Librarian: Configuring Problem-Solving Applications in Protégé":

Problem-Solving Methods (PSMs) are reusable algorithms that encode domain-independent strategies to realize common knowledge-intensive tasks by processing domain knowledge. Well-known PSMs are propose-and-revise and heuristic classification, that have been used to tackle tasks such as constraint-satisfaction, classification and diagnosis in domains such as elevator configuration, ribosome-conformation prediction and diagnosis of arthritic conditions.
A PSM provides both a formal model of a generic reasoning process, and a piece of program code that implements it. As such, PSMs serve as reasoning building blocks that can be assembled with a domain knowledge base to create a running knowledge system. PSMs are typically collected in libraries, that help indexing and retrieving PSMs for different purposes and contexts. In particular, PSM libraries formalize the properties, capabilities and requirements of the PSMs available, using a specification language such as the Unified Problem-solving Method development Language (UPML). Because PSMs are domain-independent, selecting and adapting one for its reuse in a domain-specific knowledge application are non-trivial activities.
In this talk, I will present the PSM Librarian--a tool that plugs into Protégé to support users in creating knowledge-based application by integrating domain-knowledge components and PSMs. In particular, the PSM Librarian helps users in: (1) browsing and selecting a PSM suitable for the task and domain at-hand, and (2) configuring that PSM for the specific domain by declaring explicit mapping relations between domain entities and the inputs and outputs of the PSM. Used with a domain project, the PSM Librarian tab serves as a front-end interface to a set of UPML-specified PSM libraries (using a Protégé-based UPML editor) and provides navigation and edition support for the selection and configuration of PSMs available.


Chris de Vaney & Ben Warsop:
"
A Storyboard Interface to Business Processes using Protégé":

A detailed business process knowledge base has been developed using Protégé. To support organizing and disseminating the process informartion, a storyboarding interface has been developed to allow narratives based on the use of the processes to be developed.
This presentation describes the business process and storyboard models, and provides a demonstration based on the business model for a real-world 3rd-generation mobile telephone platform.


Vladimir Diatlov :
"Organization Analysis with Protégé: Towards “Living” Theorizing":

As a part of organization studies review and qualitative meta-analysis of research done, an ontology of concepts was compiled and unified with the framework of organizational structure, process, and boundary (Pettigrew, 2000) that becomes accepted in the field of organization theory. Then, such library of organization analysis elements was populated with information from an empirical case study ­ relations, authority, and other elements of social structure were codified as instances of concepts.
While such application of the Protégé is helpful in ordering, visualising, and questioning knowledge structure of a field (new relations among concepts, frameworks rearrangement), advantages and perils for the use of Protégé for theorizing, collaboration on research knowledge, and as a R&D tool in general are yet to be established.
Key discussion issues appear as the following: (1) need of significant knowledge of field analysed as well as theory behind Protégé (object-oriented design, ontology, frame-based systems, OKBC); (2) putting rich and often broad or loosely-coupled theoretical frameworks is compromised with the object-oriented paradigm and ontology format, for instance, a typical choice, ontology design issue, would be between a child and a parent concept (object); (3) keeping track of meanings behind constructs combined and remembering heuristics and conventions made for encoding of information in one rather than other way; (4) lack of opportunity to reframe and export research data for further statistical analysis; (5) use of Protégé for transfer of empirical data in form of some standard ontology.


Cláudio Gottschalg Duque & Marcello Peixoto Bax:
"Use of Protégé in SABiO, a Brazilian Information Retrieval System for indexing Electronic Information Science Documents":

This article treats of the electronic administration of Web documents, we present an index proposal whose approach involves the application of specific linguistic theories and of Protégé to develop an ontology based in the terms extracted from texts.
The collection of the Web documents is no problem anymore, but how to treat them, how to store and to index them, how to permit access in an efficient way to the user are yet open problems. The alternatives to solve them, or at least to understand them, reside in the intersections of the areas of the human knowledge. Linguistic, Computer Science and Information Science (Cognitive Sciences) can help to develop a robust Information Retrieval System that uses linguistic and knowledge information extracted from texts that are indexing, allowing more efficiency in answer user’s queries.


Christophe Dupriez & Mélanie Roland :
"Multilingual Vocal Access to Databases: the Central Role of Ontology Management in VocaBase":

An Ontology permits to structure (human) terminology into normalised concepts and also to build generalisation hierarchies from these concepts. The "individuals" in the Ontology may provide "solutions" to the users, solutions made of explanatory texts and transaction possibilities. The solutions are indexed using the concepts hierarchies and VocaBase generates the most pertinent choices to the users to navigate in these hierarchies. We present a vocal application based on this design.


Joe Edelman :
"Some
Protégé Extensions for Scientific Data Management":

We have added support for data objects like scientific measurements and recordings, made the Protégé UI more usable for our audience through adaptive tree renderings, and are looking at a deeper level of backend and UI support for the ontology version reconcilliation process.


Henrik Eriksson :
"JessTab: Using Jess together with
Protégé":

The tutorial contains an introduction to JessTab, its design philosophy, integration model, and practical use. After the tutorial you should be able to intall JessTab, take advantage of the mapping of instances to facts, use utility functions to manipulate Protégé ontologies and knowledge bases from Jess, and handle metalevel objects in Protégé from Jess. Also, the tutorial will include a discussion of the JessTab implementation and how you can extend it.


Neil Ernst :
"Adoption Issues for Protégé
":

Adoption-centric knowledge engineering is the design of knowledge engineering tools and knowledge engineering processes to ensure the widest possible adoption. Protégé, with 8000 registered users, has a relatively large user base for this community, but evolving web standards for the Semantic Web promise to greatly increase this community. This talk looks at some issues involved in making the tools and techniques of Protégé and its plugins accessible and available to a wider audience. In turn, wide adoption will benefit the projects that choose to focus on it by increasing the user-base. One way to make a tool more adoption-centric is to provide increased cognitive support to the potential user.



Ray Fergerson & Mark Musen:
"
The Protégé Project & Protégé-2000: Past, Present and Future":

We will review the history of the Protégé project and specifically discuss the previous workshops and what was discussed and accomplished at each. We will also discuss our successes and failures since the previous workshop two years ago and briefly discuss our plans for the future.



Ray Fergerson :
"
Future Directions & Open Issues":

This will be a very interactive session to allow users to tell the Protégé team what is important and what we should be working on. I will first summarize our current plans as well as my understanding of what the important issues are that have been raised at the workshop. We will then have an open discussion of these and other items, possible priorities, and other issues that anyone would like to bring up.



Yongsheng Gao:
"
Flexible Sets of Distinctions for Multiple Paradigms - a Means of Constructing Ontologies":

Ontologies are recognized as a key technology for knowledge sharing. A key task for health informatics is to model health care domain knowledge into ontologies. Within health care it is inevitable that we must deal with heterogeneous types and sources of knowledge, information and data. When incorporating multiple paradigms into ontologies we are faced not only with differences in underlying knowledge, but also with conflicts in the terminology used. In this presentation we identify the difficulties of modelling heterogeneous domain knowledge within a single ontology and explore ways of overcoming those difficulties. Within the presentation we will explore an ontology for a unified traditional Chinese medical language system which is being designed in China. Protégé is being used both to construct the ontology and in knowledge acquisition. We will make comparisons with possible representations for Orthodox Western Medicine. Both Chinese Medicine and Orthodox Western Medicine deal with the human health condition. Much could be gained from effective knowledge sharing. However, despite superficial similarities in terminology, these two disciplines come from very different paradigms. Sowa recognised that no fixed collection of distinctions or categories is likely to be adequate for describing all things for all time ­ even under a single paradigm. We go further by arguing that different flexible sets of distinctions or categories are needed for modelling domain knowledge from different paradigms.



John Gennari :
"
YAX PIP: Yet Another Xml Plug-In for
Protégé":

I will describe the (arduous) process of designing and building a Protégé backend for an XML Schema representation. To build our backend, we made a number of design decisions about the form of the XML Schema. It is still not clear whether we choose the "right" solutions; I would hope that the Protégé community and specifically this workshop will help shed light on the possible uses of an XML output, as well as pros and cons of different designs.



Andreas Gehrmann & Syouhei Ishizu :
"Protege In Modeling Management Systems
":

Our research at Aoyama Gakuin University, Japan is about ontological modeling of generic management system (gms) standards (like ISO 9001, BS 7799-2 and ISO 14001) and their implementation in an individual context of organizations. Protégé is used in building ontologies for representing complex relationship between generic and individual organizational concepts. We developed an upper management ontology by the use of ISO 9000:2000 concepts, so the semantic relationship between key organizational concepts is mapped in to a Protégé model.
Benefits of our approach can be expected in all phases of organizational design, e.g. from organizational design, for group brain-storming in problem solving, integration of IT requirements with other management systems requirement to the evaluation of consistency in models.
Currently we are formulating models with the Object-Process Methodology (OPM). We starting to implement OPM as an ontology, which can be used in general system design. Our presentation would summarize the status of our research and demonstrate how Protégé can be used in connection with OPM for building organizational models.


Jim Hendler :
"
OWL and the National Cancer Institute's Cancer Ontology":

The University of Maryland has been working with the National Cancer Institute's BioInformatics branch to map the cancer metathesaurus into the OWL web ontology langauge. This ontology is now freely available from http://www.mindswap.org/2003/CancerOntology/ and is available for experimentation by Protégé researchers. This talk presents this ontology and explores some challenges for Protégé inherent in this work.


Mike Hobbs:
"
Dynamic Models - A case study in Developing Curriculum Regulation and Conformity using
Protégé":

This paper reports on a case study that seeks to use a document driven information system to validate organisational business rules as they change. Protégé is being used to create an ontology to support student selection of appropriate modules to create valid programmes of study in higher education. This will be linked to a knowledge based system that will implement rules derived from the documentation describing modules and academic rules. The aim is that with the appropriate semantic mark-up and ontological support changes to the documentation describing the rules can be automatically checked by the validation system.
The core questions to be answered by this research are:
Representational flexibility - what kinds of, and degrees of, change can be supported?
Domain coverage - can the informal, implicit rules of thumb of the domain be systematised?
Practical use - can the ontological and semantic resources be used to drive knowledge based systems that provide practical alternatives to static documentation?

This project is being funded by the School of Applied Sciences Research Committee at Anglia Polytechnic University.


Hong-Gee Kim & Kyung-Mo Park :
"
Representing Approximate Concepts Using Metaclass Frames in Protégé":

The frame-based knowledge structure in Protégé is very intuitive thus easily understood and modeled by non-trained ontology users. Reification from metaclass hierarchies in Protégé provides not only more flexible but also more economical means of knowledge representation than traditional frame-based techniques. A metaclass can be considered as a sort of approximate entity in the sense that it does not contain unnecessary details. In this presentation we will propose a way of representing three types of reified relations using metaclass frames in Protégé to model the medical domain knowledge: namely, mereotopological relations, functional relations, and property-value-type relations. We will also address the issue of defeasible reasoning for the propagation of features represented in hierarchies among metaclasses and classes.



Holger Knublauch :
"Editing Semantic Web (OWL) Ontologies with Protégé
":

The vision of the Semantic Web has recently fuelled efforts in the definition of a new ontology modeling language called OWL. This
language is based on Description Logics, which extend Protégé's traditional frame-based modeling approach and allow to describe classes by logical expressions. Using these expressions, reasoning components are able to derive the classification (hierarchy) in an ontology almost automatically. We are currently developing a complex Protégé plugin that will enable users to edit OWL ontologies and access classifiers. This talk will present our goals and plans and demonstrate an initial prototype.



Kai Kumpf :
"A Protein Ontology for Large-scale Textmining
":

We propose a protein ontology partially based on existing biological hierarchies. It will serve as a) a structural backbone for a protein database, the contents of which will iteratively be enhanced by adding results from textmining in publicly available abstracts and b) a reasoning tool to check the plausibilty of the database's contents. Conflicting records (as to function, structure, etc.) will thus either be avoided or can serve to restructure the ontology itself. Although the emphasis is on single protein classification, the ontology makes provisions for the inclusion of interactions, whose validity cannot be reliably checked without knowing the interaction partner's class first.



Sue-sen Lin, Fong-hao Liu & Shion-fu Loe:
"Building A knowledge Base of IEEE/EAI 12207 and CMMI with Ontology
":

To elevate the quality of software , it becomes critical for enterprise to follow the international standard to improve their business rule, and to increase software development capability.Thses are two questions for software development : 1. how to guarantee enterprise devolope software in a correct process and 2. how to inspect the capability of software development. The international standard IEEE/EIA 12207 provides a guide to life cycle process and CMMI can inspect the capability of software Maturity. when
apply these standards to the enterprise, it always take quite a few effort to undestand and use correctly. Therefore we make use of Prot¨¦g¨¦ to define IEEE/EIA 12207 and CMMI ontology, and then integrate theses two knowledges into a knowledge base to help enterprise apply these knoledge to develope software efficiently and correctly .



Feng-How Liu & Shiang Fu Luo:
"
The Study on Ontology Intergtaing and Applying the Ontologies of IEEE/EIA 12207 ,CMMI,Workflow and J2EE to Web Service Development Environment":

As researchers continue to construct ontology in the hope of porblem solving, they still lack consensus on integration of ontologies . We first describe how to integrate the different domain ontology with mergy and allignment operation theoretically.
Then, the Protégé-2000, a tool for ontology development and knowledge acquisition, can be adapted for Web Service Development Environment to obtain the integrated knowledge of the software development standars ,the workflow of organization and the implementation process of J2EE technology.
Third,we use a ontology repository to hold the integrated ontology for resue and to maintain the relations of ontology for integration.
Finally, a inference engin, the Jess, is used to provide implementation suggestion and working template of J2EE technology for building web service development environment.



Raimundo Lozano :
"
Relational Support for Protégé":

As part of the european project SCOPE, scientific information relating to Gastroenterology and Hepatology is structured and tagged conforming an ontology. We use Protégé as ontology edition tool and the Unified Medical Language System (UMLS) as source of concepts. To achieve this, two main developments have been integrated with Protégé through the use of plug-in extensions: a relational storing system for RDF and a relational database for UMLS.



Olga Medvedeva & Rebecca S. Crowley:
"
Protégé as Professor: Development of an Intelligent Tutoring System with Protégé-2000":

This presentation will describe the integration of Protégé-2000 into SlideTutor - an Intelligent Tutoring System in Pathology. SlideTutor requires a flexible and modular knowledge representation to teach problem solving skills and declarative knowledge in a complex visual diagnostic task. Based on the classification problem solving model approach we constructed a set of ontologies for domain and case knowledge representation, and for pedagogic modeling. The problem solving methods are incorporated into the expert model written in Jess, and JessTab provides the transformation of the domain knowledge into Jess facts. We conclude that Protégé-2000 provides a flexible framework for the development of intelligent tutoring systems. Co-Author: Rebecca S. Crowley.



Pericles A. Mitkas & Ioannis Athanasiadis:
"Combining
Protégé with Agent Academy Use for Building Agent-based Software Applications":

The Agent Academy (AA) platform is an integrated high-level abstraction tool that supports the development and design of agent-based applications and provides a mechanism for embedding essential rule-based reasoning into agents. A certain procedure for building agent communities is adopted by AA, starting with the design of application ontologies with Protégé and finishing with the deployment of a multi-agent system. In this talk, Agent Academy approach for building both single-agents and agent-based systems, based on Protégé ontologies, will be discussed. Technical details and a short demo will be presented too.


Mike Pearson :
"Connecting Mathematics
":

On the use of Protégé to create, visualise and publish a multilingual map of mathematical concepts for educational use within schools.


Michael Power, Bob Sugden & Sharon Smart:
"Assuring Safety and Quality in Clinical Practice Guidelines Authored Using Protégé:

The issue of assuring safety and quality in computerised clinical practice guidelines has received little attention in the literature (apart from raising awareness of the problem). We present our "hazop" analysis of potential hazards from computerised clinical guidelines and discuss our plans for minimizing the hazards.

Back to Workshop Schedule


Alberto González Ramírez & Caroll Zuleima Joaqui Trujillo:
"Sielan: Local Area Network Design Expert System":

Thesis Project: "SIELAN: LOCAL AREA NETWORK DESIGN EXPERT SYSTEM".
SIELAN is an expert system developed in Java Language, that use Protégé as knowledge Base editor and JESS as inference engine. SIELAN use fuzzy logic for the representation of the main design variables and decisions . (Prototype)

The Electronics Engineering and Telecommunications Department of the University of Cauca confered a "Honourable Mention" to the SIELAN project developers. This mention only have been confered to very few students in the history of the Department.
For the SIELAN development an Ontology in Protégé was created. This Ontology include the main concepts about the Local Area Network Design: technologies, Structured Wiring, information system, and other components for a medium size desing project.
Besides, for the SIELAN implementation it was created Bfuzzy, a software tool for the creation of fuzzy rules blocks compatible whit JESS and CLIPS. Bfuzzy use an Ontology in Protégé developed for the representation of the main concepts related with Fuzzy Logic.

Back to Workshop Schedule


Alan Rector :
"Why I want a Hybrid Environment: The case for combining Frames and Description Logics":

On the face of it, description logic languages and tools such as OWL/OilEd and frame based languages and tools such as OKBC and Protege appear to be addressing very similar tasks. However, on closer examination their goals and assumptions are more different than it at first seems. Most often, what is required is not just an ontology but an "Knowledge base anchored in an ontology". Description logics can provide the hierarchy of domain categories (classes) based on their intensions but not the knowledge base of contingent facts about the members of those categories nor can they provide the metaknowledge and annotations required to make the ontology usable and maintainable.
Description logics provide a formal semantics for reasoning about universal conditions concerning domain categories based on first order logic. All their statements are of the form: in any possible world, all instances of this category must satisfy this condition. Using such statements, they allow complex polyhierarchies of domain categories to be inferred and managed more easily and reliably than manual methods. These hierarchies are useful precisely because they are independent of any particular database or "world".
However, this is all description logics do. They provide no means either of stating facts about the categories themselves (metaknowledge) or of stating contingent facts concerning the instances of those categories in some particular closed database or "world". To provide intuitive access to metaknowledge, knowledge about instances, reasoning about this a specific world under a 'closed world' hypothesis, or managing defaults and exceptions in a way which allows sensible software engineering, requires supplementation. The traditional frame representations provide a basis for such hybrid reasoning. Some applications may require till further reasoning and computational mechanisms for specific functions but mechanisms such as Protege's plug-and-play structure provide a natural means of extension for such separable but closely coupled mechanisms.
This is not a theoretical problem. Practical applications require a wide range of knowledge, some of it universally true, some of it contingent. They require metaknowledge in several forms - editorial information, documentation for users, and meta-knowledge about the categories themselves - e.g. that they are "well accepted" or "controversial". On the other hand, maintaining a complex polyhierarchy which satisfies the needs of multiple domains so as to be widely re-usable has proved a daunting task without the support provided by the strong semantics of description logics or similar mechanisms.
This talk will explore a series of specific cases in which the hierarchy of domain categories is best managed using DL tools such as OWL, but other knowledge and the "Ontology Anchored Knowledge Base" is based captured in a more frame like way.

Back to Workshop Schedule


Harold R. Solbrig & Christopher G. Chute:
"Using Protégé to Bridge the Terminology/Information Model Boundary":

Terminology describes how information can be classified, while information models describe how information can be represented. The boundary between terminology and information models can be quite fluid, with the same assertion represented as a terminological element in one application and in an information structure in the second. Decisions about where this boundary occurs depend on the primary purpose of the application, what terminology was available at the time the decisions were made and an assortment of other technology and performance constraints.
Applications such as automated guidelines and decision support modules need to represent information in a representational neutral fashion. While they need to be able to accurately and explicitly describe the information needed to drive the logic, the decisions about what is terminology and what is information model must be postponed until the guidelines are applied to an actual information system.
This presentation describes a way that Protégé can be used to blur this distinction by representing terminological constructs to Protégé as imported slots and classes while, at the same time, allowing new classes and slots authored in Protégé to be represented externally as terminological elements.

Back to Workshop Schedule


Margaret-Anne Storey:
"Visualization Requirements for Knowledge-Based Systems":

Visualization can be an effective cognitive aid for managing ontologies and knowledge representations.  First, we will discuss how we have integrated an interactive visualization user interface (Jambalaya) with Protégé.  Then, based on our experiences in this area, as well as results from other related research, we propose some requirements for visualizing ontologies and knowledge bases.    Using these requirements as a guide, we discuss how Jambalaya and other visualization techniques available in Protégé can be applied to ontology authoring and knowledge acquisition.    We welcome suggestions and feedback from participants on their experiences using visualization in Protégé.

Back to Workshop Schedule


Maria Taboada :
"
Experiences in Reusing Knowledge Sources Using Protégé and Prompt. An Application in a Medical Domain":

Current trends in knowledge engineering advise to carry out the knowledge base development by extending domain ontology with specific knowledge of each particular application. In addition, reusing pre-existing knowledge sources may give rise to knowledge bases endowed with a common and standard terminology. But, reuse often requires an enormous effort, so few knowledge-based applications have arisen following this approach.
In this work, we analyze some problems that make reuse difficult, as well as types of activities and current available resources
that may contribute to improve this process. In particular, we present the characteristics of a case study of the development of a KB for a specific medical domain. The purpose of the KB is to support an electronic guideline in ophthalmology taken from the American Academy of Ophthalmology (AAO) Web site (http://www.aao.org/aao/education/library/ppp.cfm). So, we need a sufficiently general methodology for both the modeling of an ontology (the core of our KB) and its extension with specific details.
1) We have used the tool Protégé-2000, which is an ontology design and knowledge acquisition tool.
2) For modeling the core of the KB, we have found the use of more than one ontology very useful. In order to simplify the merging of these ontologies, we have chosen PROMPT for several reasons. Firstly, it allows merging of source ontologies into a resulting ontology. We have found this feature very useful, as development and merging time is reduced, on the contrary to those tools that only provide similarities and differences as a result. Secondly, it analyses source concepts, properties (including restriction on value properties) and relationships. Thirdly, it is interactive with the user, allowing us to accept or reject the suggestions.
3) We have also revised and selected specific concepts from the Unified Medical Language System (UMLS). This server has been developed and is maintained by the U.S.A. National Library of Medicine. The embodiment of these specific concepts in our KB is not only a process consisting of the specialization of the core ontology (i.e., consisting of the adding instances). In general, it is a development process similar to ontology design.


Samson Tu :
"
Workbench and Wizard: Making Protégé Usable as a Workbench for Encoding Clinical Guidelines":

Protégé originated as a knowledge-acquisition tool for a decision-support system that used clinical trials in its knowledge base. Since then Protégé has evolved into a general purpose ontology and knowledge modeling environment. In this presentation, I will outline our most recent effort to make Protégé-2000 into a workbench for encoding clinical guidelines. I will describe the requirements we established, the approach we've taken to model the knowledge-acquisition process, the help system we created, and the libraries of knowledge assets that are needed to make an effective guideline workbench.


C.J. Wroe, C. Puleston, M. Bada, R. Stevens & C. A. Goble :
"Using A Description Logic Approach To Help Maintain And Use The Gene Ontology
":

The Gene Ontology (GO http://www.geneontology.org) is a structured precisely defined common controlled vocabulary for describing the roles of genes and gene products in any organism. It has undergone rapid development and in a short time has become widely adopted in the molecular biology community. Its 14,000 concepts have been used to annotate over 850,000 gene product instances. Success however brings challenges in (i) maintaining complex multiple hierarchies (ii) guarding against a combinatorial explosion of new concepts (iii) making concept definitions machine interpretable. The Gene Ontology Next Generation project (GONG http://gong.man.ac.uk) is a collaboration with members of the Gene Ontology consortium, which is demonstrating how description logic (DL) and in particular the DL language DAML+OIL can help in maintaining GO. By producing formal DAML+OIL definitions for GO concepts, we have shown how DL reasoning can be used to critique the existing GO structure and suggest amendments subsequently accepted by GO editors. We will present the methodology for creating these definitions and specific examples of the results of reasoning.
Producing DAML+OIL definitions requires detailed knowledge of the language, yet we cannot expect domain experts to acquire such knowledge. Therefore GONG is also developing simplified context specific views of the underlying DAML+OIL to encapsulate this knowledge. We will demonstrate how such a view enables domain experts to record definitions in a more straightforward manner.
Many existing GO concepts are the composition of two of more elemental concepts to form a more complex phrase e.g. ‘heparin metabolism’. If a users wants to describe a gene product instance as metabolising some new class of chemical, they must ask the Gene Ontology editors to add that new combination. DL based ontologies provide the capability to build compositions as required. The Gene Ontology Annotation Tool (GOAT http://goat.man.ac.uk) project builds on the work of GONG. The project is developing tools, which allow users to extend the Gene Ontology as required by forming new compositional concepts from smaller building blocks. We will demonstrate how these tools may be used.

 

Back to Workshop Schedule

Back to Workshop Home Page

Last Updated: 07/03/03 by webmaster@smi.stanford.edu
Copyright © 2003 Stanford Medical Informatics. All rights reserved.