OnToKnowledge: Ontology-based Tools for Knowledge Management.
Dieter Fensel, Frank van Harmelen, Jos Akkermans, Michel Klein, Jeen Broekstra, Christiaan Fluit, van der Meer, J., Hans-Peter Schnurr, Rudi Studer, John Davies, Hughes, J., U. Krohn, Robert Engels, B. Bremdahl, Fredrik Ygge, Ulrich Reimer, Ian Horrocks
On-To-Knowledge (EU-IST project No. 10132) builds an ontology-based tool environment to speed up knowledge management, dealing with the large numbers of heterogeneous, distributed, and semi-structured documents typically found in large company intranets and the World-Wide Web. Results include a toolset for semantic information processing, OIL (an ontology-based inference layer), an associated methodology, and industrial validation.
Introduction: The Challenge of Information Overload
We are currently drowning in information but starving for knowledge. Traditional keyword-based search engines often fail due to homonyms (returning irrelevant results) and synonyms (missing relevant ones). For large organizations, this "information overload" is a productivity killer. Technical specifications, CVs, and reports are often distributed and weakly organized. The OnToKnowledge project addresses this by providing Semantic Access: a way for systems to process the meaning of online documents automatically using ontologies—formal specifications of a shared understanding of a domain.
Methods and Architecture
The project proposes a multi-layered architecture centered around ontologies to facilitate knowledge sharing and reuse.
The Technical Backbone: OIL
The core contribution is the Ontology-based Inference Layer (OIL). OIL unifies three distinct paradigms:
Description Logics (DL): For formal semantics and efficient reasoning.
Frame-based Systems: For modeling primitives (classes and attributes) familiar to engineers.
Web Standards (XML/RDF): For portability and web compatibility.
The OnToKnowledge system architecture, illustrating the flow from user queries through the QL engine to the OIL ontology and annotated data repositories.
Spectacle/WebMaster: For generating semantic sitemaps and visualizations.
SilRI: An RDF query engine.
Corporum: For extracting concepts from free text to automate annotation.
Industrial Case Studies
The framework was evaluated through three major industrial applications:
Swiss Life: Managing organizational memory, including skills databases and insurance documentation.
British Telecom (BT): Enhancing call center efficiency by linking technical specs and customer data.
EnerSearch: A virtual organization sharing R&D results in the energy sector.
Key Results: Visualizing Knowledge
A major achievement of the project is "shielding the user" from document-specific formatting. By querying through an ontology, users communicate in domain-specific terms. Furthermore, the project utilizes spatial proximity in visualizations, where the physical distance between nodes in a "semantic map" correlates to their conceptual relationship.
Semantic structure maps generated for the EnerSearch website, showing clusters of related research topics like 'Electronic Commerce' and 'Agents'.
The framework successfully integrated structured databases with unstructured free-text documents under a single ontological umbrella, using the FaCT reasoning engine to automatically derive classification taxonomies.
Limitations and Future Outlook
Despite its successes, the "annotation bottleneck" remains a challenge; transitioning raw text to rich semantic data requires significant effort. Additionally, scaling reasoning to millions of RDF triples in real-time is an ongoing engineering task. Future work focuses on refining the business case methodology to help organizations realize the ROI of moving to ontology-based management.
Summary updated
Abstract
On-To-Knowledge, the European EU-IST project No. 10132, builds an ontology-based tool environment to speed up knowledge management, dealing with the large numbers of heterogeneous, distributed, and semi-structured documents typically found in large company intranets and the World-Wide Web.Results aimed for by the project are: (1) a toolset for semantic information processing and user access; (2) OIL, an ontologybased inference layer on top of the World-Wide Web; (3) an associated methodology; (4) validation by industrial case studies.This paper gives an overview of the On-To-Knowledge approach to knowledge management. Heterogeneous Information Resources Need Semantic AccessSupport for information and knowledge exchange is a key issue in the Information Society.The exponential growth of online information on intranets and the Web leads to information overload.To cut down on the time wasted in searching and browsing, and reduce associated user frustration, much more selective user access is needed.This is possible by automatic meaning-directed or semantic information processing of online documents.The European IST project On-To-Knowledge builds tools that achieve this.Ontologies [1,2] are the key technology used to describe the semantics of information exchange.Defined as "specifications of a shared conceptualization of a particular domain", they provide a shared and common understanding of a domain that can be communicated across people and application systems, and thus facilitate knowledge sharing and reuse.Ontologies will play a key role in growth areas such as knowledge management [3,4] and electronic commerce.In the US, funding agencies have recognized the importance of these issues by setting up the DAML program (http://www.darpa.mil/iso/ABC/BAA0007PIP.htm) that aims at machine-processable semantics of information resources accessible to agents.The World-Wide Web (WWW) has drastically boosted the availability of electronic information.Already the present first generation of the web has changed our daily practice, and these changes will become even more significant in the near future.However, the Web itself has to change if it is to reach the next level of user service [5].Currently, the Web is
BibTeX
@inproceedings{fensel2000ontoknowledge,
author = {Dieter Fensel and Frank van Harmelen and Jos Akkermans and Michel Klein and Jeen Broekstra and Christiaan Fluit and van der Meer, J. and Hans-Peter Schnurr and Rudi Studer and John Davies and Hughes, J. and U. Krohn and Robert Engels and B. Bremdahl and Fredrik Ygge and Ulrich Reimer and Ian Horrocks},
title = {{OnToKnowledge: Ontology-based Tools for Knowledge Management.}},
booktitle = {Digital},
year = {2000},
}