SEKT will provide the next generation of knowledge management tools, based on semantic technology. These tools are being incorporated into prototypes in three application areas: intelligent content management; knowledge reuse; and decision support within the legal sector. In addition, these tools will be provided with APIs to enable their incorporation into a wide range of applications, such as business intelligence and customer relationship management.

The goal of SEKT is to use semantic technology to create an information-centric approach to knowledge management, rather than today's document-centric approach. This means using SEKT's technologies to extract information from unstructured text, rather than simply searching for text strings to identify documents which may contain items of relevance. Information from heterogeneous sources is merged and presented to the user in the form most suited to his or her requirements, e.g. as natural language.

Underlying all the work of SEKT is the use of ontologies. An ontology is a way of representing knowledge such that it can be shared and also reasoned about formally. Once created, an ontology can be used to describe documents, e.g. web pages, and also entities within documents. SEKT also uses a knowledge base, associated with a particular ontology. Thus an ontology may contain generalisations about companies, people and roles within companies, whilst the associated knowledge base will contain information about specific companies, people and their roles.

Previously, the hurdle to this approach was the high degree of overhead required to generate and maintain the ontology and the knowledge base, and to use these to annotate documents. A key innovation of SEKT is to use techniques from knowledge discovery and human language technology to semi-automate these processes. Techniques for the management and evolution of ontologies are also a key research area for SEKT.

It is likely that distributed knowledge bases will contain inconsistencies. SEKT's advanced features are dependent on the use of reasoning. Yet such inconsistencies pose a severe problem for classical reasoning algorithms. SEKT is investigating ways to reason in the presence of inconsistencies, e.g. by identifying a consistent subset of a knowledge base.

Complementing its research into the core technologies, SEKT is developing knowledge access tools which will make full use of the power of these technologies, e.g. by using SEKT's reasoning capability to enable complex queries. On the other hand, users will have the option of a simple interface, but still exploiting SEKT's sophisticated technologies.

SEKT is also investigating the business and human aspects of using the new technology. A methodology is being developed to support the deployment of an ontology in an organisation. User requirements capture and usability testing are being used to ensure that the interface to the user is appropriate to his needs.

During its first two years SEKT has produced a significant number of research results and software deliverables. The latter includes software for:

• automatic ontology generation
• ontology-based information extraction
• extracting concepts from text in order to update an ontology
• automatically suggesting a mapping between two different ontologies

In addition, a distributed and extensible integrated platform, SIP (SEKT Integration Platform) has been developed to enable components to be integrated into an end-user system.

In the first year, each of the three case studies has undertaken a comprehensive user requirements analysis. During 2005 the case studies have developed prototypes, using the SEKT components as appropriate.

For further information, please visit:

• http://www.sekt-project.com
• http://cordis.europa.eu/ist/kct/sekt_synopsis.htm