Knowledge Organization Systems and Semantic Technology

The rapidly increasing amount of information in every knowledge domain poses challenges to the retrieval, integration, and reuse of information relevant to a specific context. To cope with this situation, methods of knowledge modelling and representation play an increasingly important role. Knowledge organisation systems (KOS’s), a term summarizing knowledge structures such as controlled vocabularies, taxonomies, thesauri and ontologies, provide the basis for describing (possibly quite complex) knowledge domains. For representing KOS’s, standards such as SKOS (Simple Knowledge Organisation System) have been developed by the World Wide Web Consortium (W3C). Standardization allows creation and maintenance of KOS over their lifecycle, enables linking between KOS’s, and allows publication of KOS’s on different media on the Web as well as in traditional document formats. By expressing KOS’s with standards, the organisation systems may now be interpreted by machines, while still being comprehensible by humans.

For managing knowledge organization systems which follow W3C standards, a tool has been acquired which allows to create KOS’s, transform existing ones into the SKOS format and import them, edit KOS’s, and publish them on several media, most prominently the Web. In addition to the basic functionality of managing KOS’s, the tool offers features for interlinking of KOS concepts, and linking the concepts to external data. This opens up the world of Linked Data, the Internet of data in which data are treated in a similar way to documents. Obeying the naming conventions of each single term by URI’s (Unified Resource Identifiers), data may be interlinked with other, forming a knowledge network. The number of information sources which are published in this way grows continuously (see figure). This allows the construction of extensive knowledge graphs as e.g. used by Google to enable “discovery” of knowledge, meaning that a search will return not only results by full text search, but also related items. For representing knowledge domains, this technology, often referred to as “semantic technology”, allows the generation of content-rich knowledge models, providing the user with a navigable knowledge map which may be queried with a specific query language (SPARQL), and which are often supported by visualization and refined semantic search tools.

In the NKM Section, KOS’s have already been in use for a fairly long time, e.g. taxonomies for describing Nuclear Accidents, Fast Reactors and VVER’s. These legacy KOS’s have been imported into the KOS management tool. In addition, KOS’s developed in several IAEA departments have been imported and interlinked, most notably the International Nuclear Information System (INIS) in a preliminary, reduced version. Within the frame of the International Terminology Repository project (see also “Applying Semantic Technology in Nuclear”), KOS’s from other nuclear organizations have been converted to SKOS and imported. As KOS’s are increasingly deployed in describing knowledge domains and for the development of knowledge based applications, the number of KOS’s will continuously rise in future.

For more information please contact the Scientific Secretary M.Gladyshev at: NKM-Contact