D-Lib Magazine

November/December 2011
Volume 17, Number 11/12

Report on the 10th European Networked Knowledge Organization Systems/Services (NKOS) Workshop

Philipp Mayr
GESIS — Leibniz Institute for the Social Sciences
philipp.mayr@gesis.org

doi:10.1045/november2011-mayr

Abstract

The 10th European Networked Knowledge Organization Systems/Services (NKOS) Workshop took place from September 28-29, 2011, in Berlin, Germany, as part of the International Conference on Theory and Practice of Digital Libraries (TPDL 2011, formerly ECDL). NKOS is a community of researchers, developers and practitioners seeking to enable knowledge organization systems as networked interactive terminology and knowledge organization services via the Internet. The main topics of the workshop were "SKOS and Linked Data" and "Relation between Ontologies and (other) KOS".

Introduction

The 10th European Networked Knowledge Organization Systems/Services (NKOS) Workshop, organized by Traugott Koch, Philipp Mayr, Douglas Tudhope and Marianne Lykke, took place from September 28-29, 2011, in Berlin, Germany, as part of the International Conference on Theory and Practice of Digital Libraries (TPDL 2011, formerly ECDL). The workshop was attended by 35 colleagues from 17 countries on 4 continents. All presentations and abstracts are available on the workshop website. NKOS is a community of researchers, developers and practitioners seeking to enable knowledge organization systems (KOS) (e.g., classifications, gazetteers, lexical databases, ontologies, taxonomies and thesauri) as networked interactive terminology and knowledge organization services via the Internet.

The workshop at TPDL 2011 builds on the well-attended NKOS workshops at ECDL 2000, 2003, 2004, 2005, 2006, 2007, 2008, 2009 and 2010. This year, the workshop covered topics spread over the main topics in the Call for Papers. The following interrelated topics, "SKOS and linked data"; "Relation between KOS and ontologies"; and "From KOS to formal ontologies and back? Repurposing and reengineering of KOS", were focused on in the call for papers.

Traugott Koch started the workshop with a topical retrospection to the first European NKOS workshop at ECDL 2000 in Lisbon. Three of the NKOS 2000 workshop organizers attended the 2011 workshop. Ontologies already played a role in the first NKOS workshop in 2000 (see e.g. Bechhofer's talk "OIL: The Ontology Inference Layer"). Mutual introductions of all participants and their areas of interest followed.

Sessions

"SKOS / Linked Data", was the theme of the first (and most extensive) session, which was chaired by Traugott Koch.

Imma Subirats, FAO, Italy, reported about the FAO VocBench, an editing and management framework for thesauri and concept schemes. The FAO VocBench is an integral part of the current FAO Linked Open Data activities, with the multilingual AGROVOC thesaurus as a core vocabulary. Today the AGROVOC has been aligned with seven other thesauri. In addition, the presentation showed the re-modeling of AGROVOC from OWL back to RDF SKOS/SKOS-XL.

Christian Mader, University of Vienna, Austria, presented a short paper about typical quality aspects for KOS and how these criteria (identification of loose concepts, cyclic relations in hierarchical relations, weakly connected components, ambiguous labeling, Linked Data availability) can be best controlled automatically. The authors proposed a set of criteria for computing quality metrics on a given input vocabulary and showed preliminary results from an analysis of a collection of very heterogeneous SKOS vocabularies.

Benjamin Zapilko, GESIS, Germany, showed how thesauri from different domains (TheSoz and AGROVOC) can be matched automatically and which matching approaches are most promising for this task. He showed preliminary results of an initial matching algorithm (applied Levenshtein distance). Further research will be the comparison with some existing matching tools (i.e. FALCON-AO, COMA++, Silk).

Johan De Smedt, TenForce, Belgium, reported about a project that develops the ESCO taxonomy, the European Standard Classification of Occupations that links occupations to skills and qualifications. The paper gave specific examples related to ESCO to illustrate the need for extending the SKOS mapping properties and proposed requirements to the SKOS mapping as described in ISO 25964 Part 2 (e.g. the skos:narrowMatch property and compound mappings).

Mark Tomko, Simmons College, USA, presented an approach to translate biological data sets (e.g. Pfam and UniProt) into Linked Data via the SKOS vocabulary. His proof-of-concept was to represent protein sequences as a SKOS graph in order to improve the ability of researchers in the biological sciences to share and process scientific data.

At the end of the first session the workshop encouraged participants to give lightning talks. The most interesting talk was by Martin Doerr who suggested the workshop attendees think about the various misuses of the widely adopted SKOS vocabulary. In his view, too many misuses or misapplied SKOS and LOD utilizations are online today. A rethinking about the proper use and the limitations of SKOS should be undertaken by the community.

"Ontology special session: Relation between Ontologies and (other) KOS", was the theme of the morning session of the second workshop day, which was chaired by Douglas Tudhope.

Daniel Kless, University of Melbourne, Australia, talked about similarities and difference between thesauri and ontologies. A focus of his talk was the relata, the entities related by relationships in thesauri and ontologies. Some of his results are that hierarchical part-of relations in thesauri match with transitive ontological part-of relations and thesaurus associations generally match with intransitive ontological relations.

Antoine Isaac, Vrije Universiteit Amsterdam, Netherlands, presented a collaborative work in the domain of semantic alignments, in the context of the Europeana network. He talked about multilingual issues and strategy finding in alignment evaluations. Amalgame, a mapping tool for linked data resources, was introduced as a possible part of a solution.

Douglas Tudhope, University of Glamorgan, United Kingdom, discussed the relationship between formal ontologies and thesauri (and related KOS). He brought up examples of key differences between thesauri and ontologies from two current projects, STAR and STELLAR. The projects aimed to provide a degree of semantic interoperability between diverse archaeological datasets and archaeological reports, extracted from the OASIS grey literature library.

"Further topics", was the theme of the last session, which was chaired by Philipp Mayr.

Wilko van Hoek, GESIS, Germany, presented an experimental approach to estimate the quality of search term recommendations derived from the Normalized Google Distance (NGD) compared with a traditional pLSA implementation. NGD was used to calculate pairwise the semantic similarity of random search terms and terms from a thesaurus (the Thesaurus for the Social Sciences, TheSoz). The preliminary results showed that NGD was not a very accurate measure in this use case.

Denise Bedford, Kent State University, USA, proposed a new and sustainable approach to social tagging in a knowledge architecture. The idea was to enable end users to select tags from a semantically generated cloud of concepts. She found that: a) in 90% of the time, the semantic engine when powered by a KOS will promote the core topical term, b) the approach, when powered only by a topic-focused KOS, will promote 45% of all the terms suggested by end users.

Stella Dextre Clarke and Johan De Smedt, United Kingdom / Belgium, talked about the ISO 25964-1, the newly published ISO standard and the XML schema that was developed for exchanging thesaurus data. The presentation discusses how the XML schema (in comparison to SKOS) deals with several thesaurus aspects, including some custom extensions on the basic formal framework.

After each session, the workshop had extra time for lengthy and highly intensive discussions, with broad participation from the audience.

Additional Information

The main forum for all related activities is the international NKOS discussion list. Information and documentation about (past and future) events and publications, how to subscribe to the list and to share the work, is available at the international NKOS website.

Further information on this 10th European NKOS workshop along with links to all presentations can be found on the workshop website at http://www.comp.glam.ac.uk/pages/research/hypermedia/nkos/nkos2011/.

About the Author

Philipp Mayr is a postdoctoral researcher and project lead at the GESIS department Knowledge Technologies for the Social Sciences (WTS). Since October 2009 he has served as a visiting professor for knowledge representation at University of Applied Sciences in Darmstadt, Department of Information Science & Engineering. Dr. Mayr is a graduate of the Berlin School of Library and Information Science at Humboldt University Berlin where he finished his doctoral research in 2009. He studied LIS, computer science and sociology. Since November 2004 he has been working in the internationally recognized projects "Competence Center Modelling and Treatment of Heterogeneity" (KoMoHe) and "Value-Added Services for Information Retrieval" (IRM) as a researcher and PI. He has presented at several NKOS workshops and is a member of the NKOS network. He has published in the areas Informetrics, Information Retrieval and Digital Libraries, and is a member of the editorial board of the journals Scientometrics and Information Wissenschaft & Praxis. He serves as a reviewer for various journals and international programme committees. His research interests are non-textual ranking in digital libraries, bibliometric methods, evaluation of information systems, and applied informetrics.