Knowledge Graphs - Working Group Charter (NFDI section-metadata)

Section (Meta)data, Terminology, Provenance Working Group Charter Knowledge Graphs Name of the working group Knowledge Graphs Acronym section-metadata-wg-kg Contact persons Renat Shigapov (Universitätsbibliothek Mannheim, BERD@NFDI), [email protected] Lozana Rossenova (TIB – Leibniz-Informationszentrum Technik und Naturwissenschaften und Universitätsbibliothek, NFDI4Culture), [email protected] Moritz Schubotz (FIZ Karlsruhe - Leibniz Institute for Information Infrastructure, MaRDI), [email protected] Authors Markus Stocker (0000-0001-5492-3212), Lozana Rossenova (0000-0002-51901867), Renat Shigapov (0000-0002-0331-2558), Noemi Betancort (0000-0002-01563556), Stefan Dietze (0009-0001-4364-9243), Bridget Murphy (0000-0002-13542381), Christian Bölling (0000-0002-6544-1363), Moritz Schubotz (0000-0001-71414997), Oliver Koepler (0000-0003-3385-4232), Andreas Czerniak (0000-0003-38834169), Fidan Limani (0000-0002-5835-2784) Version 1.3 DOI: https://doi.org/10.5281/zenodo.17748725 2025-11-28 Abstract Knowledge Graphs are a key technology for implementing the FAIR principles in data infrastructures by ensuring interoperability for both humans and machines. The Working Group "Knowledge Graphs" in Section "(Meta)data, Terminologies, Provenance" of the German National Research Data Infrastructure (Nationale Forschungsdateninfrastruktur (NFDI) e.V.) aims to promote the use of knowledge graphs in all NFDI consortia, to facilitate cross-domain data interlinking and federation following the FAIR principles, and to contribute to the joint development of tools and technologies that enable transformation of structured and unstructured data into semantically reusable knowledge across different domains. 1. Motivation Knowledge Graphs (KG) are a widely used database technology for the flexible description of entities and their linking enabling advanced information modeling, exploration, and use. Since the establishment of the Google Knowledge Graph1, these technologies have gained momentum in industry and research, where KGs are used to organize and publish information about entities like articles, datasets, software, samples, instruments, contributors, organizations, projects, funders and their interrelations in a machine-readable way. Projects around these entities already exist and provide services to users from different communities. OpenAIRE Research Graph2, GESIS Knowledge Graph Infrastructure3, TIB Open Research Knowledge Graph4, Springer Nature SciGraph5, Research Graph Foundation6, the GND.network7, and PID Graph8 are such established infrastructures that connect research data and research-related entities and make it accessible. Additionally, the Wikidata knowledge graph9 contains a large number of research-related entities10 and several domain specific Wikibase11 instances are used by various research communities12. 1 https://blog.google/products/search/introducing-knowledge-graph-things-not 2 https://graph.openaire.eu 3 https://www.gesis.org/forschung/angewandte-informatik/knowledge-graph-infrastruktur 4 https://orkg.org 5 https://www.springernature.com/gp/researchers/scigraph 6 https://researchgraph.org 7 https://gnd.network 8 https://commons.datacite.org 9 https://www.wikidata.org 10 https://scholia.toolforge.org 11 https://wikiba.se 12 https://portal.mardi4nfdi.de/wiki/Links#Other_wikibase_instances Individual NFDI consortia develop services for the discovery of their assets, e.g., datasets, services, organizations, etc. For the research community served by a particular NFDI consortium, these services act as the primary entry point. However, given the high heterogeneity of resources and their interdisciplinary (re)use (e.g., datasets, methods or models), adopting the FAIR principles13 for the metadata of the resources across individual disciplines and their dependencies is crucial and lays the foundation for multiand crossdisciplinary discovery. The widespread use of computational methods (e.g., NLP or ML models) across various disciplines and their dependencies on data resources is an example where having such a discovery is required. This leads to a genuine need for FAIR and machine-readable representations of resources and their dependencies across the NFDI. Some NFDI consortia are already developing the KGs to support their data infrastructures, such as the NFDI4Culture KG14 and the MaRDI Wikibase KG15. KGs are a key technology for the implementation of the FAIR principles by ensuring interoperability for both machine and human access through the use of a common data model, adoption of established schemas and vocabularies, and reliance on Persistent Identifiers (PIDs)16 and data management policies across the entire data space. The motivation of this working group is two-fold: 1. Promote the use of KGs in all NFDI consortia and facilitate cross-domain data interlinking and federation following the FAIR Principles; and 2. Contribute to the joint development of tools and technologies that enable transformation of structured and unstructured data into semantically reusable knowledge across different domains represented by NFDI consortia. 2. Objectives ● Enable transdisciplinary connection between research data assets from different consortia via an overarching KG approach. 13 Wilkinson, M. D., et al. (2016). The FAIR Guiding Principles for scientific data management and stewardship. In Scientific Data (Vol. 3, Issue 1). Springer Science and Business Media LLC. 14 https://nfdi4culture.de/resources/knowledge-graph.html 15 https://portal.mardi4nfdi.de/wiki/Portal 16 We acknowledge that the term ‘PID’ is currently used to reference a wide range of resource identification approaches. We will work with other relevant NFDI WGs to come to a common understanding of the term. Furthermore, the work of relevant external communities such as the PID IG of the RDA will also be taken into consideration. ● Conduct an assessment of the current state of KG technology implementation across the different NFDI projects and identify the possible factors blocking wider adoption of KGs within individual consortia. ● Agree on an approach to support a single point of access for federated information of research assets in the NFDI, across all NFDI initiatives. ● Collaborate closely with the WG Ontology Mapping and Harmonization17 to establish common identifiers and connecting properties required to facilitate a unified access to heterogeneous resources across the different consortia and domains. ● Demonstrate how individual consortia can adopt KGs to make their data (and other structured data resources) interoperable with other consortia (via federated access) in practice (cf. WG Cookbook(s), Guidance and Best Practices18). ● Reach out and collaborate with open source communities developing and maintaining KG technologies to ensure that particular requirements of the NFDI research infrastructure are taken into consideration and necessary steps are taken towards implementing additional software customizations and extensions as required. ● Define the core requirements for a flexible and easy-to-use entry point solution for adding data to the overall knowledge graph. 3. Work Plan Q4 2022: Preparing to evaluate the state of the art ● Define criteria to assess the current state of KG technology adoption within the NFDI consortia. Prepare questions for a survey. ● Evaluate the use of existing integration and federation approaches (e.g., the PID Graph and its GraphQL-based implementation by DataCite19, or crosslinking via external identifiers in Wikidata). Q4 2022 - Q4 2024: Communication and collaboration activities ● Provide regular updates about WG activities at the Section Metadata meetings. ● Conduct knowledge sharing in NFDI WGs and consortia, and relevant external working groups.20 ● Communication and collaboration with KG tool maintainer groups (e.g. open source communities around Wikidata, Wikibase, among others) and other stakeholders. 17 See WG Charter: https://zenodo.org/record/6726519 18 See WG Charter: https://zenodo.org/record/6758256 19 This is an outcome of the FREYA H2020 Project. 20 See the section: “Connected to other working groups outside NFDI” further in this document. ● Prepare for and participate in NFDI Conference 2023, EOSC Conference 2024 and ongoing Section Workshops. Q1 2022 - Q4 2024: Cross-consortial exchange on KG topics ● Compile an overview of KG implementations in NFDI consortia and their respective domains. ● Organize short presentations on best practices and ongoing challenges from individual consortia during monthly meetings. ● Identify best practices and/or factors blocking the implementation of KG technology for individual consortia so that lessons learned can inform the cross-domain work as well. Q1 2023 - Q4 2024: Interaction with Base services ● Define potential NFDI Base Services21 based on the results of the preparation and communication work. ● Coordinate closely with the WG Ontology Mapping and Harmonization as well as the NFDI Base Services teams in the implementation of a pilot cross-domain KG access point. This will involve any necessary data integration and/or schema linking activities not covered by other WGs. ● Coordinate closely with the WG Terminology Services22 as well as Base4NFDI consortium and its respective task areas teams in the further development and improvement of KG-supporting infrastructure (including tools for data modeling and storage, terminology mapping, and application profile definition) in use by individual consortia. Q2 2023 - Q4 2024: Dissemination and documentation of outcomes ● Publication of a paper with an overview of knowledge graphs in NFDI. ● Talks, tutorials and workshops to WGs, NFDI consortia, and other groups of interests. Specific tasks ● Initiate regular bi-monthly and, at a later point, monthly meetings for the WG. ● Evaluate existing KG tools and technologies in use within the NFDI consortia (and beyond). ● Assess existing approaches to achieve transdisciplinary data discovery. ● Identify and prioritize concrete needs for cross-domain KG access, by sub-working groups involving all consortia. 21 See joint statement on basic services: https://zenodo.org/record/6091657#.YgvB8t8xk2w 22 See WG Charter: https://zenodo.org/record/6759325 ● Provide recommendations to individual consortia on the use of specific KG infrastructure setups, as well as recommendations for the overarching integrated KG approach. ● Provide example implementations of both domain-specific and cross-domain KG infrastructure setups. ● Collaborate with other relevant WGs to deliver the identifier mappings and common properties that can facilitate integrated access to a single discovery endpoint for all NFDI. 4. Membership List Members are those who participated in the regular meetings of the WG KGs at least once. Name Institution NFDI consortia or NFDI offices 1 Lozana Rossenova TIB NFDI4Culture 2 Markus Stocker TIB NFDI4DataScience, NFDI4Earth 3 Renat Shigapov UB Mannheim BERD@NFDI 4 Noemi Betancort Uni Bremen KonsortSWD 5 Felix Engel TIB NFDI4Ing 6 Moritz Schubotz FIZ MaRDI 7 Rebecca Wilm IDS Text+ 8 Heinrich Widmann DKRZ NFDI4Earth 9 Claus Weiland Senckenberg NFDI4Earth, FAIRagro 10 Katrin Moeller MLU Halle-Wittenberg NFDI4Memory 11 Sonja Schimmler Fraunhofer FOKUS NFDI4DataScience, NFDI4Cat 12 Jakob Voß VZG NFDI4Objects 13 Ziyad ZB MED NFDI4Microbiota 14 Jürgen Kett DNB Text+, NFDI4Culture 15 Konrad Förstner ZB MED NFDI4Microbiota 16 Stefan Dietze GESIS, HHU NFDI4DataScience, BERD@NFDI 17 Cord Wiljes NFDI-GS 18 Marie Annisius DNB Text+ 19 Bridget Murphy CAU DAPHNE4NFDI 20 Sven Nahnsen EKUT GHGA DataPlant 21 Christian Bölling MfN Berlin NFDI4Biodiversity 22 Gerhard Heyer SAW Leipzig Text+ 23 Thomas Eckart SAW Leipzig Text+ 24 Daniel Mietchen FIZ MaRDI 25 Benjamin Zapilko GESIS NFDI4DataScience 26 Fidan Limani ZBW KonsortSWD, NFDI4DataScience, BERD@NFDI 27 Felix Helfer SAW Leipzig Text+ 28 Erik Körner SAW Leipzig Text+ 29 Jürgen Diet Bayerische Staatsbibliothek NFDI4Culture 30 Nils Geißler CCeH Text+ 31 Tobias Arera-Rütenik KDWT Uni-Bamberg NFDI4Objects 32 Matthäus Zloch GESIS NFDI4DataScience 33 Saurav Karmakar GESIS NFDI4DataScience 34 David Linke LIKAT NFDI4Cat 35 Ricardo Usbeck Uni Hamburg NFDI4DataScience 36 Andreas Frech LMU NFDI4Culture 37 Sirko Schindler DLR NFDI4Earth 38 Tilahun Abedissa Taffa Uni Hamburg NFDI4DataScience 39 Dorothea Iglezakis Uni Stuttgart NFDI4Ing, MaRDI 40 Giacomo Lanza PTB NFDI4Ing, NFDI4Chem 41 Vera Clemens ZB MED NFDI4Health 42 Jonas Grieb Senckenberg NFDI4Earth, FAIRagro 43 Jan-Karl Haug DLR NFDI4Earth 44 Julia Sasse ZB MED NFDI4Health 45 Christian Langenbach DLR NFDI4Ing 46 Thorsten Trippel Uni Tübingen Text+ 47 Bianca Wentzel Fraunhofer FOKUS NFDI4DataScience, NFDI4Cat 48 Canan Hastik TU Darmstadt NFDI4Ing 49 Matthias Löbe IMISE U Leipzig NFDI4Health 50 Andreas Czerniak Bielefeld University Library 51 Victoria Tokareva KIT PUNCH4NFDI 52 Atif Latif ZBW BERD@NFDI 53 Auriol Degbelo TU Dresden NFDI4Earth 54 Norbert Riefler IWT Bremen NFDI-MatWerk 55 Ludwig Hülk RLI NFDI4Energy 56 Philipp Ost KIT NFDI4Ing 57 Anja Gerber Klassik Stiftung Weimar NFDI4Objects 58 Olaf Simons Gotha Research Centre NFDI4Memory 59 Andreas Christ CAU Kiel 60 Mattis thor Straten CAU Kiel NFDI4Objects 61 Kanishka Silva GESIS NFDI4DataScience 5. Adoption Plan Once the assessment of the current state of KG adoption across the consortia is complete and concrete requirements for a cross-domain KG access point have been specified, feedback to the consortia and the scientific communities they represent is going to be crucial for the success of the WG. Given the NFDI’s bottom-up approach, the Section Metadata and the WG KGs are the forums to discuss the best KG practices and to exchange knowledge about KGs across all NFDI domains. The WG KGs is going to support the consortia in deciding which KG services and related data integration strategies should be used and how they should be modified and adopted. Furthermore, the WG will provide the expertise required in specifying the NFDI Base Services concerning KGs and, subsequently, provide support to the teams developing them. Adoption plans for individual cases of KG implementation (including the cross-domain access point) may differ, depending on the state of the art within each domain. They are going to be agreed on a case-by-case basis within each consortium and with guidance from this WG as required. Represented Consortia 1. NFDI4Culture 2. BERD@NFDI 3. NFDI4DataScience 4. NFDI4Earth 5. KonsortSWD 6. NFDI4Ing 7. MaRDI 8. Text+ 9. NFDI4Chem 10. NFDI4Memory 11. NFDI4Cat 12. NFDI4Objects 13. NFDI4Microbiota 14. DAPHNE4NFDI 15. GHGA 16. DataPlant 17. NFDI4Health 18. FAIRagro 19. PUNCH4NFDI 20. NFDI4Biodiversity 21. NFDI-MatWerk 22. NFDI4Energy Connections to other NFDI WGs 1. WG “Ontology Mapping and Harmonization” of Section “(Meta)data, Terminologies, Provenance” 2. WG “Terminology Services” of Section “(Meta)data, Terminologies, Provenance” 3. WG “Search and Harvesting” of Section “(Meta)data, Terminologies, Provenance” 4. WG “Cookbooks, Guidance, Best Practices” of Section “(Meta)data, Terminologies, Provenance”