Read this abbreviated version of a presentation from the Ontotext and Partners Knowledge Graph Forum 2021 titled "Semantic Knowledge Organization Systems” by David Clarke, CEO…
Knowledge graph applications are quickly and continuously growing in number and variety.
In this post, you will learn about how our RDF database for knowledge graphs GraphDB was used across different domains and later cited in several scientific papers. You will also have a glimpse of where research effort is headed in the vast territory of knowledge representation and data management.
The first paper we want to highlight is Towards a Knowledge Graph-specific Definition of Digital Transformation: An Account Networking View for Auditing by Florina Livia Covaci, Robert Andrei Buchmann and Radu Dragos.
The paper presents an experimental Digital Transformation project, which aims to improve the auditing capabilities of a legacy accounting system. Enterprise Resource Planning systems that are based on traditional relational databases struggle to aggregate data from disparate sources, represent multihop relationships, etc.
To address these issues, the authors chose a knowledge graph based approach, adopting an RDF graph on top of existing ledger data. They populated a GraphDB instance through the OntoRefine plug-in (now Ontotext Refine) for turning tabular data into RDF graphs. After that, they employed semantic queries and SPARQL-based reasoning patterns to build a network of accounting relationships and interactions, which was easy to navigate and analyze.
With its ability to put data in context via linking and metadata, the RDF graph captures the rich accounts relationships, including some of the implicit knowledge that previously resided only in the heads of the accountants and the auditors. It also applies inference to mimic the reasoning and data navigation patterns of an accounting professional. Mapping accounting information to an RDF graph improves data analysis and makes it easy to locate errors or investigate patterns. Click To Tweet
The paper concludes that the project’s strengths lie in defining “a method of mapping existing accounting transactions and accountant’s knowledge to semantic query patterns based on RDF graphs and the RDF-star extension”. It sees further opportunities for knowledge graphs in the Financial Industry to become “a key enabler for cross-organization auditing if such technology is adopted at large”.
The second paper we would like to talk about is Applying the FAIR principles to data in a hospital: challenges and opportunities in a pandemic by Núria Queralt-Rosinach, Rajaram Kaliyaperumal, César H. Bernabé, Qinqin Long, Simone A. Joosten, Henk Jan van der Wijk, Erik L.A. Flikkenschild, Kees Burger, Annika Jacobsen, Barend Mons, Marco Roos, BEAT-COVID Group & COVID-19 LUMC Group, published in Journal of Biomedical Semantics, 25 April 2022.
This paper focuses on the problems for Healthcare and research organizations worldwide in collecting data and making it available to researchers quickly. The COVID-19 pandemic made it clear that it was challenging to prepare patient data for efficient analysis across multiple hospitals in multiple countries. Click To Tweet
To deal with these challenges, the authors present an approach that makes COVID-19 observational patient data more FAIR (Findable, Accessible, Interoperable and Reusable for both humans and machines). This approach was implemented in the Leiden University Medical Centre (LUMC). It is based on ontology models for data and metadata as well as a data architecture that builds on the existing data management systems. The work also involved clinicians and data managers who worked with the hospital data systems and knew the data life cycles well.
The team applied several technologies in the different steps of their method. For their data analytics with Semantic Web technologies, they used SPARQL to perform data analytics over the LUMC RDF patient data and across various external linked open data sources. To host the generated FAIR data, they used the free edition of GraphDB, which stored the data natively as RDF.
This work demonstrates that “a FAIR research data management plan based on ontological models for data and metadata, open Science, Semantic Web technologies, and FAIR Data Points is providing data infrastructure in the hospital for machine actionable FAIR Digital Objects. This FAIR data is prepared to be reused for federated analysis, linkable to other FAIR data such as Linked Open Data, and reusable to develop software applications on top of them for hypothesis generation and knowledge discovery.to help medical doctors and researchers answer research questions.”
The final paper in this selection is A Web-Based Platform for Traditional Craft Documentation by Nikos Partarakis, Voula Doulgeraki, Effie Karuzaki, George Galanakis, Xenophon Zabulis, Carlo Meghini, Valentina Bartalesi and Daniele Metilli, published 10 May, 2022.
This paper presents a web-based authoring platform for the representation of traditional crafts. The importance of traditional crafts cannot be overstated as they weave through generations, places and civilizations connecting humanity to their roots. Click To Tweet Such craft products include clothing and jewelry, costumes and props for festivals and performing arts, decorative art and ritual objects; musical instruments and household utensils, and many more. But despite their importance, traditional crafts have not benefited much from information technologies. Especially when it comes to intangible cultural assets such as “dexterity, know-how, and skilled use of tools, as well as tradition and identity of the communities in which they are, or were, practiced”.
To rectify this situation, an online collaborative documentation platform was developed in the context of the Mingei project. It makes it easy to represent cultural context through narratives and to export knowledge in multiple formats for further re-use and sharing. It also provides linkages to other relevant public knowledge bases such as Europeana, etc.
The platform architecture consists of different components with equal importance for the overall performance and functionality. One part of the asset storage component is GraphDB, which is used for storing the semantic data as RDF triples. The front end of the platform is built using the ResearchSpace project platform, which directly links to the knowledge graph. ResearchSpace was a project awarded by the British Museum to support collaborative research projects for cultural heritage scholars and the resulting platform was powered by GraphDB.
In conclusion, the authors note: “The main contribution of this work may be summarized under the phase: The representations achieved through the authoring platform are narrative-centric rather than artifact centric. To this end, the proposed authoring platform relies on a strong conceptualization and focuses on a notion of narratives that, unlike the previous approaches, exploits both sides of the representation, the semantical (fabula) and the signal-based (the narration) side, and combines these two aspects by linking semantic notions, such as events and actions, to the media objects that illustrate these notions.” They also plan to further improve the methodology and platform.
When it comes to knowledge graph technologies, there are endless opportunities to learn from academic research and borrow best practices or use existing solutions (see more scientific papers citing GraphDB in our Zotero library).
These handpicked papers about research projects provide specific examples about the versatility, efficiency and scalability of knowledge graphs. They make it easy to aggregate data from disparate sources, represent multihop relationships, navigate and analyze information, and build robust and future-proof services out of financial, healthcare and cultural heritage data.
Powered by or having integrated GraphDB in their overall solutions, these research projects not only blaze trails across various domains, but also showcase future potential uses of knowledge graphs to enable enterprises to thrive on relevant and connected information.