Logo des Repositoriums

Auflistung nach Schlagwort "Knowledge Graphs"

1 - 6 von 6
  • Konferenzbeitrag
    Creation and Utilisation of Domain Specific Knowledge Graphs (DSKG) for E-Learning
    (DELFI 2021, 2021) Meissner, Roy; Thor, Andreas
    Educational domain models are building blocks for educational software. Unfortunately, such domain models require high manual effort to be created, are quickly outdated and are thus rarely used at all. A novel way to create such models is proposed with this paper, utilising knowledge mapping algorithms, natural language processing tools, a semantic web architecture and integration with online knowledge bases. The developed processes and tools make it fast and easy to create domain specific knowledge graphs (DSKG) automatically. Apart from defining DSKGs, two use-cases are presented that show how third-party tools may utilise DSKGs: (1) improved exam generation with EAs.LiT and (2) an assessment item and educational material recommender.
  • Textdokument
    A Demonstration System towards NLP and Knowledge Driven Data Platforms for Civil Engineering
    (INFORMATIK 2022, 2022) Borst,Janos; Meinecke,Christofer; Wiegreffe,Daniel; Niekler,Andreas
    We present a demonstrator that shows a concept towards a smart document platform for Civil Engineering. It demonstrates NLP-supported organisation of documents included into a graphical user interface. The demonstrator addresses the fundamental problem of data structure that includes the internal project logic of a planning project in Civil Engineering. To this end, we build a knowledge graph that includes not only domain-specific knowledge but also standardised project structures. So far included functionalities are a navigation that allows filtering according to a logic familiar to engineers and a simplified data import via automatic classification and tagging using language technologies.
  • Konferenzbeitrag
    Educational pathways in a German labor market knowledge graph
    (INFORMATIK 2024, 2024) Kostadinovska, Katerina; Dörpinghaus, Jens
    Educational pathways play a pivotal role in fostering lifelong learning and skilled workforces. We build upon a German labor market knowledge graph that was developed from the German Labor Market Ontology (GLMO). The GLMO provides entities for qualifications, such as occupations and training programs, as well as tools and skills. To conduct research on educational pathways, we employ several graph algorithms, such as shortest paths, and present the results. It is possible to demonstrate the advantages of digital methods in sociological research, thereby advancing the boundaries of computational social sciences in labor market research.
  • Zeitschriftenartikel
    Künstliche Intelligenz zur Abbildung und Sicherung von Wissen – Nachhaltigkeit für das wichtigste Unternehmens-Asset
    (HMD Praxis der Wirtschaftsinformatik: Vol. 58, No. 1, 2021) Schirmer-Kaegebein, Ulf; Reinheimer, Stefan
    Effizientes Wirtschaften in einem Unternehmen bedingt die möglichst umfassende, langfristige und jederzeit zugängliche Ressource Wissen . Die Digitalisierung dieses Wissens – und das umfasst nicht das Einscannen von Dokumenten und ihre Bereitstellung auf zentralen Plattformen – ermöglicht es, nachhaltig mit diesem bedeutenden Unternehmens-Asset umzugehen. Komplexe Produktzusammenhänge, wie Zubehör, (zertifizierte) Kompatibilitäten, die Möglichkeit, auch bei Massenprodukten individuelle technische Dokumentation zu erstellen, Vorschläge ähnlicher Produkte unterbreiten oder technische Produktkombinationen auf Basis fachlicher Anforderungen empfehlen zu können, weisen Kompetenz nach und schaffen Vertrauen. Der klassische Wissensmanagementansatz der 90-Jahre des letzten Jahrhunderts, der sich im Wesentlichen auf die effiziente Auffindbarkeit von Dokumenten bezieht, ist hier bei weitem nicht ausreichend. Dieses Know-how unabhängig von der physischen Anwesenheit von Mitarbeitern verfügbar zu haben, ist Motivation für die langfristige und effiziente Digitalisierung und Bereitstellung des Wissens. Der Bedarf verstärkt sich durch die zu erwartende Veränderung im Personalwesen durch die kommenden Generations Y und Z, deren Verweildauer in einem Unternehmen signifikant kürzer erwartet wird als dies die aktuellen Generationen bislang vorgelebt haben. Dem Risiko, das Mitarbeiterwissen ständig zu verlieren und erneut aufbauen zu müssen, kann über moderne Digitalisierungsansätze begegnet werden – Nachhaltigkeit als Ergebnis des Einsatzes künstlicher Intelligenz (KI). Der Beitrag leitet den Bedarf ab, das Nachhaltigkeitsverständnis um den Umgang mit Wissen im Unternehmen zu erweitern. Nach einer entsprechenden theoretischen Fundierung stellt der Beitrag ein Siemens-Praxisbeispiel detailliert dar – die Anforderung sowie die Lösung mit Hilfe eines semantischen Netzes als Vertreter einer KI-Methode. Die Erkenntnis, dass diese Knowledge Graphen zur Abbildung von Wissen im Unternehmen ein wichtiger Baustein der unternehmerischen Nachhaltigkeit sind und ein Ausblick auf offene Aufgabenstellungen für die nahe Zukunft runden den Beitrag ab. Efficient corporate management requires the most comprehensive, long-term and always accessible resource knowledge. Digitizing this knowledge—and this does not mean scanning documents and making them available on centralized platforms—makes it possible to deal sustainably with this most important corporate asset. Complex product contexts, such as accessories, compatibility, the possibility to create individual technical documentation for mass-produced products, to submit proposals for similar products or to recommend technical product combinations based on functional requirements, demonstrate competence and create trust. The classic knowledge management approach of the 90s of the last century, which essentially refers to the efficient retrievability of documents, is far from sufficient here. Being able to make this knowledge available digitally and thus being independent of the physical presence of employees is the motivation for the long-term and efficient provision of knowledge. The need is intensified by the expected changes in human resources management due to the coming generations Y and Z, whose retention time in a company is expected to be significantly shorter than the current generations have demonstrated so far. The risk of constantly losing employee knowledge and having to rebuild it can be countered with modern digitization approaches—sustainability as a result of the use of artificial intelligence (AI). The article derives the need to broaden the understanding of sustainability by including the handling of knowledge within the company. After a corresponding theoretical foundation, the article presents a practical Siemens example in detail—the requirements as well as the solution of how to apply a semantic network as a representative of an AI method. The conclusion that these knowledge graphs for mapping knowledge in the company are an important component of corporate sustainability on one side and an outlook on open tasks for the near future on the other side round off the article.
  • Konferenzbeitrag
    To Graph or Not to Graph: The Missing Pieces for Knowledge Graphs in Sustainable Tourism
    (INFORMATIK 2023 - Designing Futures: Zukünfte gestalten, 2023) Neubig, Stefan; Rebholz, Dominik; Hein, Andreas; Keller, Robert; Krcmar, Helmut
    Sustainability is a critical challenge in modern tourism, exacerbated by climate change and globalization. Thanks to digitization, data-driven approaches constitute a key technology for addressing related issues, such as overtourism. However, the overarching complexity of the touristic data landscape, amplified by the interplay of diverse digital platform ecosystems, poses considerable challenges to both data owners and consumers. To mitigate such issues, knowledge graphs (KGs) have received significant attention. KGs focus on data quality by employing unified data models and continuous data refinements, making them well-suited for data-driven applications. Although promising, many challenges must be addressed to make KGs useful in practice. This paper overviews the state of the art of the field and identifies avenues for future research, explicitly focusing on touristic value and sustainability. Following our results, future research should focus on different areas, notably real-time knowledge graph population, distributed and parallelized processes, and ontologies for dynamic data types.
  • Konferenzbeitrag
    Unlocking sustainable reporting: Leveraging knowledge graphs for ESG metrics extraction: The role of knowledge graphs in sustainability reporting
    (INFORMATIK 2024, 2024) Driller, Jana; Trang, Simon Thanh-Nam
    This study investigates the use of knowledge graphs to address challenges in ESG metrics extraction and reporting. However, collecting and processing ESG data from diverse IT systems is complex. Knowledge graphs offer a solution by enabling semantic modelling and data integration. This research employs a design science approach to develop a prototype knowledge graph, enhancing the accuracy and consistency of ESG reporting. The study contributes both theoretically, by advancing knowledge graph frameworks, and practically, by providing a tool for efficient ESG data management and reporting.