![](https://cdn.prod.website-files.com/61e6de519074a0baaf971be9/668d283c6fc99cd624c8751d_conweaver-LinkedDataDays-logo.png)
![](https://cdn.prod.website-files.com/61e6de519074a0baaf971be9/668d283c6fc99cd624c8751d_conweaver-LinkedDataDays-logo.png)
ARTIKEL
9. CONWEAVER Linked Data Day in Darmstadt
AI, industrial metaverse and the knowledge graph as an enabler
New and sometimes surprising perspectives on the major topics and challenges of our time, vivid reports from the field and outlooks into the future – the 9th CONWEAVER Linked Data Day in Darmstadt was an established expert forum with ample space for exchange, discussion and knowledge transfer on the industrial networking of data landscapes along processes, organizations and lifecycles.
The continuous linking of data is becoming increasingly important. By overcoming their own data silos, companies are increasing their efficiency and gaining competitive advantages. With projects like Catena-X, there are now also approaches for cross-company and at the same time secure data spaces. This development is being fueled by the rapid advances in artificial intelligence (AI), which is dependent on reliable sources. Knowledge graphs can represent a company's knowledge without the AI hallucinating.
This is how CONWEAVER CEO Dr. Thomas Kamps outlined the thematic spectrum of the 9th Linked Data Day in his welcoming address. In close proximity to the Technical University and with an atmospheric view of Darmstadt Castle, more than 60 experts exchanged ideas at the Darmstadtium conference center and reported from the field on science, research and daily application in business. They also had plenty of opportunity to make new contacts for potential joint projects.
Professor Dr.-Ing. Christoph Legat from Augsburg University of Applied Sciences kicked off the event the evening before at an informal get-together. The co-founder of several deep-tech start-ups transported his audience to a fairytale world with his witty and entertaining talk entitled "A magical story of curious graphs and sub-symbolic magic in the realm of data". Using knights, magicians and the European Queen Ursula, he illustrated the current challenges. The problems in the field of tension between ontologies and graphs on the one hand and GenAI on the other could only be solved if there was a dream marriage between GenAI and knowledge graphs in the end. "We have to bring these things together," demanded Legat.
In his keynote "We have more than we think – how digital transformations succeed!", independent business consultant Udo Jauernig highlighted the prerequisites and ways for successful change in the development and production environment. Jauernig, who has more than three decades of international leadership experience in the automotive supply industry at Bosch, emphasized the great value of data. However, the rapidly growing amount of data, the often poor quality and the high proportion of unknown and unused data (dark data) are major challenges. Existing systems would cement data silos. Digital transformations are currently taking place at a functional, company and industry level. These are running synchronously and influencing each other. To remain capable of acting, a holistic approach is needed. The commitment of company management is crucial for success. The image of IT as a mere service provider is outdated. "IT has a seat at the table to shape the future," said Jauernig.
Two customers demonstrated the practical applications of CONWEAVER's solutions. Michael Mauer, Head of Quality Data Management at Vitesco Technologies, first gave insights into how the leading supplier of electric drives uses semantic data layers for product lifecycle management (PLM). With CONWEAVER's Linksphere they have created a product-centered data model that enables a holistic view of data. This allows information to be gathered in just a few minutes, which previously took weeks in some cases.
As Senior Consultant for Data Architecture & Management, Dr. Hannes Wolf from Robert Bosch GmbH provided a practical perspective on how business processes are continuously renewed based on integrated data and information as part of the digital transformation. The world's largest automotive supplier has been working with CONWEAVER since 2013. Today, more than 180 million documents are stored there with around one billion links between them. Many walls between the data silos had to be torn down for this. But it was worth it. "Finding information works much faster today," said Wolf. Connecting domains creates understanding and knowledge.
Lead Developer Jens Sauer and Vice President Engineering Tiziano Degaetano from CONWEAVER used a live prototype to demonstrate how a chatbot can query knowledge graphs with company knowledge very effectively without the GenAI hallucinating. The decisive factor here is that the graph represents reliable company knowledge, which is only queried by the chatbot. The chatbot itself does not generate any new knowledge.
Joachim Stratmann, Custom Success Lead from the German AI company Aleph Alpha, emphasized the topic of "Explainable AI" in his lecture. "We no longer want to rely on American companies when it comes to AI, but must create our own counterbalance in Germany and Europe," he said. Explainability, i.e. the ability to understand how dynamic and non-linearly programmed systems achieve results, is a key success factor for generative AI applications. "We are transparent and don't hide anything," said Stratmann. "We show where the answer comes from." The human remains the decision-maker, while AI simplifies his work.
The event was rounded off with short pitches from CONWEAVER's partners. Here :em engineering methods AG (Dr. Marcus Krastel), Software AG (Jan Lübberstedt), Threedy GmbH - instant3Dhub (Christian Stein) and T-Systems International (Susanne Brand, Carolin Tröster) presented their companies and services.
by Stephan Köhnlein 10/07/2024