AI Research Projects
Competitive advantage through funded AI innovation
For us, funded AI research is not an end in itself, but a head start for tomorrow's customer projects. In research and joint projects, we push the boundaries of what is possible, rigorously validate new methods, and translate the results into ready-to-use building blocks for practice – so our customers can benefit early from reliable, modern AI.
Why funded AI research?
Early access
Benefit from cutting-edge AI methods before they become mainstream
Validated innovation
Research results are scientifically robust and practice-ready
Funding instruments
Public funding instruments support your digital transformation
Transfer know-how
We translate research into productive, deployable solutions
DigiSemNet
ZIM-funded R&D joint project
Platform for digital engineering with networked data
In the development of complex systems, many engineering artifacts are created across mechanics, electrical/electronics, and software. This information is highly interdependent but often distributed and consequently becomes inconsistent quickly when changes occur. DigiSemNet develops a platform that semantically links engineering data, makes inconsistencies verifiable, and simplifies access to distributed engineering knowledge through natural language search and AI-assisted support.
What is being built in the project?
Semantic networking (Digital Thread)
A central linking graph maps relationships between relevant data and model elements – as a foundation for traceability across disciplines and tools.
Consistency checking
Rule-based checks help identify contradictions between related artifacts early.
"Chat with engineering data"
Interactive, natural language semantic search across linked artifacts; including result preparation and prioritization.
AI agent network
Multi-stage queries and (partial) automation of repetitive tasks based on linked data.
Confidentiality focus
Consideration of protection requirements so that use remains practical even with sensitive engineering data.
Contribution from talsen team
In the project, talsen team primarily shapes the building blocks that make distributed engineering knowledge findable and automatically reusable through the use of AI and AI agents:
- →Natural language, AI assisted search ("chat with engineering data"): We develop the search so that linked engineering information can be queried in natural language.
- →Present and prioritize results comprehensibly: Results are prepared so that relevance and context are more quickly recognizable – especially when many artifacts are linked together.
- →Use also via interfaces (API): In addition to interactive use, connection via a programmatic interface is considered so that search queries can also be initiated from tools or processes.
- →Adapt AI-models for the engineering context: This also includes adapting/optimizing the models used (including embeddings and language models) for domain-specific artifacts and terminology.
- →AI agent approaches for multi-stage queries: In addition, we are working on deploying agents that support multi-step queries and automate repetitive tasks when working with linked data to the greatest extent possible.
Validation in case studies
- ✓Machinery & plant engineering: Case study with an industrial partner
- ✓Automotive: Case study "fortissimo" vehicle with engineering models from various tools
Frequently Asked Questions
Learn more about AI research projects?
Get in touch to discuss project ideas and collaboration opportunities.