.svg)
GENIUS Lab
A collaboration between Delft University of Technology, Maastricht University, dsm–firmenich, and Kickstart AI.
The GENIUS (Generative Enhanced Next-Generation Intelligent Understanding Systems) Lab is a research lab that seeks to extend and enhance state-of-the-art Artificial Intelligence (AI) methods for semantic knowledge engineering, human-centered AI, and crowd computing. GENIUS Lab is a collaboration between Delft University of Technology, Maastricht University, dsm–firmenich, and Kickstart AI.
The GENIUS lab focuses on how humans and AI can collaborate in knowledge management and knowledge discovery within large organisations. The lab will develop human-centred approaches that involve humans in extracting, organising and accessing stored knowledge.
This research aims to accelerate how AI is used in science, research, innovation and operations, where knowledge management and decision making through AI-based systems might have far-reaching implications for many industries. These implications will be showcased in real-world use cases.
The lab will co-develop generative, human-AI collaborative knowledge engineering methods and techniques for distinct yet highly interweaved knowledge engineering steps: collaborative knowledge synthesis; integration and linking of distributed knowledge fragments; integration of structured knowledge in generative AI models; trustworthy conversational AI using FAIR data and services; and hybrid human-AI ethics. These methods and techniques are designed, developed, experimented with, evaluated and validated in a variety of contexts.
The lab and its partners have a shared long-term vision to improve human-AI collaboration to increase the scalability, trustworthiness and efficiency of AI- supported content-creation and decision–making systems.
Research projects
Collaborative knowledge synthesis. Reliable and resilient knowledge-graph representation and synthesis from large language models and semantic knowledge bases through algorithms, interfaces, and systems for evidence-based human-AI collaboration in the context of food systems and translational science (real-world evidence-based claims).
Integration of distributed knowledge fragments. Integration and linking of distributed knowledge fragments: resilient and reliable integration of fragmented knowledge in knowledge graphs through repeatable algorithms and methodologies in distributed and incomplete contexts, possibly applying federated learning principles in the context of food systems.
Integration of structured knowledge in generative AI models. Improving the reliability of generative AI through human-in-the-loop, knowledge-based interpretation of AI behavior and neuro-symbolic approaches for integrating structured knowledge into generative AI.
Trustworthy Conversational AI using FAIR Data and Services. Improving the accuracy and trustworthiness of conversational AI by incorporating FAIR data and services, neurosymbolic reasoning, and user interaction.
Human-AI ethics and responsible innovation. Address epistemic and ethical challenges of human-AI collaboration with generative AI models in the context of food systems.
People
PHD Students
Partners
GENIUS Lab brings contributions to the reliability and accuracy criteria from the context of Collaborative Knowledge Engineering, validated in a realistic industrial setting offered by dsm-firmenich and Kickstart AI. The resulting reliable and accurate knowledge bases are a central component in subsequent knowledge-driven AI systems. By combining the scientific expertise from TU Delft and University of Maastricht with respect to collaborative knowledge engineering for semantic data management, collaborative knowledge management, and responsible data science, the lab provides a clear and effective addition to the entire range of scientific innovations.
Partners:
- DSM-firmenich
- Kickstart AI
- TU Delft
- Maastricht University
Newsletter
Stay in the loop
