The TransfAIr project focuses on developing transfer approaches that enable companies to quickly and cost-effectively apply previously developed AI solutions to new systems and product variants.

In contrast to existing research projects, we consider two different aspects of AI transferability in combination: 1) Its implementation on an execution platform and 2) Its training and execution data.

The approaches and methods developed throughout the project will be made available as freely accessible, collaboratively usable tools. Additionally, we aim to explore the limits of transferability in this context.

The project is funded by the BMBF and carried out by the University of Duisburg-Essen in close collaboration with industrial partners.