Symbolic approaches, for example based on ontologies or knowledge graphs, use logical reasoning to uncover implicit knowledge in data and thus transform it into explicit knowledge.
In a different way, sub-symbolic AI methods, such as neural networks and deep learning, use so-called training data and algorithms to detect and learn patterns and relationships in data. This mechanism of detecting regularities can be used, for example, to identify anomalies or classify the condition of a machine.

In the joint project between HEKUMA and the THA, both approaches are being investigated and evaluated to find out which methods and their combination are best suited to improve requirements management for special purpose engineered solutions, especially in the early project stage. Either way, it is believed that the application of artificial intelligence in requirements management can help increase efficiency and quality.

This project, which is to be intensified in the coming months, achieves an optimal transfer of knowledge and exchange between science and practice and thus promotes innovative solution finding. It is planned to present the results to a broad professional audience at internationally renowned conferences, thus stimulating further discussions.

All those involved are looking forward to the successful continuation of the collaboration and are keen to see the results of the project.