Artificial intelligence consumes a great deal of energy. Our project EMULAITE therefore aims to develop a sustainable and energy-efficient software solution for training AI models.

The start-up GEMESYS has developed an AI training method based on biological principles that emulates electrical circuits. The method reduces computational complexity, improves the accuracy of AI models and makes them accessible on resource-constrained devices. Based on this training method, EMULAITE aims to develop a software solution for training AI models.

In addition, the project will investigate how, in the case of technology-oriented inventions, market, customer and user needs can be incorporated into the development process at an early stage. To this end, different approaches to requirements analysis will be tested and the results fed back into the technical development process. The method-driven translation of customer and user needs into requirements specifications plays a central role in this process.