The project quantifies and evaluates high-dimensional concentration phenomena in data, which are often associated with the term "dimension curse". This phenomenon occurs when machine learning methods are applied to high-dimensional data. So far, this phenomenon cannot yet be calculated with algorithms. It is therefore an open question to what extent it has decisively influenced results of scientific applications.

So far, the aspect of dimension curse is often only anecdotally grasped, which has led to a multitude of empirically derived recommendations that are, however, mathematically unfounded as well as contradictory. In order to ensure the (scientific) use of artificial intelligence methods in future applications, which are expected to be high-dimensional in nature, it is necessary to recognise and quantify the dimension curse. Hence, the goal of the "Dimension Curse Detector" project is to develop computable approximations of the concentration phenomenon in machine learning applications. This way DCD contributes to the overall goal of Explainable AI and enables that decisions drawn in high-dimensional AI applications are made on the basis of data transparence and understanding. In this sense, the Dimension Curse Detector should become a tool for the design of socially desirable AI applications.