Andreas Mazur studied (B.Sc.) Cognitive Computer Science (2020) and (M.Sc.) Intelligent Systems (2022) at Bielefeld University. Now he is a Ph.D. student in the machine learning group at the Center for Cognitive Interaction Technology.

In the course of priority program 2422 by the german research foundation (DFG) he is researching on explainable machine learning surrogate models for multi-stage punch-bending, a pivotal industrial process for manufacturing large amounts of products that can take the shape of various geometries.

His research focuses on intrinsic mesh convolutional neural networks (IMCNNs), which operate on Riemannian manifolds and expand the application of convolutions to non-Euclidean data such as curved surfaces. In addition, he uses methods from explainable AI (XAI), such as dimensionality reduction techniques and counterfactual explanations, to investigate the trustworthiness of developed machine learning models.