Sort by
Refine Your Search
-
the Department of European and Comparative Literature and Language Studies, characterized by a broad range of research interests within Finno-Ugric studies in the wide sense of the word and by intensive
-
approximation theory can be automated by a (neural network) guided search over the action space of standard tools (e.g., Hölder inequalities, Sobolev embeddings, ...). Certain proofs in these fields require
-
of machine learning. Our ideal candidate will have prior exposure to modern developments in theoretical machine learning and deep neural networks and will be able to take part in research project related
-
tools—such as neural networks, spline functions, or Gaussian random fields. The core challenge? Developing methods that are not only provably stable and reliable, but also efficient in high dimensions
-
developing adaptive numerical schemes powered by advanced nonlinear approximations—like Gaussian mixtures and neural networks. The key challenge? Designing robust and stable numerical schemes that remain