Florian Steinke

Max Planck Institute for Biological Cybernetics, Max Planck Society

Papers

Total Citations

H-Index

About

Florian Steinke’s research lies at the intersection of machine learning, signal processing, and geometric data analysis. His most cited work, “Non-parametric Regression Between Manifolds” (2008, 56 citations), introduces a novel algorithmic framework for learning mappings between Riemannian manifolds—a fundamental challenge in fields as diverse as computer vision, robotics, and computer graphics. By tackling the problem of non-parametric regression on curved spaces, Steinke provides a principled way to handle data that naturally lives on nonlinear structures, such as shapes, poses, or orientations. This contribution is particularly valuable for applications requiring smooth, structure-preserving transformations, such as motion capture or 3D object recognition. Though his later paper on the same topic (2009) has fewer citations, it refines the theoretical foundations and algorithmic details, underscoring his commitment to rigorous, applicable mathematics. Steinke’s work has helped bridge the gap between abstract differential geometry and practical machine learning, offering tools that enable more accurate and robust modeling of complex, high-dimensional data. His research continues to influence students and researchers working on geometric deep learning and manifold-based inference.