Kathlén Kohn

Kathlén Kohn is an Associate Professor in Mathematics at KTH Royal Institute of Technology in Stockholm, Sweden (since December 2024). She holds a PhD from Technische Universität Berlin (2018) and dual Master's degrees in Mathematics and Computer Science from Paderborn University (2015). Her research bridges algebraic geometry, geometric deep learning, and computer vision, focusing on algebraic structures in neural networks and geometric problems in AI.

Education:

• PhD in Mathematics, TU Berlin (2015–2018)
• Master of Science in Mathematics & Computer Science, Paderborn University (2013–2015)
• Bachelor of Science in Mathematics & Computer Science, Paderborn University (2009–2013)

Research: Kohn explores neural algebraic geometry, applying algebraic techniques to analyze deep learning architectures like polynomial neural networks and self-attention mechanisms. Her work includes minimal problems in computer vision (e.g., PLMP framework), metric algebraic geometry, and invariant theory connections to maximum likelihood estimation. She co-authored the book Metric Algebraic Geometry (2024) and leads the WASP-funded project on 3D scene perception.

Recent Articles: Focus on geometric neural network analysis, self-attention mechanisms, and structure-from-motion problems. Key venues include ICML, ICLR, CVPR, and SIAM Journal on Applied Algebra and Geometry.

Awards: Wallenberg Prize (2025), SIAM SIGEST Award (2024), Swedish L'Oréal-Unesco Award (2023), Göran Gustafsson Prize (2021), and multiple fellowships including Marie Skłodowska-Curie.

Teaching & Outreach: Lectures on algebraic vision, nonlinear algebra, and cryptography. Active in promoting gender equality in STEM through ELLIS and Swedish Young Academy.