Kathlén Kohn is a WASP Assistant Professor at the Department of Mathematics at KTH Royal Institute of Technology. Dr Kohn was recruited by WASP in 2019.
WASP provides great opportunities right in the middle between pure and applied math, which is where I feel at home.
WASP incentivizes research crossing scientific boundaries, which is much needed to understand the complex structures in modern data science. I have broadened my scientific network due to WASP, and my research has gotten a wider context with more visibility.
I am an algebraic geometer. In my research, I investigate the geometric structures in data science problems to shed light on their intrinsic difficulty. I study those geometric structures with algebraic tools. My research focuses on 3 data science applications: theory of machine learning with neural networks, 3D reconstruction in computer vision, and maximum likelihood estimation in statistics.
My research distinguishes between the intrinsic difficulty of a data science problem and the issues that are only caused by algorithms (instead of the problem itself). This can inform engineers on which issues they can overcome by changing the established method, and which issues are inherent to the problem at hand. I hope that my research will improve the design of neural networks, unravel the mysteries of deep learning, and enable 3D reconstruction from images in settings that are currently out of reach.