Zachary Teed

Papers

Total Citations

H-Index

About

Zachary Teed is a researcher whose work sits at the intersection of 3D computer vision, geometry, and deep learning. His primary research areas focus on enabling neural networks to reason about 3D space, particularly through the lens of Lie groups and differentiable optimization. Teed’s most notable contribution is his work on "Tangent Space Backpropagation for 3D Transformation Groups," which addresses a fundamental challenge in 3D vision and robotics: how to perform gradient-based learning over non-Euclidean spaces like SO(3), SE(3), and Sim(3). By providing a principled method for backpropagating through these smooth manifolds, his work has enabled more stable and geometrically consistent training of networks for tasks like pose estimation and SLAM. This highly cited paper (34 citations) has become a key reference for researchers building differentiable pipelines in 3D. Teed’s contributions are particularly impactful for students and engineers working at the frontier of end-to-end learning for spatial reasoning, offering a rigorous yet practical toolkit for handling the geometry that underpins modern robotics and augmented reality.