Why a 2020 Paper on Rotation Representations Is Called the Biggest Kinematics Breakthrough in a Decade
andrew_n_carr · x · 2026-10-22
Andrew Carr shared the 2020 paper "On the Continuity of Rotation Representations in Neural Networks" (USC & Adobe Research), calling it the paper that has "done more good for kinematics than essentially any other in the last decade."
The paper showed that popular 3D/4D rotation representations like quaternions and Euler angles are suboptimal for neural network regression when full rotation space is needed—converged networks still produce large errors at certain angles. The authors formalize a definition of continuous representation and advocate rotation matrices, a result that shaped how robotics and 3D deep learning handle rotations ever since.
More from Embodied
- Agility Robotics expands in Pittsburgh with $300M+ in Digit 5 orders and East Coast service hub — jonstephens85 · 2026-10-22
- Robert Downey Jr. demos Snap's new SPECS AR glasses — CodeByPoonam · 2026-10-22
- The $5,000 Unitree R1 is impressive but 'not good for much' beyond education, says analyst — jonstephens85 · 2026-10-22
- World model + APG: single-GPU 1.24h reproduction of SuperMimic humanoid control — ChongZzZhang · 2026-10-22
- Open-source robotics harness reproduces Gemini zero-shot bimanual handover — berkeley_ai · 2026-10-22
- SCOUT Adapts Sim-Trained Robot Policies to the Real World in a Few Trials — davheld · 2026-10-22