Yingda Yin
Papers
2
Total Citations
9
H-Index
2
About
Yingda Yin is a rising researcher whose work sits at the intersection of computer vision, graphics, and robotics, with a particular focus on probabilistic modeling of rotations. His major contribution lies in pioneering the application of discrete normalizing flows on the SO(3) manifold—a fundamental challenge in representing uncertainty over 3D rotations. In his highly cited 2023 paper, Yin introduced a novel framework that leverages normalizing flows to construct expressive, trackable probability distributions directly on the rotation manifold, overcoming limitations of traditional Gaussian or von Mises-Fisher models. This work has already garnered significant attention, with citations accumulating rapidly, reflecting its immediate impact on fields requiring robust rotation estimation, such as 6D object pose tracking, camera localization, and robotic manipulation. By enabling more accurate and flexible uncertainty quantification for rotations, Yin’s research bridges a critical gap between geometric deep learning and probabilistic inference. His approach not only advances theoretical understanding of manifold-valued distributions but also provides practical tools for real-world applications where rotation ambiguity is pervasive. As his citation count continues to grow, Yin is establishing himself as a key voice in probabilistic geometric learning, with his SO(3) normalizing flow framework poised to become a standard tool for researchers tackling rotation-related uncertainty.
Research Focus
Key Achievements
Top Papers
- 1
- 2