Yunchun Chen

University of Toronto

Papers

3

Total Citations

40

H-Index

2

About

Yunchun Chen is a researcher at the forefront of robotic manipulation and geometric deep learning, with a focus on enabling machines to autonomously interact with and assemble physical objects. His work bridges the critical gap between perception and action, developing self-supervised learning frameworks that allow robots to understand shape, contact, and motion without extensive human annotation. Chen’s most influential contribution is the "Neural Shape Mating" framework (2022), which redefines part assembly not as semantic reconstruction but as a precise geometric mating problem—a fundamental skill for automated manufacturing and repair. By integrating adversarial shape priors, this work achieves robust, generalizable assembly from partial observations, garnering over 34 citations and establishing a new paradigm for object assembly research. He further advanced the field with "Neural Motion Fields" (2022), which encodes entire grasp trajectories as implicit value functions, compressing the complex pipeline of grasp detection, inverse kinematics, and motion planning into a single, learned representation. This innovation enables more fluid, closed-loop robotic grasping. Chen’s work is distinguished by its elegant synthesis of implicit neural representations with practical robotic tasks, making him a rising leader in autonomous assembly and dexterous manipulation.

Research Focus

Key Achievements

2
H-Index
3
Papers
40
Total Citations
13
Avg Citations/Paper
🏆 Most Cited Paper
Neural Shape Mating: Self-Supervised Object Assembly with Adversarial Shape Priors
34 citations · 2022
📈 Most Prolific Year: 2022 (3 Papers)
🤝 Key Collaborators: 8
🏛 Institutions: University of Toronto

Top Papers

  1. 1
  2. 2
  3. 3

Key Collaborators

Contact & Links

Available for collaboration
Content generated · 13 days ago