Jansen Wong

Papers

Total Citations

H-Index

About

Jansen Wong is a rising star in robotics and computer vision, whose work focuses on bridging the gap between 2D visual understanding and 3D robotic manipulation. His most notable contribution, "Distilled Feature Fields Enable Few-Shot Language-Guided Manipulation" (2023, 9 citations), tackles a fundamental challenge: how to equip robots with the rich semantic knowledge of 2D image models while also giving them the precise 3D geometric awareness needed for real-world tasks. By distilling features from self-supervised and language-supervised models into a 3D field, Wong’s method allows robots to understand and manipulate objects based on natural language instructions with only a handful of demonstrations. This work is pioneering in its ability to combine semantic and geometric reasoning, enabling more generalizable and sample-efficient robotic learning. Though early in his career, Wong’s research has already garnered attention for its clever integration of large-scale pre-trained models with 3D scene representations, pointing toward a future where robots can seamlessly follow human commands in unstructured environments. His work is essential reading for anyone interested in language-guided robotics and 3D scene understanding.