Jon Long
Papers
1
Total Citations
3
H-Index
1
About
Jon Long is a robotics researcher whose work centers on computer vision for robot manipulation, with a particular focus on enabling intuitive, data-efficient programming paradigms. His most notable contribution is the development of "FlowControl," a pioneering method for optical flow-based visual servoing that makes one-shot imitation learning practical for real-world manipulation tasks. Rather than requiring extensive coding or thousands of demonstrations, Long’s approach leverages modern learning-based optical flow to allow a robot to replicate a task from a single human demonstration in real time. This breakthrough directly addresses a fundamental challenge in robotics: how to bridge the gap between human demonstration and robot execution with minimal data. While his 2020 paper has accumulated 3 citations, its conceptual impact lies in demonstrating that dense visual correspondence can serve as an effective intermediate representation for skill transfer. Long’s work sits at the intersection of computer vision, imitation learning, and control, offering a compelling vision for how robots might one day learn new skills as effortlessly as showing them once.
Research Focus
Key Achievements
Top Papers
- 1FlowControl: Optical Flow Based Visual Servoing3 citations · 2020