Anran Zhang
Papers
1
Total Citations
3
H-Index
1
About
Anran Zhang is an emerging researcher at the forefront of robot learning and computer vision, with a particular focus on bridging the gap between human activity understanding and autonomous robotic manipulation. Their most notable work, "VidBot: Learning Generalizable 3D Actions from In-the-Wild 2D Human Videos for Zero-Shot Robotic Manipulation" (2025), tackles one of the field's most pressing challenges: enabling robots to learn versatile manipulation skills without requiring extensive physical robot training, which is notoriously difficult to scale. By leveraging the rich information embedded in everyday 2D human videos, Zhang's research proposes an innovative pathway to zero-shot robotic manipulation — allowing robots to perform household tasks they have never been explicitly trained on. This approach addresses the embodiment gap, a longstanding obstacle in transferring human motion knowledge to robotic systems. Although early in citation accumulation with 3 citations, the work has already attracted attention within the robotics and machine learning communities. Zhang's research represents a promising direction for developing generalizable, data-efficient robotic systems, positioning them as a researcher to watch in the rapidly evolving field of embodied artificial intelligence.
Research Focus
Key Achievements
Top Papers
- 1