Jinzhe Xue
Papers
1
Total Citations
3
H-Index
1
About
Jinzhe Xue is a rising researcher at the forefront of robot learning, with a focus on enabling machines to acquire complex manipulation skills directly from unstructured human demonstrations. His work addresses a fundamental challenge in robotics: how to bridge the gap between raw human video data and deployable robot policies. In his highly-cited 2024 paper, "Contrast, Imitate, Adapt: Learning Robotic Skills From Raw Human Videos," Xue introduces a novel framework that circumvents the limitations of traditional behavior cloning and reward function learning. By combining contrastive learning for representation, imitation for policy initialization, and adaptation for domain transfer, his method allows robots to learn from diverse, unlabeled human videos without requiring costly robot-specific action labels. This work has already garnered significant attention in the community, with early citations reflecting its impact. Xue's contributions are particularly notable for addressing the scalability and generalization problems that have long hindered real-world robot learning, positioning him as a key innovator in the push toward more autonomous and data-efficient robotic systems.
Research Focus
Key Achievements
Top Papers
- 1Contrast, Imitate, Adapt: Learning Robotic Skills From Raw Human Videos3 citations · 2024