Yucheng Hu
Papers
1
Total Citations
9
H-Index
1
About
Yucheng Hu is a rising researcher at the forefront of embodied AI and robot learning, with a focus on bridging large vision-language models (VLMs) with low-level robotic control. Their most-cited work, "Improving Vision-Language-Action Model with Online Reinforcement Learning" (2025), tackles a critical bottleneck in the field: while supervised fine-tuning (SFT) on expert datasets yields powerful vision-language-action (VLA) models, these models often struggle to generalize beyond static demonstrations. Hu’s key contribution is pioneering an online reinforcement learning framework that enables VLA models to improve through trial-and-error interaction, directly addressing the distribution shift between training data and real-world deployment. This work has already garnered 9 citations in its first year, signaling its timely impact. By showing that large VLA models can be refined without costly human annotations, Hu is helping to democratize advanced robotic control. Their research sits at the intersection of computer vision, natural language processing, and reinforcement learning, offering a scalable path toward more adaptive, autonomous robots. For students and researchers, Hu’s work exemplifies how combining foundation models with online learning can unlock new capabilities in embodied intelligence.
Research Focus
Key Achievements
Top Papers
- 1