Seraph

Links

RoboMaster competition — on-robot debugging (© RoboMaster)

I’m Xian Nie, an undergraduate in Computer Science at Shanghai Jiao Tong University. My research interests include VLA models, tactile & force, and robotics.

At RethinkLab, I work with Prof. Junchi Yan and Prof. Xiaosong Jia on robust VLA policy learning. I also intern at TARS, on tactile and force-guided manipulation.

On Team Jiao Loong, I led sentinel decision-making and the competition simulator. Our team won the RoboMaster University Championship in 2024 and 2025.

GuidedVLA: Specifying Task-Relevant Factors via Plug-and-Play Action Attention Specialization

Xiaosong Jia*, Bowen Yang*, Zuhao Ge*, Xian Nie*, Yuchen Zhou*, Cunxin Fan*, et al. Equal contribution.

Robotics: Science and Systems (RSS), 2026

We guide VLA action decoders to focus on task-relevant factors by assigning dedicated attention heads to object grounding, temporal skill logic, and spatial geometry.

Project arXiv Code 机器之心

TacForeSight: Force-Guided Tactile World Model for Contact-Rich Manipulation

Yujie Zang*, Yuhang Zheng*, Xian Nie*, Yupeng Zheng, Shuai Tian, Songen Gu, Chen Gao, Zining Wang, Shuicheng Yan, Wenchao Ding

IEEE Robotics and Automation Letters (RA-L), 2026

A force-guided tactile world model for contact-rich robotic manipulation.

Project Paper Code 机器之心

IPR-1: Interactive Physical Reasoner

M Zhang, L Zhuo, T Tan, G Xie, X Nie, Y Li, R Zhao, Z He, Z Wang, J Cai, et al.

CVPR, 2026

An interactive physical reasoning model for embodied understanding and manipulation.

Project arXiv

Contact Trajectory Prompting: In-Context Transfer of Contact-Rich Behaviors from a Single Demonstration

Xian Nie*, Yujie Zang*, Yuhang Zheng*, Yupeng Zheng, Songen Gu, Wendi Chen, Chuan Wen, Cewu Lu, Wenchao Ding, Junchi Yan, Shuicheng Yan. Equal contribution.

IEEE International Conference on Robotics and Automation (ICRA), under review, 2026

We encode motion and contact information from a single demonstration into an ordered prompt, enabling a fixed policy to transfer contact-rich behaviors to unseen objects using online sensory feedback.

Project arXiv Code