Yinxi Wang | 王印熹

Hi there! I'm a second-year undergraduate student at Peking University, majoring in Artificial Intelligence. Currently, I am a research intern at the HMI Lab at Peking University, advised by Prof. Shanghang Zhang. My research focuses on humanoid loco-manipulation and dexterous manipulation toward generalizable physical intelligence.

Email / Scholar / Twitter / Github

News

[2026.10] Introducing RepresentNav: see what I need, dream what I want.
[2026.10] Introducing Physical Words: a vocabulary for latent physical reasoning.
[2026.06] Introducing LaST-HD: learning physical reasoning from human demonstrations.
[2026.04] Introducing LaST-R1: reinforcement learning for adaptive physical reasoning in robot manipulation.
[2026.01] Joined the HMI Lab as a research intern.

Selected Publications

I study how robots can coordinate whole-body movement and skillful object interaction, with a focus on humanoid loco-manipulation and dexterous manipulation.

^∗ Equal contribution Project lead

	See What I Need, Dream What I Want: A Navigation Foundation Model on a Self-Constructed Representation Space Jiawei Chen^, Yinxi Wang^∗†, Zihan Yang^, Meijun Gao^, Yaowei Liu^, Shuming Zhang^, Zecheng Li, Xiangju Mi, Yuxuan Liu, Runzhong Huang, Tianlin Li, Xun Zhang, Zhengyang Wang, Xu Miao, Tong Yang arXiv preprint*, 2026 Project / Paper RepresentNav learns what to see and predicts the world ahead, enabling goal-directed navigation, instruction following, and visual tracking in cluttered, dynamic environments.
	Physical Words: A Vocabulary for Structuring Latent Reasoning in Robot Manipulation Jiaming Liu^, Xinghao Zhu^, Yiming Zhang^, Shanghang Zhang^, Yinxi Wang, Yichu Yang, Zhuoyang Liu, Dantong Niu, Soha Pouya, Juekun Li, Yan Chang arXiv preprint, 2026 arXiv / Project Connects atomic action words with visual and tactile dynamics to support precise, contact-rich robot manipulation across bimanual and dexterous-hand systems.
	LaST-HD: Learning Latent Physical Reasoning from Scalable Human Data for Robot Manipulation Jiaming Liu^, Yinxi Wang^, Chenyang Gu^, Siyuan Qian^, Xiangju Mi^, Hao Chen, Jiawei Chen, Qingpo Wuwu, Xiaoqi Li, Nuowei Han, Yiming Zhang, Xuheng Zhang, Yang Yue, Yeqing Yang, Lei Wang, Peng Jia, Hao Tang, Shanghang Zhang arXiv preprint*, 2026 arXiv / Project / Code Aligns human-hand and robot demonstrations in a shared latent space through an action-conditioned world model, enabling low-cost, data-efficient transfer of physical reasoning to robot manipulation.
	LaST-R1: Reinforcing Robotic Manipulation via Adaptive Physical Latent Reasoning Hao Chen^, Jiaming Liu^, Zhonghao Yan^, Nuowei Han^, Renrui Zhang, Chenyang Gu, Jialin Gao, Ziyu Guo, Siyuan Qian, Yinxi Wang, Peng Jia, Shanghang Zhang, Pheng-Ann Heng NeurIPS, 2026 arXiv / Project / Code A reinforcement-learning framework that lets a VLA policy perform adaptive latent chain-of-thought reasoning before acting, sharply improving manipulation generalization and execution stability.

Education

Peking University
B.S. in Artificial Intelligence
2025.09 - Present

Research Experience

	NVIDIA 2026.07 - present Robotics
	Simplexity Robotics 2026.01 - 2026.09 Research on Embodied AI and Robot Manipulation
	Peking University HMI (Human Machine Intelligence) Lab 2026.01 - Present Embodied AI Research Advisor: Prof. Shanghang Zhang

Design adapted from Jon Barron's template.