Zhexi Luo

I am a final-year undergraduate in Computer Science at Sun Yat-sen University, where I conduct research at ISEE@SYSU under the supervision of Wei-Shi Zheng and work with Yi-Lin Wei. Currently, I am a research intern at LinS Lab, advised by Lin Shao, where I work with Zhixuan Xu.

My current research focuses on learning generalizable dexterous manipulation from human demonstrations.

🏄 Hobbies: Surfing, hiking, badminton, skiing, backpacking, and spending time with my dog.

🎓 I am actively seeking Fall 2027 PhD positions. For relevant opportunities or any thoughts related to my research, feel free to reach out!

📧 Email / 🎓 Google Scholar / 💻 Github / 💬 WeChat

1 / 3

📰 News

2026.9 🎉 T(R,O) Grasp++ — CoRL 2026!
2026.9 🎉 AXIS — IROS 2026 PWMS!
2026.7 🎉 One paper — ECCV 2026 Oral!
2026.6 ✈️ Attended ICRA 2026 in Vienna, enjoying the beautiful scenery and meeting wonderful people!
2026.4 🚀 OmniDexGrasp code released! GitHub
2026.2 🎉 OmniDexGrasp — ICRA 2026!
2026.2 🎉 DriftTrace — IEEE T-IFS, 2026!
2026.1 🤝 Participated in the development of RoboVerse V2
2025.11 🔬 Started research internship at LinS Lab@NUS, advised by Lin Shao
2025.10 🤖 Attended CoRL 2025, met many fascinating robots!
2025.9 📝 Completed my first work in robotics! Media coverage
2025.4 🚀 Started my journey in robotics research

📚 Research

^* Equal contribution · ^† Corresponding author

	DexMimic: A Generalist Dexterous Hand–Object Tracking Policy via Value-Guided Reset Allocation Zhexi Luo, Zhixuan Xu, Rui Yang, Zeyu Xiao, Yiwen Hou, Lin Shao^† *Under Review · Sep. 2026* arXiv (under preparation) / project page One generalist policy for diverse dexterous skills from human demonstrations, with efficient learning and growing generalization as training scales.
	OmniDexGrasp: Generalizable Dexterous Grasping via Foundation Model and Force Feedback Yi-Lin Wei^, Zhexi Luo^, Yuhao Lin, Mu Lin, Zhizhao Liang, Shuoyu Chen, Wei-Shi Zheng^† *ICRA 2026* arXiv / project page / code Dexterous grasping across diverse user prompts, robot hands, and tasks, all within one generalizable framework.
	T(R,O) Grasp++: Cross-Embodiment Dexterous Grasping in Cluttered Environments at Scale Xin Fei^, Zhixuan Xu^, Huaicong Fang, Fangjun Chen, Mu Zhaoyu, Yang Rui, Zhexi Luo, Lin Shao^† *CoRL 2026* PDF / project page Dexterous grasping across robot hands in cluttered scenes, with performance improving as models and data scale.
	AXIS: A Growable Community-Driven Data Engine for Scalable Robot Manipulation AXIS Robotics Team *IROS2026@PWMS* arXiv / project page Community-driven robot learning from browser-collected demonstrations, with a continually growing pool of manipulation data.
	Generalizable Humanoid Loco-Manipulation via Spatial Perception and Decision-Making of Multi-Agent Large Model Zhizhao Liang^, Yi-Lin Wei^, Xuhang Chen^, Mu Lin, Yi-Xiang He, Zhexi Luo, Jun-Hui Liu, Kun-Yu Lin, Wei-Shi Zheng^† ECCV 2026 (Oral)* arXiv / project page Generalizable whole-body mobile manipulation for humanoid robots, without task-specific real-robot data.
	DriftTrace: Combating Concept Drift in Security Applications through Detection and Explanation Yuedong Pan, Lixin Zhao, Tao Leng, Zhexi Luo, Lijun Cai, Aimin Yu, Dan Meng^† *IEEE Transactions on Information Forensics and Security (T-IFS), 2026* IEEE Xplore Detecting and explaining concept drift in security applications, helping models adapt to changing data.