Yimou Wu

Hello! I'm Yimou (Kevin) Wu, a M.Phil. student at CUHK, advised by Prof. Zheng Li. I study how robots can perceive, understand, and interact with the physical world — bridging geometry-grounded perception, memory-augmented learning, and real-time control. My research focuses on embodied AI and robot learning, with an emphasis on scene representation, spatial memory, and action.

Previously, I received my B.Eng. from CUHK, and I was also a Visiting Student at UC Berkeley.

CV / Email / GitHub / Google Scholar / X / LinkedIn

Publications

* equal contribution † corresponding author Representative papers are highlighted.

VPGGT: Deep Prior Learning for Embodied Perception

Yimou Wu, Jiaxin Guo, Yun-hui Liu^†, Zheng Li^†

arXiv 2026

Website arXiv Code

OmniPiano: Diverse Dexterous Piano-Playing Challenges for Standard, Robust, Safe, and Multi-Agent RL

Fan Xu, Fan Yang, Yifan Xu, Ruiyuan Zhang, Yimou Wu, Nuo Yi, Muning Wen, Guang Chen, Costas Spanos, Shangding Gu^†

arXiv 2026

Website arXiv Code

RoboMemory: A Brain-inspired Multi-memory Agentic Framework for Interactive Environmental Learning in Physical Embodied Systems

Mingcong Lei^*, Honghao Cai^*, Yuyuan Yang^*, Yimou Wu^*, Jinke Ren, Zezhou Cui, Liangchen Tan, Junkun Hong, Gehan Hu, Shuangyu Zhu, Shaohan Jiang, Ge Wang, Junyuan Tan, Zhenglin Wan, Zheng Li, Zhen Li, Shuguang Cui, Yiming Zhao, Yatong Han

CoRL 2026 (NeurIPS 2025 Workshop)

Website Paper

Particle importance analysis: A fast localization method under indeterminate magnetic sources

Ruoyu Xu, Yichong Sun, Yimou Wu, Yehui Li, Zheng Li

IJRR 2026

Paper Video

Neural Network Aided Kalman Filtering with Model Predictive Control Enables Robot-Assisted UAV Recovery and Suspended Payload Capture on a Wavy Surface

Yimou Wu, Mingyang Liang, Chongfeng Liu, Zhongzhong Cao, Ruoyu Xu^†, Huihuan Qian^†

arXiv 2025

Paper Video

This work was completed but has not been submitted to any conference or journal. The decision reflects timing constraints and considerations among the authors. The arXiv version is a preliminary draft containing unverified sections; the released video and methodology are provided for reference only.

Publications

Blog

Teaching

Awards & Honors