Biography
I am currently a Researcher at Tencent Hunyuan, focusing on Reinforcement Learning for 3D Foundation Models and Embodied AI.
Email: wenw27958@gmail.com
Experience
- July 2026 - Present: Researcher, Tencent Hunyuan (Qingyun Program)
- RL for 3D Generation Large Models
- Sept 2021 - June 2026: Ph.D., Chinese Academy of Sciences
- Institute of Automation, Chinese Academy of Sciences (CASIA) & School of Artificial Intelligence, University of Chinese Academy of Sciences (UCAS) & State Key Laboratory of Multimodal Artificial Intelligence Systems (MAIS)
- 2D/3D Multimodal Understanding and Generation, Robotic Mobile Manipulation, Large Multimodal Models
- May 2025 - Dec 2025: Research Intern, Tencent Hunyuan (Qingyun Program)
- RL for 3D Generation Large Models
- Aug 2024 - Feb 2025: Research Intern, Baidu
- Perception and Understanding of Large Vision-Language Models
- Sept 2017 - June 2021: B.E., Hunan University (HNU)
- School of Electrical and Information Engineering, Automation
Selected Publications
 | Mesh-Pro: Asynchronous Advantage-guided Ranking Preference Optimization for Artist-style Quadrilateral Mesh GenerationZhen Zhou†, Jian Liu†, Biwen Lei, Jing Xu, Haohan Weng, Yiling Zhu, Zhuo Chen, Junfeng Fan, Yunkai Ma, Dazhao Du, Song Guo, Fengshui Jing, Chunchao Guo (+ Equal contribution) IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), 2026 [PDF] |
 | QuadGPT: Native Quadrilateral Mesh Generation with Autoregressive ModelsJian Liu, Chunshi Wang, Song Guo, Haohan Weng, Zhen Zhou, Zhiqi Li, Jiaao Yu, Yiling Zhu, Jing Xu, Biwen Lei, Zhuo Chen, Chunchao Guo International Conference on Learning Representations (ICLR), 2026 [PDF] |
 | LIRA: Reasoning Reconstruction via Multimodal Large Language ModelsZhen Zhou, Tong Wang, Yunkai Ma, Xiao Tan, Fengshui Jing International Conference on Computer Vision (ICCV), 2025 |
 | Hunyuan3D Studio: End-to-End AI Pipeline for Game-Ready 3D Asset GenerationCore Contributor Technical Report, 2025 [PDF] |
 | EPRecon: An Efficient Framework for Real-Time Panoptic 3D Reconstruction from Monocular VideoZhen Zhou, Yunkai Ma, Junfeng Fan, Shaolin Zhang, Fengshui Jing, Min Tan IEEE International Conference on Robotics and Automation (ICRA), 2025 |
 | OBSeg: Accurate and Fast Instance Segmentation Framework Using Segmentation Foundation Models with Oriented Bounding Box PromptsZhen Zhou, Junfeng Fan, Yunkai Ma, Sihan Zhao, Fengshui Jing, Min Tan Machine Intelligence Research (MIR), 2025 |
 | Linear Gaussian Bounding Box Representation and Ring-Shaped Rotated Convolution for Oriented Object DetectionZhen Zhou, Yunkai Ma, Junfeng Fan, Zhaoyang Liu, Fengshui Jing, Min Tan Pattern Recognition (PR), 2024 |
 | WeldNet: A Deep Learning Based Method for Weld Seam Type Identification and Initial Point GuidanceYunkai Ma, Junfeng Fan, Zhen Zhou, Sihan Zhao, Fengshui Jing, Min Tan Expert Systems with Applications (ESWA), 2024 [PDF] |
 | DeepKP: A Robust and Accurate Framework for Weld Seam Keypoint Extraction in Welding RobotsSihan Zhao, Yunkai Ma, Junfeng Fan, Zhen Zhou, Hongliang Wang, Fengshui Jing, Min Tan IEEE Transactions on Instrumentation and Measurement (TIM), 2024 [PDF] |
Selected Projects
基于语言引导实例三维重建的机器人移动操作第一作者 基于“手-眼-足-脑”高度协同的具身移动操作机器人，首先根据涉及隐式及长程空间关系推理的语言指令进行目标实例三维重建，然后自主完成移动操作任务，例如视频中展示了机器人连续完成“1. Put the textbook on the lumbar-support chair by my wooden table. 2. Additionally, fetch a drink (a small one is fine) and place it on the table. 3. Lastly, put the digestive medicine on the lumbar-support chair too.”三个任务。完整视频demo见 https://www.bilibili.com/video/BV1HiVH6NEBW/ | |
 | 国家商用飞机制造工程-基于双目结构光视觉的机器人末端法向找正技术研究第一核心完成人 完成国家商用飞机制造工程-”基于双目结构光视觉的机器人末端法向找正技术研究”项目的全部内容，包括完成机器人系统的硬件（机械设计与组装、电路设计与焊接）和软件（基于深度学习的双目结构光视觉感知算法、机器人规划与控制算法、前端人机交互界面）的全流程设计与测试。 |
Service
- Reviewer: CVPR, ICCV, ICML, ICRA, IROS, TNNLS, TIP, TII, TIE, TIM