Zhihan Zhu 朱旨函 (Zhihan Zhu)
Undergraduate Student
Dept. of Electronic and Electrical Engineering
Southern University of Science and Technology (SUSTech)
Email: zhuzh2023 [at] mail.sustech.edu.cn
Dept. of Electronic and Electrical Engineering
Southern University of Science and Technology (SUSTech)
Email: zhuzh2023 [at] mail.sustech.edu.cn
I am a third-year undergraduate at the Artificial Intelligence Lab of SUSTech, advised by Prof. Zhihai He (IEEE Fellow). My research primarily focuses on Generative AI, Diffusion Models, and Flow Models. I am currently exploring Video understanding and memory, and I also maintain a keen interest in Agentic RL.
我是南方科技大学人工智能实验室的三本科生，导师为 何志海教授（IEEE Fellow）。我的研究方向聚焦于 Generative AI、Diffusion Models 与 Flow Models。目前我正在探索 Video understanding and memory，同时也对 Agentic RL 充满兴趣。
Education 教育经历
-
Sept. 2023 – Expected June 2027
Southern University of Science and Technology (SUSTech)
B.S. in Electronic and Electrical Engineering 电子与电气工程系 工学学士
GPA: 3.86/4.0
Publications 论文
Latent Bias Alignment for High-Fidelity Diffusion Inversion in Real-World Image Reconstruction and Manipulation
Under Review · IEEE Transactions on Circuits and Systems for Video Technology, 2026
Runge-Kutta Approximation and Decoupled Attention for Rectified Flow Inversion and Semantic Editing
arXiv preprint arXiv:2509.12888, 2025
We propose a high-order inversion method for rectified flow models using a Runge–Kutta solver, enabling state-of-the-art fidelity and precise semantic control via Decoupled Diffusion Transformer Attention (DDTA).
提出了一种基于 Runge–Kutta 求解器的 Rectified Flow 模型高阶反演方法。通过引入解耦扩散 Transformer 注意力机制（DDTA），实现了极高的重建保真度与精确的语义控制。
Generative Semantic Coding for Ultra-Low Bitrate Visual Communication
arXiv preprint arXiv:2510.27324, 2025
Developed a generative semantic coding framework integrating image generation with deep compression via rectified flow models, pushing the boundaries of ultra-low bitrate communication.
开发了一种生成式语义编码框架，通过 Rectified Flow 模型将图像生成与深度压缩相结合，实现了超低比特率的视觉通信。
Patents 发明专利
- Zhihai He, Weiming Chen, Zhihan Zhu, et al. "Image Editing Method Based on Attention Decoupling..." "基于注意力解耦的图像编辑方法..." (CN 202511281645.1)
- Zhihai He, Weiming Chen, Yijia Wang, Zhihan Zhu. "Template-Replacement Image Compression & Reconstruction..." "基于模板替换的图像压缩与重建方法..." (CN 202511281423.X)
Honors & Awards 荣誉与奖项
- 2025 Gold Award, China International College Students’ Innovation Competition (Guangdong) 省金奖，中国国际大学生创新大赛（广东赛区）
- 2025 Outstanding Student, SUSTech 优秀学生，南方科技大学
- 2025 Second-Class Scholarship for Academic Excellence, SUSTech 二等奖学金（学业优秀），南方科技大学
Technical Skills 技术能力
Languages & Frameworks: Python, PyTorch, Hugging Face Diffusers, CUDA, NumPy, OpenCV
Models & Algorithms: Stable Diffusion, DiT, FLUX, RLVR, VeRL
Tools & Systems: Git, Linux, HPC (LSF, Multi-GPU Training), LaTeX, STM32
Models & Algorithms: Stable Diffusion, DiT, FLUX, RLVR, VeRL
Tools & Systems: Git, Linux, HPC (LSF, Multi-GPU Training), LaTeX, STM32