Zhihan Zhu 朱旨函

Undergraduate Student, Department of Electronic and Electrical Engineering, SUSTech 南方科技大学电子与电气工程系 本科生

Email: 12312326 [at] mail.sustech.edu.cn

I am a third-year undergraduate student at the Artificial Intelligence Lab of Southern University of Science and Technology, supervised by Professor Zhihai He (IEEE Fellow). My research interests lie in Generative AI, Diffusion Models, and Flow Models. I am currently exploring video understanding and memory, and I am also interested in Agentic RL.

我是南方科技大学 电子与电气工程系 大三本科生，现于 人工智能实验室 开展科研工作，导师为 何志海教授 （IEEE Fellow）。我的研究方向主要包括 Generative AI、Diffusion Models 和 Flow Models。目前我正在探索视频理解与记忆，同时也对 Agentic RL 感兴趣。

GitHub Google Scholar Curriculum Vitae 个人简历
Zhihan Zhu

Education & Honors 教育背景与荣誉

Education 教育经历

Southern University of Science and Technology (SUSTech) 南方科技大学（SUSTech）

B.S. in Electronic and Electrical Engineering
Expected Graduation: June 2027
GPA: 3.86 / 4.0

电子与电气工程专业 工学学士
预计毕业时间：2027 年 6 月
GPA：3.86 / 4.0

Honors & Awards 奖项与荣誉

  • 2025 Gold Award, China International College Students’ Innovation Competition (Guangdong) 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 二等奖学金（学业优秀奖）

Publications 学术论文

Paper 3

Latent Bias Alignment for High-Fidelity Diffusion Inversion in Real-World Image Reconstruction and Manipulation

Weiming Chen, Qifan Liu, Siyi Liu, Yijia Wang, Zhihan Zhu, Zhihai He

Under Review · IEEE Transactions on Circuits and Systems for Video Technology, 2026

Under review.

审稿中。

Paper 1

Runge-Kutta Approximation and Decoupled Attention for Rectified Flow Inversion and Semantic Editing

Weiming Chen, Zhihan Zhu, Yijia Wang, Zhihai He

arXiv:2509.12888 (September 2025)

We propose a high-order inversion method for rectified flow models based on a Runge–Kutta solver, achieving strong reconstruction fidelity and controllable semantic editing with Decoupled Diffusion Transformer Attention.

本文提出了一种基于 Runge–Kutta 求解器的 rectified flow 高阶反演方法，并结合 Decoupled Diffusion Transformer Attention，实现了较高的重建保真度与可控语义编辑。

Paper 论文 Code 代码
Paper 2

Generative Semantic Coding for Ultra-Low Bitrate Visual Communication

Weiming Chen, Yijia Wang, Zhihan Zhu, Zhihai He

arXiv:2510.27324 (October 2025)

This work develops a generative semantic coding framework that integrates image generation and learned compression through rectified flow models for ultra-low bitrate visual communication.

本文提出了一种生成式语义编码框架，通过 rectified flow 模型融合图像生成与深度压缩，以支持超低比特率视觉通信。

Paper 论文 Code (Coming Soon) 代码（即将开源）

Patents 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 and Reconstruction...” “基于模板替换的图像压缩与重建方法...” CN 202511281423.X

Technical Skills 技术技能

Python PyTorch Diffusion Models Flow Models Hugging Face Diffusers CUDA NumPy OpenCV Linux Git LaTeX Stable Diffusion DiT FLUX Multi-GPU Training STM32 Matplotlib RLVR VeRL