👋 About Me

I am Zexin Feng, a Ph.D. student at the Mobility Transformation Lab, The University of Hong Kong (HKU), Hong Kong SAR, China, supervised by Prof. Xintao Yan. My research interests include Physical AI, Autonomous Driving, and Machine Learning.

Previously, I earned my B.Eng. in Computer Science from the Southern University of Science and Technology (SUSTech), Shenzhen, China (expected June 2025). During my undergraduate studies, I was a long-term member of the iMED Intelligent Medical Imaging Lab led by Prof. Jiang Liu, and I also worked with Prof. Xiaohui Xie at the University of California, Irvine on AI for healthcare.

If you are interested in my work, please contact me at zexinfeng@connect.hku.hk or mark4551124015@outlook.com.

📖 Educations

• Southern University of Science and Technology (SUSTech)
  • B.Eng. in Computer Science*, 09/2022 - 06/2025
• The University of Hong Kong (HKU)
  • PhD Student, 09/2025 - Present

🔬 Research

Physical AI & Robot Manipulation

Generative Simulation & Autonomous Driving

CVPR 2026 · SAD Workshop

Beyond Binary Metrics: Unveiling the Safety Illusion in Autonomous Driving Simulation

Zexin Feng, Lingyu Xiao, Xintao Yan. CVPR 2026 Workshop on Simulation for Autonomous Driving (SAD).

Paper · Project Page

• Composite Collision Metric (CCM) measures collision severity and tail risk beyond binary collision indicators.

Medical Imaging

ICASSP 2024

Flattening Singular Values of Factorized Convolution for Medical Images

Zexin Feng, N. Zeng, J. Fang, X. Wang, X. Lu, H. Meng, J. Liu

Conference

• Proposed SFConv for efficient CNN optimization, achieving 98% parameter compression while maintaining performance on public datasets.

🔥 News

• 2026.06:  🎉 Long-term Traffic Simulation via Structured Autoregressive Modeling was accepted to ECCV 2026. Project Page
• 2026.03:  🎉 Beyond Binary Metrics: Unveiling the Safety Illusion in Autonomous Driving Simulation was accepted to the CVPR 2026 Workshop on Simulation for Autonomous Driving (SAD). Project Page
• 2024.04:  🎉 Flattening Singular Values of Factorized Convolution for Medical Images was accepted to ICASSP 2024.

🎖 Honors and Awards

• Best Undergraduate Thesis Award (2025), for outstanding research performance.
• Best Graduate Student Award (2025).
• BYD Scholarship (2024).
• China National Scholarship (2024), Top 0.2%.
• School Motto Scholarship “Truth” Special Award (2024), awarded to students with outstanding academic performance.
• Outstanding Student (2023), Top 20%.
• Best Progress Award (2022), recognizing significant academic improvement.

🛠️ PROJECTS

• CS205 Cpp Program Design Libtensor Designed with C/C++ and CUDA, Zexin Feng, Yupeng Su, Xiaoqun Liu.

• CS301 Embedded System Touchscreen Framework STM32F103, Zexin Feng, Zongqi Yang, Chunhui Xu, Haoxian Liu.

• CS202 Computer Organization 5-Stage Pipelined MIPS CPU, Zexin Feng, Beiye Gou, Xiaoqun Liu.

• CS102 Java Programming JavaFX Game Engine - Reversi Game, Zexin Feng, Rui Ji

🧑‍🏫 Teaching

• Student Lecturer for C Programming Assisting Course
  SUSTech, CSE, Spring 2024
  Shenzhen, China

• Student Assistant for CS305 Computer Networks
  SUSTech, CSE, Fall 2023
  Shenzhen, China

💻 Experience

• Mobility Transformation Lab, The University of Hong Kong, Hong Kong SAR, China
  PhD Student, supervised by Prof. Xintao Yan. (09/2025-Present)

• iMED Intelligent Medical Imaging, Shenzhen, China
  Undergraduate Research Assistant, supervised by Prof. Jimmy Liu (05/2023-09/2023)

• Xiaohui Xie’s Lab, UC Irvine, USA
  Undergraduate Research Assistant, supervised by Prof. Xiaohui Xie (09/2024-01/2025)