Biography
Hi, my name is Fengyuan Liu, an incoming PhD student of The Chinese University of Hong Kong, Shenzhen, advised by Prof. Mengnan Du. I am an undergraduate student in University of Science and Technology of China.
Previously, I was a research intern at WeChat Vision Team, Tencent.
My research interests include Generative Modeling, Representation Learning and Data-Centric Learning.
In my spare time, I write some blogs about generative models, applied mathematics and PyTorch technics, which I will update simultaneously with the Chinese blogs on Zhihu (you can see the Chinese-Version blogs here).
Experience
Research Engineer Intern @ Tencent
Sep. 2025 - Dec. 2025
Visual Modeling
Intern Student @ University of Pennsylvania (Penn)
May 2025 - Sep. 2025
Visual Modeling
Education
Ph.D. @ The Chinese University of Hong Kong, Shenzhen (CUHK-SZ)
Sep. 2026 - Now
Data Science
BSc @ University of Science and Technology of China (USTC)
Sep. 2022 - Jul. 2026
Data Science
Publications
#equal contribution, *corresponding author.
Blogs
Flow Matching Principles
This article mainly introduces the background and mathematical principles of Flow Matching.
Read MoreRectified Flow Introduction
This article primarily introduces and interprets the seminal work on the Flow-Matching method.
Read MoreOptimal Transport in Generative Modeling --an Introduction
Optimal Transport (OT) theory originated from Gaspard Monge's 18th-century transportation problem, which sought to move mass...
Read More