Jordan Lin

👋🏼  My name is lín林kuān寬héng恒. I go by Jordan and publish under Kuan Heng Lin.

I'm a first-year PhD student at Columbia University, where I am fortunate to be advised by Aleksander Hołyński and Micah Goldblum. I am interested in computer vision, multimodal generative models, and machine learning.

I was a Research Intern at Netflix Eyeline Labs, where I worked on 4D-grounded video generation and scene editing with Ning Yu. I was also a Research Engineering Intern at Snap Research, where I worked on 2D parallelism for scaling up video diffusion training (Snap Video V2) with Willi Menapace.

I did my undergrad at UCLA in Computer Science (+ Math minor), during which I had a great time working with Bolei Zhou and Sicheng Mo on controllable image generation.

I think pixels are cool, and I do 📷 film photography and used to do 🎥 filmmaking.

Research

Vista4D: Video Reshooting with 4D Point Clouds
Kuan Heng Lin, Zhizheng Liu, Pablo Salamanca, Yash Kant, Ryan Burgert, Koichi Namekata, Yiwei Zhao, Bolei Zhou, Micah Goldblum, Paul Debevec, Ning Yu
CVPR 2026 Highlight
paper website code
Reshoot a video by moving the camera and manipulating the scene in a 4D point cloud. Robust to real-world 4D reconstruction artifacts to bridge the train-inference gap and enable cool applications.
Ctrl-X: Controlling Structure and Appearance for Text-To-Image Generation Without Guidance
Ctrl-X: Controlling Structure and Appearance for Text-To-Image Generation Without Guidance
Kuan Heng Lin*, Sicheng Mo*, Ben Klingher, Fangzhou Mu, Bolei Zhou
NeurIPS 2024
paper website code
Training-free, guidance-free control of structure and appearance in text-to-image diffusion models via semantically-weighed feature injection. Structure control with any modality, fast, and plug-and-play.
FreeControl: Training-Free Spatial Control of Any Text-to-Image Diffusion Model with Any Condition
FreeControl: Training-Free Spatial Control of Any Text-to-Image Diffusion Model with Any Condition
Sicheng Mo*, Fangzhou Mu*, Kuan Heng Lin, Yanli Liu, Bochen Guan, Yin Li, Bolei Zhou
CVPR 2024
paper website code
Training-free spatial control for any text-to-image diffusion model by guiding generation to align with any input spatial condition via PCA-based feature coefficients.

Not just research

Teaching

I'm very passionate about CS & ML education!

Volunteering

Awards

Academic service