About Me
I am a fifth-year CS PhD student at Columbia University, advised by Prof. Changxi Zheng. Before coming to Columbia, I received my MSc in Computer Science from University of Toronto in 2021, advised by Prof. David I.W. Levin, and my B.Eng. in Computer Science from Zhejiang University in 2019.
My research lies at the intersection of computer graphics/vision, generative AI, optimization, and physical simulation.
I develop machine learning and optimization techniques to make simulation of the physical world and creative workflows more efficient, accurate and reliable. I've had the fortune to intern at Adobe, Meta, NVIDIA, and Microsoft. My research is supported by a Roblox Graduate Fellowship.
✨ I'm currently on the job market for full-time opportunities starting in 2026. Feel free to contact me 😊
Publications
-
ViPS: Video-informed Pose Spaces for Auto-Rigged Meshes
In submission
Training a feed-forward diffusion model that distills motion generative priors from video models and VLMs into a plausible pose space for auto-rigged meshes.
Trust-Region Eigenvalue Filtering for Projected Newton
SIGGRAPH Asia 2024 (Conference Track)
Paper (38MB) | Paper (10MB) | Project | Code
A two-line change that stabilizes and accelerates the physical simulation of deformable objects using the idea of spectral trust regions.
Stabler Neo-Hookean Simulation: Absolute Eigenvalue Filtering for Projected Newton
SIGGRAPH 2024 (Conference Track)
Paper (65MB) | Paper (15MB) | Project | Code
A one-line change that makes near-incompressible material simulation more stable and efficient.
Implicit Neural Spatial Representations for Time-dependent PDEs
ICML 2023
Paper (37MB) | ArXiv (6MB) | Project | Code
Solving time-dependent PDEs by evolving an implicit neural spatial representation over time.
Local Deformation for Interactive Shape Editing
SIGGRAPH 2023 (Conference Track)
Paper (81MB) | ArXiv (8MB) | Project | Code
A localized physical deformation method for artifact-free interactive shape editing (used in Adobe Fresco 🎨).
The Shape Matching Element Method: Direct Animation of Curved Surface Models
SIGGRAPH 2021
Paper (80MB) | Paper (13MB) | Project | Video
A meshless method for direct physics-based animation of volumetric NURBS models.
Chordal Decomposition for Spectral Coarsening
SIGGRAPH Asia 2020
Paper (81MB) | ArXiv (18MB) | Project | Talk | Code
A solver reducing geometric operators while preserving low-frequency spectra and controllable sparsity.
Teaching
-
Introduction to Optimization for Simulation
Honglin Chen
Symposium on Computer Animation, 2024
How to pronounce my name?
My first name is pronounced as "Hong-Lin". "Hong" is pronounced the same way as the first half of "Hong Kong", and "Lin" is pronounced the same as "Lynn".
I am happy to accept any variation of the pronunciation of my name, so feel free to use your favorite approximation. :)
Outside of work, I enjoy long-distance running, drawing, photography, visiting museums and art exhibitions, and exploring the world. 💫