Yifei Li

李一飞 | CS PhD Student

Massachusetts Institute of Technology

I am a fifth-year PhD student at MIT CSAIL, advised by Prof. Wojciech Matusik. I received my bachelor’s degree in Computer Science from Carnegie Mellon University, where I was fortunate enough to be advised by Prof. Jessica Hodgins.

I was fortunate to receive the 2020 MIT Stata Family Presidential Fellowship. I’ve previously spent wonderful summers interning at Boston Dynamics AI Institute (2024), Meta Reality Labs (2023), NVIDIA (2022) and Meta AI (2020) as a research intern, Google (2019) as a SWE intern and Activision Blizzard (2018) working on the Call of Duty: Black Ops 4 game engine development.

Outside of research, I enjoy ballet, bouldering and piano. I am a member of the Harvard Ballet Company.

Interests

Computer Graphics
(Differentiable) Physical Simulation
Computational Design & Fabrication
Machine Learning

Education

Massachusetts Institute of Technology 2020 - Present

Ph.D. in EECS
Massachusetts Institute of Technology 2020 - 2022

M.S. in EECS
Carnegie Mellon University 2016 - 2020

B.S. in Computer Science , minor in Machine Learning

Academic Service

Reviewer: SIGGRAPH, SIGGRAPH Asia, CVPR, ICCV, ECCV, NeurIPS, ICLR, Computer Graphics Forum, Pacific Graphics

News

[2024/9] NeuralFluid is accepted to NeurIPS 2024.
[2024/5] Selected as WiGRAPH Rising Stars in Computer Graphics. See you in Denver at SIGGRAPH !
[2024/2] DiffAvatar is accepted to CVPR 2024 .
[2023/5] Starting research internship at Meta Reality Labs, working with Tuur Stuyck.
[2023/3] Paper on Animated Drawings is accepted by TOG and will be presented at SIGGRAPH 2023.
[2022/12] AnisotropicStokes is covered by MIT News and spotlighted on the MIT homepage.
[2022/10] The source code for DiffCloth is released. Feel free to contact me for questions.
[2022/8] AnisotropicStokes is accepted to the Journal Track of SIGGRAPH Asia 2022.
[2022/8] Gave a talk at GAMES Webinar (in Chinese).
[2022/5] Joining NVIDIA’s simulation technology group as a research intern this summer.
[2022/3] Gave a talk at Toronto Geometry Colloquium.
[2022/3] DiffCloth is accepted by TOG. JoinABLe is accepted by CVPR 2022.

Publications

NeuralFluid: Nueral Fluidic System Design and Control with Differentiable Simulation

Yifei Li, Yuchen Sun, Pingchuan Ma, Eftychios Sifakis, Tao Du, Bo Zhu, Wojciech Matusik

Conference on Neural Information Processing Systems (NeurIPS) 2024 September 2024

Neural control and design of complex fluidic systems with dynamic solid boundaries using differentiable simulation.

DiffAvatar: Simulation-Ready Garment Optimization with Differentiable Simulation

Yifei Li, Hsiao-yu Chen, Egor Larionov, Nikolaos Sarafianos, Wojciech Matusik, Tuur Stuyck

Computer Vision and Pattern Recognition (CVPR) 2024 December 2023

Physics-based digital avatar body and garment shape and material joint optimization using differentiable simulation

A Method for Automatically Animating Children’s Drawings of the Human Figure

Harrison Jesse Smith, Somya Jain, Yifei Li, Qingyuan Zheng, Jessica Hodgins

ACM Transactions on Graphics (presented at SIGGGRAPH 2023) February 2023

A framework for bringing children’s drawings of human figures to life. Try out our Public Demo (which was used by millions of users) to animate your own drawing, or explore our Dataset of ~180,000 animated drawings for your own research.

Fluidic Topology Optimization with an Anisotropic Mixture Model

Yifei Li, Tao Du, Sangeetha Grama Srinivasan, Kui Wu, Bo Zhu, Eftychios Sifakis, Wojciech Matusik

ACM Transactions on Graphics (Proceedings of SIGGRAPH Asia 2022) September 2022

A topology optimization method with an anisotropic and differentiable constitutive model to design Stokes-flow fluidic devices.

DiffCloth: Differentiable Cloth Simulation with Dry Frictional Contact

Yifei Li, Tao Du, Kui Wu, Jie Xu, Wojciech Matusik

ACM Transactions on Graphics (presented at SIGGGRAPH 2022) April 2022

A differentiable cloth simulator that uses a fast and novel method for deriving gradients in PD-based cloth simulation with dry frictional contact.

JoinABLe: Learning Bottom-up Assembly of Parametric CAD Joints

Karl D.D. Willis, Pradeep Kumar Jayaraman, Hang Chu, Yunsheng Tian, Yifei Li, Daniele Grandi, Aditya Sanghi, Linh Tran, Joseph G. Lambourne, Armando Solar-Lezama, Wojciech Matusik

Computer Vision and Pattern Recognition (CVPR) 2022 March 2022

A learning-based method that assembles parts together to form joints.

Algorithmic Quilting Pattern Generation for Pieced-Quilts

Yifei Li, David E. Breen, Jim McCann, Jessica Hodgins

Graphics Interface 2019 May 2019

A computational framework to algorithmically design and fabricate quilts with decorative patterns.

Experience

Research Intern

Boston Dynamics AI Institute

Managed by David Watkins

Jun 2024 – Aug 2024 Cambridge, MA

Ongoing research project on robot gripper finger design and optimization.

Research Scientist Intern

Meta Reality Labs

Managed by Tuur Stuyck

May 2023 – Sep 2023 Sausalito, CA

Research on physics-based digital avatar body and garment shape and material joint optimization using differentiable simulation, which resulted in a paper that is currently in submission.

Research Intern

NVIDIA

Managed by Miles Macklin , Jonathan Leaf

May 2022 – Aug 2022 (remote) Cambridge, MA

Research intern in the Simulation Technology group, where I worked on neural surrogate simulation model for cloth-body dynamics.

Research Intern

Facebook AI Research

Managed by Jessica Hodgins

May 2020 – Aug 2020 Pittsburgh, PA

Research on animating children’s humanoid sketches. The project resulted in a publication in ACM Transactions on Graphics, a public demo with millions of users, which in turn resulted in an annotated dataset of ~180,000 children drawings.

Software Engineering Intern

Google (GeoAR Team)

Managed by Carlos Correa

May 2019 – Aug 2019 Mountain View, CA

Designed, implemented and and created a demo for photorealistic rendering of geo-located outdoor and indoor AR objects. The concept of the demo is now revealed in Google Map immersive AR Navigation, where AR widgets are rendered realistically by estimating the lighting model of the environment.

Software Engineering Intern

Activision Blizzard

Managed by Michael Vance , Wade Brainerd

May 2018 – Aug 2018 Portland, ME

Implemented and optimized (~7ms/frame on PS4 ) a procedural grass system that runs on GPU for the core game engine of Call of Duty Black Ops 4, which was shipped with the game’s battle royale mode across all platforms in October 2018. You can see the procedural grass system in the BO4’s Battle Royale mode trailer.