Zilu Li

Hi! I'm a first-year CS PhD student at UC San Diego, co-advised by Prof. Ravi Ramamoorthi and Prof. Tzu-Mao Li. My research focus revolves around computer graphics, and neural fields. Previously, I was an undergrad at Cornell where I was fortunate to work with Prof. Bharath Hariharan, Prof. Steve Marschner, and Prof. Gordon Wetzstein.

Email: zil124 [at] ucsd [dot] edu

Email  /  Google Scholar  /  Twitter  /  Github

profile photo

Research

Neural Control Variates with Automatic Integration
Zilu Li*, Guandao Yang*, Qingqing Zhao, Xi Deng, Leonidas Guibas, Bharath Hariharan, Gordon Wetzstein
SIGGRAPH, 2024

We present a method that uses arbitrary neural network architectures as control variates with automatic differentiation to create unbiased, low-variance, and numerically stable Monte Carlo estimators for various problem setups.

Neural Cache for Monte Carlo Partial Differential Equation Solver
Zilu Li*, Guandao Yang*, Xi Deng, Christopher De Sa, Bharath Hariharan, Steve Marschner
SIGGRAPH Asia, 2023
[Page]

Using neural field caches to reduce variance in Monte Carlo PDE solvers. Our method excels when computing is limited and enables effective bias and variance trade-off.

BT^2: Backward-compatible Training with Basis Transformation
Yifei Zhou*, Zilu Li*, Abhinav Shrivastava, Hengshuang Zhao, Antonio Torralba, Taipeng Tian, Ser-Nam Lim
ICCV, 2023
[Paper]

Introducing a Basis Transformation Method to avoid backfilling in learning new representations for modern retrieval systems.

Miscellaneous

Outside of research, I spend most time rocking with my band Cornell Interlude and random city-walking:D

Teaching

I served as a teaching assistant for the following courses.

Spring 2024: CS 4670 Computer Vision

Spring 2023, Fall 2023: CS 4780 Machine Learning for Intelligent Systems


Website Template: Jon Barron's website.