Hi! I am a undergrad senior at UC Berkeley studying CS and applied math. Currently, I am working on 3D computer vision research with Prof. Angjoo Kanazawa. My interest is at the intersection of vision and graphics. Recently, I am particularly interested in volume rendering and neural representations.
Over the past few years I have explored several distinct areas of computer science. Last spring, I tried my hands on some problems in complexity theory, specifically in query complexity, advised by Prof. Avishay Tal. Please check out this table. Before that, I worked on some human and hand pose tracking at the FHL Vive Center and anomaly detection models for SETI.
Prior to Berkeley, I lived in Vancouver, Canada for many years. I was born in Zhejiang province, China.
Summer 2021 | Incoming Research Intern, Adobe Research
Advisor: Oliver Wang
Fall 2020- | Undergraduate Researcher, Kanazawa AI Research @ BAIR
Working on Neural Radiance Fields
Fall 2019 | Teaching Assistant, CS 61A @ UC Berkeley
Taught discussion and lab. Responsible for the traditional Hog Contest for several semesters.
Summer 2019 | SWE Intern, Google
Built banking features for Google Assistant
Spring 2019 | URAP Apprentice, SETI: Breakthrough Listen
Created visualization tools and explored ML anomaly detection models for Search for Extraterrestrial Intellegence (SETI) working with Dr. Gerry Zhang.
Fall 2017-Fall 2019 | Research Assistant, FHL Vive Center
Worked on human reconstruction and hand detection associated with OpenARK
PlenOctrees for Real-time Rendering of Neural Radiance Fields
We introduce a method to render Neural Radiance Fields (NeRFs) in real time using PlenOctrees, an octree-based 3D representation which supports view-dependent effects. Our method can render 800x800 images at more than 150 FPS, which is over 3000 times faster than conventional NeRFs. We do so without sacrificing quality while preserving the ability of NeRFs to perform free-viewpoint rendering of scenes with arbitrary geometry and view-dependent effects. Real-time performance is achieved by pre-tabulating the NeRF into a PlenOctree. In order to preserve view-dependent effects such as specularities, we factorize the appearance via closed-form spherical basis functions. Specifically, we show that it is possible to train NeRFs to predict a spherical harmonic representation of radiance, removing the viewing direction as an input to the neural network. Furthermore, we show that PlenOctrees can be directly optimized to further minimize the reconstruction loss, which leads to equal or better quality compared to competing methods. Moreover, this octree optimization step can be used to reduce the training time, as we no longer need to wait for the NeRF training to converge fully. Our real-time neural rendering approach may potentially enable new applications such as 6-DOF industrial and product visualizations, as well as next generation AR/VR systems. PlenOctrees are amenable to in-browser rendering as well; please visit the project page for the interactive online demo.
pixelNeRF: Neural Radiance Fields from One or Few Images
We propose pixelNeRF, a learning framework that predicts a continuous neural scene representation conditioned on one or few input images. The existing approach for constructing neural radiance fields involves optimizing the representation to every scene independently, requiring many calibrated views and significant compute time. We take a step towards resolving these shortcomings by introducing an architecture that conditions a NeRF on image inputs in a fully convolutional manner. This allows the network to be trained across multiple scenes to learn a scene prior, enabling it to perform novel view synthesis in a feed-forward manner from a sparse set of views (as few as one).
The Breakthrough Listen Search for Intelligent Life: Data Formats, Reduction and Archiving
Breakthrough Listen is the most comprehensive and sensitive search for extraterrestrial intelligence (SETI) to date, employing a collection of international observational facilities including both radio and optical telescopes. During the first three years of the Listen program, thousands of targets have been observed with the Green Bank Telescope (GBT), Parkes Telescope and Automated Planet Finder. At GBT and Parkes, observations have been performed ranging from 700 MHz to 26 GHz, with raw data volumes averaging over 1PB / day... In this paper, we describe the hardware and software pipeline used for collection, reduction, archival, and public dissemination of Listen data.
- Math 128A: Numerical Analysis
- Math 185: Complex Analysis
- CS 184: Computer Graphics
- CS 294-173: Learning for 3D Vision
- Math 202A: Intro to Topology and Analysis
- CS 271: Randomized Algorithms
- CS 294-92: Analysis of Boolean Functions
- EE 229A: Information Theory and Coding
- Math 104: Real Analysis
- Math 113: Abstract Algebra
- CS 61C: Great Ideas in Computer Architecture
- CS 294-137: Virtual Reality and Immersive Computing
- CS 375: CS Education
- CS 270: Graduate Algorithms
- EECS 126: Probability and Random Processes
- CS 172: Computability and Complexity
- CS 170: Algorithms
- CS 189: Machine Learning
- Math 110: Linear Algebra
- Math 53: Multivariable Calculus
- EE 16B: Designing Information Devices and Systems II
- CS 61B: Data Structures
- Math 54: Linear Algebra and Differential Equations
- CS 61A: Structure and Interpretation of Computer Programs
- CS 70: Discrete Mathematics and Probability Theory