Ziang Li

Ziang Li

CS PhD Student at Georgia Tech

Ziang Li (李子昂)

I am a CS PhD student at School of Computational Science and Engineering, Georgia Institute of Technology, advised by Prof. Yunan Luo. My research interests include computational biology and graph representation learning. Previously, I obtained a B.E. in computer science from Tsinhua University.

During my undergraduate, I was honored to have worked with Prof. Maria Brbic, Prof. Jure Leskovec at Stanford and Prof. Jie Tang at Tsinghua University. In the industry, I used to intern at Tencent Quantum Lab, working with Dr. Chang-Yu Hsieh and at Helixon, working with Prof. Jian Peng and Prof. Jianzhu Ma.

My first name is actually Zi’ang, pronounced as /z/ (pause) and /ʌŋ/ :). [Pronunciation]

Publications

Rethinking the Setting of Semi-supervised Learning on Graphs

Ziang Li *, Ming Ding *, Weikai Li, Zihan Wang, Ziyu Zeng, Yukuo Cen, Jie Tang
Proceedings of the 31st International Joint Conference on Artificial Intelligence (IJCAI) 2022 （full paper）(short oral)

Propose ValidUtil to explore the limit of over-tuning hyperparameters for semi-supervised learning on graphs. With ValidUtil, even GCN can easily get high accuracy of 85.8% on Cora.
To avoid over-tuning, we construct an i.i.d. graph benchmark (IGB) consisting of 4 datasets which is proved to be a more stable benchmark than previous datasets for semisupervised learning on graphs.
[pdf] | [code]

Modeling Protein Using Large-scale Pretrain Language Model

Yijia Xiao, Jiezhong Qiu, Ziang Li, Chang-Yu Hsieh, Jie Tang
The International Workshop on Pretraining: Algorithms, Architectures, and Applications (Pretrain@KDD) 2021

Pretrained the second largest protein language model in the world, Wen Su, which exceeded baseline in protein folding contact prediction by 39%.
[pdf] | [code]

Membership

  1. Stanford Undergraduate Visiting Research Program (UGVR): Chinese student visit Stanford and research in the summer. A professor accepts only one student from twelve candidates researching in corresponding field nationwide (1/12). I was advised by Prof. Jure Leskovec.
  2. Spark Innovative Talent Cultivation Program: Academic and scientific research talents training program hold by Tsinghua University, only two students major in computer science are selected each year (2/250).

Awards

  1. Tsinghua University Research Excellence Award (1/250), 2022
  2. Tsinghua University Research Excellence Award (1/250), 2021
  3. Tsinghua University Academic Excellence Award, 2020
  4. Tsinghua University Freshmen Scholarship (top 3%), 2019
  5. Chinese Chemistry Olympiad (Final) Gold Medal (17/500), 2018

More about me

  1. My favorite sport is basketball and I usually play as a power forward (PF). I can shoot the three, so I can also serve as a center in a small lineup. And I have been a fan of LeBron James for more than 10 years!
  2. I want to be an optimistic, cheerful and hard-working person with strong belief like Naruto Uzumaki.