Hi, I'm Fang WU!

Welcome to my personal web page! I am a Ph.D. student at Stanford Computer Science, affiliated with the Arc Institute and Laboratory of Evolutionary Design. I have great fortune to work with Prof. James Zou and Prof. Brian Hie during the first-year rotation. Previously, I was a research engineer at Tsinghua University advised by Jinbo Xu. I obtained my Master's degree at Columbia University, advised by Dragomir Radev. It is a profound loss for me to lose Prof. Dragomir on March 29, 2023 (in memoriam). My research focuses on deep learning algorithms for scientific problems — in particular, 3D geometric networks, generative AI, domain adaptation, and other applications in chemistry and structural biology.

                 
Email: fangwu97 [at] stanford [dot] edu
Address: Stanford, CA, Stanford
Last update time: 2024.10.01

News and Highlights

   [2024/09] One paper on semi-supervised learning is accepted by NeurIPS 2024.
   [2024/03] One paper on GNN bottleneck is accepted by IEEE TKDE.
   [2024/03] One paper is accepted by ICML 2024 and another paper is accepted by IJCAI 2024.
   [2023/10] One paper is accepted by Nature Communications.
   [2023/09] One paper is accepted by NeurIPS 2023.
   [2023/06] One paper is accepted by Communications Biology.
   [2023/04] One paper and one co-authored paper are accepted by ICML 2023.
   [2023/02] One work on molecular domain adaptation is accepted by Cell Patterns.
   [2023/01] A talk on M2D2 invited by MILA and Valence.
   [2022/12] Two papers are accepted by AAAI 2023.
   [2022/10] One work on 3D pretraining is accepted by Advanced Science.

       Read more

Research Summary

* represents equal contribution and co-first authorship.

Graph Neural Networks (GNNs)

   Rethinking Explaining Graph Neural Networks via Non-parametric Subgraph Matching
   Fang Wu, Siyuan Li, Dragomir Radev, Stan Z. Li
   ICML 2023
   [Paper]    [Code]

   Discovering and Explaining the Representation Bottleneck of Graph Neural Networks from Multi-order Interactions
   Fang Wu*, Siyuan Li*, Dragomir Radev, Stan Z. Li
   IEEE TKDE
   [Paper]    [Code]

Generative Models for Drug Design

   A Hierarchical Training Paradigm for Antibody Structure-sequence Co-design
   Fang Wu, Stan Z. Li
   NeurIPS 2023, MLHC 2023
   [Paper]

   A Survey of Generative AI for de novo Drug Design: New Frontiers in Molecule and Protein Generation.
   Xiangru Tang*, Howard Dai*, Elizabeth Knight*, Fang Wu,, Yunyang Li, Tianxiao Li, Mark Gerstein
   Briefings in Bioinformatics
   [Paper]    [Github Repo.]

Biological Language Models (a.k.a. LLM for Science)

   Integration of Pre-trained Protein Language Mdels into Geometric Deep Learning Networks.
   Fang Wu, Liong Wu, Dragomir Radev, Jinbo Xu, Stan Z. Li
   Communications Biology
   [Paper]    [Code]

3D Geometric Deep Learning

   Molformer: Motif-based Transformer on 3D Heterogeneous Molecular Graphs.
   Fang Wu, Dragomir Radev, Stan Z. Li
   AAAI 2023
   [Paper]    [Code]

   Direct Prediction of Gas Adsorption via Spatial Atom Interaction Learning.
   Jiyu Cui*, Fang Wu*, Wen Zhang*, Lifeng Yang*, Jianbo Hu, Yin Fang, Peng Ye, Qiang Zhang, Xian Suo, Yiming Mo, Xili Cui, Huajun Chen, Huabin Xing
   Nature Communications
   [Paper]    [Code]

Pretraining & Semi-supervised Learning

   Surface-VQMAE: Vector-quantized Masked Auto-encoders on Molecular Surfaces.
   Fang Wu, Stan Z. Li
   ICML 2024
   [Paper]    [Code]

   A Semi-supervised Molecular Learning Framework for Activity Cliff Estimation.
   Fang Wu
   IJCAI 2024
   [Paper]    [Code]

   Instructor-inspired Machine Learning for Robust Molecular Property Prediction.
   Fang Wu*, Shuting Jin*, Siyuan Li, Stan Z. Li
   NeurIPS 2024
   [Paper]    [Code]

Domain Adaptation in Biochemistry

   Metric Learning-enhanced Optimal Transport for Biochemical Regression Domain Adaptation
   Fang Wu*, Nicolas Courty*, Shuting Jin*, Stan Z. Li
   Patterns
   [Paper]    [Code]

Molecular Dynamics (MD) Simulations

   Pre-training of Equivariant Graph Matching Networks with Conformation Flexibility for Drug Binding
   Fang Wu*, Shuting Jin*, Yinghui Jiang*, Xurui Jin, Bowen Tang, Zhangming Niu, Qiang Zhang, Xiangxiang Zeng, Stan Z. Li
   Advanced Science
   [Paper]    [Code]

   DiffMD: A Geometric Diffusion Model for Molecular Dynamics Simulations
   Fang Wu, Stan Z. Li
   AAAI 2023 (Oral)
   [Paper]

Others (e.g., NLP, CV)

   InsertGNN: Can Graph Neural Networks Outperform Humans in TOEFL Sentence Insertion Problem?
   Fang Wu, Stan Z. Li
   EMNLP 2024 Findings
   [Paper]    [Data]

   Architecture-Agnostic Masked Image Modeling: From ViT back to CNN
   Siyuan Li*, Di Wu*, Fang Wu, Zelin Zang, Kai Wang, Lei Shang, Baigui Sun, Hao Li, Stan Z. Li
   ICML 2023
   [Paper]    [Code]

   SemiReward: A General Reward Model for Semi-supervised Learning
   Siyuan Li*, Weiyang Jin*, Zedong Wang, Fang Wu,, Zicheng Liu, Cheng Tan, Stan Z. Li
   ICLR 2024
   [Paper]    [Code]

Education

   Stanford University, 2024-now
   • Ph.D. in Computer Science

   Columbia University, 2019-2021
   • Master of Science
   • GPA: 3.51/4.0

   Central University of Economics and Finance, 2015-2019
   • Bachelor of Science
   • GPA: 3.85/4.0

        Read more

Research Experience

Before joining Stanford University, I feel fortunate to be a research assistant/engineer advised by Jinbo Xu at Tsinghua University and Stan Z. Li at Westlake University, and recieved guidance as a visiting student from Huajun Chen, Xiang Bai and Danny Lan.

   Research Student (2024.09-now)
   • Arc Institute
   • Advised by Brian Hie

   Research Engineer (2022.08-2023.05)
   • Tsinghua University
   • Advised by Jinbo Xu

   Research Assistant (2021.11-2022.07)
   • Westlake University
   • Advised by Stan Z. Li


   Visiting Student (2021.03-2021.10)
   • Zhejiang University
   • Hosted by Huajun Chen

        Read more

Professional Services

Reviewer: ICLR 2024-2025, NeurIPS 2023-2024, AISTATS 2025, ML4H 2023, IEEE TNNLS

Acknowledgement

My study cannot be possible without the support from my awesome friends, mentors, and collaborators! Check out some of them:

Prof. Brian Hie, Dr. Francois Chaubard at Stanford University.
Prof. Dragomir Radev, Dr. Xiangru Tang at Yale University. R.I.P. to Dr. Dragomir.
Prof. Stan Z. Li, Prof. Danny Lan, Dr. Siyuan Li, Dr. Lirong Wu, Dr. Haitao Lin, Dr. Yufei Huang at Westlake University.
Prof. Jinbo Xu, Prof. Xianyun Zhan, Dr. Shikun Feng at Tsinghua University, and Prof. Wenbing Huang at Renmin University.
Prof. Huajun Chen, Prof. Qiang Zhang, Prof. Huabing Xing, Dr. Jiyu Cui at Zhejiang University.
Prof. Xiangxiang Zeng at Hunan University and Prof. Shuting Jin at Wuhan Technology University.
Prof. Nicolas Courty at University Bretagne Sud.
Prof. Buyong Ma, Prof. Shuangjia Zheng , Prof. Junchi Yan at Shanghai Jiaotong University
Prof. Xiang Bai at Huazhong University of Science and Technology.

        Read more

Aside from university collaborations, I also collaborated with many industrial AIDD companies, including MindrankAI, MoleculeMind, and Biomap

Dr. Zhangming Niu, Dr. Xurui Jin, and Dr. Yinghui Jiang at MindrankAI.
Dr. Xiaoyang Jing, Dr. Tenglong Wang, Dr. Wuwei Tan at MoleculeMind.
Prof. Le Song at BioMap.

        Read more