Gang Liu

Gang Liu

Gang Liu / 刘网

Assistant Professor, Department of Data Science, Hong Kong Institute of AI for Science, City University of Hong Kong

Gang.Liu@cityu.edu.hk GitHub Google Scholar πŸ€— Hugging Face LinkedIn

Hi! I am an assistant professor at the City University of Hong Kong. I work on Generative AI, (Multimodal) Large Language Models, Foundation Models, and AI for Science. I will be recruiting multiple PhD students and Research Assistants. Please contact me at Gang.Liu@cityu.edu.hk if you are interested.

I received my Ph.D. from the University of Notre Dame, where I worked with Prof. Meng Jiang on Data-Centric Learning and Foundation Models for molecular discovery. My work has been published in KDD, NeurIPS, ICLR, IEEE TKDE, ACM TKDD, and Cell Reports Physical Science. Our work has been featured by MIT News, Notre Dame Engineering, and Snap Research. I am the lead or co-author of two books on deep learning for polymers and the creator of torch-molecule, an open-source toolkit for molecular discovery.

My awards include the IBM Ph.D. Fellowship, OpenAI Researcher Access Program, and Kaggle Competition Research Grant. I lead the NeurIPS 2025 Open Polymer Challenge with 10,000+ entrants / 50,000+ submissions from 100+ countries.

Latest News

2026

  • 2026 08 Joined City University of Hong Kong as an Assistant Professor. I will be traveling to KDD at Jeju. I am open to any conversation about research and collaboration opportunities!
  • 2026 01 Two papers accepted by ICLR 2026: (1) DemoDiff with Byte-Pair Encoding / In-Context Learning for molecules and (2) GeoBPE for protein structure tokenization.

2025

2024

2023

2022

Research by Area

View All Publications β†’
Active Generative & Foundation Models Data-Centric Learning Multimodal Learning Representation Learning Scientific Applications Books

1. Generative and Foundation Models for Inverse Molecular Design

We develop generative foundation models for multi-conditional, multi-modal, and synthesizable molecular design. We develop Graph Diffusion Transformers and multimodal large language models (Llamole) to tell us what a molecule looks like and how to synthesize the molecule (retrosynthetic planning).

Related Publications:

Multimodal Large Language Models for Inverse Molecular Design with Retrosynthetic Planning

Gang Liu, Michael Sun, Wojciech Matusik, Meng Jiang, Jie Chen

The International Conference on Learning Representations (ICLR), 2025 [paper] [code] [models] [MIT News]

Graph Diffusion Transformers for Multi-Conditional Molecular Generation

Gang Liu, Jiaxin Xu, Tengfei Luo, Meng Jiang

Conference on Neural Information Processing Systems (NeurIPS), 2024 (Oral) [paper][code]
Llamole Multimodal LLM Demo
Graph Diffusion Transformer

2. Data-Centric Learning for Virtual Screening

We use interpretable rationales (e.g., substructures in molecules) to overcome limited supervision, address imbalanced learning, and advance transfer learning.

Related Publications:

Data-Centric Learning from Unlabeled Graphs with Diffusion Model

Gang Liu, Eric Inae, Tong Zhao, Jiaxin Xu, Tengfei Luo, Meng Jiang

Conference on Neural Information Processing Systems (NeurIPS), 2023 [paper] [code]

Semi-Supervised Graph Imbalanced Regression

Gang Liu, Tong Zhao, Eric Inae, Tengfei Luo, Meng Jiang

Graph Rationalization with Environment-based Augmentations

Gang Liu, Tong Zhao, Jiaxin Xu, Tengfei Luo, Meng Jiang

ACM SIGKDD, 2022 [paper][code]

✨ Try our [online tool] for polymer property prediction!

Data-Centric Transfer Learning
Semi-Supervised Graph Imbalanced Regression
Graph Rationalization with Environment-based Augmentations

3. Translational Impacts

Our research leads to the discovery of new polymer materials for applications in sustainable gas separation. Our research reaches broader communities by open-source toolkits and competitions.

Related Publications:

Superior Polymeric Gas Separation Membrane Designed by Explainable Graph Machine Learning

Jiaxin Xu*, Agboola Suleiman*, Gang Liu*, Michael Perez, Renzheng Zhang, Meng Jiang, Ruilan Guo, Tengfei Luo

Cell Reports Physical Science, 2024 [paper][patent]

✨ New sustainable materials discovered by SGIR (KDD'23) and GREA (KDD'22) for heat-free gas separation! Featured by [Notre Dame News].

Related Competitions:

NeurIPS - Open Polymer Prediction 2025

Gang Liu, Sobin Alosious, Yuhan Liu, Eric Inae, Yihan Zhu, Renzheng Zhang, Jiaxin Xu, Ying Li, Tengfei Luo, Meng Jiang, ... and MarΓ­a Cruz

Accepted by NeurIPS 2025 Competition Track [website][Kaggle]
Cell Reports Physical Science Cover
open polymer challenge

Software & Toolkits

View All Software/Code β†’

Open-source tools and platforms developed for the research community

Awards & Honors

Professional Activities & Service

Media Coverage

Reviewer

  • 2026 ICLR
  • 2025 NeurIPS, KDD (Aug/Feb), ICLR, ICML, AAAI
  • 2024 NeurIPS, KDD, ICLR, ICML, AAAI, WWW, SDM
  • 2023 NeurIPS, KDD, ICML, AAAI
  • 2022 NeurIPS, ICML
  • Journals Pattern Recognition, IEEE TKDE, IEEE TCYB, Information Sciences, Digital Signal Processing

Mentorship

Experience

Appointment

2026 –
Assistant Professor
City University of Hong Kong

Education

2021 – 2026
Ph.D. in Computer Science and Engineering
University of Notre Dame, U.S.
Advisor: Prof. Meng Jiang
I also worked closely with Prof. Tengfei Luo on deploying ML models and developing foundation models for polymers.
2017 – 2021
B.E. in Software Engineering
Southwest University, China
Advisor: Prof. Yong Deng
Advisor: Prof. Fuyuan Xiao

Internships

05 – 08, 2024
Research Intern, MIT-IBM Watson AI Lab, Boston, U.S.
Worked on LLM+Graph models for molecules with Dr. Jie Chen
02/2024 - 05/2025
Research Intern with Extended Affiliation, Broad Institute of MIT and Harvard, Boston, U.S.
Worked on drug/molecular representation learning with cellular response data with Dr. Anne E. Carpenter and Dr. Shantanu Singh
05 – 08, 2023
Applied Scientist Intern, Amazon, Seattle, U.S.
Worked on Transformers with attribute inputs for sequential recommendations