
Andi Han
Lecturer at USYD | Visiting Scientist at RIKEN
Postdoctoral Researcher at RIKEN AIP, 2023-2025
PhD at USYD, 2020-2023
Email: andi[dot]han[at]sydney[dot]edu[dot]au
jm3andy[at]gmail[dot]com
Github | Google Scholar | LinkedIn
About Me
Hi! I am currently a Lecturer (equiv. Assistant Professor) in the School of Mathematics and Statistics, University of Sydney.
Before joining USYD, I was a Postdoctoral Researcher at RIKEN AIP, Continuous Optimization Team, under the supervision of Prof. Akiko Takeda. I completed my PhD
in Business Analytics at USYD, where I was advised by Prof. Junbin Gao.
My research broadly covers large generative models, optimization (on manifolds),
efficiency of foundation models and
graph neural networks with applications to biology and chemistry.
I am looking for motivated students and interns. Feel free to reach out via email!
News
- [2025.05] Four papers accpeted to ICML 2025, and one paper accepted to COLT 2025.
- [2025.01] Four papers on Diffusion model feature learning, Transformer optimization, DMD for GNN, GNN hyperparameter tuning with diffusion model accpeted to ICLR 2025.
- [2024.12] We are organizing a workshop Deep Generative Model in Machine Learning: Theory, Principle and Efficacy at ICLR 2025! See more details on Call for papers!
- [2024.11] Created a GitHub repo on Riemannian optimization, compiling key papers, books, and resources.
- [2024.10] One paper on Protein sequence generation accpeted to IEEE BIBM 2024.
- [2024.09] Four papers on Riemannian bilevel optimization, Parameter and memory efficient pretraining, Theoretical comparisons between single and multi-modal contrastive learning, In context learning with multi-concept word semantics accpeted to NeurIPS 2024.