I am an assistant professor in the Department of Computer Science at The University of Hong Kong and The HKU Musketeers Foundation Institute of Data Science. I am generally interested in machine learning, stochastic optimization, and graph learning, with a special focus on the theoretical/empirical understanding (or Physics) of deep learning (especially foundation models). I am also particularlly interested in devloping the AI/ML methods for practical problems in other area, such as signal processing, intelligent transportation, and math problems.
Previously, I obtained my Ph.D. in the Computer Science department at the University of California, Los Angeles (UCLA), supervised by Prof. Quanquan Gu. I obtained my master degree in electrical engineering and bachelor degree in applied physics, both in Unversity of Science and Technology of China (USTC).
News
- Multiple openings for PhD, Postdoc, and RA. Please drop me an email with your CV and transcript (optional) if you are interesed in joining my research group. For interested PhD candidates in 2025, please submit your application in https://i.cs.hku.hk/~gradappl/index.html and inform me via email.
[2024-06] One paper on faster diffusion inference/sampling is accepted to SPIGM @ ICML workshop as an oral presentation and wins the best paper award!
[2024-05] One paper on faster non-log-concave sampling is accepted to COLT 2024.
[2024-05] Four papers on handling spurious correlation via better group classifier, transformer expressive power with varying layers, faster rate of stochastic proximal sampler, and benign overfitting for XOR data are accepted to ICML 2024.
[2024-04] One paper on continual learning for GNN is accepted to CoLLAs 2024.
- [2024-01] Three papers on SGD with large learning rate, scalable training of dynamic GNN, and finite sample in-context learning are accepted to ICLR 2024.
[2023-05] Our paper on the implicit bias of batch normalization is accepted to COLT 2023.
[2023-03] Two manuscripts on explaning the advantages of Mixup and Gradient Regularization in training neural networks are online.
[2023-03] I will serve as the Area Chair in NeurIPS 2023.
[2023-01] Our paper on the generalization separation between Adam and GD has been accepted by ICLR 2023.
[2022-09] Two papers accepted by NeurIPS 2022. The first paper studies the generalization of multi-pass SGD for over-parameterized least squres; the second paper demonstrates the power and limitation of pretraining-finetunning for linear regression with distribution shift.
[2022-08] Dr Difan Zou just joined HKU CS as an assistant professor.