Peng Wang
 |
Peng Wang |
About Me
Currently, I am an assistant professor at the Department of Artificial Intelligence and Department of Computer Science at the University of Macao. Before this, I am a postdoc research fellow advised by Professors Laura Balzano and Qing Qu at the University of Michigan. I got my Ph.D. under the supervision of Professor Anthony Man-Cho So at The Chinese University of Hong Kong.
Open Positions (Fall 2027)
I am actively seeking PhD students with a strong background in mathematics or programming. Research assistants and postdoc positions are also available. If you are interested in any of these positions, please email me at pengw@um.edu.mo with the subject line Application for the PhD/RA/Postdoc Position with your CV. Please visit here for more information! I am particularly interested in students working in the following research areas:
Nonconvex optimization theory of deep neural networks, including local geometry, landscape, and convergence
Memorization, generalization, and reasoning abilities of deep generative models, including continuous and discrete diffusion models and language models
Scalable and efficient algorithms for pretraining and fine-tuning foundation models
AI for science (AI4S)
Research Interests
Broadly speaking, my research interests lie in the intersection of optimization, deep learning, and artificial intelligence. Currently, I am devoted to understanding the mathematical foundations of deep learning and generative AI models, including diffusion models and large language models. A unifying theme of my research is to understand how low-complexity structures in data and representations interact with model architectures and over-parameterization to shape optimization and learning. This perspective enables me to move beyond worst-case analyses, explain model capabilities and behavior, and design efficient learning algorithms.
Feel free to email me if you are interested in my research. Remote collaboration is also welcome!
What's New
[September 2026] Our paper on the optimization theory of maximal coding rate reduction is accepted by NeurIPS 2026.
[August 2026] I will give a talk on diffusion models in Workshop at the Institute for Advanced Study in Mathematics in Hangzhou!
[August 2026] I will give a tutorial lecture on diffusion models in Workshop on Mathematical Foundations of AI Models at National University of Singapore!
[August 2026] Our NeurIPS 2026 tutorial proposal Principles of Deep Representation Learning via Low-Dimensional Models is accepted!
[August 2026] Our paper on the error bound of deep linear networks is accepted by Mathematics of Operations Research!
[July 2026] Our paper on the distillation of diffusion models with PDE constraints won Best Paper Runner-up Award (3/230) in ICML2026 FoGen Workshop!
[June 2026] Our poster on the analysis of loss landscape and error bound of deep linear networks won the Best Poster Award in 2026 International Workshop on Modern Optimization and Applications!
[June 2026] Our paper on the loss landscape of quadratic programs with orthogonality constraints and its application to HePPCA is posted!
[June 2026] Our paper on data-dependent concentration inequalities is posted!
[June 2026] Our paper on distriution learning of diffusion models is accepted by Journal of Machine Learning Research!
[June 2026] Three papers are accepted by ICML 2026 Workshop on Foundations of Deep Generative Models: Understanding Memorization, Generalization, and Reasoning! Thanks to all my collaborators.
[May 2026] Our paper on the loss landscape of the regularized deep linear networks is accepted by SIAM Journal on Optimization!
[May 2026] I received the Gold Reviewer Award of ICML 2026!
[Apr 2026] Our paper on the distillation of diffusion models is accepted by ICML 2026!
[Mar 2026] Our workshop on deep generative models is accepted by ICML 2026! Thanks to all organizers!
[Feb 2026] Our paper is accepted by IEEE Signal Processing Magazine Special Issue on the Mathematics of Deep Learning. Thanks to all my collaborators!
[Jan 2026] One paper is accepted by AISTATS 2026!
[Dec 2025] I will chair a session at the IEEE International Conference on Big Data in Macao!
[Nov 2025] I will give a talk at the Conference on Mathematics and AI Science in Dongguan, Guangdong!
[Oct 2025] I will give a talk at the HKUST(GZ) & UM Data Science Workshop at the University of Macau!
[Oct 2025] I will give a presentation at the seminar hosted by the Department of Statistics and Data Science at the Chinese University of Hong Kong!
[Sep 2025] One paper on understanding feature learning of diffusion models is accepted by NeurIPS 2025!
[Aug 2025] Our paper on understanding hierarchical representation learning of deep neural networks is accepted by Journal of Machine Learning Research!
[Aug 2025] Our new book Learning Deep Representations of Data Distributions is posted online! Thanks for my wonderful collaborators Dr. Sam Buchanan, Druv Pai, and Prof. Yi Ma!
[Jun 2025] One paper on the global loss landscape analyais of deep matrix factorization is posted!
[May 2025] One paper on understanding the mechenism of transformers has been accepted by ICML 2025!
Award
Best Paper Runner-up Award (3/230) in ICML2026 FoGen Workshop!
Best Poster Award in 2026 International Workshop on Modern Optimization and Applications
Gold Reviewer Award of ICML 2026
Rising Star Award at Conference on Parsimony and Learning (CPAL), 2024.
Teaching Courses
CISC7402 Mathematics for Artificial Intelligence (Fall 2025, 2026, Syllabus)
CISC3023 Machine Learning (Spring 2026, Syllabus)
Service
Program Chair of ICML Workshop on Foundations of Deep Generative Models: Understanding Memorization, Generalization, and Reasoning (2026)
Publication Chair of CPAL (2027)
Area Chairs of ICML (2027), ICLR (2026, 2027), CPAL (2024-Present)
Reviewers of SIAM Journal on Optimization, Operation Research, Mathematics of Operation Research, Journal of the American Statistical Association, Journal of Machine Learning Research, IEEE Transactions on Information Theory, Nature Machine Intelligence, IEEE Transactions on Pattern Analysis and Machine Intelligence, IEEE Transactions on Image Processing, Information and Inference: A Journal of the IMA, Journal of Global Optimization, etc.
Reviewers of ICML (2021-2026), ICLR (2023-2025), NeurIPS (2022-Present), COLT (2026)