I am currently a fifth year PhD candidate from the Department of Statistics and Data Science at Yale University, and I am so honored to be supervised by Prof. Harrison Zhou.
Previously, I obtained my bachelor’s degree from School of Mathematical Sciences, Peking University (2015.9-2019.6). I have wide research interests including statistics, data science and generative AI. My research goal is to design efficient and provable algorithms, to deepen the statistical understanding, and to provide new perspective for practical machine learning problems. Recently, my main research interest lies in score-based diffusion models.
Selected Publications
2024
- From Optimal Score Matching to Optimal Sampling
Zehao Dou, Subhodh Kotekal, Zhehao Xu, Harrison H. Zhou
Arxiv Preprint (Under Review by the Annals of Statistics) - Think Twice Before You Act: Improving Inverse Problem Solving With MCMC
Yaxuan Zhu, Zehao Dou, Haoxin Zheng, Yasi Zhang, Ying Nian Wu, Ruiqi Gao
Arxiv Preprint - Diffusion Transformer Captures Spatial-Temporal Dependencies: A Theory for Gaussian Process Data
Hengyu Fu, Zehao Dou, Jiawei Guo, Minshuo Chen, Mengdi Wang
Arxiv Preprint - Theory of Consistency Diffusion Models: Distribution Estimation Meets Fast Sampling
Zehao Dou, Minshuo Chen, Mengdi Wang, Zhuoran Yang
ICML 2024
2023
- Diffusion Posterior Sampling for Linear Inverse Problem Solving: A Filtering Perspective
Zehao Dou, Yang Song
ICLR 2024 - Learning Narrow One-Hidden-Layer ReLU Networks
Sitan Chen, Zehao Dou, Surbhi Goel, Adam R Klivans, Raghu Meka
COLT 2023
2022
- Rates of Estimation for High-dimensional Multi-reference Alignment
Zehao Dou, Zhou Fan, Harrison Zhou
Annals of Statistics (AoS) - Understanding Value Decomposition Algorithms in Deep Cooperative Multi-Agent Reinforcement Learning
Zehao Dou, Jakub Grudzien Kuba, Yaodong Yang
arxiv preprint
2021
- On the One-sided Convergence of Adam-type Algorithms in Non-convex Non-concave Min-max Optimization
Zehao Dou, Yuanzhi Li
arxiv preprint - Gap Dependent Bounds for Two-Player Markov Games
Zehao Dou, Zhuoran Yang, Zhaoran Wang, Simon S.Du
AISTATS 2022 - Diff–ResNets for Few-shot Learning – an ODE Perspective
Tangjun Wang, Zehao Dou, Chenglong Bao, Zuoqiang Shi
Transactions on Pattern Analysis and Machine Intelligence (TPAMI)
2020
- A Priori Estimates of the Generalization Error for Autoencoders
Zehao Dou, Weinan E, Chao Ma
ICASSP 2020 - Making Method of Moments Great Again? – How can GANs learn distributions
Yuanzhi Li, Zehao Dou
arxiv preprint
Before 2019
- Improving Word Embeddings for Antonym Detection Using Thesauri and Sentiwordnet
Zehao Dou, Wei Wei, Xiaojun Wan
NLPCC 2018 - Mathematical Analysis of the Adversarial Attacks
Zehao Dou, Stanley Osher, Bao Wang
arxiv preprint
Experiences
6/2022 - 9/2022: Worked as a student researcher at Google AI in the Lux Group advised by Ankur Parikh, Ran Tian, Lei Tao and Behnam Neyshabur. During the internship, I studied a new version of T5 encoder-decoder language pretraining models, and apply them to natural language generation and protein generation, showing much better performance.
- 9/2018 - 12/2018: Participated in an internship of Microsoft Research Asia in the machine learning group advised by Weidong Ma. Using reinforcement learning and imitation learning to solve combinatorial optimization and transportation management problem.
- 9/2018 - 6/2019: Researched, as a research assistant, the “environmental policies analysis” program and being in charge of coding and natural language processing. Collaborated with Dr. Yingdan Lu from Department of Communication, Stanford University and Prof. Xiao Tang from Tsinghua University.
Selected Awards and Scholarships
2020
- Selected as a Nathan Hale Associates Scholar for the 2020-2021 academic year.
- Awarded the Fan Family Fellowship for the 2020-2021 academic year.
- Silver Medal (rank 7/50000) in Alibaba World Mathematics Contest.
2019
- Received Outstanding Graduate Award of Peking University.
- Received an annual department scholarship, university scholarship and Excellent Students Awards every year during undergrad study. Awarded Guanghua Scholarship in Freshman, Fangzheng Scholarship in Sophomore and Yizheng Scholarship in Junior.
2017
- In S-T. Yau College Student Mathematics Contests, ranked 3rd and awarded silver medal in the “Applied Mathematics and Computational Mathematics” category, also awarded gold medal in Team Event (rank 1st).
2016
- First Prize in the Chinese Mathematics Competitions (rank 9/100000).
- Selected as a Degree 3 table tennis player.
- Awarded Degree 9 in Sudoku.
Before 2015
- Gold Medal in Chinese Mathematics Olympics, and being selected into the Team Selection Test.
- Being invited to Russian Mathematics Olympics, and winning Gold Medal.