About Me

I am a researcher at AI Theory Group of Huawei Noah’s Ark Lab. I obtained my Ph.D. degree in Probability and Mathematical Statistics from Academy of Mathematics and Systems Science, Chinese Academy of Sciences in 2022, under the supervision of Prof. Zhi-Ming Ma. Before that, I received my Bachelor’s Degree in Math from Central China Normal University in 2017.

Hiring: I am looking for research interns in Huawei Noah’s Ark Lab with strong machine learning and mathematical background. Please drop me an email if you are interested in machine learning theory, optimization, or generative model. Base: Beijing.

Research Interests

Optimization Theory

Statistical Learning

Deep Learning

Working Experience

2018.9-2020.3 Research Intern, Machine Learning Group of Microsoft Research Asia, supervisor Wei Chen

2020.4-2021.2 Research Intern, Speech & Language Processing Group of Huawei Noah’s Ark Lab, supervisor Lu Hou

2021.2-2022.5 Research Intern, AI Theory Group of Huawei Noah’s Ark Lab, supervisor Zhenguo Li

Publications

“*” means equally contribution, “†” means corresponding author

Continuous-time Riemannian SGD and SVRG Flows on Wasserstein Probabilistic Space
Mingyang Yi, Bohan Wang
Preprint

On the Generalization of Diffusion Model
Mingyang Yi, Jiacheng Sun, Zhenguo Li
Preprint

SA-Solver: Stochastic Adams Solver for Fast Sampling of Diffusion Models
Shuchen Xue, Mingyang Yi†, Weijian Luo, Shifeng Zhang, Jiacheng Sun, Zhenguo Li, Zhi-Ming Ma 37th Conference on Neural Information Processing Systems (NeurIPS 2023)

Breaking Correlation Shift via Conditional Invariant Regularizer
Mingyang Yi, Ruoyu Wang, Jiacheng Sun, Zhenguo Li, Zhi-Ming Ma
11th International Conference on Learning Representations (ICLR 2023)

Towards the Generalization of Contrastive Self-supervised Learning
Weiran Huang*, Mingyang Yi*, Xuyang Zhao*, Zihao Jiang
11th International Conference on Learning Representations (ICLR 2023)

Characterization of Excess Risk for Locally Strongly Convex Population Risk
Mingyang Yi*, Ruoyu Wang*, Zhi-Ming Ma
36th Conference on Neural Information Processing Systems (NeurIPS 2022)

Accelerating Training of Batch Normalization: A Manifold Perspective
Mingyang Yi
38th Uncertainty in Artificial Intelligence (UAI 2022).

Out-of-distribution Generalization with Causal Invariant Transformations
Ruoyu Wang*, Mingyang Yi*, Zhitang Chen, Shengyu Zhu
40th Computer Vision and Pattern Recognition (CVPR 2022)

Improved OOD Generalization via Adversarial Training and Pre-training
Mingyang Yi, Lu Hou, Jiacheng Sun, Lifeng Shang, Xin Jiang, Qun Liu, Zhi-Ming Ma
38th International Conference on Machine Learning (ICML 2021)

Reweighting Augmented Samples by Minimizing the Maximal Expected Loss
Mingyang Yi, Lu Hou, Lifeng Shang, Xin Jiang, Qun Liu, Zhi-Ming Ma
9th International Conference on Learning Representations (ICLR 2021)

BN-invariant Sharpness Regularizes the Training Model to Better Generalization
Mingyang Yi, Huishuai Zhang, Wei Chen, Zhi-Ming Ma, Tie-Yan Liu
29th International Joint Conference on Artificial Intelligence (IJCAI 2019)

Stablize Deep ResNets with A Sharp Scaling \tau
Huishuai Zhang, Da Yu, Mingyang Yi, Wei Chen, Tie-Yan Liu
Journal of Machine Learning

Improving Deep Learning by Regularized Scale-Free MSE of Representations
Xufang Luo, Mingyang Yi, Yunhong Wang
26th International Conference on Neural Information Processing (ICONIP 2019)

Awards

AMSS Presidential Scholarship, 2019

Excellent Intern at Huawei Noah’s Ark Lab, 2021

Chinese National Scholarship, 2021

ZhuLiYueHua Scholarship of the Chinese Academy of Sciences, 2022

Outstanding PhD Thesis of the Chinese Academy of Sciences, 2023

Academic Services

Conference Reviewer: ICML, ICLR, NeurIPS, AISTATS

Journal Reviewer: IJCV, TNNLS