Online Summer School of Deep Learning Theory


This online summer school aims to introduce the progress of deep learning theory in last several years, including mean field theory, neural tangent kernel, function space theory, implicit regularization, frequency principle etc. These interactions among different viewpoints may shed light on the development of deep learning theory.

Audience: Senior undergraduates, graduates, and young researchers.


July 16~July 22,2020


Online Zoom
Detailed Zoom Info could be found in the Schedule.

Application and Registration

Please register online.

Preference is given to, but not limited to, applicants with a basic understanding of ordinary differential equations and probability. No registration fee.




Date Time Title Lecturer Zoom ID
Jul 16, 2020 09:30-12:00 Accuracy and trainability of neural networks: a mean-field perspective on approximation and optimization Grant M. Rotskoff 962-308-15257
Jul 16, 2020 15:00-17:30 Neural tangent kernel: convergence and generalization in neural networks Arthur Jacot 681-823-42064
Jul 17, 2020 09:30-12:00 A function space theory and generalization error estimates for neural network models Chao Ma 912-899-29245
Jul 18, 2020 09:30-12:00 Towards understanding the implicit regularization in deep learning Lei Wu 645-275-76523
Jul 19, 2020 09:30-12:00 Frequency principle: linear model and general theory Luo Tao, Yaoyu Zhang 999-355-46463
Jul 22, 2020 09:30-12:00 Neural tangent kernel made practical Wei Hu 969-766-38784

Zoom Password: 738669

If the above Zoom IDs does not work, we would shift to a backup Zoom ID: 266-664-3379 with the same password.



Contact us

Zhiqin Xu:xuzhiqin@sjtu.edu.cn