2021 Spring Lecture Notes Reading

11. Covering Numbers and Its Application on Deep Neural Networks

Time: 2:00 - 4:00 pm, May 13, 2021

Speaker: Xianli Zeng

Zoom Link: https://cuhk-edu-cn.zoom.com.cn/j/98480113666?pwd=ME9ZVjdONXJnb2hRK2duTHpNK094QT09

Reference:

• Chapter 20-21 of [Deep learning theory lecture notes] by Matus Telgarsky

10. Rademacher Complexity in Deep Networks

Time: 2:00 - 4:00 pm, May 6, 2021

Speaker: Haoyu Wei

Zoom Link: https://cuhk-edu-cn.zoom.com.cn/j/98480113666?pwd=ME9ZVjdONXJnb2hRK2duTHpNK094QT09

Reference:

• Chapter 18-19 of [Deep learning theory lecture notes] by Matus Telgarsky

9. Concentration Inequalities For Statistical Learning

Time: 2:00 - 4:00 pm, Apr 20, 2021

Speaker: Huiming Zhang

Zoom Link: https://cuhk-edu-cn.zoom.com.cn/j/98480113666?pwd=ME9ZVjdONXJnb2hRK2duTHpNK094QT09

Reference:

• Chapter 17 of [Deep learning theory lecture notes] by Matus Telgarsky

8. Margin Maximization and Implicit bias

Time: 2:00 - 4:00 pm, Apr 15, 2021

Speaker: Ziniu Li

Zoom Link: https://cuhk-edu-cn.zoom.com.cn/j/98480113666?pwd=ME9ZVjdONXJnb2hRK2duTHpNK094QT09

Reference:

• Chapter 15 of [Deep learning theory lecture notes] by Matus Telgarsky

7. Stochastic Gradients Nonsmoothness, Clarke differentials, and positive homogeneity

Time: 2:00 - 4:00 pm, Apr 1, 2021

Speaker: Yujun Li

Zoom Link: https://cuhk-edu-cn.zoom.com.cn/j/98480113666?pwd=ME9ZVjdONXJnb2hRK2duTHpNK094QT09

Reference:

• Chapter 12, 14 of [Deep learning theory lecture notes] by Matus Telgarsky

6. Optimization Toolbox for Deep Learning: The Convergence Analysis of Gradient Descent and Gradient Flow

Time: 2:00 - 4:00 pm, Mar 25, 2021

Speaker: Yushun Zhang

Zoom Link: https://cuhk-edu-cn.zoom.com.cn/j/98480113666?pwd=ME9ZVjdONXJnb2hRK2duTHpNK094QT09

Reference:

• Chapter 9, 10 and 11 of [Deep learning theory lecture notes] by Matus Telgarsky

5. Function Space Norms, and Neural Tangent Kernel

Time: 2:00 - 3:30 pm, Mar 18, 2021

Speaker: Chendi Wang

Zoom Link: https://cuhk-edu-cn.zoom.com.cn/j/98480113666?pwd=ME9ZVjdONXJnb2hRK2duTHpNK094QT09

Reference:

• Chapter 8 of [Deep learning theory lecture notes] by Matus Telgarsky

4. Error Bounds for Approximations with Deep ReLu Networks

Time: 2:00 - 3:30 pm, Mar 11, 2021

Speaker: Zhiying Fang

Zoom Link: https://cuhk-edu-cn.zoom.com.cn/j/98480113666?pwd=ME9ZVjdONXJnb2hRK2duTHpNK094QT09

Reference:

• Chapter 7 of [Deep learning theory lecture notes] by Matus Telgarsky

3. Benefits of Depth

Time: 2:00 - 3:30 pm, Mar 04, 2021

Speaker: Jingyi Cui

Zoom Link: https://cuhk-edu-cn.zoom.com.cn/j/98480113666?pwd=ME9ZVjdONXJnb2hRK2duTHpNK094QT09

Reference:

• Chapter 6 of [Deep learning theory lecture notes] by Matus Telgarsky

2. Sampling from Infinite Width Networks

Time: 2:00 - 3:30 pm, Feb 23, 2021

Speaker: Qing Yang

Zoom Link: https://cuhk-edu-cn.zoom.com.cn/j/98480113666?pwd=ME9ZVjdONXJnb2hRK2duTHpNK094QT09

Reference:

• Chapter 5 of [Deep learning theory lecture notes] by Matus Telgarsky

1. Background and Universal Approximation

Time: 10:30 - 11.30 am, Feb 18, 2021

Speaker: Ziyi Fang

Zoom Link: https://purdue-edu.zoom.us/j/5339424029?pwd=YmdVUm55ZnphdldsU0JvZ3ZLVlFHQT09

Reference:

• Chapter 1-3 of [Deep learning theory lecture notes] by Matus Telgarsky