# 2023.02.22 TIGP-C2 course: Convolutional Neural Networks, Recurrent Neural Networks
> Instructor: Prof. Che Lin \<[e-mail]\>
> TA: Yi-Hsien Hsieh \<[e-mail]\>, Le-Yin Hsu \<[e-mail]\>
## Pre-class Readings & Videos
* **Note that there will be a quiz about the pre-class materials at the beginning of the class**
* This quiz consists of three questions and will take about ten minutes via google form
* This quiz accounts for 5% of this week's score
* Let's have a brief overview of deep learning first
* Coursera: *[Feed-Forward Neural Networks](https://pt.coursera.org/lecture/machine-learning-applications/feed-forward-neural-networks-Gy5JW)*
* **Convolutional Neural Networks (CNN)**
* Stanford University Lecture Collection | *[Convolutional Neural Networks](https://www.youtube.com/watch?v=bNb2fEVKeEo)* | [slide](https://drive.google.com/file/d/1SZXu0tkvnLI2eQ1MeVScGRVi-ofNFemE/view?usp=sharing)
* (Supplementary) *[The history of CNN for visual recognition](https://www.youtube.com/watch?v=vT1JzLTH4G4)*

<!--
* *[Deep Learning](https://www.deeplearningbook.org/)*
* Chapter 9.2 (Motivation)
* Chapter 9.3 (Pooling)
* Chapter 9.7 (Data Types)
* *[Coursera: Convolutional Settings - Padding and Stride](https://www.coursera.org/lecture/deep-learning-reinforcement-learning/convolutional-settings-padding-and-stride-YUl5F)*
* *[Coursera: CNN Example](https://www.coursera.org/lecture/convolutional-neural-networks/cnn-example-uRYL1)*
-->

* **Recurrent Neural Networks (RNN)**
* [AIIM](https://nol.ntu.edu.tw/nol/coursesearch/print_table.php?course_id=942%20U0780&class=&dpt_code=9450&ser_no=72403&semester=111-2) week 10: *[RNN basics (01:44:25-)](https://youtu.be/EQCVFQ8s-ME)* | [slide](https://drive.google.com/file/d/1VTZkPmf2cMPuiEbBvhdBnD3NzqwPAPv3/view?usp=share_link)
* [AIIM](https://nol.ntu.edu.tw/nol/coursesearch/print_table.php?course_id=942%20U0780&class=&dpt_code=9450&ser_no=72403&semester=111-2) week 11: *[Gated models (01:07:00-01:37:00)](https://youtu.be/XOmFGeB1Byc)* | [slide](https://drive.google.com/file/d/18MnvwsqDKmSOEVug-7N4Z--8teNNzA_J/view?usp=share_link)

## In-class Discussion
* **For each student, please prepare at least one question about CNN & RNN, respectively**
* CNN
* Questions from students
* Case study: DTI prediction
* More advanced models for image data
* RNN
* Questions from students
* Can we replace the recurrent operation? -> the *Transformer*
* More advanced models for sequential data

## Homework
* [TIGP_C2_course_230222_HW.pdf](https://drive.google.com/file/d/1t5dx2bdRlBTJISkIxSfW5aENoQy4H5C0/view?usp=share_link)
* (Important) [Grading policy](https://docs.google.com/spreadsheets/d/1ihyhLQbDy53FL-Bn90t-55F1-c3BNwq0f9svlTt8QEo/edit?usp=sharing)