# 7/30/24 Meeting Notes #1

- [ ] Understanding the basic knowledge of deep learning
- [ ] Decide on a title for the project

Title:

- [ ] Start working on the project right after the semester starts
- [ ] Watch YouTube videos and books (recommended by TA)

Book recommendations:
1. Deep Learning：用Python進行深度學習的基礎理論實作
2. Deep Learning 2: 用Python進行自然語言處理的基礎理論實作
3. Deep Learning 3: 用Python進行深度學習框架的開發實作
4. Deep Learning with Python by Francois Chollet
5. Introducing Python: Modern Computing in Simple Packages

Video recommendations:
【機器學習2022】 教學 (台大 李宏毅 教授)
https://www.youtube.com/playlist?list=PLJV_el3uVTsPM2mM-OQzJXziCGJa8nJL8
Machine Learning Foundations (機器學習基石) (林軒田 教授)
https://www.youtube.com/playlist?list=PLXVfgk9fNX2I7tB6oIINGBmW50rrmFTqf
Machine Learning Techniques (機器學習技法)
https://www.youtube.com/playlist?list=PLXVfgk9fNX2IQOYPmqjqWsNUFl2kpk1U2

**IDEAS**
-
Idea 1 chatgpt API
- Using arduino or a hardware and using chatgpt API for the project. Will be a physical device for chatgpt.
- Train model to extract the information through websites.
- Device recommended for hotels, schools etc. To provide information about the place or events near the area.
- TA's recommendation/ feedback:
-- Tecnhical contribution might be insufficient if it's just about integrating other people's models.
-- Connecting programs together for more contribution. Best to include some projects that include our own training, such as detection tasks, to ensure that contribution is sustainable enough.

Idea 2 self driving car
- Brain of a self driving car
- Camera on dashboard to recognize other vehicles, person, and traffic light color.
- Another 2 sensor on the side of the car to keep the car inside the lane.
- 1 sensor in front of the car to keep the distance of the car infront for cruising and sudden breaking.
- TA's recommendation/ feedback:
-- find a way to apply the idea

Idea 3 monitoring for injured cats
- Using camera to identify different cats (e.g. fur colour, size, fur pattern).
- Detect if a cat is hurt or sick (e.g. limping, vomiting, obvious bleeding).
- Notify humans which cat is injured or sick and where is the cat.
- Best used in shelters as there are many cats there and humans are unable to keep track of the cat 24/7.
- TA's recommendation/ feedback:
-- Need more datasets on cats to back it up.

**Keywords to look up:**

- [ ] detection
- [ ] chat box
- [ ] Lidar
- [ ] 3d detection ( diffusion model, generative model )
- [ ] CVPR, AAAI, ECCV, ICCV

---
Next: [8/11 Meeting Notes](https://hackmd.io/@emps-113up/meeting2)
Full Content List [here](https://hackmd.io/@emps-113up/full-list)