How Do Computers See Images? - Neural Networks

Описание: Hello everyone, this is a video that I made in a little over a week and I had originally planned to have all 3 parts into one. However, due to time constraints on #SoME2, I had to push most of the exciting and "less talked about" stuff into part 2 and 3.

Note that this mainly serves as a refresher on neural networks to lead to part 2 and 3, but I think it could also work quite well as a first introduction if paired with more detailed resources.

If you have any feedback, I'd really appreciate it if you comment it as I'll try to incorporate it into my next videos!!

Enjoy the video :D

Chapters:
0:00 - intro to the problem of computer vision
1:11 - overview of the series
1:57 - single unit neural network
4:18 - gradient descent
6:29 - generalization to multilayer network
7:25 - notation and affine transforms
8:49 - non-linear transforms
9:41 - backpropagation
10:40 - overview of upcoming videos

Resources for Backpropagation:
Lecture 6: Backpropagation by Justin Johnson -    • Lecture 6: Backpropagation  
What is backpropagation really doing? | Chapter 3, Deep learning by Grant Sanderson -    • Backpropagation, intuitively | Deep Learni...  
Let the legendary Andrej Karpathy walk through his mini autograd engine :D    • The spelled-out intro to neural networks a...  

This is the article I used for the small demonstration of momentum in gradient descent: https://distill.pub/2017/momentum/ This is really worth a read if you're a bit familiar with machine learning.