How to build a simple Neural Network from scratch without frameworks. Just Math and Python. (With lots of animations and code)

Hi ML community!

I've made a video (at least to the best of my abilities lol) for beginners about the origins of neural networks and how to build the simplest network from scratch. Without frameworks or libraries, just using math and python, with the objective to get people involved with this fascinating topic!

I tried to use as many animations and manim as possible in the making of the video to help visualizing concepts :)

The video can be seen here Building the Simplest AI Neural Network From Scratch with just Math and Python - Origins of AI Ep.1 (youtube.com)

It covers:

• The origins of neural networks
• The theory behind the Perceptron
• Weights, bias, what's all that?
• How to implement the Perceptron
• How to make a simple Linear Regression
• Using the simplest cost function - The Mean Absolute Error (MAE)
• Differential calculus (calculating derivatives)
• Minimizing the Cost
• Making a simple linear regression

I tried to go at a very slow pace because as I mentioned, the video was done with beginners in mind! This is the first out of a series of videos I am intending to make. (Depending of course if people like them!)

I hope this can bring value to someone! Thanks!