Building a Perceptron From SCRATCH (no frameworks, only math and python) - The Origins of AI - Ep.1

I sure love AI! And there is nothing better than when you learn how something works from absolute scratch. In this video series, I would like you to join me on a journey to the origins of AI and Machine Learning. We are using no frameworks, and in this particular video not even numpy. Just pure math, python and well science really, to explain how AI came to be.

There are many forms of machine learning, but neural networks are hands down the most evolved and for sure, mediatic. But before the popular transformers, we had many other types and architectures of machine learning such as CNN, RNN, MLP and of course... the perceptron aka the single artificial neuron or the simplest form of neural network.

In this video, we cover the biological science that inspired neural networks, that mathematical functions that make them possible, topics like backpropagation and gradient descent and we do it all from scratch!

Hope you enjoy!

Source Code on my github - https://github.com/fmiguelmmartins/or...

Chapters
00:00 - Intro
01:43 - The Perceptron
02:11 - What is a function
04:50 - Building the perceptron
06:04 - The Weights
08:40 - The Cost Function
10:20 - The Learning Rate
11:16 - Cost Minimization
12:30 - The Bias
15:15 - Optimizing the Weights and Bias
15:48 - Gradient Descent and Backpropagation
23:19 - Outro








#neuralnetworks #machinelearning #artificialintelligence