Introduction

When people hear the term Neural Networks, it often sounds mysterious. Almost magical.

But under the hood, neural networks are not magic.

They are simply:

• numbers
• matrix multiplications
• weighted sums
• error calculations
• repeated adjustments

This document explains neural networks from the ground up : the exact same ideas covered in the presentation : but in a more detailed and intuitive way.

By the end, you should understand:

• what a neuron actually does
• how neural networks learn
• why activation functions are necessary
• what loss functions are
• how backpropagation works
• why gradient descent matters
• and how these simple ideas scale all the way to systems like GPT-4