Gabriel NN

Introduction

Introduction

First Neural Network

Introducing Our First Problem
Neural Network with One Weight
How Our Network Learns (Backpropagation Made Easy)
Learning Through Repetition (Epochs)

Understanding Weight Updates

Introducing Our Second Problem
Trouble With Our Weight Updates
Fixing the Update Rule With Inputs

Gradient Descent, Loss and Cost functions

Let's Get More Theoretical
Loss and Cost
Slope
Using Slope to Update
Unlocking Derivatives
Gradient Descent

The Role of Bias

Third Problem
Why We Need Bias
How Does Bias Work — And How Do We Update It?
Implementing Bias in Code

Code to Cortex (artificial vs. biological neurons)

From Biology to Code
Why Activation Functions Matter

Introducing Activation Functions

The XOR Challenge
Activation to the Rescue
Adjusting Weights with Activation in Mind
Let’s Derive the Sigmoid Together(optional)
Introducing Hidden Layer
The Math Behind Hidden Layer Learning
Backprop Begins Here
From Theory to XOR

Building a Multi-Layer Neural Network

Build a Network Workshop
A Matrix Makeover
Making a Modular Network
Let’s Test It
Try It By Yourself!

From Numbers to Images

Stepping Into Images
What Is An Image?
Getting the Data
Data Normalization
Data Augmentation
Training and Testing

Conclusion & References

Final Remarks
References

Conteúdo

Aqui será exibido o conteúdo dos módulos selecionados.