How to Create a Neural Network and Train It to Identify Doodles

Dive into the world of neural networks with this comprehensive 55-minute video tutorial. Learn how to program a neural network from scratch in C# and train it to recognize doodles and images. Explore key concepts such as decision boundaries, weights, biases, hidden layers, activation functions, cost calculation, and gradient descent. Follow along as the tutorial progresses from basic principles to advanced topics like backpropagation and practical applications in digit recognition, fashion item classification, and doodle identification. Gain hands-on experience by coding your own neural network and testing it on various datasets, including MNIST digits, fashion items, and Google Quick Draw doodles. Conclude with a final challenge to solidify your understanding of neural network creation and training.

Introduction
The decision boundary
Weights
Biases
Hidden layers
Programming the network
Activation functions
Cost
Gradient descent example
The cost landscape
Programming gradient descent
It's learning! slowly
Calculus example
The chain rule
Some partial derivatives
Backpropagation
Digit recognition
Drawing our own digits
Fashion
Doodles
The final challenge