Scott Logic goes through the basics of the Layered Neural Network, a simple introduction to some of the core concepts.

All neural networks are made up of a network of neurons, much like the brain that they model, but what is a neuron?

A neuron is made up of its inputs, weights, a bias, an activation function and a singular output. It was originally intended to replicate the biological neuron.

There is a one-to-one ratio of inputs to weights, which when combined gives a series of weighted inputs that get summed together. To this sum a bias gets added.

In neural networks, the neurons can be arranged in a variety of ways. For the basic neural network the neurons are arranged into several distinct layers.

There are three categories

1. Input Layer
2. Hidden Layers
3. Output Layer

Check out the complete description in the post here.