Deep Learning Dictionary - Lightweight Crash Course

Artificial Neural Networks - Deep Learning Dictionary

An artificial neural network (ANN) is a computing system that is comprised of a collection of connected processing units called neurons or nodes that are organized into layers, which are connected to each other via weights.

The network accepts data as input and maps the supplied inputs to defined outputs.

The general architecture behind these networks is loosely inspired by the structure and function of the brain's physiological neural networks, and as such, is where the name artificial neural networks comes from.

The first layer in a network is called the input layer, the last layer is called the output layer, and any layers in between the input and output layers are called hidden layers. If a network has more than one hidden layer, it is considered deep.

After passing input data to the network, processing will occur on that data by each neuron, and each weight connecting neurons transmits a signal about the data from one neuron to the other until reaching the output layer of the network.

We can think of the output from the network as a prediction about the data. For example, if we pass an image to the network for classification, the output of the network will give us the predicted label that the network assigns to this image.

Generally, the type of input and output supplied to and from the network will depend on the type of data contained in the dataset and the task for which the network is being used.