A Short Introduction To Neural Networks

Neural networks (NN) or Artificial neural networks (ANN) seems to be a primarily discussed topic in the realms of Artificial Intelligence (AI). So what exactly are NN or ANNs?

Just like neurons which helps with the transmitting of nerve impulses in our brains, an ANN can be seen as an information processing system which imitates the impulses or operations of the human mind.

Compared to computers which are capable of performing thousands or not millions of computations per second, our human brain’s rate of operation is somewhat slower. However, this does not make us utterly obsolete as we would triumph in performing complex tasks such as hearing, navigating a car or making difficult decisions. A level that AI and ANN’s have not yet accomplished.

Structure of an ANN

Most ANN has a similar structure when broken down. They consist of an Input, Output and Hidden layers. An ANN which has multiple Hidden layers can be further classified as a Deep neural network (DNN).

Just for illustration purposes, I’ve decided to include a sample output of how an ANN would look like once plotted in R below. You can still see the Input, Hidden and the Output layers.

In this instance, I’ve decided to use three hidden layers in the computation of this ANN.