Deep learning is a subset of Machine learning, which makes use of artificial neural networks to solve complex PROBLEMS. The artificial neural NETWORK is a concept inspired by information processing and distributed communication nodes, neurons, which are present in human brains. It gives deep learning the power to look at a problem and solve it like a human brain in that situation would. The word ‘deep’ in deep learning means the number of hidden layers in the neural network. Deep learning models are developed in such a way that they are capable of training and managing themselves.

In the above figure, the deep neural network takes input through an input layer. The algorithm's input and output are separated by a hidden layer, in which the function applies weights to the inputs and guides them through an activation function as the output. Activation functions of a Deep neural network may VARY. For example, a Sigmoid Function can take any input and produce the output between 0 and 1. The output layer is the network's LAST layer that takes the information gleaned from the hidden layer and turns it into a final value. 

In a nutshell, the hidden layers conduct nonlinear changes on the network's inputs. The function of the neural network determines the hidden layers, and the layers themselves may vary depending on their associated weights.