Training a Neural Network

So far, we know that once an input is provided to a neural network, it enters the input layer which is an interface that exists to pass on the input to the next layer. If a hidden layer is present, then the inputs are sent to the activation nodes of the hidden layer via weighted connections. The weighted sum of all the inputs received by the activations nodes is calculated by multiplying the inputs with their respective weights and adding these values up along with the bias. The activation function generates an activation value from the weighted sum and this is passed on to the nodes in the next layer. If the next layer is another hidden layer, then it uses the activation values from the previous hidden layer as inputs and repeats the activation process. However, if the proceeding layer is the output layer, then the output is provided by the neural network.

From all of this information, we can conclusively say that there are three parts of the deep learning model...