- Feed Forward Neural Network(FFNN)
An extension of the logistic regression algorithm. FFNN are stacking multiple hidden layers (any layers that's not the input or output layer) followed by non linear activation, such as the sigmoid or softmax activations.
FFNN are only made of fully connected layers, where each neuron in one layer is connected to all the neurons in the previous layer.
y=f*(x)
y : output
f : intermediate computation
x : input
y maps an input x to a category y.
It defines a mapping y = f(x;θ) and learns the value of the parameters θ
There are no feedback connections in which outputs of the model are fed back into itself.
- Recurrent Neural Network(RNN)
It is not only from the input later to the output layer but also recur in the hidden layer.
It is a class of neural networks that allow previous outputs to be used as inputs while having hidden states. The data flows in reverse direction during output generation, it would form a cycle and the output could be generated.
- Fully connected Neural Network(FC, FCNNs, Dense layer)
All neurons are connected to all neurons in the previous layer.