1. Different types of neural networks

ANNs are inspired by biological neural networks, and are made up of artificial neurons interconnected by synapses.

Each neuron integrates information from the synapses connected to it, and activates or deactivates its output. Each synapse transmits information from one neuron to another by multiplying a factor called weight, memorized by the network. The output of each neuron is calculated as a non-linear function of the sum of its inputs. The first representation of an artificial neuron, called a perceptron, was proposed by Rosenblatt and is still used in many ANNs. Indeed, neural networks such as single-layer or multi-layer perceptron networks...