- 84 Downloads
The abstract neuron is the building block of any neural network. It is a unit that mimics a biological neuron, consisting of an input (incoming signal), weights (synaptic weights), and activation function (neuron firing model). This chapter introduces the most familiar types of neurons (perceptron, sigmoid neuron, etc.) and investigates their properties.