Glowing violet and cyan dots, joined by thin lines against a deep blue background. You’ve seen images like this in almost every article about AI. It’s a stylised neural network, and it’s a surprisingly good way to understand how machine learning works.
Nodes and connections
Each dot is a node, or artificial neuron: it takes in numbers, combines them and passes a result on. Each line is a connection with a weight that controls how much one node influences the next. Larger, brighter nodes in the picture hint that some parts of a network matter more for a given task.
How the network learns
Training a model means adjusting those weights, millions or billions of them, until the network’s outputs match the examples it is shown. No single connection knows anything on its own; the knowledge lives in the pattern of all of them together.
