Perceptron (Glossary)

The term Perceptron is at home in the field of Artificial Intelligence and Automation.

A perceptron is the simplest model of an artificial nerve cell. It was developed to teach computers how to learn, much like how our brains process information. Imagine a perceptron as a filter: it takes in different pieces of information and, based on simple rules, decides how to categorise them.

For example, a perceptron receives numbers that describe certain features – such as the shape and size of objects in photos. It weighs these features and then makes a decision, for example: „There is an apple in this picture“ or „not an apple“. If the perceptron makes mistakes, it learns from these mistakes and adjusts its rules. This way, it can make the right decisions better and better over time.

Today, perceptrons are building blocks of modern artificial intelligences, used, for example, in speech recognition or quality control in production. The basic idea – simple learning by trial and error – laid the foundation for today's intelligent computer systems.