Unsupervised anomaly learning is a term from the fields of artificial intelligence, big data and smart data, and automation. It describes a method by which computer systems automatically detect unusual patterns or deviations in large amounts of data – without humans needing to specify beforehand exactly what to look for.
Imagine an automated quality control process in a factory: thousands of components are produced daily. With unsupervised anomaly learning, AI continuously analyses data from these components, such as dimensions, weight, or material properties. Initially, it doesn't know exactly what is normal and what is faulty. However, it independently learns what the usual process is and raises an alarm if a component suddenly deviates significantly from it. This allows, for example, faulty products to be detected and sorted out more quickly.
This is particularly practical when errors are rare or very different. Companies benefit from this through improved quality, less scrap, and reduced monitoring effort. Unsupervised anomaly learning makes it possible to automatically monitor vast amounts of data in a short period and react to undesirable deviations – ideal for modern, automated production processes and data analyses.
Deutsch 
English (UK)