Cross-domain transfer learning is a term from the fields of Artificial Intelligence, Big Data and Smart Data, and Digital Transformation. It describes a method where knowledge from one area of application (domain) is transferred to another. While this sounds complicated at first, it is very practical: This way, an artificial intelligence doesn't have to start from scratch but can utilise already learned knowledge from another sector.<\/p>\n
Imagine an AI has learned to detect tumours on X-ray images in medicine. This knowledge can be used across domains to detect defects on industrial X-ray images in manufacturing \u2013 without complete retraining. The AI uses the skills already acquired in one domain and \u201etransfers\u201c them to the other.<\/p>\n
Cross-domain transfer learning saves time, costs and computing power. This enables companies to bring innovations to different areas more quickly as they can utilise existing expertise more intelligently. This is particularly exciting for companies that want to efficiently utilise diverse data from different sources in order to make processes more intelligent and flexible.<\/p>","protected":false},"excerpt":{"rendered":"