Active learning is a term used in the fields of artificial intelligence, big data and automation. It describes a method in which computer programmes or algorithms learn in a targeted manner - by actively asking for the most important information. The special thing about this is that the system itself decides which data it needs to improve.
Imagine you have a company that automatically sorts emails according to "spam" and "no spam". Instead of presenting the programme with thousands of emails, it selects the most difficult or unclear cases itself and asks specifically for feedback. As a result, the system learns much more efficiently and becomes more accurate in recognising spam.
Active learning is used wherever large amounts of data are available, but not all information is equally important or easy to evaluate. This saves time, money and increases the accuracy of automated systems, be it in recognising fraud in the financial sector, in image analysis in medicine or in customer service. In short, active learning makes machines smarter by asking themselves when and where they need to learn something new.
Deutsch 
English (UK)