One-shot learning is a term used in the fields of artificial intelligence and automation. It describes a special ability of computers or algorithms: They can learn from a single piece of sample information and apply this knowledge to new, similar cases.
In traditional machine learning, a computer often needs thousands of examples to reliably recognise an object, for example. One-shot learning, on the other hand, saves a lot of time and effort because a single example image is sufficient. The system "remembers" the most important features and recognises the object the next time it sees it.
Imagine showing an intelligent computer a photo of your favourite coffee cup. Thanks to one-shot learning, the artificial intelligence can now recognise your cup in other images from different angles - without having to analyse hundreds of other photos.
One-shot learning is particularly useful in areas such as facial recognition, the automation of work processes and rapidly changing data volumes. This enables companies to make their applications more efficient and reduce the costs of collecting and "feeding" data.
Deutsch 
English (UK)