One-shot learning is a term from the fields of artificial intelligence and automation. It describes a special ability of computers or algorithms: they can learn from a single example of information and apply this knowledge to new, similar cases.
In classical machine learning, a computer often needs thousands of examples to reliably recognise an object, for instance. One-shot learning, on the other hand, saves a lot of time and effort, as a single example image is sufficient here. The system „remembers“ the most important features and recognises the object again the next time.
Imagine showing an intelligent computer a photo of your favourite coffee mug just once. Thanks to one-shot learning, the artificial intelligence can now recognise your mug in other pictures, from different angles, without having to analyse hundreds more photos of it.
One-shot learning is particularly useful in areas such as facial recognition, process automation, or even with rapidly changing data volumes. This allows companies to make their applications more efficient and reduce costs for collecting and „feeding“ data.















