The term Early Stopping originates from the field of Artificial Intelligence, particularly in machine learning, and also plays a role in the context of Big Data and Smart Data. Early Stopping is a method used to prevent overly long training periods for computer models, for example, in image recognition or self-learning systems.
Imagine you're teaching software to distinguish between dogs and cats in photos. The system learns from many examples. If trained for too long, it memorises too many specific details (like the colour of a particular dog) and makes poorer decisions on new, unknown images. Therefore, you monitor how well the system performs on new examples. As soon as no further improvement is seen, Early Stopping halts the training – this keeps the model flexible and helps it make better decisions on new data.
This method saves time and computing power, and ensures that results are more usable in practical everyday life later on. Early stopping is therefore an important building block in making artificial intelligence application-ready and economically efficient.
Deutsch 
English (UK)