Data labelling is a key term within the fields of Artificial Intelligence, Big Data and Smart Data, as well as automation. For machines to learn complex tasks or recognise patterns, they require large amounts of data – and this is precisely where data labelling comes into play.
In data labelling, data such as images, text or audio files are provided with clear information (labels). These can be simple descriptions, such as „dog“ or „cat“ in a photo. Without these clear assignments, algorithms would not be able to independently distinguish what is visible in the images. The quality of the labels therefore significantly influences how well artificial intelligence works in practice.
A clear example: For an app that automatically recognises photos of whether an apple or a banana is visible, many images with the respective fruit must first be correctly labelled as „apple“ or „banana“. Only through this data labelling can the app later determine the correct fruit in new photos.
So, data labelling is the foundation of many modern technologies – from voice assistants to facial recognition – and is crucial for success in data-driven projects.
Deutsch 
English (UK)