Labels

Definition

The final output, category, or answer that a Supervised Machine Learning model is trying to predict. It is the known 'truth' attached to the training data.

Explain Like I'm New

The answer key on a flashcard. The front of the flashcard has the picture (the Features). The back of the flashcard says 'Dog' (the Label).

Real World Example

In medical imaging AI, doctors look at 10,000 X-Rays (Features) and manually tag them with 'Cancer' or 'No Cancer' (Labels). The AI studies this dataset to learn how to predict the label itself.

Common Use Cases

•Supervised Learning
•Data annotation

Interview Questions

basic

Do Unsupervised Learning models use Labels?

intermediate

What is a 'Data Labeling' farm?

Flash Cards

Question

Do unsupervised use labels?

Click to reveal answer

Answer

No. By definition, unsupervised learning relies on unlabeled data, forcing the algorithm to find its own groupings.

Question

Data Labeling farm?

Click to reveal answer

Answer

Because AI needs millions of labeled examples to learn, tech companies often hire thousands of human workers (often in developing countries) to sit at computers all day manually drawing boxes around stop signs and pedestrians in videos. These are the 'Labels' fed to self-driving cars.