Loading…
Glossary term
Learning that combines a small amount of labelled data with a large amount of unlabelled data.
Semi-supervised learning mixes a small set of labelled examples with a large pool of unlabelled ones. The model learns the overall shape of the data from the unlabelled part and uses the labels to name what it found. It helps when labelling is expensive but raw data is cheap.
It stretches a limited labelling budget, which is common in medicine, law and other expert domains.