AI Glossary

Hard negative

What does “Hard negative” mean?

Hard negative: A negative example that is easy for a model to mistake for a positive.

Hard negative visual guide: A negative example that is easy for a model to mistake for a positive.
Visual guide Open full size (new tab)

What is an example of “Hard negative”?

A search model sees a page about the wrong product with nearly the same name.

Keep learning

See also

DatasetTermTraining dataTermData labelingTermOversamplingTermUndersamplingTerm

Same topic

Active learningTermWeak supervisionTermHyperparameterTermGradient descentTermInstruction tuningTerm