GlossarIA
Open AI glossary for companies
← Back to glossary

Dropout

Regularization technique that randomly deactivates neurons during training to prevent overfitting, forcing the model to be more robust.

Intermediate regularizacion entrenamiento overfitting

Full definition

Regularization technique that randomly deactivates neurons during training to prevent overfitting, forcing the model to be more robust.

Example in a business context

Applying dropout of 0.5 in dense layers so the model doesn't memorize training data.