Overfitting
Overfitting happens when a model learns its training data too closely, including noise and quirks, so it performs well on that data but poorly on new data.
What is Overfitting?
An overfitted model has in effect memorised rather than learned the underlying pattern. The warning sign is a large gap between high training accuracy and much lower validation or test accuracy. It is more likely with small datasets and very flexible models.
How it works
Developers reduce overfitting with more or more varied data, simpler models, regularisation, dropout and early stopping when validation performance stops improving.
← All Artificial Intelligence & Superintelligence definitions · ← Back to the A–Z definitions