Overfitting

ML

When a machine learning model learns the training data too well, including its noise and quirks, and fails to generalize to new data. Signs include high training accuracy but poor test accuracy, addressed through regularization or more diverse training data.