Overfitting
When a model memorizes its training data — noise and all — instead of learning the general pattern, so it aces training but flops on new data. The classic sign is great training scores and poor test scores.
Think of It Like This
A student who memorizes the practice answers but can't handle a reworded question.
An over-powerful model, or one trained too long on too little data, starts fitting quirks that don't generalize. You spot it when training loss keeps dropping while validation loss turns back upward. The usual fixes are regularization, more data, and early stopping — all different ways to stop the model from clinging to noise it should be ignoring.