Considering a scenario where you have labeled examples of emails $($spam or not spam$),$ explain how supervised learning utilizes this data during training to make predictions on new, unseen emails.

Group of answer choices

By providing the labeled examples as input during the prediction phase

By directly using the labeled examples to test the model's performance

By memorizing the labeled examples to identify specific patterns in the emails

By training the model to generalize patterns from the labeled examples to classify new emails