Supervised learning
ML where the model is trained on labelled data — every example has a "right answer".
"Here are 10,000 images, which are cats?": the model learns the answer. Different from unsupervised (find structure on its own) and reinforcement (learn via reward).