Supervised Learning
Learning from labeled input-output examples.
Supervised learning trains a model on examples paired with correct answers, adjusting it to predict the labels. It is the most common ML paradigm and includes the fine-tuning step that turns base models into task-specific ones.