Supervised Learning is the most widely used approach in machine learning. The model trains on labeled data where every input has a known correct output. You show the algorithm thousands of images labeled "car" and thousands labeled "bicycle." It analyzes the pixels, finds patterns like wheels, windshields, or handlebars that correlate with each label, and builds an internal model. After training, it can classify new, unlabeled images it has never seen before. This exact methodology powers your email spam filter. It was trained on millions of emails that humans labeled as "spam" or "not spam." It drives medical diagnostics where an AI trained on thousands of labeled X-rays can flag early signs of tumors or pneumonia. Supervised learning is extremely powerful, but its primary bottleneck is that it requires massive volumes of high-quality, human-annotated data. Getting those labels is expensive and time-consuming, which is why techniques like semi-supervised learning and self-supervised learning exist.
Back to Glossary