Supervised learning is the most common form of machine learning, where a model is trained on examples that each come with a correct answer, called a label. The “Machine Learning Basics” chapter of Goodfellow, Bengio, and Courville’s “Deep Learning” describes supervised learning as observing examples of an input together with an associated target and learning to predict the target from the input.
For instance, to build a spam filter, you show the model many emails each marked “spam” or “not spam.” The model adjusts itself until its predictions match the labels, then applies what it learned to new emails it has never seen. Common supervised tasks are classification (choosing a category) and regression (predicting a number).
The quality of supervised learning depends heavily on the labeled training data. Accurate, representative labels lead to accurate predictions, while biased or noisy labels lead the model astray.
Why business readers should care: Supervised learning powers many everyday business tools, including credit scoring, demand forecasting, and document classification. Its main practical cost is obtaining good labeled data, which often requires human effort and is frequently the biggest bottleneck in a project.