Supervised learning is a machine learning paradigm where models learn from labeled training data to make predictions on unseen data. Each training example consists of an input-output pair, enabling the algorithm to learn a mapping function from inputs to outputs. The two primary tasks are classification (predicting discrete categories) and regression (predicting continuous values). The fundamental challenge lies in balancing the bias-variance tradeoff: simple models underfit (high bias, low variance), complex models overfit (low bias, high variance), and the goal is finding a model that generalizes well by minimizing total error on unseen data.