Supervised learning
Supervised learning is like a student learning with an answer key; the computer is shown many examples of problems along with their correct solutions so it can learn to solve new, unseen problems on its own.
A machine learning approach where a model is trained on a dataset consisting of input-output pairs, adjusting its internal parameters to minimize the difference between its predictions and the provided ground-truth labels.
A learning paradigm defined by the optimization of a hypothesis function $h$ that maps input space $X$ to output space $Y$, derived from a training set of i.i.d. samples $(x_i, y_i)$ by minimizing a loss function $L(y_i, h(x_i))$ over the empirical distribution.
evolution
- 1958 · historyPerceptron Invention
Frank Rosenblatt developed the Perceptron, the first algorithm capable of learning to classify inputs based on labeled training data.
- 1986 · historyBackpropagation Popularization
Rumelhart, Hinton, and Williams demonstrated that backpropagation could efficiently train multi-layer neural networks using labeled error signals.
- 1995 · historySupport Vector Machines
Cortes and Vapnik introduced SVMs, providing a mathematically rigorous framework for supervised classification and regression.
- 2012 · historyAlexNet Breakthrough
The success of AlexNet in the ImageNet competition proved that deep supervised learning on large labeled datasets could achieve state-of-the-art performance.