Classical ML foundations
Four detailed lessons, mathematical definitions, worked examples and limitations.
Explore the lessons
Concept 05Supervised learning and riskA model learns a mapping from inputs to labels by minimizing empirical risk with appropriate validation.Concept 06Optimization and gradientsGradient descent, stochastic updates and adaptive optimizers tune trainable parameters.Concept 07Generalization and model selectionCapacity, regularization, bias–variance trade-offs and cross-validation determine out-of-sample reliability.Concept 08Kernel methods and feature spacesKernel methods learn linear predictors in implicit feature spaces; their quality depends on geometry and regularization.