What machine learning is, optimization viewpoints, gradient methods, second-order approximation, and Newton’s method.

Linear regression, multivariable least squares, robust alternatives, model assessment, weighting, and probabilistic interpretation.

Binary and multiclass classification, logistic and cross-entropy losses, perceptrons, support vector machines, performance metrics, and mini-batch optimization.

Projection, principal component analysis, SVD, linear autoencoders, and k-means clustering.

Feature preparation and selection, nonlinear feature maps, model complexity, cross-validation, and kernel methods.

Network architecture, the chain rule, backpropagation, and practical neural-network training.

Decision trees, random forests, and gradient boosting.