Covers the fundamentals of machine learning tools used in materials science. Application areas include surrogate models for atomistic simulations as well as machine-learning tools for 3-dimensional image analysis and the analysis of experimental data. The course starts with an overview of probability theory and Bayesian inference followed by treatments of regularized regression, Gaussian process models and neural networks. Concepts of invariance and equivariance in the context of spatial transformations will be developed and neural network architectures that are equivariant to translational and rotational transformations will be analyzed.