Foundations
Supervised learning
Learning a function from labelled examples with nearest neighbours, linear regression and logistic regression, and telling whether the model will hold up on data it has not seen. The train, validation and test split used here is the same one used to judge a language model.