Machine learning courses and concepts

Machine learning is the branch of AI in which a system learns to do a task by finding patterns in examples, rather than being programmed with a rule for every case. It is the layer almost everything else rests on — the models behind generative tools and agents are themselves machine-learned — and it remains a working field in its own right wherever the job is prediction, classification or ranking. Learning it turns out to be less about algorithms than about the discipline around them: what you trained on, and how you know the result is any good.

The vocabulary of machine learning

A striking share of these terms are units of measurement. Accuracy, precision and recall, and the various ways a set of predictions gets laid out for inspection — this is the language of judging whether a model is any good, and it is the part a course outline assumes you already have. It is also what lets you assess a number somebody else shows you: a model reported as accurate can still be useless, and this vocabulary is how you tell.

Courses on machine learning

The dividing question is whether your problem is framing the work or building the model. One group covers the decisions made before any model exists — whether machine learning is the right tool at all, what data would have to exist, what a good outcome would even look like; the other is the build itself, from preparing data through training to keeping a deployed model honest, and the two are rarely taught together.

Learn AI — the ideas, without building anything

Use AI — use trained models and services

Develop AI — build, train and run models

Machine learning is the common ancestor of most of the rest of AI, so it tends to be a subject you return to rather than finish. Learn AI puts it in context alongside the newer things built on top of it; Use AI covers applying models and services without training anything yourself; Develop AI is the hands-on route — building, evaluating and running models as a job.

← Browse all courses