Google’s Machine Learning Crash Course

About the resource

The Machine Learning Crash Course is a free course prepared by Google to introduce machine learning in a practical, concise way. The company used it to train its own staff before opening it to everyone, and it combines short videos, explanatory text, interactive visualisations in which you change values yourself and see the effect at once, and programming exercises. That interactive approach is one of its greatest strengths, because many machine learning ideas are hard to grasp in words alone and become clear when you watch them change in front of you.

What you will find in the course

• The core ideas: what a model is, how it learns from data, what a loss function is, and how gradient descent works.
• Linear regression, logistic regression and classification, and ways of measuring a model’s quality.
• Working with data: cleaning it, splitting it into training and test sets, and avoiding overfitting.
• An introduction to neural networks, and more recent topics related to large models.
• Units on responsibility and fairness in machine learning systems, and how projects move from experiment to real use.

Who it suits

It suits university students in technical and scientific subjects, and graduates and professionals who want a solid understanding of machine learning. You need basic Python programming and some secondary-school maths, such as functions, graphs and averages. The course is mainly in English, and some of its pages may appear automatically translated depending on your browser’s language.

How to use it, step by step

1. If you do not know Python yet, start with that first, then come back to this course.
2. Follow the units in order, and do not skip the interactive visualisations; change the values until you understand the effect of each one.
3. Do the programming exercises yourself rather than just reading the solutions.
4. After each unit, explain what you learned to a friend or write it in a short paragraph; if you cannot explain it simply, repeat the unit.
5. At the end, choose a simple dataset and apply a model to it, from start to evaluating its results.

Before you start

The course is a “crash course”, as its name says: it covers a lot in a short time, and it is normal to return to some units more than once. To complete the picture, read “A beginner’s map for learning AI” on the academy’s blog to see where this course sits in the learning journey. In the artificial intelligence course at Badr Academy Education we apply these ideas through projects with the instructor’s follow-up. The content and its rights belong to Google.

Links

• The resource in the library
• The original source – https://developers.google.com/machine-learning/crash-course