# Welcome toAn Interactive Journey into Machine Learning** Warning: this is a work in progress, still incomplete and possibly including errors

## In words...

This webbook proposes an interactive journey into Machine Learning, in which each notion is first described with words, then illustrated with pictures and finally formalized in mathematical terms. Most of the pictures include some sort of interaction so that you can play with and explore various settings to understand what is really going on.

## In pictures...

This webbook illustrates each topic with pictures and interactive demos...

Double-click on a bubble to be taken to its topic directly.

## In maths...

This webbook should also give you the required mathematical details to understand what is going on in the pictures.

Click here to learn how we can predict $y\in\mathcal{Y}$ given $x\in\mathcal{X}$ and a set of pairs $(x_i, y_i)$...

## Notes

### Why this webbook?

Some people need to hear a lot of words to get an idea of a subject, others need to see images before they can picture it for themselves, and yet others will not be satisfied until they see a clear mathematical formulation. And often people actually need these three media together to fully assimilate a topic. This is what this webbook tries to provide in an attempt to make it easy for most people eager to learn machine learning.

How is that different from a classical book?

A classical book offers words, pictures and maths, but not systematically. For instance, some topics are mostly discussed in words, while others are only detailed in maths; and often only a few very important or difficult points are illustrated by pictures. On the other hand, this book has been written with a single hard constraint: provide words, pictures and maths on a single page for all the points discussed in the book, independently of the importance or difficulty of the points.
Another difference is that this book is a webbook, which is interactive, making the pictures react to the reader input. Thus, an infinite number of illustrated examples can be generated to help the reader fully understand the highlighted phenomenon.

How is that different from Wikipedia?

This format of the book is close to what could be found on Wikipedia, with hyperlinks between notions and pictures. But a major difference with a collection of pages on Wikipedia is that, as for all standard books, the notations and presentation is consistent and homogeneous throughtout all the pages, including the Appendix discussing matters that can be far from the main subject of the book. This should make it easier to go from one topic to the other without having to delve into new notations every time.

### What this webbook is not

• a comprehensive encyclopedia of machine learning (many topics are not covered by the book)
• a course (though a linear reading track is depicted in the Table of contents, it is not the main purpose of the book; also, the book offers very few exercises and no teacher)