Rust for Machine Learning

Chapter 2. The Search for Structure

A Note for Early Release Readers

With Early Release ebooks, you get books in their earliest form—the author’s raw and unedited content as they write—so you can take advantage of these technologies long before the official release of these titles.

This will be the 2nd chapter of the final book. Please note that the GitHub repo will be made active later on.

If you’d like to be actively involved in reviewing and commenting on this draft, please reach out to the editor at [e-mail].

All models are wrong, some are useful.

George Box

In the last chapter we’ve implemented the perceptron, an algorithm that learned a separating line from a handful of data points. We saw how the perceptron adjusted its parameters every iteration until it found a boundary that correctly classified the data.

The data we used were the truth tables for AND and OR, two ubiquitous boolean functions. In fact, when we are first learning logic, the boolean functions are many times defined in terms of their truth tables. In this case, data and function coincide. The four examples we’ve seen tell us everything we need to know about these functions. There is nothing left unknown.

But that is rarely the case in the real world.

When we speak about “data out there in the world,” we usually imagine measurements like heights of animals, contract values, pixel intensities, sensor readings. Yet from a modeling perspective, data is not merely a list of numbers. It is the visible ...