1. What is this?

This is a set of lecture notes on numerical methods for machine learning. This is not a book. A book is generally comprehensive and can stand alone. Again, this is not a book.

Lecture notes generally follow the order of topics in a class; consequently, topics may jump around. Moreover, these notes are often just examples of methods with minimal written content. Most importantly, there is no expectation that the presentation of the material be complete—the material is intended to be paired with in-class activities.

If you are in my numerical methods course, then you know that an abundance of freely available textbooks have been assigned (in case you really want that book).

1.1. Who is This For?

These notes are for students taking my numerical methods for machine learning course. Each student is expected to have some familiarity with common machine learning methods and be able to code in Python. Mathematics is necessary, so students should have familiarity with single and multivariable calculus and basic linear algebra.

1.2. What is Numerical Methods for Machine Learning?

Numerical methods for machine learning is the “nuts and bolts” of machine learning—the “what goes on underneath the hood,” the “how it works.”

Machine learning methods are mathematical algorithms written into code that extract information from data. This class is about learning how those mathematical algorithms work and how to write them “from scratch” into code.

Your previous experience with machine learning may have involved using imported libraries like sklearn or TensorFlow. A good chunk of your effort was likely preparing data for the model and analyzing the output. This class is different. Our focus is on the model itself. To understand the model, we don’t use an imported library—instead, we build our model ourselves.

The table of contents lists the topics covered in these notes. Go ahead and have a look at that now. I hope you are excited to learn “how this all works.”

1.3. Under Construction

These notes are under construction. Expect continuous updates. Additionally, your feedback—whether catching mistakes or making suggestions—is valuable and welcomed. Thank you!