8
Matrices and Graphs

Now that we have gotten past the hard part (that is, the singular value decomposition and other matrix factorizations), it’s time to finish our journey through linear algebra with a bang. In my teaching experience, one of students’ most common concerns is the apparent disconnect between practice and theory. Among machine learning practitioners and software engineers, there’s often a reluctance to touch anything that is not immediately valuable in practice.

As a mathematician, I completely get where this dread comes from. We are often taught arcane topics of no practical importance, taking valuable time away from hacking and slashing our way through data.

In this chapter, we’ll look at a subject that is not immediately useful for your machine learning practice, but will pay serious dividends in the future. Considering how beautiful it is, it might be the inspiration for your next genius idea. (No promises, though.)

Let me introduce you...