A huge part of the AI revolution is the fact that we can now interact with these models using natural language -- through text and speech.

But computers don't know what language is, what words are, and what they represent (objects, emotions and so on). So how do we find ways to represent words and language in a computer so that it can be manipulated by them in AI applications?

Well it might surprise you to learn that basically words are turned into lists of numbers. AI systems manipulate those lists in different ways and then turn them back into words.

In this chapter we go over exactly how those lists are created. We show how they can be used in practice for a simple spam classification problem by introducing the first machine learning algorithm in this course: nearest neighbours.

We also have some fun with text generation and take a look at hallucination, which is a big topic for these large AI language models. This chapter lays a big foundation in understanding how the modern Large Language Models work.