Tai-Danae Bradley: Category Theory and Language Models

Ever wanted to learn about category theory? Did you know it can offer insights into language models of today’s AI rage? Now you can with my latest podcast guest, the wonderful Tai-Danae Bradley over at the Cartesian Cafe. We talk about the basic constructions and definitions of category theory, providing lots of definitions as well as analogies inspired by computer science. We then delve into the Yoneda Lemma, a fundamental result in category theory, and apply it to give insights into language models by clarifying the notions of syntax, word probabilities, and semantics.