Information, Geometry, and Physics Seminar

The success of large language models is striking, especially given that the training data consists of raw, unstructured text. In this talk, we will propose that category theory can provide a natural framework for investigating this passage from texts, and probability distributions on them, to a more semantically meaningful space. Equipped with motivation from an analogy between linear algebra and category theory, we will define a category of expressions in language enriched over the unit interval and afterwards pass to enriched copresheaves on that category. We will see that the latter setting has rich mathematical structure and comes with ready-made tools in which to explore that structure.