Categorical Stochastic Processes and Likelihood
This work provides a theoretical framework for understanding statistical models through category theory, which is incremental in applying existing categorical concepts to stochastic processes.
The paper tackles the relationship between probabilistic modeling and function approximation using category theory, defining extensions of function composition for stochastic processes and applying the Para construction to parameterized statistical models, with a demonstration linking Maximum Likelihood Estimation to a functor between categories.
In this work we take a Category Theoretic perspective on the relationship between probabilistic modeling and function approximation. We begin by defining two extensions of function composition to stochastic process subordination: one based on the co-Kleisli category under the comonad (Omega x -) and one based on the parameterization of a category with a Lawvere theory. We show how these extensions relate to the category Stoch and other Markov Categories. Next, we apply the Para construction to extend stochastic processes to parameterized statistical models and we define a way to compose the likelihood functions of these models. We conclude with a demonstration of how the Maximum Likelihood Estimation procedure defines an identity-on-objects functor from the category of statistical models to the category of Learners. Code to accompany this paper can be found at https://github.com/dshieble/Categorical_Stochastic_Processes_and_Likelihood