Neural Network Symmetrisation in Concrete Settings

TL;DR

This paper reviews Cornish's 2024 general theory of neural network symmetrisation within Markov categories and discusses its practical implications for symmetrising deterministic functions and Markov kernels.

Contribution

It provides a high-level overview of the theoretical framework and explores its concrete applications in neural network symmetrisation.

Findings

Theoretical framework for neural network symmetrisation in Markov categories

Implications for symmetrising deterministic functions

Implications for symmetrising Markov kernels

Abstract

Cornish (2024) recently gave a general theory of neural network symmetrisation in the abstract context of Markov categories. We give a high-level overview of these results, and their concrete implications for the symmetrisation of deterministic functions and of Markov kernels.