Presenting Neural Networks via Coherent Functors

Matthew Pugh, Jo Grundy, Corina Cirstea, Nick Harris · 2026

Read Presenting Neural Networks via Coherent Functors by Matthew Pugh, Jo Grundy, Corina Cirstea, Nick Harris (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2604-15100.txt.

Description

Pugh and colleagues recast neural networks as models of formal theories, treating machine learning models as databases. They show any functorial database schema carries a coherent theory whose models match its instances, and data can be hard-coded into a coherent category. With these tools, any dense feed-forward network over floating-point numbers becomes a coherent category whose models are that architecture's networks. Inference then appears as a functor along a coherent map, unifying networks with logic.

Catalog record

CREATORMatthew Pugh, Jo Grundy, Corina Cirstea, Nick Harris
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE68849
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2604.15100

Browse the full library · About the Tower of Babel