---
title: "Accelerating Machine Learning Systems via Category Theory: Applications to Spherical Attention for Gene Regulatory Networks"
authors:
  - "Vincent Abbott"
  - "Kotaro Kamiya"
  - "Gerard Glowacki"
year: 2025
venue: "AGI 2025, Springer LNCS"
arxiv: null
doi: "10.1007/978-3-032-00686-8_1"
source: "https://doi.org/10.1007/978-3-032-00686-8_1"
topics:
  - "category-theory"
  - "spherical-attention"
  - "ml-systems"
seed_rank: 936
seed_batch: "user-add-2026-08-13"
reviewed: "2026-08-13"
pool: "languages"
relevance_score: 8
lineage: categorical-deep-learning
cites:
  - title: "Neural Circuit Diagrams: Robust Diagrams for the Communication, Implementation, and Analysis of Deep Learning Architectures"
    url: "https://arxiv.org/abs/2402.05424"
    year: 2024
    arxiv: "2402.05424"
see:
  - "934-neural-circuit-diagrams"
---

# Accelerating Machine Learning Systems via Category Theory: Applications to Spherical Attention for Gene Regulatory Networks

## One-sentence takeaway

Category-theory constructions are used to accelerate an ML system, with spherical attention on gene-regulatory networks as the worked non-grid geometry.

## Why it matters here

Abbott-lab systems paper sitting between NCDs and the 2026 weaves framework — same algebra, different domain. Shows the diagram/term story is not locked to Euclidean grids.

## Key ideas

- AGI 2025 proceedings Part I, pp. 1–11; DOI `10.1007/978-3-032-00686-8_1`.
- Treats architecture acceleration as a categorical construction rather than a one-off kernel.
- Spherical attention: a geometry without a grid that still has to broadcast, so the axis/weaving story still applies.
- Short conference paper; the algebra is developed more fully in Weaves 2026.

## Caveats

## Links

- DOI: [10.1007/978-3-032-00686-8_1](https://doi.org/10.1007/978-3-032-00686-8_1)