Neural Discovery in Mathematics: Do Machines Dream of Colored Planes?

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Mundinger, Konrad, Zimmer, Max, Kiem, Aldo, Spiegel, Christoph, Pokutta, Sebastian
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915326231314432
author Mundinger, Konrad
Zimmer, Max
Kiem, Aldo
Spiegel, Christoph
Pokutta, Sebastian
author_facet Mundinger, Konrad
Zimmer, Max
Kiem, Aldo
Spiegel, Christoph
Pokutta, Sebastian
contents We demonstrate how neural networks can drive mathematical discovery through a case study of the Hadwiger-Nelson problem, a long-standing open problem at the intersection of discrete geometry and extremal combinatorics that is concerned with coloring the plane while avoiding monochromatic unit-distance pairs. Using neural networks as approximators, we reformulate this mixed discrete-continuous geometric coloring problem with hard constraints as an optimization task with a probabilistic, differentiable loss function. This enables gradient-based exploration of admissible configurations that most significantly led to the discovery of two novel six-colorings, providing the first improvement in thirty years to the off-diagonal variant of the original problem. Here, we establish the underlying machine learning approach used to obtain these results and demonstrate its broader applicability through additional numerical insights.
format Preprint
id arxiv_https___arxiv_org_abs_2501_18527
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Neural Discovery in Mathematics: Do Machines Dream of Colored Planes?
Mundinger, Konrad
Zimmer, Max
Kiem, Aldo
Spiegel, Christoph
Pokutta, Sebastian
Machine Learning
Combinatorics
We demonstrate how neural networks can drive mathematical discovery through a case study of the Hadwiger-Nelson problem, a long-standing open problem at the intersection of discrete geometry and extremal combinatorics that is concerned with coloring the plane while avoiding monochromatic unit-distance pairs. Using neural networks as approximators, we reformulate this mixed discrete-continuous geometric coloring problem with hard constraints as an optimization task with a probabilistic, differentiable loss function. This enables gradient-based exploration of admissible configurations that most significantly led to the discovery of two novel six-colorings, providing the first improvement in thirty years to the off-diagonal variant of the original problem. Here, we establish the underlying machine learning approach used to obtain these results and demonstrate its broader applicability through additional numerical insights.
title Neural Discovery in Mathematics: Do Machines Dream of Colored Planes?
topic Machine Learning
Combinatorics
url https://arxiv.org/abs/2501.18527