Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2502.10335

Studying number theory with deep learning: a case study with the Möbius and squarefree indicator functions

David Lowry-Duda
View original source ↗Read source document ↗Export metadata ↓

Abstract

Building on work of Charton, we train small transformer models to calculate the Möbius function $μ(n)$ and the squarefree indicator function $μ^2(n)$. The models attain nontrivial predictive power. We apply a mixture of additional models and feature scoring to give a theoretical explanation.

Explore related subjects

math.NTcs.LG

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

David Lowry-Duda. 2025-07-25. Studying number theory with deep learning: a case study with the Möbius and squarefree indicator functions. https://arxiv.org/abs/2502.10335

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Asymptotic density of k-almost primes

Landau's well known asymptotic formula $$N_k(x):=\ \mid\{n\leq x : Ω(n)=k\}\mid \ \sim \left( \frac{x}{\log x} \right) \frac{(\log\log x)^{k-1}}{(k - 1)!}\ \ (x \rightarrow \infty),$$ which also holds for $$π_k(x):=\ \mid\{n\leq x : ω(n)=k\}\mid,$$ is known to be fairly poor for $k > 1$, and when $k$ is allowed to tend to infinity with $x$, the study of $N_k(x)$ and $π_k(x)$ becomes very technical [1, Chapter II.6, $§$ 6.1, p.200]. I hope to show that the method described below provides not only a more accurate approach, but rather increases in its asymptotic accuracy as $k$ tends to infinity.

math.NT↗
arXiv2026

Transcendence Meets Normality: Construction of Transcendentally Normal Numbers

In this work, we study real numbers $x$ for which $p(x)$ is (absolutely) normal for every non-constant integer-valued polynomial $p$. We call such numbers transcendentally normal. We prove that almost every real number is transcendentally normal and provide an explicit construction of such a number, based on Sierpinski's covering method and novel ideas involving the so-called stretch function. In the next step, we transform this construction into an algorithm that computes the digits of a t-normal number recursively in all integer bases. Moreover, we extend our covering approach to construct and compute LIL-normal numbers whose discrepancies are of the order predicted by the law of the iterated logarithm. We also take the opportunity to discuss several interesting open problems.

math.NT↗