Complexity of One-Dimensional ReLU DNNs

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Kogan, Jonathan, Jananthan, Hayden, Kepner, Jeremy
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866908699669299200
author Kogan, Jonathan
Jananthan, Hayden
Kepner, Jeremy
author_facet Kogan, Jonathan
Jananthan, Hayden
Kepner, Jeremy
contents We study the expressivity of one-dimensional (1D) ReLU deep neural networks through the lens of their linear regions. For randomly initialized, fully connected 1D ReLU networks (He scaling with nonzero bias) in the infinite-width limit, we prove that the expected number of linear regions grows as $\sum_{i = 1}^L n_i + \mathop{o}\left(\sum_{i = 1}^L{n_i}\right) + 1$, where $n_\ell$ denotes the number of neurons in the $\ell$-th hidden layer. We also propose a function-adaptive notion of sparsity that compares the expected regions used by the network to the minimal number needed to approximate a target within a fixed tolerance.
format Preprint
id arxiv_https___arxiv_org_abs_2512_08091
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Complexity of One-Dimensional ReLU DNNs
Kogan, Jonathan
Jananthan, Hayden
Kepner, Jeremy
Machine Learning
We study the expressivity of one-dimensional (1D) ReLU deep neural networks through the lens of their linear regions. For randomly initialized, fully connected 1D ReLU networks (He scaling with nonzero bias) in the infinite-width limit, we prove that the expected number of linear regions grows as $\sum_{i = 1}^L n_i + \mathop{o}\left(\sum_{i = 1}^L{n_i}\right) + 1$, where $n_\ell$ denotes the number of neurons in the $\ell$-th hidden layer. We also propose a function-adaptive notion of sparsity that compares the expected regions used by the network to the minimal number needed to approximate a target within a fixed tolerance.
title Complexity of One-Dimensional ReLU DNNs
topic Machine Learning
url https://arxiv.org/abs/2512.08091