Signal-to-Noise Ratio and Sample Size Govern Representational Alignment in Neural Networks

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Umar, Ali Hussaini, Laio, Alessandro
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866913164064456704
author Umar, Ali Hussaini
Laio, Alessandro
author_facet Umar, Ali Hussaini
Laio, Alessandro
contents Neural networks are known to develop latent representations that are $aligned$, namely structurally similar across networks trained with different architectures, training protocols, or training datasets. We study this phenomenon in a controlled setting, where we train an ensemble of networks on regression and classification tasks using training sets perturbed by independent realizations of a noise process. We show that the signal-to-noise ratio (SNR) and the training sample size influence the alignment in qualitatively similar ways in networks trained on real-world datasets and in an extremely simple $linear$ network with a single hidden layer, for which the alignment can be estimated analytically. Across linear and nonlinear networks, regression and classification tasks, and both synthetic and real-world data, we consistently observe that alignment varies monotonically with SNR but non-monotonically with training sample size. In particular, the alignment is minimized near the interpolation threshold, and a stronger alignment does not necessarily correspond to better generalization error. These findings reveal a non-trivial dependence of alignment on data quality and quantity, decoupled from generalization performance.
format Preprint
id arxiv_https___arxiv_org_abs_2605_26973
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Signal-to-Noise Ratio and Sample Size Govern Representational Alignment in Neural Networks
Umar, Ali Hussaini
Laio, Alessandro
Machine Learning
Disordered Systems and Neural Networks
Neural and Evolutionary Computing
Neurons and Cognition
Neural networks are known to develop latent representations that are $aligned$, namely structurally similar across networks trained with different architectures, training protocols, or training datasets. We study this phenomenon in a controlled setting, where we train an ensemble of networks on regression and classification tasks using training sets perturbed by independent realizations of a noise process. We show that the signal-to-noise ratio (SNR) and the training sample size influence the alignment in qualitatively similar ways in networks trained on real-world datasets and in an extremely simple $linear$ network with a single hidden layer, for which the alignment can be estimated analytically. Across linear and nonlinear networks, regression and classification tasks, and both synthetic and real-world data, we consistently observe that alignment varies monotonically with SNR but non-monotonically with training sample size. In particular, the alignment is minimized near the interpolation threshold, and a stronger alignment does not necessarily correspond to better generalization error. These findings reveal a non-trivial dependence of alignment on data quality and quantity, decoupled from generalization performance.
title Signal-to-Noise Ratio and Sample Size Govern Representational Alignment in Neural Networks
topic Machine Learning
Disordered Systems and Neural Networks
Neural and Evolutionary Computing
Neurons and Cognition
url https://arxiv.org/abs/2605.26973