Saved in:
Bibliographic Details
Main Authors: Mendes, Vicente Conde, Bardone, Lorenzo, Koller, Cédric, Moreira, Jorge Medina, Erba, Vittorio, Troiani, Emanuele, Zdeborová, Lenka
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2602.10680
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915791349219328
author Mendes, Vicente Conde
Bardone, Lorenzo
Koller, Cédric
Moreira, Jorge Medina
Erba, Vittorio
Troiani, Emanuele
Zdeborová, Lenka
author_facet Mendes, Vicente Conde
Bardone, Lorenzo
Koller, Cédric
Moreira, Jorge Medina
Erba, Vittorio
Troiani, Emanuele
Zdeborová, Lenka
contents Many real-world datasets contain hidden structure that cannot be detected by simple linear correlations between input features. For example, latent factors may influence the data in a coordinated way, even though their effect is invisible to covariance-based methods such as PCA. In practice, nonlinear neural networks often succeed in extracting such hidden structure in unsupervised and self-supervised learning. However, constructing a minimal high-dimensional model where this advantage can be rigorously analyzed has remained an open theoretical challenge. We introduce a tractable high-dimensional spiked model with two latent factors: one visible to covariance, and one statistically dependent yet uncorrelated, appearing only in higher-order moments. PCA and linear autoencoders fail to recover the latter, while a minimal nonlinear autoencoder provably extracts both. We analyze both the population risk, and empirical risk minimization. Our model also provides a tractable example where self-supervised test loss is poorly aligned with representation quality: nonlinear autoencoders recover latent structure that linear methods miss, even though their reconstruction loss is higher.
format Preprint
id arxiv_https___arxiv_org_abs_2602_10680
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle A solvable high-dimensional model where nonlinear autoencoders learn structure invisible to PCA while test loss misaligns with generalization
Mendes, Vicente Conde
Bardone, Lorenzo
Koller, Cédric
Moreira, Jorge Medina
Erba, Vittorio
Troiani, Emanuele
Zdeborová, Lenka
Machine Learning
Disordered Systems and Neural Networks
Many real-world datasets contain hidden structure that cannot be detected by simple linear correlations between input features. For example, latent factors may influence the data in a coordinated way, even though their effect is invisible to covariance-based methods such as PCA. In practice, nonlinear neural networks often succeed in extracting such hidden structure in unsupervised and self-supervised learning. However, constructing a minimal high-dimensional model where this advantage can be rigorously analyzed has remained an open theoretical challenge. We introduce a tractable high-dimensional spiked model with two latent factors: one visible to covariance, and one statistically dependent yet uncorrelated, appearing only in higher-order moments. PCA and linear autoencoders fail to recover the latter, while a minimal nonlinear autoencoder provably extracts both. We analyze both the population risk, and empirical risk minimization. Our model also provides a tractable example where self-supervised test loss is poorly aligned with representation quality: nonlinear autoencoders recover latent structure that linear methods miss, even though their reconstruction loss is higher.
title A solvable high-dimensional model where nonlinear autoencoders learn structure invisible to PCA while test loss misaligns with generalization
topic Machine Learning
Disordered Systems and Neural Networks
url https://arxiv.org/abs/2602.10680