GSVD for Geometry-Grounded Dataset Comparison: An Alignment Angle Is All You Need

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Marques, Eduarda de Souza, da Rocha, Arthur Sobrinho Ferreira, Paixao, Joao, Mirandola, Heudson, Menasche, Daniel Sadoc
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866908878050951168
author Marques, Eduarda de Souza
da Rocha, Arthur Sobrinho Ferreira
Paixao, Joao
Mirandola, Heudson
Menasche, Daniel Sadoc
author_facet Marques, Eduarda de Souza
da Rocha, Arthur Sobrinho Ferreira
Paixao, Joao
Mirandola, Heudson
Menasche, Daniel Sadoc
contents Geometry-grounded learning asks models to respect structure in the problem domain rather than treating observations as arbitrary vectors. Motivated by this view, we revisit a classical but underused primitive for comparing datasets: linear relations between two data matrices, expressed via the co-span constraint $Ax = By = z$ in a shared ambient space. To operationalize this comparison, we use the generalized singular value decomposition (GSVD) as a joint coordinate system for two subspaces. In particular, we exploit the GSVD form $A = HCU$, $B = HSV$ with $C^{\top}C + S^{\top}S = I$, which separates shared versus dataset-specific directions through the diagonal structure of $(C, S)$. From these factors we derive an interpretable *angle score* $θ(z) \in [0, π/2]$ for a sample $z$, quantifying whether z is explained relatively more by $A$, more by $B$, or comparably by both. The primary role of $θ(z)$ is as a *per-sample geometric diagnostic*. We illustrate the behavior of the score on MNIST through angle distributions and representative GSVD directions. A binary classifier derived from $θ(z)$ is presented as an illustrative application of the score as an interpretable diagnostic tool.
format Preprint
id arxiv_https___arxiv_org_abs_2603_10283
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle GSVD for Geometry-Grounded Dataset Comparison: An Alignment Angle Is All You Need
Marques, Eduarda de Souza
da Rocha, Arthur Sobrinho Ferreira
Paixao, Joao
Mirandola, Heudson
Menasche, Daniel Sadoc
Machine Learning
Geometry-grounded learning asks models to respect structure in the problem domain rather than treating observations as arbitrary vectors. Motivated by this view, we revisit a classical but underused primitive for comparing datasets: linear relations between two data matrices, expressed via the co-span constraint $Ax = By = z$ in a shared ambient space. To operationalize this comparison, we use the generalized singular value decomposition (GSVD) as a joint coordinate system for two subspaces. In particular, we exploit the GSVD form $A = HCU$, $B = HSV$ with $C^{\top}C + S^{\top}S = I$, which separates shared versus dataset-specific directions through the diagonal structure of $(C, S)$. From these factors we derive an interpretable *angle score* $θ(z) \in [0, π/2]$ for a sample $z$, quantifying whether z is explained relatively more by $A$, more by $B$, or comparably by both. The primary role of $θ(z)$ is as a *per-sample geometric diagnostic*. We illustrate the behavior of the score on MNIST through angle distributions and representative GSVD directions. A binary classifier derived from $θ(z)$ is presented as an illustrative application of the score as an interpretable diagnostic tool.
title GSVD for Geometry-Grounded Dataset Comparison: An Alignment Angle Is All You Need
topic Machine Learning
url https://arxiv.org/abs/2603.10283