Fiber bounds for Gaussian tensors with arbitrary variance profiles

For a centered real Gaussian tensor of order r with independent entries and arbitrary deterministic variances, let F_j be the expected largest Euclidean norm of a fiber in direction j. We prove that the expected injective norm lies between max_j F_j and an explicit universal constant times r^(1/4) ∑_j F_j, uniformly over all rectangular formats and variance profiles. The resulting all-profile comparison factor is at most C r^(5/4); dense Gaussian tensors give a lower bound of order √(r log(er)). The upper estimate follows from a comparison of Gaussian increments that retains the full one-coordinate marginals, together with a bound for Gaussian widths indexed jointly by a Euclidean ball and a set of nonnegative multipliers. A finite family shows that the squared increment comparison incurs a loss of order √r even when the tensor norm itself stays comparable to the largest fiber. The optimal dependence on r for arbitrary profiles is not determined.

Authors

Publication Details

Journal
Zenodo (CERN European Organization for Nuclear Research)
Published
2026-11-18
DOI
https://doi.org/10.5281/zenodo.24330189
Primary Topic
Tensor decomposition and applications
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Fiber bounds for Gaussian tensors with arbitrary variance profiles

Fomalhaut
2 links
Zenodo (CERN European Organization for Nuclear Research)
Tensor decomposition and applications
preprint

Fiber bounds for Gaussian tensors with arbitrary variance profiles

Fomalhaut
preprint en

Abstract

For a centered real Gaussian tensor of order r with independent entries and arbitrary deterministic variances, let F_j be the expected largest Euclidean norm of a fiber in direction j. We prove that the expected injective norm lies between max_j F_j and an explicit universal constant times r^(1/4) ∑_j F_j, uniformly over all rectangular formats and variance profiles. The resulting all-profile comparison factor is at most C r^(5/4); dense Gaussian tensors give a lower bound of order √(r log(er)). The upper estimate follows from a comparison of Gaussian increments that retains the full one-coordinate marginals, together with a bound for Gaussian widths indexed jointly by a Euclidean ball and a set of nonnegative multipliers. A finite family shows that the squared increment comparison incurs a loss of order √r even when the tensor norm itself stays comparable to the largest fiber. The optimal dependence on r for arbitrary profiles is not determined.

Zenodo (CERN European Organization for Nuclear Research)
Tensor decomposition and applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon