Generalization analysis of deep nonlinear matrix completion
We provide generalization bounds for matrix completion with Schatten $p$ quasi-norm constraints, which is equivalent to deep matrix factorization with Frobenius constraints. In the uniform sampling regime, the sample complexity scales like $\widetilde{O}\left( rn\right)$ where $n$ is the size of the...
Saved in:
Main Authors: | , |
---|---|
Format: | text |
Language: | English |
Published: |
Institutional Knowledge at Singapore Management University
2024
|
Subjects: | |
Online Access: | https://ink.library.smu.edu.sg/sis_research/9303 https://ink.library.smu.edu.sg/context/sis_research/article/10303/viewcontent/ledent24a.pdf |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Institution: | Singapore Management University |
Language: | English |
id |
sg-smu-ink.sis_research-10303 |
---|---|
record_format |
dspace |
spelling |
sg-smu-ink.sis_research-103032024-09-21T15:30:39Z Generalization analysis of deep nonlinear matrix completion LEDENT, Antoine ALVES, Rodrigo We provide generalization bounds for matrix completion with Schatten $p$ quasi-norm constraints, which is equivalent to deep matrix factorization with Frobenius constraints. In the uniform sampling regime, the sample complexity scales like $\widetilde{O}\left( rn\right)$ where $n$ is the size of the matrix and $r$ is a constraint of the same order as the ground truth rank in the isotropic case. In the distribution-free setting, the bounds scale as $\widetilde{O}\left(r^{1-\frac{p}{2}}n^{1+\frac{p}{2}}\right)$, which reduces to the familiar $\sqrt{r}n^{\frac{3}{2}}$ for $p=1$. Furthermore, we provide an analogue of the weighted trace norm for this setting which brings the sample complexity down to $\widetilde{O}(nr)$ in all cases. We then present a non-linear model, Functionally Rescaled Matrix Completion (FRMC) which applies a single trainable function from $\R\rightarrow \R$ to each entry of a latent matrix, and prove that this adds only negligible terms of the overall sample complexity, whilst experiments demonstrate that this simple model improvement already leads to significant gains on real data. We also provide extensions of our results to various neural architectures, thereby providing the first comprehensive uniform convergence PAC analysis of neural network matrix completion. 2024-07-01T07:00:00Z text application/pdf https://ink.library.smu.edu.sg/sis_research/9303 https://ink.library.smu.edu.sg/context/sis_research/article/10303/viewcontent/ledent24a.pdf http://creativecommons.org/licenses/by-nc-nd/4.0/ Research Collection School Of Computing and Information Systems eng Institutional Knowledge at Singapore Management University Matrix Completion Schatten p Quasi Norms Learning Theory Databases and Information Systems |
institution |
Singapore Management University |
building |
SMU Libraries |
continent |
Asia |
country |
Singapore Singapore |
content_provider |
SMU Libraries |
collection |
InK@SMU |
language |
English |
topic |
Matrix Completion Schatten p Quasi Norms Learning Theory Databases and Information Systems |
spellingShingle |
Matrix Completion Schatten p Quasi Norms Learning Theory Databases and Information Systems LEDENT, Antoine ALVES, Rodrigo Generalization analysis of deep nonlinear matrix completion |
description |
We provide generalization bounds for matrix completion with Schatten $p$ quasi-norm constraints, which is equivalent to deep matrix factorization with Frobenius constraints. In the uniform sampling regime, the sample complexity scales like $\widetilde{O}\left( rn\right)$ where $n$ is the size of the matrix and $r$ is a constraint of the same order as the ground truth rank in the isotropic case. In the distribution-free setting, the bounds scale as $\widetilde{O}\left(r^{1-\frac{p}{2}}n^{1+\frac{p}{2}}\right)$, which reduces to the familiar $\sqrt{r}n^{\frac{3}{2}}$ for $p=1$. Furthermore, we provide an analogue of the weighted trace norm for this setting which brings the sample complexity down to $\widetilde{O}(nr)$ in all cases. We then present a non-linear model, Functionally Rescaled Matrix Completion (FRMC) which applies a single trainable function from $\R\rightarrow \R$ to each entry of a latent matrix, and prove that this adds only negligible terms of the overall sample complexity, whilst experiments demonstrate that this simple model improvement already leads to significant gains on real data. We also provide extensions of our results to various neural architectures, thereby providing the first comprehensive uniform convergence PAC analysis of neural network matrix completion. |
format |
text |
author |
LEDENT, Antoine ALVES, Rodrigo |
author_facet |
LEDENT, Antoine ALVES, Rodrigo |
author_sort |
LEDENT, Antoine |
title |
Generalization analysis of deep nonlinear matrix completion |
title_short |
Generalization analysis of deep nonlinear matrix completion |
title_full |
Generalization analysis of deep nonlinear matrix completion |
title_fullStr |
Generalization analysis of deep nonlinear matrix completion |
title_full_unstemmed |
Generalization analysis of deep nonlinear matrix completion |
title_sort |
generalization analysis of deep nonlinear matrix completion |
publisher |
Institutional Knowledge at Singapore Management University |
publishDate |
2024 |
url |
https://ink.library.smu.edu.sg/sis_research/9303 https://ink.library.smu.edu.sg/context/sis_research/article/10303/viewcontent/ledent24a.pdf |
_version_ |
1814047875167944704 |