A Probabilistic Model for Non-Contrastive Learning

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Fleissner, Maximilian, Esser, Pascal, Ghoshdastidar, Debarghya
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916758317694976
author Fleissner, Maximilian
Esser, Pascal
Ghoshdastidar, Debarghya
author_facet Fleissner, Maximilian
Esser, Pascal
Ghoshdastidar, Debarghya
contents Self-supervised learning (SSL) aims to find meaningful representations from unlabeled data by encoding semantic similarities through data augmentations. Despite its current popularity, theoretical insights about SSL are still scarce. For example, it is not yet known whether commonly used SSL loss functions can be related to a statistical model, much in the same as OLS, generalized linear models or PCA naturally emerge as maximum likelihood estimates of an underlying generative process. In this short paper, we consider a latent variable statistical model for SSL that exhibits an interesting property: Depending on the informativeness of the data augmentations, the MLE of the model either reduces to PCA, or approaches a simple non-contrastive loss. We analyze the model and also empirically illustrate our findings.
format Preprint
id arxiv_https___arxiv_org_abs_2501_13031
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A Probabilistic Model for Non-Contrastive Learning
Fleissner, Maximilian
Esser, Pascal
Ghoshdastidar, Debarghya
Machine Learning
Self-supervised learning (SSL) aims to find meaningful representations from unlabeled data by encoding semantic similarities through data augmentations. Despite its current popularity, theoretical insights about SSL are still scarce. For example, it is not yet known whether commonly used SSL loss functions can be related to a statistical model, much in the same as OLS, generalized linear models or PCA naturally emerge as maximum likelihood estimates of an underlying generative process. In this short paper, we consider a latent variable statistical model for SSL that exhibits an interesting property: Depending on the informativeness of the data augmentations, the MLE of the model either reduces to PCA, or approaches a simple non-contrastive loss. We analyze the model and also empirically illustrate our findings.
title A Probabilistic Model for Non-Contrastive Learning
topic Machine Learning
url https://arxiv.org/abs/2501.13031