Guardado en:
Detalles Bibliográficos
Autores principales: Simai, Idan, Talmon, Ronen, Shaham, Uri
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:https://arxiv.org/abs/2506.18104
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866915354894139392
author Simai, Idan
Talmon, Ronen
Shaham, Uri
author_facet Simai, Idan
Talmon, Ronen
Shaham, Uri
contents In this paper, we argue that viewing VICReg-a popular self-supervised learning (SSL) method--through the lens of spectral embedding reveals a potential source of sub-optimality: it may struggle to generalize robustly to unseen data due to overreliance on the training data. This observation invites a closer look at how well this method achieves its goal of producing meaningful representations of images outside of the training set as well. Here, we investigate this issue and introduce SAG-VICReg (Stable and Generalizable VICReg), a method that builds on VICReg by incorporating new training techniques. These enhancements improve the model's ability to capture global semantics within the data and strengthen the generalization capabilities. Experiments demonstrate that SAG-VICReg effectively addresses the generalization challenge while matching or surpassing diverse state-of-the-art SSL baselines. Notably, our method exhibits superior performance on metrics designed to evaluate global semantic understanding, while simultaneously maintaining competitive results on local evaluation metrics. Furthermore, we propose a new standalone evaluation metric for embeddings that complements the standard evaluation methods and accounts for the global data structure without requiring labels--a key issue when tagged data is scarce or not available.
format Preprint
id arxiv_https___arxiv_org_abs_2506_18104
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Enhancing VICReg: Random-Walk Pairing for Improved Generalization and Better Global Semantics Capturing
Simai, Idan
Talmon, Ronen
Shaham, Uri
Computer Vision and Pattern Recognition
Machine Learning
In this paper, we argue that viewing VICReg-a popular self-supervised learning (SSL) method--through the lens of spectral embedding reveals a potential source of sub-optimality: it may struggle to generalize robustly to unseen data due to overreliance on the training data. This observation invites a closer look at how well this method achieves its goal of producing meaningful representations of images outside of the training set as well. Here, we investigate this issue and introduce SAG-VICReg (Stable and Generalizable VICReg), a method that builds on VICReg by incorporating new training techniques. These enhancements improve the model's ability to capture global semantics within the data and strengthen the generalization capabilities. Experiments demonstrate that SAG-VICReg effectively addresses the generalization challenge while matching or surpassing diverse state-of-the-art SSL baselines. Notably, our method exhibits superior performance on metrics designed to evaluate global semantic understanding, while simultaneously maintaining competitive results on local evaluation metrics. Furthermore, we propose a new standalone evaluation metric for embeddings that complements the standard evaluation methods and accounts for the global data structure without requiring labels--a key issue when tagged data is scarce or not available.
title Enhancing VICReg: Random-Walk Pairing for Improved Generalization and Better Global Semantics Capturing
topic Computer Vision and Pattern Recognition
Machine Learning
url https://arxiv.org/abs/2506.18104