001034964 001__ 1034964
001034964 005__ 20250203103412.0
001034964 0247_ $$2doi$$a10.48550/arXiv.2412.03314
001034964 0247_ $$2datacite_doi$$a10.34734/FZJ-2025-00072
001034964 037__ $$aFZJ-2025-00072
001034964 1001_ $$0P:(DE-Juel1)190396$$aWang, Qin$$b0$$eCorresponding author
001034964 245__ $$aEquivariant Representation Learning for Augmentation-based Self-Supervised Learning via Image Reconstruction
001034964 260__ $$barXiv$$c2024
001034964 3367_ $$0PUB:(DE-HGF)25$$2PUB:(DE-HGF)$$aPreprint$$bpreprint$$mpreprint$$s1736774017_16364
001034964 3367_ $$2ORCID$$aWORKING_PAPER
001034964 3367_ $$028$$2EndNote$$aElectronic Article
001034964 3367_ $$2DRIVER$$apreprint
001034964 3367_ $$2BibTeX$$aARTICLE
001034964 3367_ $$2DataCite$$aOutput Types/Working Paper
001034964 520__ $$aAugmentation-based self-supervised learning methods have shown remarkable success in self-supervised visual representation learning, excelling in learning invariant features but often neglecting equivariant ones. This limitation reduces the generalizability of foundation models, particularly for downstream tasks requiring equivariance. We propose integrating an image reconstruction task as an auxiliary component in augmentation-based self-supervised learning algorithms to facilitate equivariant feature learning without additional parameters. Our method implements a cross-attention mechanism to blend features learned from two augmented views, subsequently reconstructing one of them. This approach is adaptable to various datasets and augmented-pair based learning methods. We evaluate its effectiveness on learning equivariant features through multiple linear regression tasks and downstream applications on both artificial (3DIEBench) and natural (ImageNet) datasets. Results consistently demonstrate significant improvements over standard augmentation-based self-supervised learning methods and state-of-the-art approaches, particularly excelling in scenarios involving combined augmentations. Our method enhances the learning of both invariant and equivariant features, leading to more robust and generalizable visual representations for computer vision tasks.
001034964 536__ $$0G:(DE-HGF)POF4-5111$$a5111 - Domain-Specific Simulation & Data Life Cycle Labs (SDLs) and Research Groups (POF4-511)$$cPOF4-511$$fPOF IV$$x0
001034964 536__ $$0G:(DE-HGF)POF4-5112$$a5112 - Cross-Domain Algorithms, Tools, Methods Labs (ATMLs) and Research Groups (POF4-511)$$cPOF4-511$$fPOF IV$$x1
001034964 536__ $$0G:(DE-Juel1)Helmholtz-SLNS$$aSLNS - SimLab Neuroscience (Helmholtz-SLNS)$$cHelmholtz-SLNS$$x2
001034964 588__ $$aDataset connected to DataCite
001034964 650_7 $$2Other$$aComputer Vision and Pattern Recognition (cs.CV)
001034964 650_7 $$2Other$$aFOS: Computer and information sciences
001034964 7001_ $$0P:(DE-Juel1)129347$$aKrajsek, Kai$$b1
001034964 7001_ $$0P:(DE-Juel1)129394$$aScharr, Hanno$$b2
001034964 773__ $$a10.48550/arXiv.2412.03314$$tarXiv$$y2024
001034964 8564_ $$uhttps://juser.fz-juelich.de/record/1034964/files/Equivariant%20Representation%20Learning%20for%20Augmentation-based%20Self-Supervised%20Learning%20via%20Image%20Reconstruction%20ArXiv%202024.pdf$$yOpenAccess
001034964 909CO $$ooai:juser.fz-juelich.de:1034964$$pdnbdelivery$$pdriver$$pVDB$$popen_access$$popenaire
001034964 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)190396$$aForschungszentrum Jülich$$b0$$kFZJ
001034964 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)129347$$aForschungszentrum Jülich$$b1$$kFZJ
001034964 9101_ $$0I:(DE-588b)5008462-8$$6P:(DE-Juel1)129394$$aForschungszentrum Jülich$$b2$$kFZJ
001034964 9131_ $$0G:(DE-HGF)POF4-511$$1G:(DE-HGF)POF4-510$$2G:(DE-HGF)POF4-500$$3G:(DE-HGF)POF4$$4G:(DE-HGF)POF$$9G:(DE-HGF)POF4-5111$$aDE-HGF$$bKey Technologies$$lEngineering Digital Futures – Supercomputing, Data Management and Information Security for Knowledge and Action$$vEnabling Computational- & Data-Intensive Science and Engineering$$x0
001034964 9131_ $$0G:(DE-HGF)POF4-511$$1G:(DE-HGF)POF4-510$$2G:(DE-HGF)POF4-500$$3G:(DE-HGF)POF4$$4G:(DE-HGF)POF$$9G:(DE-HGF)POF4-5112$$aDE-HGF$$bKey Technologies$$lEngineering Digital Futures – Supercomputing, Data Management and Information Security for Knowledge and Action$$vEnabling Computational- & Data-Intensive Science and Engineering$$x1
001034964 9141_ $$y2024
001034964 915__ $$0StatID:(DE-HGF)0510$$2StatID$$aOpenAccess
001034964 920__ $$lyes
001034964 9201_ $$0I:(DE-Juel1)IAS-8-20210421$$kIAS-8$$lDatenanalyse und Maschinenlernen$$x0
001034964 9201_ $$0I:(DE-Juel1)JSC-20090406$$kJSC$$lJülich Supercomputing Center$$x1
001034964 980__ $$apreprint
001034964 980__ $$aVDB
001034964 980__ $$aUNRESTRICTED
001034964 980__ $$aI:(DE-Juel1)IAS-8-20210421
001034964 980__ $$aI:(DE-Juel1)JSC-20090406
001034964 9801_ $$aFullTexts