[2603.26657] Tunable Soft Equivariance with Guarantees

Computer Science > Computer Vision and Pattern Recognition arXiv:2603.26657 (cs) [Submitted on 27 Mar 2026] Title:Tunable Soft Equivariance with Guarantees Authors:Md Ashiqur Rahman, Lim Jun Hao, Jeremiah Jiang, Teck-Yian Lim, Raymond A. Yeh View a PDF of the paper titled Tunable Soft Equivariance with Guarantees, by Md Ashiqur Rahman and 4 other authors View PDF HTML (experimental) Abstract:Equivariance is a fundamental property in computer vision models, yet strict equivariance is rarely satisfied in real-world data, which can limit a model's performance. Controlling the degree of equivariance is therefore desirable. We propose a general framework for constructing soft equivariant models by projecting the model weights into a designed subspace. The method applies to any pre-trained architecture and provides theoretical bounds on the induced equivariance error. Empirically, we demonstrate the effectiveness of our method on multiple pre-trained backbones, including ViT and ResNet, across image classification, semantic segmentation, and human-trajectory prediction tasks. Notably, our approach improves the performance while simultaneously reducing equivariance error on the competitive ImageNet benchmark. Subjects: Computer Vision and Pattern Recognition (cs.CV); Machine Learning (cs.LG) Cite as: arXiv:2603.26657 [cs.CV]   (or arXiv:2603.26657v1 [cs.CV] for this version)   https://doi.org/10.48550/arXiv.2603.26657 Focus to learn more arXiv-issued DOI via DataCite (pending reg...