Saved in:
Bibliographic Details
Main Authors: Schmidt, Tobias, Schneider, Steffen, Bethge, Matthias
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2510.21706
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914112433291264
author Schmidt, Tobias
Schneider, Steffen
Bethge, Matthias
author_facet Schmidt, Tobias
Schneider, Steffen
Bethge, Matthias
contents We propose Equivariance by Contrast (EbC) to learn equivariant embeddings from observation pairs $(\mathbf{y}, g \cdot \mathbf{y})$, where $g$ is drawn from a finite group acting on the data. Our method jointly learns a latent space and a group representation in which group actions correspond to invertible linear maps -- without relying on group-specific inductive biases. We validate our approach on the infinite dSprites dataset with structured transformations defined by the finite group $G:= (R_m \times \mathbb{Z}_n \times \mathbb{Z}_n)$, combining discrete rotations and periodic translations. The resulting embeddings exhibit high-fidelity equivariance, with group operations faithfully reproduced in latent space. On synthetic data, we further validate the approach on the non-abelian orthogonal group $O(n)$ and the general linear group $GL(n)$. We also provide a theoretical proof for identifiability. While broad evaluation across diverse group types on real-world data remains future work, our results constitute the first successful demonstration of general-purpose encoder-only equivariant learning from group action observations alone, including non-trivial non-abelian groups and a product group motivated by modeling affine equivariances in computer vision.
format Preprint
id arxiv_https___arxiv_org_abs_2510_21706
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Equivariance by Contrast: Identifiable Equivariant Embeddings from Unlabeled Finite Group Actions
Schmidt, Tobias
Schneider, Steffen
Bethge, Matthias
Machine Learning
We propose Equivariance by Contrast (EbC) to learn equivariant embeddings from observation pairs $(\mathbf{y}, g \cdot \mathbf{y})$, where $g$ is drawn from a finite group acting on the data. Our method jointly learns a latent space and a group representation in which group actions correspond to invertible linear maps -- without relying on group-specific inductive biases. We validate our approach on the infinite dSprites dataset with structured transformations defined by the finite group $G:= (R_m \times \mathbb{Z}_n \times \mathbb{Z}_n)$, combining discrete rotations and periodic translations. The resulting embeddings exhibit high-fidelity equivariance, with group operations faithfully reproduced in latent space. On synthetic data, we further validate the approach on the non-abelian orthogonal group $O(n)$ and the general linear group $GL(n)$. We also provide a theoretical proof for identifiability. While broad evaluation across diverse group types on real-world data remains future work, our results constitute the first successful demonstration of general-purpose encoder-only equivariant learning from group action observations alone, including non-trivial non-abelian groups and a product group motivated by modeling affine equivariances in computer vision.
title Equivariance by Contrast: Identifiable Equivariant Embeddings from Unlabeled Finite Group Actions
topic Machine Learning
url https://arxiv.org/abs/2510.21706