Group-invariant Coresets for Data-efficient Active Learning
Not on the map. Why?
Abstract (arXiv)
Active learning reduces labeling cost by querying the most informative unlabeled samples, but standard coreset methods ignore known data symmetries and can waste budget on transformed versions of the same instance. We propose GRINCO, a group-invariant coreset framework that performs acquisition in the quotient space induced by a transformation group, so that selection operates on orbits rather than raw samples. The method uses either canonical representatives or learned orbit-separating invariant embeddings to define practical quotient metrics, and combines quotient-space k-center selection with invariant training through an orbit-averaged loss. We further derive a generalization bound that relates excess orbit-averaged risk to quotient-space coverage, label uncertainty, and intra-orbit variability. Experiments on synthetic scale-invariant data and image benchmarks with rotation-induced redundancy show that GRINCO improves orbit coverage and achieves stronger label efficiency than conventional coreset baselines, especially when group-induced redundancy is substantial.
Key facts. No citing papers in the map yet (data through 7 Oct 2026). Builds on 31 of its 66 references that are in the map. First posted on arXiv 1 Jul 2026.
Builds on
31 of 66 references.
References · 31
- 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
- 2021Emerging Properties in Self-Supervised Vision Transformerscited by 6,893
- 2018Unsupervised Representation Learning by Predicting Image Rotationscited by 1,586
- 2017Deep Setscited by 1,394
- 2016Group Equivariant Convolutional Networkscited by 1,154
- 2021Geometric Deep Learning: Grids, Groups, Graphs, Geodesics, and Gaugescited by 1,131
- 2019Deep Batch Active Learning by Diverse, Uncertain Gradient Lower Boundscited by 602
- 2020A Survey of Deep Active Learningcited by 547
- 2016Harmonic Networks: Deep Translation and Rotation Equivariancecited by 293
- 2020Generalizing Convolutional Neural Networks for Equivariance to Lie Groups on Arbitrary Continuous Datacited by 248
- 2023A Survey on Self-supervised Learning: Algorithms, Applications, and Future Trendscited by 205
- 2022Active Learning Through a Covering Lenscited by 62
19 more
- 2024A Survey on Deep Active Learning: Recent Advances and New Frontierscited by 53
- 2021Reducing Label Effort: Self-Supervised meets Active Learningcited by 42
- 2022Unsupervised Learning of Group Invariant and Equivariant Representationscited by 38
- 2022Low Dimensional Invariant Embeddings for Universal Geometric Learningcited by 38
- 2025A Coreset Selection of Coreset Selection Literature: Introduction and Recent Advancescited by 28
- 2022In Defense of Core-set: A Density-aware Core-set Selection for Active Learningcited by 26
- 2022Equivariant Representation Learning via Class-Pose Decompositioncited by 17
- 2023Label-Efficient Learning in Agriculture: A Comprehensive Reviewcited by 16
- 2024Generalized Coverage for More Robust Low-Budget Active Learningcited by 15
- 2017Estimation under group actions: recovering orbits from invariantscited by 14
- 2022Group invariant machine learning by fundamental domain projectionscited by 11
- 2025Learning (Approximately) Equivariant Networks via Constrained Optimizationcited by 10
- 2023Fair Wasserstein Coresetscited by 8
- 2024Integrating Deep Metric Learning with Coreset for Active Learning in 3D Segmentationcited by 7
- 2024Bridging Diversity and Uncertainty in Active learning with Self-Supervised Pre-Trainingcited by 6
- 2023Diffusion Maps for Group-Invariant Manifoldscited by 4
- 2023Learning Symmetrization for Equivariance with Orbit Distance Minimizationcited by 4
- 2024Small coresets via negative dependence: DPPs, linear statistics, and concentrationcited by 4
- 2025Geometric Median Matching for Robust k-Subset Selection from Noisy Datacited by 1
Built on by
Not cited yet.