Research map / Label-efficient and robust learning / Recognition and long-tail
A Group-Theoretic Framework for Data Augmentation
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Average an estimator over group transformations | General estimators · Exact-invariance setting | mean squared error | Helped |
| Average the loss over group transformations | ERM · Exact-invariance setting | Rademacher complexity | Helped |
| Average the loss over group transformations | ERM · Exact-invariance setting | generalization bound | Helped |
| Average gradients over group transformations | Strongly convex ERM · Exact-invariance setting | parameter estimation error bound | Helped |
| Average gradients over group transformations | M-estimators · Exact-invariance setting | asymptotic covariance | Helped |
| Use orbit averaging when transformations are only approximately invariant | General estimators · Approximate-invariance setting | mean squared error | Mixed |
| Use augmented MLE rather than ordinary MLE | MLE · Exact-invariance setting | asymptotic covariance | Helped |
| Use augmented MLE when the invariant parameter set contains an open neighborhood of the true parameter | MLE · Exact-invariance setting | asymptotic variance reduction | No clear effect |
| Use augmented MLE instead of constrained MLE | MLE · Invariant parameter set is a singleton | parameter recovery | Hurt |
Closest papers by meaning
- Tradeoffs in Data Augmentation: An Empirical Study (ICLR 2021 · similarity 0.89)
- How Much Data Are Augmentations Worth? An Investigation into Scaling Laws, Invariance, and Implicit Regularization (ICLR 2023 · similarity 0.89)
- Revisiting Data Augmentation in Deep Reinforcement Learning (ICLR 2024 · similarity 0.87)
- Characterizing the Generalization Error of Random Feature Regression with Arbitrary Data-Augmentation (NeurIPS 2026 · similarity 0.86)
- First-Order Manifold Data Augmentation for Regression Learning (ICML 2024 · similarity 0.85)
- How Data Augmentation Shapes Neural Representations (NeurIPS 2026 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.