AWhat changed
AutoAugment searches operation/probability/magnitude subpolicies using validation rewards; RandAugment replaces the separate proxy search with uniform sampled operations and a small count/magnitude parameter space.
Scientific scope
AutoAugment searches operation/probability/magnitude subpolicies using validation rewards; RandAugment replaces the separate proxy search with uniform sampled operations and a small count/magnitude parameter space. These related original/efficient-extension methods form one family with separate source bylines. Image operations and baseline augmentations are inherited. RandAugment still tunes count and magnitude, typically by grid search, and is not parameter-free. The original AutoAugment v1 has 16 candidate operations including SamplePairing, unlike later variants.
Evidence
The linked original sources describe the stated contribution and its evidence limits. WRS has not independently rerun the experiments or certified a proof. These related original/efficient-extension methods form one family with separate source bylines. Image operations and baseline augmentations are inherited. RandAugment still tunes count and magnitude, typically by grid search, and is not parameter-free. The original AutoAugment v1 has 16 candidate operations including SamplePairing, unlike later variants.
Contributors
- Source-listed authors and technical contributors · SCIENCE: Source-listed authors or explicitly credited technical contributors for the cited versions. Family entries retain source-specific attribution; a combined list is not one original-paper byline or equal individual work. Names alone do not establish resolved identities.
- AutoAugment: Learning Augmentation Policies from Data
- RandAugment: Practical data augmentation with no separate search
Provisional WRS assessment, edition 2026-10-16.1. Independent expert calibration is pending. This grade does not certify correctness or assign ownership.
Full scientific record · Browse contribution grades · Request a correction