Papers of the day   All papers

RandAugment: Practical automated data augmentation with a reduced search space

Comments

Nov 19 2019 Barret Zoph

*New paper* RandAugment: a new data augmentation. Better & simpler than AutoAugment. Main idea is to select transformations at random, and tune their magnitude. It achieves 85.0% top-1 on ImageNet. Paper: https://arxiv.org/abs/1909.13719 Code: https://git.io/Jeopl https://t.co/equmk59K2i
5 replies, 622 likes


Nov 19 2019 Quoc Le

RandAugment was one of the secret sources behind Noisy Student that I tweeted last week. Code for RandAugment is now opensourced.
3 replies, 454 likes


Nov 19 2019 hardmaru

RandAugment: Practical automated data augmentation with a reduced search space Decreasing the search space in a clever way avoids the need to perform highly expensive computation search. ie. NAS→EfficientNets, AutoAugment→RandAugment Might be useful for domain randomization.
1 replies, 135 likes


Nov 20 2019 Delip Rao

Not a comment about this paper but, in computer vision, there is a genre of papers — “doing Y achieves X% top-1 on ImageNet”. In practice, many of those Ys don’t apply for the computer vision problem _you_ are trying to solve.
3 replies, 52 likes


Nov 19 2019 Ekin Dogus Cubuk

RandAugment has a significantly smaller search space, which allows it to be optimized on the model and dataset of interest (instead of having to use a smaller proxy task). It works on CIFAR-10/100, SVHN, ImageNet, and COCO. https://t.co/EcQJH3yotr
1 replies, 28 likes


Nov 19 2019 Sean J. Taylor

Has anyone seen a good paper on data augmentation applied to modeling/forecasting time series data?
2 replies, 26 likes


Nov 19 2019 Sharon Zhou

Glad this is out!! This is also exciting b/c RandAugment makes it easier to incorporate new data augmentation methods ("distortions") than AutoAugment. And, the UDA (Unsupervised Data Augmentation) paper was updated a couple months back to use RandAugment instead of AutoAugment.
0 replies, 9 likes


Nov 19 2019 Joan Serrà

Interesting work: 1) "demonstrate that the optimal strength of a data augmentation depends on the model size and training set size" 2) "introduce a vastly simplified search space for data augmentation"
0 replies, 7 likes


Nov 16 2019 arXiv CS-CV

RandAugment: Practical automated data augmentation with a reduced search space http://arxiv.org/abs/1909.13719
0 replies, 3 likes


Oct 01 2019 cs.CV Papers

RandAugment: Practical data augmentation with no separate search. Ekin D. Cubuk, Barret Zoph, Jonathon Shlens, and Quoc V. Le http://arxiv.org/abs/1909.13719
1 replies, 0 likes


Oct 01 2019 Brundage Bot

RandAugment: Practical data augmentation with no separate search. Ekin D. Cubuk, Barret Zoph, Jonathon Shlens, and Quoc V. Le http://arxiv.org/abs/1909.13719
1 replies, 0 likes


Content