[2018.06.14] AutoAugment: Learning Augmentation Policies from Data

[unizard]

Institute: Google Brain URL: https://arxiv.org/pdf/1805.09501.pdf Keyword: Data Augmentation, AutoML, ReinforceLearning Interest: 5 Code: https://github.com/DeepVoltaire/AutoAugment GoogleBlog: https://ai.googleblog.com/2018/06/improving-deep-learning-performance.html

Summary In this paper, we take a closer look at data augmentation for images, and describe a simple procedure called AutoAugment to search for improved data augmentation policies. On ImageNet, we attain a Top-1 accuracy of 83.54%. On CIFAR-10, we achieve an error rate of 1.48%, which is 0.65% better than the previous state-of-the-art. On reduced data settings, AutoAugment performs comparably to semi-supervised methods without using any unlabeled examples.

역시 생각만하면 지는군 #멋져요 구글 브레인 짱짱맨

[unizard]

Results

• Validation set Top-1 / Top-5 error rates (%) on ImageNet. Lower is better.

• Test set Top-1 error rates (%) on FGVC datasets. AutoAugment results use the policy found on ImageNet to train an Inception v4 from scratch.

[unizard]

• List of all image transformations, AutoAugment policy found on reduced ImageNet.

1	2