NeurIPS 2020

RandAugment: Practical Automated Data Augmentation with a Reduced Search Space


Meta Review

This paper got mixed reviews. The original ratings are 6,5,5,6. On the positive side, reviewers think the paper solves an important problem. Data augmentation is recognized to be an important step for improving machine learning model performance. However, existing auto data augmentation methods are typically very costly. This paper solves a timely question and has good practical value. On the negative side, reviewers feel the novelty of this paper is somehow limited. Authors did a good job in the response, two reviewers increase their ratings. One reviewer still has concerns on the novelty. AC reads the paper and agree with the reviewers on the positive side. The method proposed in this paper -- reducing the augmentation space -- is useful and this paper will benefit the community. Thus, AC recommends acceptance.