This paper was reviewed by three experts in the field and received 677 recommendations. The reviewers found the proposed approach straightforward and incremental in terms of novelty, but all appreciated its effectiveness demonstrated in the experiments. R1 liked the particular implementation of using GAN for data augmentation work. R3 additionally liked the investigation into BigGAN-based data augmentation. Based on the reviewers' feedback, the decision is to recommend the paper for acceptance to NeurIPS 2019. The reviewers did raise some valuable concerns, for example the additional experiments suggested by R2. These questions should be addressed in the final camera-ready version of the paper. The authors are encouraged to make the necessary changes to the best of their ability. We congratulate the authors on the acceptance of their paper!