BriefGPT.xyz
May, 2019
快速自动增强
Fast AutoAugment
HTML
PDF
Sungbin Lim, Ildoo Kim, Taesup Kim, Chiheon Kim, Sungwoong Kim
TL;DR
本文提出了一种名为Fast AutoAugment的算法,通过基于密度匹配的更高效的搜索策略找到有效的数据扩充策略,相比于AutoAugment,该算法在各种模型和数据集上加速了搜索时间,并实现了可比较的图像识别性能。
Abstract
data augmentation
is an indispensable technique to improve generalization and also to deal with imbalanced datasets. Recently,
autoaugment
has been proposed to automatically search augmentation policies from a da
→