English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Data augmentation is a dominant method for reducing model overfitting andimproving generalization. Most existing data augmentation methods tend to finda compromise in augmenting the data, \textit{i.e.}, increasing the amplitude ofaugmentation carefully to avoid degrading some data too much and doing harm tothe model performance. We delve into the relationship between data augmentationand model performance, revealing that the performance drop with heavyaugmentation comes from the presence of out-of-distribution (OOD) data.Nonetheless, as the same data transformation has different effects fordifferent training samples, even for heavy augmentation, there remains part ofin-distribution data which is beneficial to model training. Based on theobservation, we propose a novel data augmentation method, named\textbf{DualAug}, to keep the augmentation in distribution as much as possibleat a reasonable time and computational cost. We design a data mixing strategyto fuse augmented data from both the basic- and the heavy-augmentationbranches. Extensive experiments on supervised image classification benchmarksshow that DualAug improve various automated data augmentation method. Moreover,the experiments on semi-supervised learning and contrastive self-supervisedlearning demonstrate that our DualAug can also improve related method. Code isavailable at\href{https://github.com/shuguang99/DualAug}{https://github.com/shuguang99/DualAug}.

DualAug: Exploiting Additional Heavy Augmentation with OOD Data Rejection