AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

DivAug: Plug-In Automated Data Augmentation With Explicit Diversity Maximization

Zirui Liu, Haifeng Jin, Ting-Hsiang Wang, Kaixiong Zhou, Xia Hu

ICCV 2021 · Label-efficient and robust learning › Semi-supervised learning · approach: Self-training theory · 20 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
replace k-means++ selection with random selection of augmented images (i.e. add k-means++ selection) (ablation)DivAug (WRN-28-10) · CIFAR-10accuracy: 97.7 -> 98.1Improved
replace k-means++ selection with random selection (i.e. add k-means++ selection) (ablation)DivAug (WRN-28-10) · CIFAR-100accuracy: 83.3 -> 84.2Improved
use DivAug instead of RandAugment as perturbation in UDAUDA · CIFAR-10 SSL (1000 labels)error rate: 7.37 -> 6.94Improved
use DivAug instead of RandAugment as perturbation in UDAUDA · CIFAR-10 SSL (2000 labels)error rate: 6.50 -> 6.26Improved
use DivAug instead of RandAugment as perturbation in UDAUDA · CIFAR-10 SSL (4000 labels)error rate: 5.44 -> 5.40No clear effect
use DivAug instead of RandAugmentWide-ResNet-28-10 · CIFAR-10accuracy: 97.3 -> 98.1Improved
use DivAug instead of Fast AutoAugmentWide-ResNet-28-10 · CIFAR-10accuracy: 97.3 -> 98.1Improved
use DivAug instead of AutoAugmentWide-ResNet-40-2 · CIFAR-100accuracy: 79.3 -> 81.3Improved
use DivAug instead of Adversarial AutoAugmentResNet-50 · ImageNet valaccuracy: 79.4 -> 78.0Hurt
use DivAug instead of Adversarial AutoAugmentWide-ResNet-28-10 · CIFAR-100accuracy: 84.5 -> 84.2Hurt
use DivAug instead of Adversarial AutoAugmentWide-ResNet-28-10 · CIFAR-10accuracy: 98.1 -> 98.1No clear effect
train DivAug with E=8, S=4 instead of RA sampling: extra training costDivAug · CIFAR-10training cost relative to RA: 1.0 -> 4.5Mixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy