AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

Semantic Perturbations With Normalizing Flows for Improved Generalization

Oguz Kaan Yüksel, Sebastian U. Stich, Martin Jaggi, Tatjana Chavdarova

ICCV 2021 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Disentangled and discrete latent models · 7 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
add adversarial latent attacks via normalizing flowResNet-18 standard training with common DA · CIFAR-10 (5% train)test accuracy: 64.1 -> 80.4Improved
add randomized latent attacks via normalizing flowResNet-18 standard training with common DA · CIFAR-10 (5% train)test accuracy: 64.1 -> 70.1Improved
add adversarial latent attacksResNet-18 · CIFAR-10 (full train set)test accuracy: 95.2 -> 96.6Improved
use VAE-GAN on-manifold perturbation instead of noneResNet-18 · CIFAR-10 (5% train)test accuracy: 64.1 -> 58.9Hurt
use Adversarial-LA instead of MixupResNet-18 · CIFAR-10 (5% train)test accuracy: 73.4 -> 80.4Improved
add Adversarial-LA with flow trained on CIFAR-10 (transfer)ResNet-18 · CIFAR-100 (10% train)test accuracy: 36.4 -> 45.0Improved
train flow on CIFAR-10 instead of SVHN itself (ablation)ResNet-18 with Adversarial-LA · SVHN (5% train)test accuracy: 86.9 -> 90.5Improved
use latent-space attack instead of image-space PGD for trainingResNet-18 · CIFAR-10 (full train set)test accuracy: 86.3 -> 96.6Improved
add Adversarial-LA trainingResNet-18 vs Adversarial-LA attack · CIFAR-10accuracy under attack: 58.8 -> 76.4Improved
add image-space PGD trainingResNet-18 vs Adversarial-LA attack · CIFAR-10accuracy under attack: 58.8 -> 36.2Hurt
add Adversarial-LA without any data augmentationResNet-18 · CIFAR-10 (5% train, no DA)test accuracy: 49.8 -> 58.2Improved
use VAE-GAN without data augmentationResNet-18 · CIFAR-10 (5% train, no DA)test accuracy: 49.8 -> 49.4No clear effect

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