AI Paper Atlas

Research map / Adversarial robustness and security / Certified and empirical robustness

Shift Invariance Can Reduce Adversarial Robustness

Vasu Singla, Songwei Ge, Ronen Basri, David Jacobs

NeurIPS 2021 · Adversarial robustness and security › Certified and empirical robustness · approach: Robust overfitting and regularization · 6 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
Use a shift-invariant CNN instead of an FC networkClassifier on orthogonal-vector synthetic data · Orthogonal vectors, n=50, 100, and 200average l2 adversarial distance: 0.113 -> 0.033; 0.083 -> 0.029; 0.060 -> 0.024Hurt
Use a shift-invariant CNN instead of an FC networkClassifier on orthogonal-frequency synthetic data · Orthogonal frequencies, n=50, 100, and 200average l2 adversarial distance: 0.129 -> 0.272; 0.126 -> 0.251; 0.103 -> 0.326Improved
Increase the shared class component pFC network on dataset 3 · Dataset 3, n=50, 100, and 200average l2 adversarial distance: n=50: 0.106 at p=0 -> 0.181 at p=0.3; n=100: 0.080 at p=0 -> 0.181 at p=0.3; n=200: 0.060 at p=0 -> 0.170 at p=0.3Improved
Increase shift invariance by increasing circular paddingCNN on MNIST · MNIST, PGD l2 and l∞ attacksrobust accuracyHurt
Replace the shift-invariant CNN with an FC networkClassifier on MNIST and Fashion-MNIST · MNIST and Fashion-MNIST, PGD l2 and l∞ attacksrobust accuracyImproved
Use a ResNet instead of an FC networkClassifier on SVHN · SVHN, PGD l2 and l∞ attacksrobust accuracyHurt
Use AlexNet or Sparse-MLP instead of ResNetClassifier on CIFAR-10 · CIFAR-10, adversarial attacksrobust accuracyImproved
Use a ViT instead of a BiT-ResNetImageNet classifier · ImageNet-1K, PGD l2 and l∞ attacksrobust accuracyImproved
Increase input dimension while keeping the margin from shrinkingFC network on dataset 3 · Dataset 3average l2 adversarial distanceNo 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