AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

VIP: Unified Certified Detection and Recovery for Patch Attack with Vision Transformers

Junbo Li, Huan Zhang, Cihang Xie

ECCV 2022 · Adversarial robustness and security › Attacks on text and graphs · approach: Physical adversarial patches and camouflage · 10 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 DeiT instead of original ViT as base classifierViP · ImageNet detection 2x2certified robustness: 72.0 -> 75.0Improved
use MAE pretraining instead of DeiTViP · ImageNet detection 2x2certified robustness: 75.0 -> 77.7Improved
use linear instead of quadratic mask slidingViP MAE base · ImageNet detection 2x2certified robustness: 77.7 -> 74.5Hurt
use MAE instead of ViT for dual-patch detectionViP · ImageNet dual-patch 2x2certified robustness: 22.2 -> 42.0Improved
use MAE instead of DeiT for dual-patch detectionViP · ImageNet dual-patch 4x4certified robustness: 16.6 -> 23.9Improved
replace CNN method with ViP MAE for recoveryViP MAE base · ImageNet recovery 1% pixelscertified robustness: 32.3 -> 45.0Improved
use MAE instead of DeiT for certified recoveryViP · ImageNet recovery 24x24adversarial robustness: 43.8 -> 45.0Improved
finetune 60 epochs instead of 30 (ablation)ViP MAE base · ImageNet recovery 24x24adversarial robustness: 45.0 -> 46.0Improved
use stride 1 instead of 10 (ablation)ViP MAE base · ImageNet recovery 24x24adversarial robustness: 38.8 -> 45.0Improved
scale from MAE base to MAE large (ablation)ViP · ImageNet recovery 24x24adversarial robustness: 42.5 -> 48.3Improved
widen unmasked area from 19 to 37 (ablation)ViP MAE base · ImageNet recovery 24x24adversarial robustness: 45.0 -> 45.1No clear effect

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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