AI Paper Atlas

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

GeoDetect: Geometric Adversarial Detection for VLPs

Afsaneh Hasanebrahimi, Hanxun Huang, Christopher Leckie, James Bailey, Sarah Erfani

ECCV 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: Vision-language model robustness · 0 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 geometric scores (kNN) instead of MCMCLIP-CNN · Zero-shot ImageNet1k, Sep_uniAUC (%): 86.10 -> 99.65Improved
use geometric scores (LID) instead of MCMCLIP-CNN · Zero-shot CIFAR100, Co-AttackAUC (%): 43.99 -> 100Improved
use geometric scores (Mahalanobis) instead of MCM on multimodal attackALBEF · Zero-shot CIFAR10, Sep_multiAUC (%): 47.43 -> 100Improved
kNN/KDE against multimodal-target attack on fused VLPGeoDetect-kNN · Zero-shot ImageNet1k, ALBEF Sep_multiAUC (%): 98.60 -> 51.92Hurt
add multimodal-encoder LID featuresGeoDetect-LID · Zero-shot ImageNet1k, ALBEF Sep_multiAUC (%): 51.92 -> 78.77Improved
kNN vs LID on retrieval, Sep_multiGeoDetect · Retrieval Flickr30k, TCL Sep_multiAUC (%): kNN 32.89 vs LID 84.72Hurt
adaptive attack against kNN detector (ablation)GeoDetect-kNN · Retrieval Flickr30k, CLIP-CNN Sep_uniAUC (%): 99.99 -> 67.82Hurt
adaptive attack against kNN detector (ablation)GeoDetect-kNN · Retrieval Flickr30k, ALBEF Sep_uniAUC (%): 99.75 -> 51.38Hurt
adaptive attack against LID detector (ablation)GeoDetect-LID · Retrieval Flickr30k, CLIP-ViT Sep_uniAUC (%): 99.37 -> 31.56Hurt
adaptive attack against LID detector (ablation)GeoDetect-LID · Retrieval COCO, CLIP-CNN Sep_uniAUC (%): 99.54 -> 93.14Hurt
adaptive attack against LID detector on multimodal-target attack (ablation)GeoDetect-LID · Retrieval Flickr30k, ALBEF Sep_multiAUC (%): 74.26 -> 89.83Improved

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