AI Paper Atlas

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

Robustifying Zero-Shot Vision Language Models by Subspaces Alignment

Junhao Dong, Piotr Koniusz, Liaoyuan Feng, Yifei Zhang, Hao Zhu, Weiming Liu, Xinghua Qu, Yew-Soon Ong

ICCV 2025 · Adversarial robustness and security › Attacks on text and graphs · approach: Vision-language model robustness · 2 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 intermediate adversaries, Omega alignment and antonymous prompts (ablation)AoS · 15 datasets averagePGD robust accuracy (%): 41.20 -> 43.88Improved
use MaxExp subspace distance instead of Frobenius norm (ablation)AoS · 15 datasets averagePGD robust accuracy (%): 40.86 -> 43.88Improved
use MaxExp instead of exact SVD for subspaces (ablation)AoS · 15 datasets averagePGD robust accuracy (%): 42.35 -> 43.88Improved
use universal set perturbation instead of original-sample perturbation (ablation)AoS · 15 datasets averagePGD robust accuracy (%): 41.43 -> 43.88Improved
use universal perturbation instead of instance-wise perturbation (ablation)AoS · 15 datasets averagePGD robust accuracy (%): 44.12 -> 43.88No clear effect
increase augmentations n=q from 5 to 20 (ablation)AoS · 15 datasets averagePGD robust accuracy (%): 42.95 -> 43.88Improved
increase augmentations n=q from 20 to 50 (ablation)AoS · 15 datasets averagePGD robust accuracy (%): 43.88 -> 44.07No clear effect
use subspace alignment AoS instead of sample-wise TeCoAAoS · 15 datasets averagePGD robust accuracy (%): 38.91 -> 43.88Improved
use AoS instead of FARE under larger radius 4/255AoS · 15 datasets averagePGD robust accuracy (%): 7.74 -> 9.51Improved
use AoS with visual prompt tuning instead of FARE (eps=1/255)AoS · 15 datasets average (VPT)PGD robust accuracy (%): 32.69 -> 34.38Improved
use AoS with BLIP instead of FARE for captioningAoS · Nocaps captioningrobust CIDEr: 58.7 -> 62.1Improved

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