AI Paper Atlas

Research map / Adversarial robustness and security / Black-box and transfer attacks

Low-Rank and Sparsity Are All You Need: Exploring Robust Hierarchical Latent Subspaces for Transferable Adversarial Attack

Shuangshuang Pu, Wen Yang, Min Li, Guodong Liu, Chris Ding, Di Ming

ICML 2026 · Adversarial robustness and security › Black-box and transfer attacks · approach: Transferable adversarial examples

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 hierarchical mixture of experts (S-M-L) versus no experts (ablation)LRS-Attack · Inc-v3 surrogate, 6 modelsavg ASR: 44.1 -> 49.0Improved
add hierarchical mixture of experts (S-M-L) versus no experts (ablation)LRS-Attack · ViT-B surrogate, 6 modelsavg ASR: 55.7 -> 73.8Improved
order experts sparse-middle-deep (S-M-L) instead of low-rank-mixture-sparse (L-M-S) (ablation)LRS-Attack · Inc-v3 surrogateavg ASR: 45.4 -> 49.0Improved
order experts S-M-L instead of L-M-S (ablation)LRS-Attack · ViT-B surrogateavg ASR: 72.8 -> 73.8Improved
use S-M-L instead of all-sparse experts (ablation)LRS-Attack · Inc-v3 surrogateavg ASR: 47.8 -> 49.0Improved
use S-M-L instead of all-low-rank experts (ablation)LRS-Attack · ViT-B surrogateavg ASR: 72.9 -> 73.8Improved
add LRS decomposition versus no decomposition on MI-FGSMMI-FGSM · Inc-v3 surrogate, 9 CNNsAvgCNN ASR: 51.7 -> 61.8Improved
use LRS instead of SVD low-rank attackMI-FGSM · Inc-v3 surrogate, 9 CNNsAvgCNN ASR: 55.5 -> 61.8Improved
use SVD low-rank attack instead of dense baselineDI-FGSM · Vgg-19 surrogate, 9 CNNsAvgCNN ASR: 64.6 -> 63.3Hurt
use LRS instead of SVD under MIMMIM · DeiT-B surrogate, 8 ViTsAvgViT ASR: 81.2 -> 86.6Improved
use LRS instead of SVD under MIMMIM · ViT-B surrogate, 9 CNNsAvgCNN ASR: 53.6 -> 57.7Improved
use LRS instead of dense baseline under ATTATT · ViT-B surrogateAvgDef-CNN ASR: 46.2 -> 45.4No 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