AI Paper Atlas

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

Diversifying Counterattacks: Orthogonal Exploration for Robust CLlP Inference

Chengze Jiang, Minjing Dong, Xinli Shi, Jie Gui

AAAI 2026 · Adversarial robustness and security › Certified and empirical robustness · approach: Robust overfitting and regularization · 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 orthogonal gradient augmentation (OGA) with momentum to TTC counterattack (ablation)TTC · 16-dataset avg, PGD-10robust accuracy (%): 21.43 -> 31.83Improved
add directional sensitivity score (DSS) to TTC (ablation)TTC · 16-dataset avg, PGD-10clean accuracy (%): 55.66 -> 58.23Improved
add DSS to TTC (ablation)TTC · 16-dataset avg, PGD-10robust accuracy (%): 21.43 -> 23.37Improved
add DSS on top of OGA (ablation)DOC · 16-dataset avg, PGD-10clean accuracy (%): 55.38 -> 58.27Improved
add DSS on top of OGA (ablation)DOC · 16-dataset avg, PGD-10robust accuracy (%): 31.83 -> 31.04Hurt
replace TTC with DOCCLIP ViT-B/32 · 16-dataset avg, PGD-10 eps=4/255robust accuracy (%): 21.22 -> 31.02Improved
replace TTC with DOCCLIP ViT-B/32 · 16-dataset avg, CW eps=4/255robust accuracy (%): 20.61 -> 28.18Improved
replace TTC with DOCCLIP ViT-B/32 · 16-dataset avg, AutoAttackrobust accuracy (%): 13.0 -> 17.1Improved
replace TTC with DOCCLIP ViT-B/32 · EuroSAT, PGD-10robust accuracy (%): 14.24 -> 14.49No clear effect
replace TTC with DOCCLIP ViT-B/32 · 16-dataset avg, PGD-10clean accuracy (%): 55.63 -> 58.26Improved
counterattack steps N from 1 up to 3 (ablation)DOC · 12 datasets, PGD-10 and CWrobust accuracyImproved
add DOC as plug-in to adversarially fine-tuned models (TeCoA, PMG, FARE)adversarial fine-tuning · 16 datasets, PGD-10 eps=1/255robust accuracy: 4%-5% over baselinesImproved

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