AI Paper Atlas

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

NAP-Tuning: Neural Augmented Prompt Tuning for Adversarially Robust Vision-Language Models

Jiaming Zhang, Xin Wang, Xingjun Ma, Lingyu Qiu, Yu–Gang Jiang, Jitao Sang

TPAMI 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
add TokenRefiner depth of 2 layers (versus none) (ablation)NAP-Tuning · ImageNet and Flowers, PGD-5 trainingaccuracyImproved
add Neural Augmentor module (ablation)NAP-Tuning · ImageNet and Flowers, varying prompt depthrobust accuracyImproved
add image-side and text-side TokenRefiner (dual-modal vs single) (ablation)NAP-Tuning · 11 datasets, PGD-100robust accuracyImproved
increase prompt depth from 1 to 12 layers (ablation)NAP-Tuning · 11 datasetsclean and robust accuracyMixed
increase shots from 1 to 16 (ablation)NAP-Tuning · 11-dataset averageclean and robust accuracy: 40.6 -> 73.5 clean; 14.1 -> 49.9 robustImproved
increase adversarial regularization alpha (ablation)NAP-Tuning · Flowersrobust accuracyImproved
increase context vector count 1 to 16 (ablation)NAP-Tuning · ImageNet and FlowersaccuracyNo clear effect
raise training eps from 1 to 4 (ablation)NAP-Tuning · Flowersrobust accuracy at test eps=8: 0.1 -> 11.8Improved
raise training eps from 1 to 4 (ablation)NAP-Tuning · Flowersclean accuracy: 96.1 -> 87.5Hurt
combine with DiffPure image purification (ablation)NAP-Tuning · 11 datasetsrobust accuracy: 12.9 percentage points averageImproved
freeze encoder instead of fine-tuning (NAP-Tuning vs TeCoA)CLIP ViT-B/32 · PGD-100clean accuracy: 36.8 -> 69.1Improved
use deep multi-layer prompts instead of AdvPT input-level promptCLIP ViT-B/16 · PGD-100 averagerobust accuracy: 4.4 -> 49.9Improved

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