AI Paper Atlas

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

When and Where do Data Poisons Attack Textual Inversion?

Jeremy Styborski, Mingzhi Lyu, Jiayou Lu, Nupur Kapur, Adams Wai-Kin Kong

ICCV 2025 · Adversarial robustness and security › Attacks on text and graphs · approach: Diffusion-based adversarial purification · 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 JPEG compression (quality 25) as a defense against poisonsTextual Inversion on CustomConcept101 · CustomConcept101 poison averageDINOv2 similarity: 0.21 -> 0.37Improved
restrict TI training to high timesteps (T600) (ablation)Textual Inversion on CustomConcept101 · CustomConcept101 poison averageDINOv2 similarity: 0.21 -> 0.22 (ADM+ 0.10 -> 0.25, ADM- 0.33 -> 0.27)Mixed
add loss masking to constrain learning to concept regions (ablation)Textual Inversion on CustomConcept101 · CustomConcept101 poison averageDINOv2 similarity: 0.21 -> 0.36Improved
add JPEG preprocessing before loss masking (ablation)Loss masking defense · CustomConcept101 poison averageDINOv2 similarity: 0.36 -> 0.45Improved
add JPEG preprocessing before T600 timestep restriction (ablation)T600 defense · CustomConcept101 poison averageDINOv2 similarity: 0.22 -> 0.43Improved
combine JPEG, T600 and loss masking (full SZT) versus JPEG+LM (ablation)SZT · CustomConcept101 poison averageDINOv2 similarity: 0.45 -> 0.46Improved
apply SZT to poisoned data versus nominal TI on clean dataSZT · CustomConcept101DINOv2 similarity: 0.21 -> 0.46 (clean nominal 0.46)Improved
apply PDMPure regeneration defense to clean dataTextual Inversion on CustomConcept101 · CustomConcept101 clean dataDINOv2 similarity: 0.46 -> 0.25Hurt
raise the high-timestep threshold for TI training on ADM+ data (ablation)Textual Inversion on NovelConcepts10 · NovelConcepts10 ADM+DINOv2 similarity: 0.10 -> 0.21 (t>=600) -> 0.28 (t>=700)Improved
concentrate timesteps at extremely high values (t >= 900) (ablation)Textual Inversion on NovelConcepts10 · NovelConcepts10 ADM+DINOv2 similarity: 0.33 (t>=800) -> 0.26 (t>=900)Hurt
restrict TI training to low timesteps (low threshold) instead of high (ablation)Textual Inversion on NovelConcepts10 · NovelConcepts10 clean and poisonedDINOv2 similarity: Clean 0.44 -> 0.17 (low threshold 0.1)Hurt
increase TI learning rate and training steps (ablation)Textual Inversion on NovelConcepts10 · NovelConcepts10 clean, ADM+, ADM-, EADINOv2 similarityMixed

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