AI Paper Atlas

Research map / Vision-language and multimodal / CLIP and zero-shot

SS-TPT: Stability and Suitability-Guided Test-Time Prompt Tuning for Adversarially Robust Vision-Language Models

Sunoh Kim, Daeho Um

ICML 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Test-time adaptation of VLMs · 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 SS-guided consistency loss (ablation)SS-TPT · 10 fine-grained datasetsadversarial accuracy (%): 45.5 -> 46.8Improved
add SS-weighted prediction (ablation)SS-TPT · 10 fine-grained datasetsadversarial accuracy (%): 42.0 -> 46.8Improved
add stability score on top of suitability (ablation)SS-TPT · 10 fine-grained datasetsadversarial accuracy (%): 46.3 -> 46.8Improved
add suitability score on top of stability (ablation)SS-TPT · 10 fine-grained datasetsadversarial accuracy (%): 45.4 -> 46.8Improved
use selected views instead of all views for consistency loss (ablation)SS-TPT · Caltech101adversarial accuracy (%): 79.1 -> 81.0Improved
use min-max normalization instead of z-score (ablation)SS-TPT · Caltech101adversarial accuracy (%): 78.9 -> 81.0Improved
add affine transformation to stability augmentations (ablation)SS-TPT · Caltech101adversarial accuracy (%): 77.2 -> 81.0Improved
add horizontal flip to stability augmentations (ablation)SS-TPT · Caltech101adversarial accuracy (%): 81.0 -> 80.5Hurt
use inverse mean distance instead of mean similarity for suitability (ablation)SS-TPT · Caltech101adversarial accuracy (%): 80.6 -> 81.0No clear effect
use Jensen-Shannon instead of symmetric KL divergence for stability (ablation)SS-TPT · Caltech101adversarial accuracy (%): 80.8 -> 81.0No clear effect
decrease loss balance lambda below 1 (ablation)SS-TPT · Caltech101adversarial accuracy (%)Hurt
combine stability and suitability (alpha=0.4) instead of either alone (ablation)SS-TPT · Caltech101adversarial accuracy (%)Improved

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