AI Paper Atlas

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

Think Twice: Test-Time Reasoning for Robust CLIP Zero-Shot Classification

Shenyu Lu, Zhaoying Pan, Xiaoqian Wang

ICCV 2025 · Vision-language and multimodal › CLIP and zero-shot · approach: Attribute-based zero-shot learning · 1 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
remove spurious backgrounds (core object only, black background)CLIP ViT-L/14 zero-shot · Waterbirdsworst-group accuracy: 34.13 -> 91.39Improved
add PCA then K-means localization of the core object before classification (ablation)CLIP ViT-L/14 zero-shot · Waterbirdsworst-group accuracy: 47.36 -> 74.14Improved
use UMAP for dimensionality reduction instead of PCA+K-means (ablation)CLIP ViT-L/14 zero-shot · Waterbirdsworst-group accuracy: 47.36 -> 36.01Hurt
add SAM masks to Grounding DINO boxes (ablation)CLIP ViT-L/14 zero-shot · Waterbirdsworst-group accuracy: 66.80 -> 87.79Improved
reduce to 3 dimensions with PCA before K-means instead of clustering 768-d features (ablation)TTR · Waterbirdsworst-group accuracy: 69.43 -> 74.14Improved
add TTR to zero-shot CLIP on general datasetsCLIP ViT-L/14 · General datasets averageworst-group accuracy: 54.56 -> 70.58Improved
add TTR on UrbanCarsCLIP ViT-L/14 · UrbanCarsworst-group accuracy: 8.00 -> 34.00Improved
add TTR on MetashiftCLIP ViT-L/14 · Metashiftworst-group accuracy: 89.54 -> 90.85No clear effect
use a queue of historical samples for PCA (ablation)TTR · Waterbirdsworst-group accuracyImproved
increase number of K-means clusters beyond 3 (ablation)TTR · Waterbirdsworst-group accuracyHurt
fill masked regions with zeros instead of mean or Gaussian blur (ablation)TTR · Waterbirdsworst-group accuracyImproved
apply TTR to ViT-B/32 and ViT-H/14 (ablation)CLIP · Spurious-correlation benchmarksworst-group accuracyImproved

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