AI Paper Atlas

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

Is CLIP ideal? No. Can we fix it? Yes!

Raphi Kang, Yue Song, Georgia Gkioxari, Pietro Perona

ICCV 2025 · Vision-language and multimodal › CLIP and zero-shot · approach: CLIP training recipes and data curation · 28 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 functional rows (constant vectors for spatial words) (ablation)DCSM (synthetic-trained) · WhatsUpaccuracy: 48.6 -> 62.6Improved
add functional rows (constant vectors for spatial words) (ablation)DCSM (synthetic-trained) · COCO 2-objaccuracy: 55.5 -> 70.9Improved
use a 2-layer CNN scorer instead of a small ViT (ablation)DCSM (synthetic-trained) · WhatsUpaccuracy: 29.4 -> 62.6Improved
use a 2-layer CNN scorer instead of a small ViT (ablation)DCSM (synthetic-trained) · COCO 2-objaccuracy: 47.0 -> 70.9Improved
finetune CLIP on the synthetic data (instead of DCSM) (ablation)CLIP ViT-B/16 · WhatsUpaccuracy: 30.5 -> 25.5Hurt
train an MLP scorer on CLIP embeddings (ablation)CLIP ViT-B/16 · WhatsUpaccuracy: 30.5 -> 25.4Hurt
train an MLP scorer on CLIP embeddings (ablation)CLIP ViT-B/16 · COCO 2-objaccuracy: 45.9 -> 53.6Mixed
replace cosine similarity with DCSM scoring (COCO-trained)CLIP ViT-B/16 · CLEVR-bindaccuracy: 20.2 -> 39.9Improved
replace cosine similarity with DCSM scoring (COCO-trained)CLIP ViT-B/16 · WhatsUpaccuracy: 30.5 -> 63.7Improved
replace cosine similarity with DCSM scoring (COCO-trained)CLIP ViT-B/16 · VG_attraccuracy: 61.8 -> 68.1 (NegCLIP 72.4)Mixed
train DCSM on synthetic instead of COCO dataDCSM · NegBench COCOaccuracy: 48.6 -> 38.8Hurt
add LLM-updated functional-word lookup (open vocab)DCSM · VL-Checklist VG spatialaccuracy: 50.6 -> 63.5Improved

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