AI Paper Atlas

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

CLIP-like Model as a Foundational Density Ratio Estimator

Fumiya Uchiyama, Rintaro Yanagi, Shohei Taniguchi, Shota Takashiro, Masahiro Suzuki, Hirokatsu Kataoka, Yusuke Iwasawa, Yutaka Matsuo

CVPR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Fine-grained improvements to CLIP · 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
importance-weight CLIP pretraining loss with CLIP-based density ratio from a single domain promptCLIP pretraining on CC12M · Oxford-IIIT Pet zero-shotaccuracy and F1: +2-8 points accuracy, +3-7 points F1 at epochs 4-6Improved
importance-weighted learning with domain prompts (food, pets, flowers)CLIP pretraining on CC12M · Food101, Pets, Flowers102 zero-shotaccuracy and F1Improved
filter DataComp small pool to top 25% by text D_KLCLIP pretraining on DataComp small · ImageNet-1k zero-shotaccuracy: 0.025 -> 0.0300Improved
filter DataComp small pool to top 25% by text D_KLCLIP pretraining on DataComp small · mean over 38 tasksmean metric: 0.132 -> 0.1337No clear effect
use text-based KL divergence instead of image-based KL divergence for data filtering (ablation)D_KL filtering · DataComp small, mean over 38 tasksmean metric: 0.122 -> 0.1337Improved
use CLIPScore filtering instead of KL-based filtering (ablation)DataComp small filtering · DataComp small, mean over 38 tasksmean metric: 0.1337 -> 0.173Improved
filter by image-based D_C (centered norm) instead of text-based (ablation)DataComp small filtering · ImageNet-1k zero-shotaccuracy: 0.0312 -> 0.0108Hurt
KL-guided filtering versus LAION2B filteringDataComp small filtering · DataComp small, mean over 38 tasksmean metric: 0.133 vs 0.1337No clear effect

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