AI Paper Atlas

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

PowerCLIP: Powerset Alignment for Contrastive Pre-Training

Masaki Kawamura, Nakamasa Inoue, Rintaro Yanagi, Hirokatsu Kataoka, Rio Yokota

CVPR 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: CLIP training recipes and data curation · 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
Add the triplet margin loss (ablation)PowerCLIP-S · zero-shot classification averageaverage top-1 accuracy (%): 35.1 -> 42.2Improved
Use region sets instead of individual regions (ablation)PowerCLIP-S · zero-shot classification averageaverage top-1 accuracy (%): 41.1 -> 42.2Improved
Use parse trees instead of individual tokens (ablation)PowerCLIP-S · zero-shot retrieval averageaverage R@1 (%): 45.4 -> 47.0Improved
Add R2T aggregation (ablation)PowerCLIP-S · zero-shot classification averageaverage top-1 accuracy (%): 40.8 -> 42.2Improved
Add T2R aggregation (ablation)PowerCLIP-S · zero-shot classification averageaverage top-1 accuracy (%): 41.8 -> 42.2Improved
Use SAM masks instead of random masks (M=10) (ablation)PowerCLIP · zero-shot classification and retrieval averageaverage top-1 accuracy (%): 41.5 -> 42.2Improved
Increase number of masks from 10 to 15 (ablation)PowerCLIP · zero-shot retrieval averageaverage R@1 (%): 47.0 -> 44.9 (SAM)Hurt
Use Softplus instead of ReLU in NLA-T1 (ablation)PowerCLIP-S · zero-shot classification averageaverage top-1 accuracy (%): 40.2 -> 42.2Improved
Use Tanh instead of Sigmoid in NLA-T2 (ablation)PowerCLIP-S · zero-shot classification averageaverage top-1 accuracy (%): 40.5 -> 42.2Improved
Use larger temperature 0.1 instead of 0.001 in NLAs (ablation)PowerCLIP-S · zero-shot classification averageaverage top-1 accuracy (%): 42.2 -> 40.7Hurt
Vary interpolation parameter in NLA-T2 from 0.75 to 0.50 (ablation)PowerCLIP-S · zero-shot classification averageaverage top-1 accuracy (%): 42.2 -> 42.3No clear effect
Use SAM masks instead of random masksPowerCLIP · Winoground group scoreaccuracy (%): 6.5 -> 10.2Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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