AI Paper Atlas

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

AdaBoosting Text Prompts for Vision-Language Models

Seokhee Jin, Changhwan Sung, Sunung Mun, Hoyoung Kim, Jungseul Ok

ECCV 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Prompt learning for 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
increase number of boosting rounds from M=1 to M=50 (ablation)TPB · 10-dataset avg excl. ImageNet, 16-shottop-1 accuracy: 70.13 -> 73.36Improved
increase number of boosting rounds from M=1 to M=50 (ablation)TPB · 10-dataset avg excl. ImageNet, 1-shottop-1 accuracy: 61.80 -> 67.39Improved
add image augmentation (a=4) in the boosting loop (ablation)TPB · 11-dataset avg, ViT-B/32top-1 accuracyImproved
add DCLIP descriptions to CuPL in the prompt pool, then all descriptions (ablation)TPB · 11-dataset avg, ViT-B/32top-1 accuracyImproved
use full prompt pool instead of CuPL+DCLIP (ablation)TPB · ViT-L target, 16-shottop-1 accuracy: 81.12 -> 82.07Improved
use full prompt pool instead of CuPL+DCLIP (ablation)TPB · ViT-H target, 16-shottop-1 accuracy: 82.88 -> 84.24Improved
add prompt duplication within GPC (ablation)TPB · ViT-L target, 16-shottop-1 accuracy: 81.64 -> 82.07Improved
add prompt duplication within GPC (ablation)TPB · ViT-H target, 16-shottop-1 accuracy: 83.99 -> 84.24Improved
add prompt duplication within GPC (ablation)TPB · ViT-B/32 source, 16-shottop-1 accuracy: 72.77 -> 72.87No clear effect
trade boosting rounds for augmentation at fixed M*a=200 (M/a 50/4 to 1/200) (ablation)TPB · ViT-B/32 source, 16-shottop-1 accuracy: 73.36 -> 74.42Mixed
trade boosting rounds for augmentation at fixed M*a=200 (M/a 50/4 to 1/200) (ablation)TPB · ViT-H target, 16-shottop-1 accuracy: 84.49 -> 83.12Hurt
use TPB text prompts instead of PEZ gradient-optimized prompts for transferTPB vs PEZ · RN50 to ViT-L/14, 16-shottop-1 accuracy: 53.02 -> 78.46Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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