AI Paper Atlas

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

VIP: Visual-guided Prompt Evolution for Efficient Dense Vision-Language Inference

Hao Zhu, Shuo Jin, Wenbin Liao, Jiayu Xiao, Yan Zhu, Siyue Yu, Feng Dai

ICML 2026 · Vision-language and multimodal › CLIP and zero-shot · approach: Prompt learning for CLIP · 0 citations (OpenAlex)

Read the paper

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 self-correction attention modulation to dino.txt (ablation)dino.txt · avg of VOC, Object, City, ADEmIoU: 25.4 -> 38.0Improved
add unfiltered LLM-generated aliases (ablation)dino.txt with self-correction · avg of VOC, Object, City, ADEmIoU: 38.0 -> 37.6Hurt
add visual-guided alias distillation (ablation)VIP · avg of VOC, Object, City, ADEmIoU: 37.6 -> 46.7Improved
add saliency-aware soft aggregation (ablation)VIP · avg of VOC, Object, City, ADEmIoU: 46.7 -> 50.7Improved
add LLM-evolved text templates (ablation)VIP · avg of VOC, Object, City, ADEmIoU: 50.7 -> 51.3Improved
swap GPT-5 for Gemini-2.5 or DeepSeek-R1 in alias generation (ablation)VIP · avg of VOC21, Object, City, ADEmIoU: 51.3 -> 51.1 / 50.8No clear effect
use CuPL-style category descriptors instead of aliases (ablation)dino.txt with self-correction · avg of VOC21, Object, City, ADEmIoU: 38.0 -> 34.9Hurt
apply distillation filtering to category descriptors (ablation)descriptor queries · avg of VOC21, Object, City, ADEmIoU: 34.9 -> 35.4No clear effect
extend attention modulation to two layers on CLIP (ablation)CLIP-based methods · image-text feature similaritysimilarity: 0.45 -> 0.15Hurt
extend attention modulation to two layers on dino.txt (ablation)dino.txt · image-text feature similaritysimilarity: 0.46 -> 0.57Improved
replace CLIP-based method with VIP on dino.txtVIP · remote sensing averagemIoU: 35.8 -> 44.3Improved

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