Research map / Vision-language and multimodal / CLIP and zero-shot
Knowledge-Enhanced Explainable Prompting for Vision-Language Models
Yequan Bie, Andong Tan, Zhixuan Chen, Zhiyuan Cai, Luyang Luo, Hao Chen
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add logit_IPA attention logit to KEEP (ablation) | KEEP · Medical average | accuracy (%): 85.98 -> 87.09 | Improved |
| add logit_IPA attention logit to KEEP (ablation) | KEEP · Natural average | accuracy (%): 80.16 -> 80.88 | Improved |
| add image-prompt matching contrastive loss L_IPM (ablation) | KEEP · Medical average | accuracy (%): 84.69 -> 87.09 | Improved |
| add image-prompt matching contrastive loss L_IPM (ablation) | KEEP · Natural average | accuracy (%): 79.61 -> 80.88 | Improved |
| add cross-entropy fusion loss L_CLS (ablation) | KEEP · Medical average | accuracy (%): 80.50 -> 87.09 | Improved |
| add cross-entropy fusion loss L_CLS (ablation) | KEEP · Natural average | accuracy (%): 70.25 -> 80.88 | Improved |
| use fine-grained domain-specific knowledge prompts instead of no, random, general or intervened knowledge (ablation) | KEEP · all eight datasets | accuracy | Improved |
| adapt CLIP with KEEP prompt learning | CLIP · Medical average | accuracy (%): 43.59 -> 87.09 | Improved |
| replace LASP with KEEP at 10% training data | CCBTM · CCBTM (10% data) | accuracy (%): 82.69 -> 92.89 | Improved |
| reduce CCBTM training data from 50% to 10% for KEEP | KEEP · CCBTM | accuracy (%): 94.73 -> 92.89 | No clear effect |
| use KEEP instead of BiomedCoOp (BiomedCLIP-based) | medical datasets · Medical average | accuracy (%): 79.62 -> 87.09 | Improved |
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.
- LASP — LASP: Text-to-Text Optimization for Language-Aware Soft Prompting of Vision & Language Models (CVPR 2023) · better on 5 of 5 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
- IPO: Interpretable Prompt Optimization for Vision-Language Models (NeurIPS 2024 · similarity 0.91)
- Domain-Agnostic Mutual Prompting for Unsupervised Domain Adaptation (CVPR 2024 · similarity 0.91)
- PromptKD: Unsupervised Prompt Distillation for Vision-Language Models (CVPR 2024 · similarity 0.90)
- Learning to Prompt with Text Only Supervision for Vision-Language Models (AAAI 2025 · similarity 0.90)
- ProAPO: Progressively Automatic Prompt Optimization for Visual Classification (CVPR 2025 · similarity 0.90)
- Learning to Learn Better Visual Prompts (AAAI 2024 · similarity 0.89)
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.
