Research map / Vision-language and multimodal / Cross-modal learning
Hierarchical Cross-modal Prompt Learning for Vision-Language Models
Hao Zheng, Shunzhi Yang, Zhuoxin He, Jinfeng Yang, Zhenhua Huang
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 |
|---|---|---|---|
| use bidirectional knowledge flow T->I | I->T instead of unidirectional T->I (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Harmonic mean: 80.10 -> 81.75 | Improved |
| use bidirectional knowledge flow T->I | I->T instead of unidirectional I->T (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Harmonic mean: 79.10 -> 81.75 | Improved |
| start flow from text to image rather than image to text in the bidirectional design (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Harmonic mean: 80.58 -> 81.75 | Improved |
| add bidirectional knowledge flow (BKF) with consistency loss (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Harmonic mean: 77.90 -> 81.75 | Improved |
| add bidirectional knowledge flow alone (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Base accuracy: 82.15 -> 85.96 | Improved |
| add consistency regularization alone (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Base accuracy: 82.15 -> 82.09 | No clear effect |
| add consistency regularization alone (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Novel accuracy: 74.07 -> 76.02 | Improved |
| set layer boundary k=6 instead of k=2 (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Harmonic mean: 81.37 -> 81.75 | Improved |
| set layer boundary k=6 instead of k=10 (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Harmonic mean: 81.30 -> 81.75 | Improved |
| use multi-scale mapper instead of single-scale in both directions (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Harmonic mean: 80.89 -> 81.75 | Improved |
| use LKP compression instead of equal weighting (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Harmonic mean: 81.30 -> 81.75 | Improved |
| use cosine consistency instead of MSE (ablation) | HiCroPL · Base-to-novel generalization (11 datasets avg) | Harmonic mean: 79.39 -> 81.75 | 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.
- MaPLe — MaPLe: Multi-Modal Prompt Learning (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
- Learning Hierarchical Prompt with Structured Linguistic Knowledge for Vision-Language Models (AAAI 2024 · similarity 0.91)
- IPO: Interpretable Prompt Optimization for Vision-Language Models (NeurIPS 2024 · similarity 0.90)
- Vision-Language Models Create Cross-Modal Task Representations (ICML 2025 · similarity 0.90)
- CoPL: Contextual Prompt Learning for Vision-Language Understanding (AAAI 2024 · similarity 0.90)
- Concept-Guided Prompt Learning for Generalization in Vision-Language Models (AAAI 2024 · similarity 0.90)
- COMMA: Co-articulated Multi-Modal Learning (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.
