Research map / Vision-language and multimodal / CLIP and zero-shot
Prompt Distribution Learning
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| learn a distribution over diverse prompts rather than one soft prompt | CLIP prompt adaptation · 12 downstream datasets, 1-shot average | score: +8.5% relative average performance improvement | Helped |
| learn task-adapted prompts rather than use hand-crafted prompts | CLIP · 12 downstream datasets, 1-shot average | score: +9.1% relative average result | Helped |
| add the upper-bound loss to learn the classifier-weight distribution (ablation) | ProDA · 12 downstream datasets, average, 1-shot | score: 65.8 -> 67.0 | Helped |
| vary category-name position across prompts (ablation) | ProDA · 12 downstream datasets, average, 1-shot | score: 66.6 -> 67.0 | Helped |
| add semantic orthogonality loss (ablation) | ProDA · 12 downstream datasets, average, 1-shot | score: 66.8 -> 67.0 | Helped |
| sample a prompt mini-batch rather than use all prompts per iteration (ablation) | ProDA · CIFAR-10 | accuracy: 74.6 vs 74.4; 76.4 vs 76.3; 78.3 vs 78.3; 79.6 vs 79.7; 80.9 vs 81.0 | No clear effect |
| increase the number of learnable prompts (ablation) | ProDA · CIFAR-100 | recognition score | Helped |
Closest papers by meaning
- Any-Shift Prompting for Generalization over Distributions (CVPR 2024 · similarity 0.90)
- Aggregate-and-Adapt Natural Language Prompts for Downstream Generalization of CLIP (NeurIPS 2024 · similarity 0.90)
- Make Prompts Adaptable: Bayesian Modeling for Vision-Language Prompt Learning with Data-Dependent Prior (AAAI 2024 · similarity 0.90)
- PromptKD: Unsupervised Prompt Distillation for Vision-Language Models (CVPR 2024 · similarity 0.89)
- Learning to Prompt with Text Only Supervision for Vision-Language Models (AAAI 2025 · similarity 0.89)
- 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.