Research map / Vision-language and multimodal / CLIP and zero-shot
Open-Set Fine-Grained Retrieval via Prompting Vision-Language Evaluator
Read the paperCheck 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 |
|---|---|---|---|
| Add PLEor's prompted CLIP evaluator and open-set knowledge transfer | ResNet-50 retrieval model · CUB-200-2011 | Recall@1: 72.2 -> 74.8 | Helped |
| Add PLEor's prompted CLIP evaluator and open-set knowledge transfer | ResNet-50 retrieval model · Stanford Cars 196 | Recall@1: 90.1 -> 94.4 | Helped |
| Add PLEor's prompted CLIP evaluator and open-set knowledge transfer | ResNet-50 retrieval model · FGVC Aircraft | Recall@1: 71.1 -> 86.3 | Helped |
| Add category-specific discrepancy matching (Lcdm) (ablation) | CUB-200-2011 PLEor · CUB-200-2011 | Recall@1: 66.3% -> 72.1% | Helped |
| Add category-irrelevant region matching (Lcrm) (ablation) | CUB-200-2011 PLEor · CUB-200-2011 | Recall@1: 72.1% -> 74.4% | Helped |
| Add pattern-map regularization (Lreg) (ablation) | CUB-200-2011 PLEor · CUB-200-2011 | Recall@1: 74.4% -> 75.1% | Helped |
| Add open-set knowledge transfer (Lokt) (ablation) | CUB-200-2011 PLEor · CUB-200-2011 | Recall@1: 75.1% -> 74.8% | Hurt |
| Use learnable text prompts instead of handcrafted text prompts (ablation) | CLIP evaluator · CUB-200-2011 | Recall@1: 71.5% -> 73.3% | Helped |
| Add a vision prompt to the handcrafted text prompt (ablation) | CLIP evaluator · CUB-200-2011 | Recall@1: 71.5% -> 72.4% | Helped |
| Use both learnable vision and text prompts (ablation) | CLIP evaluator · CUB-200-2011 | Recall@1: 71.5% -> 74.8% | Helped |
| Increase text prompt length from 16 to 32 vectors (ablation) | PLEor · CUB-200-2011 | Recall@1: 74.8% -> 74.6% | Hurt |
| Freeze pretrained CLIP and train prompts rather than fine-tuning CLIP (ablation) | PLEor · CUB-200-2011 | Recall@1: 65.4% -> 74.8% | Helped |
Closest papers by meaning
- From Coarse to Fine-Grained Open-Set Recognition (CVPR 2024 · similarity 0.87)
- FineCLIP: Self-distilled Region-based CLIP for Better Fine-grained Understanding (NeurIPS 2024 · similarity 0.86)
- Retrieval-Enhanced Contrastive Vision-Text Models (ICLR 2024 · similarity 0.86)
- Unified Generative and Discriminative Training for Multi-modal Large Language Models (NeurIPS 2024 · similarity 0.86)
- FG-CLIP: Fine-Grained Visual and Textual Alignment (ICML 2025 · similarity 0.86)
- GUIDED: Granular Understanding via Identification, Detection, and Discrimination for Fine-Grained Open-Vocabulary Object Detection (NeurIPS 2025 · similarity 0.86)
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.