Research map / Vision-language and multimodal / CLIP and zero-shot
Few-Shot Image and Sentence Matching via Gated Visual-Semantic Embedding
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 semantic relation modeling (attentional LSTM, Skip-Thought LSTM) to uncommon VSE (ablation) | uncommon VSE · Flickr30k 0-shot | mR: 64.9 -> 68.8 | Helped |
| add semantic relation modeling to uncommon VSE (ablation) | uncommon VSE · MSCOCO 5000 test 0-shot | mR: 50.2 -> 56.2 | Helped |
| fine-tune CNN in common VSE with two-step learning (ablation) | common VSE · Flickr30k 0-shot | mR: 55.3 -> 62.7 | Helped |
| fine-tune CNN in common VSE (ablation) | common VSE · MSCOCO 5000 test 0-shot | mR: 46.9 -> 54.6 | Helped |
| use generic pretrained representations (uncommon VSE) instead of self-learnt (common VSE) (ablation) | few-shot VSE · Flickr30k 0-shot | mR: 55.3 -> 64.9 | Helped |
| combine uncommon and common VSE by summing metrics (ablation) | uncommon VSE + relation · Flickr30k 0-shot | mR: 68.8 -> 73.2 | Helped |
| replace averaged sum with gated metric fusion (ablation) | Gated VSE · Flickr30k 0-shot | mR: 73.2 -> 73.9 | Helped |
| replace averaged sum with gated metric fusion (ablation) | Gated VSE · MSCOCO 5000 test 0-shot | mR: 62.1 -> 62.5 | Helped |
| GVSE vs best compared method in 0-shot matching | few-shot matching · Flickr30k 0-shot | mR: 62.7 -> 73.9 | Helped |
| GVSE vs SCO on conventional matching | image-sentence matching · Flickr30k | mR: 69.7 -> 78.8 | Helped |
| GVSE vs SCO on conventional matching | image-sentence matching · MSCOCO 1000 test | mR: 83.2 -> 85.0 | Helped |
| GVSE vs SCO on MSCOCO 5000 test | image-sentence matching · MSCOCO 5000 test | mR: 61.6 -> 66.9 | Helped |
Closest papers by meaning
- Asymmetric Visual Semantic Embedding Framework for Efficient Vision-Language Alignment (AAAI 2025 · similarity 0.86)
- DH-Set: Improving Vision-Language Alignment with Diverse and Hybrid Set-Embeddings Learning (CVPR 2025 · similarity 0.85)
- Bridging Vision and Language Spaces with Assignment Prediction (ICLR 2024 · similarity 0.85)
- DINOv2 Meets Text: A Unified Framework for Image- and Pixel-Level Vision-Language Alignment (CVPR 2025 · similarity 0.85)
- VT-FSL: Bridging Vision and Text with LLMs for Few-Shot Learning (NeurIPS 2025 · similarity 0.85)
- Visual-Text Cross Alignment: Refining the Similarity Score in Vision-Language Models (ICML 2024 · similarity 0.85)
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.