Research map / Vision-language and multimodal / Cross-modal learning
IMRAM: Iterative Matching With Recurrent Attention Memory for Cross-Modal Image-Text Retrieval
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 |
|---|---|---|---|
| use iterative matching with recurrent attention memory | IMRAM · Flickr8K | R@sum: 407.4 -> 420.0 | Helped |
| use iterative matching with recurrent attention memory | IMRAM · Flickr30K | R@sum: 482.6 -> 484.2 | Helped |
| use iterative matching with recurrent attention memory | IMRAM · MS COCO 1K | R@sum: 516.8 -> 516.6 | Hurt |
| use iterative matching with recurrent attention memory | IMRAM · MS COCO 5K | R@sum: 415.7 -> 416.5 | Helped |
| use Full-IMRAM rather than i-t + t-i | KWAI-AD · KWAI-AD | Text retrieval R@1: 7.3 -> 10.2 | Helped |
| increase matching steps from K=1 to K=2 (ablation) | Full-IMRAM · MS COCO 5K | Text retrieval R@1: 49.7 -> 53.1 | Helped |
| increase matching steps from K=2 to K=3 (ablation) | Full-IMRAM · MS COCO 5K | Text retrieval R@1: 53.1 -> 53.7 | Helped |
| increase matching steps from K=1 to K=3 (ablation) | Image-IMRAM · MS COCO 5K | Text retrieval R@1: 40.8 -> 53.2 | Helped |
| increase matching steps from K=1 to K=3 (ablation) | Text-IMRAM · MS COCO 5K | Text retrieval R@1: 46.2 -> 51.4 | Helped |
| add the proposed modified gated memory aggregation rather than addition (ablation) | Text-IMRAM · Flickr30K | Text retrieval R@1: 64.5 -> 68.8 | Helped |
| add the proposed modified gated memory aggregation rather than MLP aggregation (ablation) | Text-IMRAM · Flickr30K | Text retrieval R@1: 66.6 -> 68.8 | Helped |
| add the proposed modified gated memory aggregation rather than attention aggregation (ablation) | Text-IMRAM · Flickr30K | Text retrieval R@1: 66.1 -> 68.8 | Helped |
Closest papers by meaning
- Identification of Necessary Semantic Undertakers in the Causal View for Image-Text Matching (AAAI 2024 · similarity 0.88)
- Text Embedding is Not All You Need: Attention Control for Text-to-Image Semantic Alignment with Text Self-Attention Maps (CVPR 2025 · similarity 0.87)
- Recurrence-Enhanced Vision-and-Language Transformers for Robust Multimodal Document Retrieval (CVPR 2025 · similarity 0.87)
- Beyond Global Similarity: Multi-Conditional Retrieval for Fine-Grained Cross-Modal Understanding (CVPR 2026 · similarity 0.87)
- Composing Object Relations and Attributes for Image-Text Matching (CVPR 2024 · similarity 0.86)
- CoMat: Aligning Text-to-Image Diffusion Model with Image-to-Text Concept Matching (NeurIPS 2024 · 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.