Research map / Vision-language and multimodal / Cross-modal learning
Multi-Modality Cross Attention Network for Image and Sentence Matching
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 |
|---|---|---|---|
| Jointly model intra- and inter-modality relationships with self-attention and cross-attention | MMCA · Flickr30K | Image-to-text R@1: 69.1 -> 74.2 | Helped |
| Jointly model intra- and inter-modality relationships with self-attention and cross-attention | MMCA · MS-COCO 1K testing set | Image-to-text R@1: 72.7 -> 74.8 | Helped |
| Jointly model intra- and inter-modality relationships with self-attention and cross-attention | MMCA · MS-COCO 5K testing set | Image-to-text R@1: 50.4 -> 54.0 | Helped |
| Jointly model intra- and inter-modality relationships with self-attention and cross-attention | MMCA · MS-COCO 5K testing set | Text-to-image R@1: 39.0 -> 38.7 | Mixed |
| Use a 256d rather than 64d hidden space (ablation) | MMCA · Flickr30K | Image-to-text R@1: 71.4 -> 74.2 | Helped |
| Use a 256d rather than 128d hidden space (ablation) | MMCA · Flickr30K | Image-to-text R@1: 72.6 -> 74.2 | Helped |
| Use a 512d rather than 256d hidden space (ablation) | MMCA · Flickr30K | Image-to-text R@1: 74.2 -> 73.6 | Hurt |
| Use a 256d rather than 64d hidden space (ablation) | MMCA · MS-COCO 1K testing set | Image-to-text R@1: 71.7 -> 74.8 | Helped |
| Use a 512d rather than 256d hidden space (ablation) | MMCA · MS-COCO 1K testing set | Image-to-text R@1: 74.8 -> 74.1 | Hurt |
| Set cross-attention score weight alpha to 0.2 rather than 0.1 (ablation) | MMCA · Flickr30K | Image-to-text R@1: 71.7 -> 74.2 | Helped |
| Set cross-attention score weight alpha to 0.4 rather than 0.2 (ablation) | MMCA · Flickr30K | Image-to-text R@1: 74.2 -> 71.8 | Hurt |
| Set cross-attention score weight alpha to 0.8 rather than 0.2 (ablation) | MMCA · Flickr30K | Image-to-text R@1: 74.2 -> 67.1 | Hurt |
Closest papers by meaning
- CL2CM: Improving Cross-Lingual Cross-Modal Retrieval via Cross-Lingual Knowledge Transfer (AAAI 2024 · similarity 0.87)
- Cross the Gap: Exposing the Intra-modal Misalignment in CLIP via Modality Inversion (ICLR 2025 · similarity 0.87)
- Seeing is Understanding: Unlocking Causal Attention into Modality-Mutual Attention for Multimodal LLMs (ICML 2026 · similarity 0.86)
- Dynamic Adapter with Semantics Disentangling for Cross-lingual Cross-modal Retrieval (AAAI 2025 · similarity 0.86)
- Text Embedding is Not All You Need: Attention Control for Text-to-Image Semantic Alignment with Text Self-Attention Maps (CVPR 2025 · similarity 0.86)
- Remodeling Semantic Relationships in Vision-Language Fine-Tuning (AAAI 2026 · 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.