Research map / Vision-language and multimodal / Cross-modal learning
Universal Weighting Metric Learning for Cross-Modal Matching
Jiwei Wei, Xing Xu, Yang Yang, Yanli Ji, Zheng Wang, Heng Tao Shen
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| Replace SCAN's Triplet loss with Max Polynomial Loss | SCAN · Flickr30K, Image-to-Text | R@1: 67.9 -> 69.4 | Improved |
| Replace SCAN's Triplet loss with Max Polynomial Loss | SCAN · Flickr30K, Text-to-Image | R@1: 43.9 -> 47.5 | Improved |
| Replace SCAN's Triplet loss with Max Polynomial Loss | SCAN · MS-COCO 1K, Image-to-Text | R@1: 69.2 -> 71.1 | Improved |
| Replace SCAN's Triplet loss with Max Polynomial Loss | SCAN · MS-COCO 1K, Text-to-Image | R@1: 54.4 -> 56.8 | Improved |
| Replace CE's Triplet loss with Max Polynomial Loss | CE · MSR-VTT, Text-to-Video | R@1: 22.5 -> 25.0 | Improved |
| Replace CE's Triplet loss with Max Polynomial Loss | CE · ActivityNet-captions, Video-to-Text | R@1: 27.9 -> 27.9 | No clear effect |
| Vary b1 and b2 in the polynomial weighting function (ablation) | CE + Max Polynomial Loss · MSR-VTT | R@1: Video-to-Text: 34.0–36.2; Text-to-Video: 24.6–25.0 | Mixed |
| Use average rather than maximum polynomial loss (ablation) | Cross-modal matching model · MS-COCO validation | Recall@K | Mixed |
Closest papers by meaning
- Noisy Correspondence Rectification via Asymmetric Similarity Learning (AAAI 2025 · similarity 0.87)
- Robust Noisy Correspondence Learning with Equivariant Similarity Consistency (CVPR 2024 · similarity 0.86)
- Negative Pre-aware for Noisy Cross-Modal Matching (AAAI 2024 · similarity 0.85)
- Connect, Collapse, Corrupt: Learning Cross-Modal Tasks with Uni-Modal Data (ICLR 2024 · similarity 0.85)
- Learning to Rematch Mismatched Pairs for Robust Cross-Modal Retrieval (CVPR 2024 · similarity 0.85)
- Cross the Gap: Exposing the Intra-modal Misalignment in CLIP via Modality Inversion (ICLR 2025 · 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.
