Research map / Vision-language and multimodal / Cross-modal learning
Causality-Aligned Semantic Recovery for Incomplete Cross-Modal Retrieval
Haipeng Chen, Yu Liu, Xun Yang, Yuheng Liang, Yingda Lyu
Read the paperCheck your idea against 137,217 papers
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 |
|---|---|---|---|
| add MMI module (ablation) | CASR · Pascal Sentence ICMR (50%P, 25%I, 25%T), avg | mAP: 0.761 -> 0.778 | Improved |
| add MMI module (ablation) | CASR · NUS-WIDE ICMR (50%P, 25%I, 25%T), avg | mAP: 0.653 -> 0.668 | Improved |
| add ECA module (ablation) | CASR · Pascal Sentence ICMR (50%P, 25%I, 25%T), avg | mAP: 0.763 -> 0.778 | Improved |
| add ECA module (ablation) | CASR · NUS-WIDE ICMR (50%P, 25%I, 25%T), avg | mAP: 0.655 -> 0.668 | Improved |
| add MMI and ECA to baseline (ablation) | CASR · Pascal Sentence ICMR (50%P, 25%I, 25%T), avg | mAP: 0.709 -> 0.778 | Improved |
| add MMI and ECA to baseline (ablation) | CASR · NUS-WIDE ICMR (50%P, 25%I, 25%T), avg | mAP: 0.632 -> 0.668 | Improved |
| CASR vs OTPAL under severe missing | CASR · Wikipedia ICMR (10%P, 90%I, 0%T) | mAP: 0.603 -> 0.647 | Improved |
| CASR vs OTPAL under severe missing | CASR · Pascal Sentence ICMR (10%P, 90%I, 0%T) | mAP: 0.730 -> 0.744 | Improved |
| CASR vs OTPAL with complete data | CASR · Pascal Sentence ICMR (100%P, 0%I, 0%T) | mAP: 0.782 -> 0.806 | Improved |
| CASR vs OTPAL with 20% paired data | CASR · XmediaNet PACMR (20% paired data) | mAP: 0.716 -> 0.754 | Improved |
| combine CLIP4CLIP with CASR instead of OTPAL | CLIP4CLIP · MSR-VTT (10%P, 90%I, 0%T) | mAP: 329.1 -> 342.5 | Improved |
| treat unpaired data as missing modality | CASR · partial alignment | mAP | Improved |
Closest papers by meaning
- Same Concept, Different Directions: Cross-Modal Feature Heterogeneity in Sparse Autoencoders (NeurIPS 2026 · similarity 0.88)
- Cross the Gap: Exposing the Intra-modal Misalignment in CLIP via Modality Inversion (ICLR 2025 · similarity 0.88)
- Dynamic Adapter with Semantics Disentangling for Cross-lingual Cross-modal Retrieval (AAAI 2025 · similarity 0.88)
- Learning to Rematch Mismatched Pairs for Robust Cross-Modal Retrieval (CVPR 2024 · similarity 0.88)
- Beyond Global Similarity: Multi-Conditional Retrieval for Fine-Grained Cross-Modal Understanding (CVPR 2026 · similarity 0.87)
- Causality-Inspired Invariant Representation Learning for Text-Based Person Retrieval (AAAI 2024 · similarity 0.87)
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.
