Research map / Vision-language and multimodal / Cross-modal learning
DecAlign: Hierarchical Cross-Modal Alignment for Decoupled Multimodal Representation Learning
Chengxuan Qian, Shuo Xing, Li Li, Yue Zhao, Zhengzhong Tu
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 heterogeneity alignment (ablation) | DecAlign · CMU-MOSI | F1 Score: 84.03 -> 85.82 | Improved |
| add homogeneity alignment (ablation) | DecAlign · CMU-MOSI | F1 Score: 84.46 -> 85.82 | Improved |
| add heterogeneity alignment (ablation) | DecAlign · CMU-MOSEI | MAE: 0.588 -> 0.543 | Improved |
| add feature decoupling (MFD) (ablation) | DecAlign · CMU-MOSI | F1 Score: 81.56 -> 81.92 | Improved |
| add prototype-guided optimal transport (ablation) | DecAlign · CMU-MOSI | F1 Score: 84.17 -> 85.82 | Improved |
| add prototype-guided optimal transport (ablation) | DecAlign · CMU-MOSEI | MAE: 0.624 -> 0.543 | Improved |
| add cross-modal transformer (ablation) | DecAlign · CMU-MOSI | F1 Score: 84.36 -> 85.82 | Improved |
| add semantic consistency loss (ablation) | DecAlign · CMU-MOSI | F1 Score: 84.73 -> 85.82 | Improved |
| add MMD regularization (ablation) | DecAlign · CMU-MOSI | F1 Score: 84.61 -> 85.82 | Improved |
| add MMD regularization (ablation) | DecAlign · CMU-MOSEI | MAE: 0.564 -> 0.543 | Improved |
| use DecAlign instead of DMD | DecAlign · CMU-MOSI | Acc-2: 83.24 -> 85.75 | Improved |
| use DecAlign instead of CGGM | DecAlign · IEMOCAP (six-class) | WAcc: 72.25 -> 73.35 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ReconBoost — ReconBoost: Boosting Can Achieve Modality Reconcilement (ICML 2024) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- On the Value of Cross-Modal Misalignment in Multimodal Representation Learning (NeurIPS 2025 · similarity 0.93)
- AlignMamba: Enhancing Multimodal Mamba with Local and Global Cross-modal Alignment (CVPR 2025 · similarity 0.90)
- UNIALIGN: Scaling Multimodal Alignment within One Unified Model (CVPR 2025 · similarity 0.90)
- DeepAlign: Mitigating Modality Conflict through Modality-Specific Alignment (CVPR 2026 · similarity 0.89)
- Aligning Multimodal Representations through an Information Bottleneck (ICML 2025 · similarity 0.89)
- When to Align, When to Predict: A Phase Diagram for Multimodal Learning (NeurIPS 2026 · similarity 0.89)
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.
