Research map / Vision-language and multimodal / Cross-modal learning
Interact Before Align: Leveraging Cross-Modal Knowledge for Domain Adaptive Action Recognition
Lijin Yang, Yifei Huang, Yusuke Sugano, Yoichi Sato
Check 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 |
|---|---|---|---|
| Use CIA instead of Kim et al. | I3D multi-modal domain adaptation · UCF-HMDB U→H | class-wise action recognition accuracy: 84.70 -> 88.33 | Improved |
| Use CIA instead of Kim et al. | I3D multi-modal domain adaptation · UCF-HMDB H→U | class-wise action recognition accuracy: 92.80 -> 94.05 | Improved |
| Add CIA to TA3N | I3D-TRN · UCF-HMDB U→H | class-wise action recognition accuracy: 89.17 -> 91.94 | Improved |
| Use CIA instead of STCDA | EPIC-Kitchens-55 · mean across six domain transfers | class-wise action recognition accuracy: 51.2 -> 52.2 | Improved |
| Use CIA with RGB and Flow instead of MM-SADA | EPIC-Kitchens-100 · validation | action accuracy: 12.80 -> 14.25 | Improved |
| Add both self-refinement and cross-refinement (ablation) | CIA on EPIC-Kitchens-100 · validation | action accuracy: 19.19 -> 20.49 | Improved |
| Add self-refinement (ablation) | CIA on EPIC-Kitchens-100 · validation | action accuracy: 19.19 -> 19.56 | Improved |
| Add cross-refinement (ablation) | CIA on EPIC-Kitchens-100 · validation | action accuracy: 19.19 -> 19.67 | Improved |
| Add SC to MC (ablation) | CIA on EPIC-Kitchens-100 · validation | action accuracy: 19.98 -> 20.49 | Improved |
| Add MC to SC (ablation) | CIA on EPIC-Kitchens-100 · validation | action accuracy: 19.83 -> 20.49 | Improved |
| Use multi-scale rather than single-scale correlation (ablation) | SC on EPIC-Kitchens-100 · validation, domain adaptation | action accuracy: 19.62 -> 19.83 | Improved |
| Replace SC with spatial attention (ablation) | EPIC-Kitchens-100 domain adaptation · validation | action accuracy: 19.83 -> 19.39 | Hurt |
Closest papers by meaning
- Multi-Modal Domain Adaptation for Fine-Grained Action Recognition (CVPR 2020 · similarity 0.91)
- Discovering Informative and Robust Positives for Video Domain Adaptation (ICLR 2023 · similarity 0.89)
- Human-Centric Transformer for Domain Adaptive Action Recognition (TPAMI 2024 · similarity 0.89)
- XKD: Cross-Modal Knowledge Distillation with Domain Alignment for Video Representation Learning (AAAI 2024 · similarity 0.89)
- Modality-Collaborative Test-Time Adaptation for Action Recognition (CVPR 2024 · similarity 0.88)
- Cross-modal Multi-task Learning for Multimedia Event Extraction (AAAI 2025 · 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.
