Research map / Label-efficient and robust learning / Domain adaptation and generalization
Multi-Modal Domain Adaptation for Fine-Grained Action Recognition
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 |
|---|---|---|---|
| Add multi-modal self-supervision over source and target data (ablation) | MM-SADA · EPIC-Kitchens cross-domain mean (6 domain pairs) | Top-1 accuracy: 45.5 -> 47.9; +2.4% | Helped |
| Add per-modality adversarial alignment to self-supervision (ablation) | MM-SADA · EPIC-Kitchens cross-domain mean (6 domain pairs) | Top-1 accuracy: 47.9 -> 50.3 | Helped |
| Combine multi-modal self-supervision with adversarial alignment | MM-SADA · EPIC-Kitchens cross-domain mean (6 domain pairs) | Top-1 accuracy: 45.5 -> 50.3; +4.8% | Helped |
| Use MM-SADA instead of MCD | I3D action recognition · EPIC-Kitchens cross-domain mean (6 domain pairs) | Top-1 accuracy: 47.3 -> 50.3 | Helped |
| Use MM-SADA instead of AdaBN | I3D action recognition · EPIC-Kitchens cross-domain mean (6 domain pairs) | Top-1 accuracy: 47.2 -> 50.3 | Helped |
| Use MM-SADA instead of MMD | I3D action recognition · EPIC-Kitchens cross-domain mean (6 domain pairs) | Top-1 accuracy: 46.8 -> 50.3 | Helped |
| Apply MM-SADA to D3→D2 with λd=1, λc=5 | MM-SADA · EPIC-Kitchens D3→D2 | Top-1 accuracy: 56.3 -> 56.1; -0.2% | Hurt |
| Use self-supervision with λd=0, λc=5 instead of source-only (ablation) | MM-SADA · EPIC-Kitchens D3→D2 | Top-1 accuracy: 56.3 -> 57.4 | Helped |
| Reduce adversarial loss weight from λd=1 to λd=0.5 while retaining self-supervision (ablation) | MM-SADA · EPIC-Kitchens D3→D2 | Top-1 accuracy: 56.1 -> 56.9 | Helped |
| Add multi-modal self-supervision to adversarial alignment (ablation) | RGB stream · EPIC-Kitchens cross-domain mean (6 domain pairs) | Top-1 accuracy: 41.5 -> 43.9 | Helped |
| Add multi-modal self-supervision to adversarial alignment (ablation) | Flow stream · EPIC-Kitchens cross-domain mean (6 domain pairs) | Top-1 accuracy: 48.2 -> 49.3 | Helped |
| Use MM-SADA rather than source-only training (ablation) | RGB stream · EPIC-Kitchens cross-domain mean (6 domain pairs) | Top-1 accuracy: 37.4 -> 43.9 | Helped |
Closest papers by meaning
- Unsupervised Video Domain Adaptation with Masked Pre-Training and Collaborative Self-Training (CVPR 2024 · similarity 0.89)
- Modality-Collaborative Test-Time Adaptation for Action Recognition (CVPR 2024 · similarity 0.88)
- Local-Global Multi-Modal Distillation for Weakly-Supervised Temporal Video Grounding (AAAI 2024 · similarity 0.86)
- SeFAR: Semi-supervised Fine-grained Action Recognition with Temporal Perturbation and Learning Stabilization (AAAI 2025 · similarity 0.86)
- Learning Causal Domain-Invariant Temporal Dynamics for Few-Shot Action Recognition (ICML 2024 · similarity 0.85)
- Storyboard-guided Alignment for Fine-grained Video Action Recognition (NeurIPS 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.