Research map / Label-efficient and robust learning / Domain adaptation and generalization
Shuffle and Attend: Video Domain Adaptation
Jinwoo Choi, Gaurav Sharma, Samuel Schulter, Jia-Bin Huang
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 clip order prediction on source and target (ablation) | Clip DA · Kinetics->NEC-Drone | top-1 accuracy: 23.7 -> 28.5 | Improved |
| add clip order prediction on source only (ablation) | Clip DA · Kinetics->NEC-Drone | top-1 accuracy: 23.7 -> 25.9 | Improved |
| add clip order prediction on target only (ablation) | Clip DA · Kinetics->NEC-Drone | top-1 accuracy: 23.7 -> 22.4 | Hurt |
| add clip-level domain adaptation (ablation) | I3D source only · Kinetics->NEC-Drone | top-1 accuracy: 17.2 -> 23.7 | Improved |
| use video-level alignment instead of clip-level alignment (ablation) | SAVA · Kinetics->NEC-Drone | top-1 accuracy: 28.5 -> 30.3 | Improved |
| add clip attention (ablation) | SAVA · Kinetics->NEC-Drone | top-1 accuracy: 30.3 -> 31.6 | Improved |
| use GRU instead of MLP for attention module (ablation) | SAVA attention · Kinetics (attention accuracy) | attention accuracy: 72.2 -> 78.0 | Improved |
| use GRU instead of MLP for attention module (ablation) | SAVA attention · UCF (attention accuracy) | attention accuracy: 86.1 -> 78.9 | Hurt |
| use SAVA instead of TA3N (I3D-based TRN) | domain adaptation · HMDB->UCF | top-1 accuracy: 90.5 -> 91.2 | Improved |
| use SAVA instead of source only | I3D · UCF->HMDB | top-1 accuracy: 80.3 -> 82.2 | Improved |
Closest papers by meaning
- Unsupervised Video Domain Adaptation for Action Recognition: A Disentanglement Perspective (NeurIPS 2023 · similarity 0.91)
- Human-Centric Transformer for Domain Adaptive Action Recognition (TPAMI 2024 · similarity 0.89)
- VideoNet: A Large-Scale Dataset for Domain-Specific Action Recognition (CVPR 2026 · similarity 0.88)
- Unsupervised Video Domain Adaptation with Masked Pre-Training and Collaborative Self-Training (CVPR 2024 · similarity 0.88)
- Align Before Adapt: Leveraging Entity-to-Region Alignments for Generalizable Video Action Recognition (CVPR 2024 · similarity 0.88)
- Modality-Collaborative Test-Time Adaptation for Action Recognition (CVPR 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.
