Research map / Label-efficient and robust learning / Domain adaptation and generalization
CDFSL-V: Cross-Domain Few-Shot Learning for Videos
Sarinda Samarasinghe, Mamshad Nayeem Rizve, Navid Kardan, Mubarak Shah
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 |
|---|---|---|---|
| use curriculum scaling of the consistency loss instead of equal loss weighting (ablation) | cross-domain few-shot video action recognition method · Kinetics-100 source, 5-target average | accuracy: 24.90 -> 26.82 | Improved |
| add temperature sharpening of teacher pseudo-labels (ablation) | cross-domain few-shot video action recognition method · Kinetics-100 source, 5-target average | accuracy: 25.59 -> 26.82 | Improved |
| combine self-supervised and supervised curriculum instead of self-supervised training only (ablation) | cross-domain few-shot video action recognition method · Kinetics-100 source, 5-target average | accuracy: 25.11 -> 26.82 | Improved |
| combine curriculum with supervised training instead of supervised-only source training (ablation) | cross-domain few-shot video action recognition method · Kinetics-100 source, 5-target average | accuracy: 23.53 -> 26.82 | Improved |
| use self-supervised-only pretraining instead of supervised-only pretraining (ablation) | transfer baseline · Kinetics-100 source, 5-target average | accuracy: 23.53 -> 25.11 | Improved |
| use larger source dataset (Kinetics-400 instead of Kinetics-100) (ablation) | cross-domain few-shot video action recognition method · 5-target average | accuracy: 26.82 -> 47.24 | Improved |
| remove temperature sharpening (temperature 1) on Diving48 (ablation) | cross-domain few-shot video action recognition method · Diving48, Kinetics-100 source | accuracy: 16.37 -> 16.77 | No clear effect |
| increase sharpening temperature (less sharp) (ablation) | cross-domain few-shot video action recognition method · Kinetics-100 source | accuracy | Hurt |
| propose masked autoencoder + curriculum method instead of Dynamic Distillation | cross-domain few-shot video action recognition · Kinetics-400 source, 5-target average | accuracy: 43.81 -> 47.24 | Improved |
Closest papers by meaning
- Unsupervised Video Domain Adaptation with Masked Pre-Training and Collaborative Self-Training (CVPR 2024 · similarity 0.88)
- Learning Causal Domain-Invariant Temporal Dynamics for Few-Shot Action Recognition (ICML 2024 · similarity 0.88)
- Discriminative Sample-Guided and Parameter-Efficient Feature Space Adaptation for Cross-Domain Few-Shot Learning (CVPR 2024 · similarity 0.88)
- Pushing the Limit of Fine-Tuning for Few-Shot Learning: Where Feature Reusing Meets Cross-Scale Attention (AAAI 2024 · similarity 0.87)
- TAMT: Temporal-Aware Model Tuning for Cross-Domain Few-Shot Action Recognition (CVPR 2025 · similarity 0.87)
- A Closer Look at the CLS Token for Cross-Domain Few-Shot Learning (NeurIPS 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.
