Research map / Label-efficient and robust learning / Zero-/few-shot and meta-learning
Towards Universal Representation for Unseen Action Recognition
Read the paperCheck 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 |
|---|---|---|---|
| Use the proposed cross-dataset pipeline instead of GMIL+ESZSL | URL pipeline · HMDB51 cross-dataset | accuracy: 25.7 -> 51.8 | Helped |
| Use the proposed cross-dataset pipeline instead of GMIL+ESZSL | URL pipeline · UCF101 cross-dataset | accuracy: 19.8 -> 42.5 | Helped |
| Add the JSD constraint to NMF-based URL (ablation) | URL · HMDB51 cross-dataset, high dimension | accuracy: 18.0 -> 51.8 | Helped |
| Add the JSD constraint to NMF-based URL (ablation) | URL · UCF101 cross-dataset, high dimension | accuracy: 17.4 -> 42.5 | Helped |
| Add TJM semantic domain adaptation (ablation) | URL pipeline · HMDB51 cross-dataset, high dimension | accuracy: 50.5 -> 51.8 | Helped |
| Add TJM semantic domain adaptation (ablation) | URL pipeline · UCF101 cross-dataset, high dimension | accuracy: 36.6 -> 42.5 | Helped |
| Add TJM domain adaptation using target visual examples (ablation) | URL pipeline · HMDB51 transductive, high dimension | accuracy: 53.9 -> 58.2 | Helped |
| Add TJM domain adaptation using target visual examples (ablation) | URL pipeline · UCF101 transductive, high dimension | accuracy: 38.6 -> 49.9 | Helped |
| Use GMIL rather than Fisher Vector encoding (ablation) | URL pipeline · HMDB51 cross-dataset, high dimension | accuracy: 48.6 -> 51.8 | Helped |
| Use GMIL rather than Fisher Vector encoding (ablation) | URL pipeline · UCF101 cross-dataset, high dimension | accuracy: 39.7 -> 42.5 | Helped |
| Replace URL with CCA (ablation) | URL pipeline · HMDB51 cross-dataset, high dimension | accuracy: 51.8 -> 12.2 | Hurt |
| Increase UR dimensionality from low to high (ablation) | URL pipeline · HMDB51 cross-dataset | accuracy: 49.6 -> 51.8 | No clear effect |
Closest papers by meaning
- LLaVAction: evaluating and training multi-modal large language models for action understanding (ICLR 2026 · similarity 0.85)
- SeFAR: Semi-supervised Fine-grained Action Recognition with Temporal Perturbation and Learning Stabilization (AAAI 2025 · similarity 0.85)
- DVANet: Disentangling View and Action Features for Multi-View Action Recognition (AAAI 2024 · similarity 0.85)
- VA-AR: Learning Velocity-Aware Action Representations with Mixture of Window Attention (AAAI 2025 · similarity 0.84)
- Action Detail Matters: Refining Video Recognition with Local Action Queries (CVPR 2025 · similarity 0.84)
- Align Before Adapt: Leveraging Entity-to-Region Alignments for Generalizable Video Action Recognition (CVPR 2024 · similarity 0.84)
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.