Research map / Video / Video-language understanding
Unified Embedding and Metric Learning for Zero-Exemplar Event Detection
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 |
|---|---|---|---|
| use a unified visual-textual embedding instead of visual-only embedding (ablation) | model^U · MED-13 | mAP: 11.90 -> 17.86 | Helped |
| use a unified visual-textual embedding instead of visual-only embedding (ablation) | model^U · MED-14 | mAP: 10.76 -> 16.67 | Helped |
| add contrastive loss to visual-only embedding (ablation) | model^V · MED-13 | mAP: 11.90 -> 13.29 | Helped |
| add contrastive loss to visual-only embedding (ablation) | model^V · MED-14 | mAP: 10.76 -> 12.31 | Helped |
| add textual embedding to the visual-only contrastive model (ablation) | model^C · MED-13 | mAP: 13.29 -> 17.86 | Helped |
| add textual embedding to the visual-only contrastive model (ablation) | model^C · MED-14 | mAP: 12.31 -> 16.67 | Helped |
| jointly embed visual and textual inputs instead of training separate classifiers (ablation) | model^U · MED-13 | mAP: 15.60 -> 17.86 | Helped |
| jointly embed visual and textual inputs instead of training separate classifiers (ablation) | model^U · MED-14 | mAP: 13.49 -> 16.67 | Helped |
| use contrastive metric learning instead of MSE loss and cosine distance (ablation) | model^U · MED-13 | mAP: 15.92 -> 17.86 | Helped |
| use contrastive metric learning instead of MSE loss and cosine distance (ablation) | model^U · MED-14 | mAP: 14.36 -> 16.67 | Helped |
| use the proposed method rather than VideoStory with EventNet and ResNet features | MED-14 · MED-14 | mAP: 15.80 -> 16.67 | Helped |
Closest papers by meaning
- Learning Event Completeness for Weakly Supervised Video Anomaly Detection (ICML 2025 · similarity 0.84)
- MultiVENT 2.0: A Massive Multilingual Benchmark for Event-Centric Video Retrieval (CVPR 2025 · similarity 0.83)
- Mitigating Semantic Collapse in Partially Relevant Video Retrieval (NeurIPS 2025 · similarity 0.83)
- Streaming Detection of Queried Event Start (NeurIPS 2024 · similarity 0.83)
- Bridging the Gap: A Unified Video Comprehension Framework for Moment Retrieval and Highlight Detection (CVPR 2024 · similarity 0.83)
- HyrCap: Hybrid Rank-Calibration of Action Proposals for Temporal Event Understanding (NeurIPS 2026 · similarity 0.83)
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.