Research map / Video / Video-language understanding
Semi-Supervised Video Paragraph Grounding With Contrastive Encoder
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 SVPTR instead of DepNet | Video paragraph grounding · ActivityNet-Caption | Recall@1, IoU=0.5: 55.91 -> 61.70 | Helped |
| Use SVPTR instead of DepNet | Video paragraph grounding · Charades-CD-OOD | mIoU: 29.30 -> 36.01 | Helped |
| Use SVPTR instead of DepNet | Video paragraph grounding · TACoS | Recall@1, IoU=0.3: 41.34 -> 47.89 | Helped |
| Train SVPTR with only 10% labeled data instead of DepNet@10% | Video paragraph grounding · ActivityNet-Caption | Recall@1, IoU=0.5: 45.14 -> 56.72 | Helped |
| Train SVPTR with only 30% labeled data instead of DepNet@30% | Video paragraph grounding · Charades-CD-OOD | mIoU: 28.09 -> 32.13 | Helped |
| Add multimodal encoding (ablation) | VPTR · ActivityNet-Caption | mIoU: 50.02 -> 50.34 | Helped |
| Add contrastive loss to the multimodal-encoding variant (ablation) | VPTR · ActivityNet-Caption | mIoU: 50.34 -> 50.14 | Hurt |
| Add consistency loss to the sentence-query variant (ablation) | VPTR · ActivityNet-Caption | mIoU: 50.02 -> 51.04 | Helped |
| Add contrastive loss to the consistency-loss variant (ablation) | VPTR · ActivityNet-Caption | mIoU: 51.04 -> 50.42 | Hurt |
| Add feature perturbations (ablation) | SVPTR · ActivityNet-Caption | mIoU: 51.70 -> 51.98 | Helped |
| Remove sentence-level contextual information (ablation) | VPTR · ActivityNet-Caption and Charades-CD-OOD | grounding precision | Hurt |
| Shuffle sentence order (ablation) | VPTR · ActivityNet-Caption and Charades-CD-OOD | grounding precision | Hurt |
Closest papers by meaning
- Siamese Learning with Joint Alignment and Regression for Weakly-Supervised Video Paragraph Grounding (CVPR 2024 · similarity 0.90)
- Learning Multi-Scale Video-Text Correspondence for Weakly Supervised Temporal Article Gronding (AAAI 2024 · similarity 0.89)
- Empower Words: DualGround for Structured Phrase and Sentence-Level Temporal Grounding (NeurIPS 2025 · similarity 0.89)
- Multi-Pair Temporal Sentence Grounding via Multi-Thread Knowledge Transfer Network (AAAI 2025 · similarity 0.87)
- Comprehensive Visual Grounding for Video Description (AAAI 2024 · similarity 0.87)
- VideoGrounding-DINO: Towards Open-Vocabulary Spatio-Temporal Video Grounding (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.